c/EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
EPA-600/2-78-199
September 1978
Research and Development
Proceedings:
Symposium/Workshop
on Petroleum Refining
Emissions (April 1978,
Jekyll Island, GA)
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental Protec-
tion Agency, have been grouped into nine series. These nine broad categories were
established to facilitate further development and application of environmental tech-
nology. Elimination of traditional grouping was consciously planned to foster technology
transfer and a maximum interface in related fields. The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECHNOLOGY
series. This series describes research performed to develop and demonstrate instrumen-
tation, equipment, and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the new or improved tech-
nology required for the control and treatment of pollution sources to meet environmental
quality standards.
REVIEW NOTICE
This report has been reviewed by the U.S. Environmental
Protection Agency, and approved for publication. Approval
does not signify that the contents necessarily reflect the
views and policy of the Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161. '"forma
-------
EPA-600/2-78-199
September 1978
Proceedings:
Symposium/Workshop
on Petroleum Refining Emissions
(April 1978, Jekyll Island, GA)
by
Susan R. Fernandes, Compiler
Radian Corporation
P. O. Box 9948
Austin, Texas 78766
Contract No. 68-02-2608
Task No. 24
Program Element No. IAB604C
EPA Project Officer: Irvin A. Jefcoat
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
TABLE OF CONTENTS
WELCOME, by Dr. John K. Burchard
HIGH ANXIETY AND THE ENVIRONMENT, by Francis N. Dawson, Jr.
PETROLEUM REFINERY FUGITIVE EMISSIONS MEASUREMENT EMISSION FACTORS AND
PROFILES, by H. J. Taback
FUGITIVE HYDROCARBON EMISSIONS - MEASUREMENT AND DATA ANALYSIS METHODS,
by Karen M. Hanzevack
DETECTION AND CLASSIFICATION OF FUGITIVE COMPONENT LEAKS, by P. R. Harrison
FUGITIVE EMISSION CONTROL STRATEGIES FOR PETROLEUM REFINERIES, by I. A.
Jefcoat, R. G. Wetherold, and W. Leigh Short (Abstract Only)
DETECTION OF VOLATILE ORGANIC COMPOUND EMISSIONS FROM EQUIPMENT LEAKS,
by K. C. Hustvedt and R. C. Weber
STATE REGULATIONS FOR CONTROL OF REFINERY EMISSIONS, by Henry E. Sievers
INSPECTION AND MONITORING CONCEPTS FOR REFINERY FUGITIVE EMISSIONS,
by John H. Nakagama
API EMISSION MEASUREMENT PROGRAMS, by J. G. Zabaga
VALVES - A POSSIBLE SOURCE OF FUGITIVE EMISSIONS IN HYDROCARBON PROCESSES,
by Alton M. Williamson
PLANS FOR ASSESSMENT OF WATER AND RESIDUALS EMISSIONS FROM THE PETROLEUM
REFINERY, by Fred M. Pfeffer
WORKSHOP DISCUSSION - PART I, Fugitive Emissions Detection and Measurement,
by Paul R. Harrison
WORKSHOP DISCUSSION - PART I, Fugitive Emissions Control, by Fred Storer
WORKSHOP DISCUSSION - PART I, Point Source Emissions, by Larry Johnson
WORKSHOP DISCUSSIONS - PART I, QUESTIONS AND ANSWERS
WORKSHOP DISCUSSIONS - PART II, Emissions from Waste Treatment Facilities,
by Fred Pfeffer
WORKSHOP DISCUSSIONS - PART II, Regulations, by Leigh Short
-------
TABLE OF CONTENTS (Cont'd)
WORKSHOP DISCUSSIONS - PART II, Equipment Considerations, by Alton Williamson
WORKSHOP DISCUSSIONS - PART II, Fugitive Emissions, by James W. Daily
-------
WELCOME
Dr. John K. Burchard
Director, Industrial Environmental Research Laboratories
Environmental Protection Agency
Research Triangle Park, North Carolina
On behalf of the Lab and EPA, I would like to welcome you to the conference
on Emissions from Petroleum Refining. It looks like the weather has cleared up
and we are going to have a nice day.
Just for your information, I would like to describe a little about the
Laboratory. We are responsible for research, development and demonstration
for air and water pollution control technology from a variety of stationary
sources. We have three divisions: (1) our Utility and Industrial Power
Division is concerned with flue gas treatment and particulate technology;
(2) our Engineering Assessment and Control Division is concerned with the
more near term effects of energy technology such as coal gasification, coal
liquefaction and combustion research; and (3) our third division is the
Industrial Process Division. We have a sister laboratory in Cincinnati with
the same name, and between the two of us, we split up the entire industrial
world as far as environmental assessment and pollution control technology.
The Chemical Processes Branch of our Industrial Process Division, headed up
by Dale Denny, is responsible for textiles, agricultural chemicals which
includes pesticides and fertilizers, and last but not least, petrochemicals
and petroleum refining. This branch, therefore, is the chief sponsor to
this meeting.
As I am sure many of you know, we are currently engaged in a fairly large
program to measure emissions from refineries. Refineries are a rather signi-
ficant part of the fuel conversion industry pollution emissions. These, of
course, may be potential health hazards. Results from this program should
allow us to aim our Research and Development funds better so that we can
determine where our money should be spent as far as control technology
development goes. It should help the refiners to reduce hydrocarbon feedstock
and product losses, and this, of course, is important from an energy standpoint
as well as from an environmental standpoint. And the results should allow the
various regulatory agencies to set more realistic standards since we will have
a better data base. I would like to make it clear that we in the Laboratory
are not part of the regulatory part of EPA, so we are the good guys.
Those of you who are attending the Symposium represent many kinds of
industry, not only the oil industry, and various government agencies. Through
our combined efforts we can identify and solve our various problems. The
program to measure refinery emissions, which I mentioned before, is a good
example of this cooperation. When we started in 1976, it was pretty obvious
that we would need good cooperation from the petroleum industry to avoid the
study just becoming another abstract effort. The staff met with people
from the American Petroleum Institute to discuss the scope of work.
-------
Dr. John K. Burchard
API formed a Task Force and together we developed a program which should provide
valuable information to both of us. Through the help of the National Petroleum
Refiners Association and the API, we have measured emissions in refineries
throughout the country. At all the sites we have found real cooperation and
valuable assistance. The managerial staff and the people in the refinery have
devoted a lot of time and effort. We recognize this contribution and I would
like to say that we are really appreciative of it. We couldn't do the study
without this kind of cooperation.
We have already sampled at seven different refineries. We have come up
with several results; we have defined some major sampling problems and developed
some innovative techniques to handle these problems. We have gathered a data
base for a variety of refineries and processes. We have identified important
sources of fugitive emissions and we have developed some rapid emperical
screening methods for leaks which, of course, are one of the major sources of
fugitive emissions. We are now just about at the half-way point in the program
and this Symposium will give us a chance to share results and get the benefit
of your comments which will help us complete the rest of the study. In the
next couple days, we will be discussing emissions data, sampling techniques,
control methods and potential regulations and control techniques.
Again, I would like to welcome you to our meeting and hope that the neKt
few days prove interesting and productive. Chairman of the first session is
Dr. Donald Rosebrook.
-------
HIGH ANXIETY AND THE ENVIRONMENT
Francis N. Dawson, Jr.
President, USA Petrochem Corporation
Santa Monica, California
ABSTRACT
A measure of uncertainty and concern persists among our neighbors
about the quality of life to which we may look forward including, especially,
air quality. Concurrently, a growing number of individuals are opposing
continued growth of our cities and industries for similar reasons. The
paper attempts to assess some of the factors which relate to industrial and
economic growth, to examine the relation between growth and personal income
and to consider briefly the impact of regulatory policy as it relates to our
future growth and economic progress.
RESUME
Our keynote speaker is Francis N. Dawson, Jr.,(Fritz) President and Chief
Executive Officer of U.S.A. Petrochem Corporation, Inc., Santa Monica,
California. He also serves as Executive Vice President of the U.S.A.
Petroleum Corporation of which U.S.A. Petrochem is a subsidiary. Dawson
first joined U.S.A. Petrochem in February of 1974 as Executive Vice President.
Resident of Southern California since 1967, he formerly held the position
of Director of Manufacturing with Douglas Oil Company which is a subsidiary
of Continental Oil in Paramount, California. He previously served with
Continental Oil in various capacities including assignments in engineering
technical services and refinery management. He served for several years with
Conoco1s International Division in New York, Europe and Latin America. He
graduated from the University of Michigan in 1955 with a B.S. and a M.S. degree
in Chemical Engineering and honors from Tau Beta Phi and Phi Lambda Epsilon.
He has served in many professional organizations in the past, including the
American Petroleum Institute, Chairman of the Western Oil and Gas Association's
Environmental Conservation Committee, and its Refining Committee. He was
also a member of WOGA's Pipeline Committee. He is a Director of the National
Petroleum Refiner's Association and the California Fertilizer Association.
He is a member of the Fertilizer Institute and the Los Angeles Petroleum
Club. As the petroleum refining and fertilizer manufacturing arm of the U.S.A.
Petroleum, U.S.A. Petrochem produces petroleum products and anhydrous ammonia
and operates a 21,000 barrel/day refinery in Ventura California. His paper
is being presented by Jim Daily from Chevron.
-------
HIGH ANXIETY AND THE ENVIRONMENT
Let me respond to a letter I received from a woman from Ojai, California, a
town which could be considered as an air receptor area from our Ventura
plant. Bessie Simon wrote to me about growth and jobs and the quality of
life. Here are a few of her comments:
"Don't you see that you are perpetuating a vicious circle. The
more jobs and supplies you produce, the more people. And the
more people, the greater demand for more jobs and supplies. And
what has been accomplished. Somebody has to take the initiative
to stop expansion until the crush of over-breeding dies down.
It should have been done long ago. There must be a stop to this
blindly expansive, ruthless and irresponsible growth."
Continuing on, she says:
"What it does to the people doesn't matter, just keep on making
liquor, cigarettes, soft drinks, junk foods, silly gadgets, nuclear
power plants. Really," she says, "we should start praying that it
does collapse before it destroys everything that is good and
beautiful. You are demanding the privilege of using our resources
in order to supply more of the very things that brought on this
sickness." She says: "What are the needs of a sick society? Is
it more jobs, is it more material things, is it more luxuries,
this philosophy of more is what has gotten us into the mess we are
in. Surely then we do not need more of more. We need to about-
face, to back-track as fast as possible. The young people should
be studying up on farming, old fashioned farming with the help of
solar technology and make a go of it as our ancestors did."
She concludes by saying that perhaps we should close our whole plant.
In a nutshell, Bessie Simon is saying:
1. We are responsible for the increased population because
we provide more jobs.
2. Perhaps we are abusing our resources.
3. We ought to revert to small farms or small businesses.
4. We are abusing nature and, she adds
5. A liberal infusion of anti-business sentiment.
But the issue she raises is growth. Should we have it and what are its con-
sequences?
-1-
-------
Francis N. Dawson, Jr.
As to whether my creation of jobs has the slightest effect on the breeding
habits of you in the audience or anyone else, well, let's face it, Bessie
Simon, either I'm a lot more influential than you think I am, or your
pencil got away from you just a little bit there. Not only am I unable
to influence the supply of employees through population control but, as a
matter of fact, the people who will be moving into the work force within
the next twenty years have already been born. The population growth pro-
blem is less serious in the developed nations where the jobs are. It is
much worse in poor countries where large families are desired to help with
the family support.
More importantly, Bessie Simon is opposed to growth. Is she alone in her
objections to growth? Most assuredly not. I will talk mostly about the
energy business because that's what I know the most about. But just in our
own industry, the anti-growth or the slow-growth advocates have been enor-
mously successful.
Cancellation of the George's Bank lease-sale earlier this year.
Denial of at least thirteen large proposed refinery projects on the U.S.
East Coast.
Potential loss of heavy crude oil production in California's central valley.
Dow's difficulty in the San Francisco Bay area.
Lease suspensions for oil shale development.
Delay of Sohio's Long Beach terminal to handle Alaskan north slope crude.
Closer to home, we experienced utter defeat for our company's proposed project
in Ventura to reactivate portions of an idle urea facility which had been in
use for twenty years, despite a commitment to reduce emissions by 90% from
their level under a former owner.
Other problems:
Conoco's Cat Canyon crude.
Exxon's twenty-eight new wells in the Santa Barbara channel.
Crude oil production from Elk Hills.
Phillips $400,000 proposed modernization in Texas.
Yes, Bessie Simon - you have many who agree with you. Why are these projects
being challenged or turned down? And these are just the energy related pro-
jects. In fact, just a few of them. You are no doubt familiar with other
examples in other industries.
Right now, I want to go right to the heart of Bessie Simon's question: Why
not forget about this growth, why not go backwards? What good is growth
anyway? The role of business is poorly understood by many of our neighbors
today.
-2-
-------
Francis N. Dawson, Jr.
I'd like to pose a little quiz and, at the same time, recommend to you and
your children a little brochure called "The American Economic System and
Your Part In It."1 In the brochure you will find that the gross national
product of our country in 1975 was a little over one and a half trillion
dollars. Most of you, I am sure, will remember that gross national product
is the total of all the goods and services that we produce in the country.
Now here's the quiz. What would you guess is the proportion of that one
and a half trillion that winds up as personal income? How many of you think
that it would be as much as ten percent of the total 6NP? Let me have a show
of hands. How many of you think it would be over twenty percent? How many
think it would be over twenty-five percent? Well, the answer is three-fourths.
Seventy-five percent, three fourths of gross national product can be accounted
for as personal income. In other words, GNP can be considered in a way as
almost synonymous with personal income.
If you divide the GNP, or at least the portion of it that's personal income,
by the population, we get something on the order of $15,000 average family
income in 1975 spread out among our 214 million people.
One hundred years before that, back in 1876, the gross national product at
that time was only 47.4 billion dollars, expressed in 1975 dollars. The
population in those days was only 44 million. That means in a hundred years,
we increased the population by five times, but the gross national product,
and with it our personal income, has gone up 32 times, all expressed again
in 1975 dollars. Therefore the real personal income has gone up, and with
it our average standard of living, by a factor of roughly six in the last
hundred years.
You see what I've done. I've said that personal income accounts for most
of our GNP, the production of goods and services. I've divided by our
current population to see how much personal income per person we have to date
as compared to how much personal income per person we had a hundred years ago.
Well, how did we do it? How did we accomplish this big increase in our average
real personal income which is how we get standard of living? First, remember
that GNP and personal income are not truly the same. GNP is what we produce.
Income simply accounts for most of the cost of what we produce. We were able
to increase our GNP and our personal income because we became more productive.
Our productivity went up. It went up because we learned how to make better
tools, spent the money to buy them, and we drove those tools with energy other
than human power. Energy used per person increased 2.5 times from 1900 to 1973.
The only time the trend was reversed in the United States was during the Great
Depression of the 1930's and, sure enough, both GNP and personal income during
that period went down on a per capita basis.
It's interesting to note that the GNP of virtually every major country in the
world relates directly and simply to the energy it uses. The United States
consumes roughly a third of the world's energy but it produces roughly a third
of the world's gross national product.
-3-
-------
Francis N. Dawson, Jr.
What are some of the by-products of this growth In personal income? For
one thing, we've greatly minimized child labor. The hours that we have
to work in a week, our standards of health, education, housing, medical
care, and the other amenities, have been achieved only because growth has
occurred at a rate faster than the population has grown. In other words,
average real income really means standard of living.
Let's take the other approach. Consider what might have occurred since,
let's say 1948, if growth had been held down to only the rate of population
growth. In 1948, we saw a disposable family income of $7,430.00, that is,
in 1975 dollars. Population then went up and employment increased from
58 million up to 84 million in 1973. That's an increase of about 1% percent
a year. If the growth had been restricted during that period to only the
same percentage of increase, that is 1% percent per year, then you and I
would still be taking home to our family the same average of $7,430.00 a
year of 1975 dollars. Compare that with what really happened — nearly
$15,000.00.2
Likewise, taxes that the Government took in in 1948 were only $1,060.00 per
family, and that could not have gone up either without growth. Instead,
taxes did go up to an average of $3,740.00 per family, some three and a half
times. The Government increased Social Security payments 2.7 times and health
expenditures 7.8 times. Expenditures for education by the Government increased
4.1 times, and housing went up 7 times. Again, all in 1975 dollars.2
Thus, not only has personal income doubled, but Government spending has greatly
increased, all adding to the quality of our lives. Had growth been held down
to a "slow growth" level of 1% percent per year from 1948 to 1973, real per-
sonal income would be one half of today's level and benefits from the Government
services would be about one fourth of what we receive today.
Even more importantly, if "no growth" had occurred, then we could have fore-
seen the same gross national product divided by the larger population increase
or a reduction in real personal income.
Well, enough of the economics lesson, and whether you remember the numbers or
not, the important thing is don't let anybody tell you that our standard of
living, our personal incomes, and the benefits we get from Government services
are not related to growth. They are directly related to growth. They are a
consequence of our growth and, more specifically, how much growing we do for
the population we have.
Now let's take a look at what can happen to our own local area. Some of you
may have seen the Wall Street Journal article on February 8th about Santa
Barbara. The article says "hundreds of residents are leaving. Plans for
developing the downtown area have been put off for years. Thousands are
attracted by the mild climate, scenic beaches, and the relaxed way of life,
and that is just the way most south coasters want the place to stay. But
preserving the status quo has its social and economic price."
-4-
-------
Francis N. Dawson, Jr.
Still quoting from the article, "Construction curbs have sent the prices of
existing homes rocketing, driving these homes out of reach for the middle
class people with growing families who want more space. People are leaving
who want to stay in the area but can't find career opportunities. Limits on
business and industrial growth coupled with the inflated housing values have
pushed property taxes skyward for those who remain." An auto mechanic
wanted to open his own auto repair shop but was forced to leave because slow-
growth country is a poor place for that kind of a project. Carpenters have
left the area. School enrollment is down. Teachers are losing their jobs.
Young people are going away. It's impossible for them to get a decent job
and even if they did, they'd have to live at home. Apparently some thought-
ful anti-growth proponents recognize an adverse impact but they argue that
too much growth would be worse. What alternative do we have? Crime? Poverty?
More minorities? Social problems? These are the things we would have to deal
with.
My question to you, ladies and gentlemen, is what effect you might foresee by
restricting growth to a level below reasonable balance. Remembering that ex-
cessive restriction leads to decline in personal income and employment, I
question whether the consequences might not be more crime, more poverty, and
more social problems than if growth is allowed to occur in an orderly way.
The history of Ventura, our community, has been one of agriculture and petro-
leum industries. Both are classic examples of endeavors in which productivity
gains have been striking. I can still remember when my grandfather first got
baling machines and combines and forage harvesters.
The same kinds of things have occurred in the petroleum industry. The industry
has been able to grow and increase its output of energy to fuel the tools of
industry and transportation.
Bessie Simon would have us go back to the small farm. "Old fashioned farming"
to restore the quality of life. Both my grandparents were farmers and I still
remember something of the farmer's life in those days. The farmer's standard
of living, work load, health and other social values have improved, as has life
for most of the rest of us since then.
Should we close our plant as Bessie Simon suggests, or prevent others from
growing, or make it impossible for our employees to live here? Are these the
answers to growing crime, poverty, minorities and social problems? As Bessie
says, Nature and Clean Air have been common denominators of man's existence
since the days of pre-history but the progress of his culture, art, music,
medicine, leisure, the time to enjoy the out-of-doors, all of this progress
has been made possible by our average income and the increased ability of each
of us to deliver more goods or services for the time we put in.
The remaining question posed by Bessie Simon is "Are we abusing nature by
abusing the use of its raw materials?" Are we wasting precious resources and
are we destroying or contaminating others? The answer to both questions is
clearly yes.
-5-
-------
Francis N. Dawson, Jr.
We have indeed been wasteful and we are now learning to be more conscious
but we are not as wasteful as you might think. I said before we use about
a third of the world's energy but we also produce about a third of the
world's GNP. We can undoubtedly save ten to perhaps twenty percent of our
energy usage. There will be a cost as well as an energy consumption to
achieve the energy conservation. Beyond twenty percent, I expect impairment
of output with concurrent reduction in our personal income and in our re-
lated quality of life.
Many of the changes in public attitude toward industry, and the petroleum
industry in particular, have come about because of abuses to the environment,
our air and our water especially.
These problems led to a rapid proliferation of regulations and legislation:
1) the creation of local APCD's, 2) the Clean Act of 1970, with Federal Air
Quality Standards, 3) the Clean Water Act, 4) the California Supreme Court
decision in the early 1970's, the Friends of Mammoth decision requiring
EIR's on private projects, 5) the Regional Water Quality Water Control Boards.
You know this history well.
Obviously there has been recognition of environmental problems. Legitimate
restraint on both emissions and growth has occurred as a consequence of these
resulting laws and regulations. The abuses are now largely controlled, however.
At least two other important restraints on growth have developed. The first
is the rise in the consumer interest groups and the power they exert.
The other restraint stems from the fact that media coverage tends not to favor
industrial or business expansion, but rather to flag the opposite view. Hit-
and-run journalism clearly leads to impossible public relations and delayed
or perhaps defeated projects. Fortunately, we now see some real changes
taking place for the better in this area.
Perhaps our case may be typical. Our company has a direct payroll of over
$1.5 million a year and we pay some $0.5 million in property taxes. We
support some two hundred local businesses, repair shops, suppliers, stationery
stores, banks, etc., in the Ventura area. We buy some $3.0 million in goods
and services to run the business from people who do business in the Ventura
area. Markets permitting and given continuing community support, we see an
active future role in Ventura. 1) Our company will not determine population
in the next twenty years but we do hope to provide employment for those who
will be needing jobs to support their families. 2) Our contribution to
society, both local and otherwise, reflects a high level of need — food and
clean fuels -- agriculture and energy. 3) We are a product of the new era
in regulatory controls, controls which demand safety to the public and to our
employees and which will not tolerate abuse of the environment. If we have
difficulty in meeting our limits, it is because no one else has such limits.
We cannot exceed ten miles per hour in a fifty-five mile zone for everyone else.
Up to now, I have talked about why reasonable growth is important, some restraints
to growth during the past decade, and the contribution companies such as ours
can make to their communities. Now what about the future?
-6-
-------
Francis N. Dawson, Jr.
Our company may well be among the last major projects we shall see for some
time. Recent amendments to the Clean Air Act provide that in any area where
air quality standards have not been attained, no new industrial plant can be
built after July 1, 1979, unless the State has adopted and the EPA has
approved an air pollution control plan that will assure full compliance by
the end of 1982, or 1987 in the case of photochemical oxidants.
In an article in the Wall Street Journal on December 28th of last year, John
Quarles, former Deputy Administrator of EPA, says this new law, like a loose
cannon on a pitching deck, threatens a path of destruction. It is likely to
stop construction of new plants or new plant expansions in Southern California,
the Great Lakes region, the Gulf Coast, and the East Coast from Virginia to
Maine.
EPA adopted and Congress supported a compromise policy of emissions trade-offs
or offsets, that is reductions in emissions from other existing sources above
and beyond requirements of existing air pollution laws. But then Congress
added the additional burden of the total State compliance.
The difficulties with this program are 1) those companies that are planning
to build or modify a plant and the workers whom it will employ are without
the power to assure completion of the State plan, 2) companies like ours
which are already doing ten miles an hour in a fifty-five mile an hour zone
have no trade-offs available, and 3) other companies who may have trade-offs
now will lose them if the APCD should tighten emission rules, or at the very
least, upon completion of their first project.
As it now stands, the new law is rigid, it has no flexibility for non-enforcement.
It is mandate for a nearly no-growth policy for over sixty percent of the
country's population. The consequences of such a mandate are obvious if you
remember the first part of my presentation about growth. There is no doubt in
my mind that the results could be disastrous and that a political resolution of
the confrontation will have to occur.
Some action will have to be taken, and it will be taken. The political pressures
would be intolerable otherwise. Assuming some resolution will occur' vis-a-vis
the SIP compliance mandate, we may still be tackling the trade-off problems
for some time to come. Even if the trade-off policy changes, your work at this
conference will have far-reaching impact on companies such as ours who seek new
projects upwind from towns like Ojai.
We went through a makeshift "new source review process" before the laws were
written using local permitting rules as the authority. The lack of data on
emissions, controls and the inability to model those emissions were tough
problems for us. At the same time, our opponents had a field day because the
claims of environmental impact were more emotionally constructed rather than
factually based.
Our experiences with photochemical oxidant suggests that even when you have
narrowed the gaps on hydrocarbon emission data, we will still have a lot of work
to do before we understand how to use them. The real story will unfold as we
use your new data and the new modelling procedures to see if we can control
environmental quality while yet allowing a balanced growth to take place in our
communities.
-7-
-------
Francis N. Dawson, Jr.
In summary, my pitch to you has been this. We have enjoyed a higher standard
of living because our personal income has gone up and our hours of work have
gone down. Our personal income accounts for 75% of our gross national product
and we have been able to increase our personal income only because we have
been able to increase our gross national product faster than our population
has grown. Continued growth to match population is a must, or we shall indeed
go backwards, and backwards in a way that we may not like. Secondly, I have
touched on some of the environmental and other problems that have been tackled
over the last few years and some of the steps that have been taken to try to
correct them. The highly regulated conditions imposed during this last decade
are typical of those which our own company must now meet. Finally, I look for
ways to minimize potential over-reaction, especially the consequences of the
recent amendments to the Clean Air Act concerning review of new sources of
emissions. I hope to alert you to the "rock-and-the-hard-place" situation
which we as industry now face. While you have a unique opportunity to make direct
input to more accurate procedures and data, I urge you also to listen to the
issues with a critical and a skeptical mind, to give industry the support it
deserves, including greater credibility, and to actively encourage as private
citizens in your own communities a policy of balanced growth.
-8-
-------
Francis N. Dawson, Jr.
REFERENCES
1 The Advertising Council, Inc., 825 Third Avenue, New York, N. Y. 10022.
2 Statistics compiled by Electric Power Research.Institute, NPRA Annual
Meeting, March 28, 1977.
-9-
-------
Francis N. Dawson, Jr.
COMMENTS
Dick Keppler, US EPA, Region I - OCS development on the Georges Bank
has not been cancelled, it has been delayed pending revisions of the
Environmental Impact Statement. Of the 13 proposals on the East Coast,
I have been involved in most of them, and there were only three serious
proposals wherein the process design was investigated and presented to us.
The remainder were proposals of what would happen if I built a refinery at
this point1. Of those three serious ones, two are still under consideration,
one at East Port, Maine and another one at Hampton Rhodes, Virginia. The
third of the three failed because of financial support. I would also like
to make a comment that as far as I know, refining is not a labor intensive
business and I would expect that the Ventura Refinery running 21,000 bbl/day
from 1948 (or whenever it started) to 1978 has reduced manpower, jobs have
been eliminated and not increased.
Comment - Mr. Chairman it sounds like there is a difference within the
Federal Government as to the cause for the 13 refineries on the East Coast
not being developed. The Department of Energy (ERDA or FEA) in a publication
did list each of those and contends that it was due to environmental grounds.
There is a difference.
Comment - My company is behind the design or did the design of all three
of these refineries and I did the environmental design for the refineries.
The one that failed financially did get all of the approvals, but did not
get the financial backing. The other two have been delayed totally because
of environmental reasons.
10
-------
PETROLEUM REFINERY FUGITIVE EMISSIONS
MEASUREMENT EMISSION FACTORS AND PROFILES
H. J. Taback
KVB, Inc.
Tustin, California
ABSTRACT
As part of an assessment of organic species emissions from stationary
sources in the greater Los Angeles area, measurements were made of fugitive
organic emissions from refinery valves, pumps, separators and cooling
towers. The measurements included emission rate and composition. The
technique used for measuring emissions and computing emission factors
is presented along with results comparing emission factors to those in
EPA document AP-42. Also included are organic species profiles (i.e.,
percent composition by weight) for fugitive emissions from valves and
pumps. Based on this limited testing, it was concluded that the emission
factors in AP-42 are reasonable estimates for average emissions and that
refinery fugitive emissions are primarily paraffins with low photochemical
reactivity.
RESUME
Hal Taback has a Bachelor's degree in Mechanical Engineering from the
University of Rochester and a Master's of Science in Aeronautical Engineering
from Princeton. He is currently the Manager of Energy and Environmental
Systems at KVB where he is responsible for emissions inventories, emission
control technology assessment, energy conservation and energy systems
reliability. He is currently conducting programs to inventory gaseous NO ,
SO , and organic species and particulate matter by size and composition
for the California South Coast Air Basin as well as measuring these same
emissions from tertiary oil recovery operations. Prior to joining KVB,
Mr. Taback was in private practice as an industrial consultant and environ-
mental contractor, and for 16 years he was associated with Aerojet General
as a rocket designer and program manager. He is a former Naval engineering
officer with four years of active duty. He has also spent three years as a
researcher at the Gugenheim Aeronautical Laboratory of Princeton University
where he was engaged in the fundamental aspects of solid rocket propellant
burning.
-------
H. Taback
PETROLEUM REFINERY FUGITIVE EMISSIONS
MEASUREMENT EMISSION FACTORS AND PROFILES
INTRODUCTION
As part of an assessment of organic species emissions in a region of
California known as the South Coast Air Basin (i.e., greater Los Angeles), a
test program and an emission inventory was conducted. The program involved
the characterization of all organic emissions in the Basin of which oil re-
fineries accounted for a little over 10%. Approximately three weeks of test-
ing with a crew of four was available in the program budget. Therefore, the
objectives of the tests were to check selected emission factors published in
the EPA document AP-42 and to perform GC/MS analyses on emission samples
to develop emission profiles (i.e., chemical composition in weight percent).
One week was spent in a small (40,000 bbl/day) asphalt refinery and two weeks
were spent in a large (180,000 bbl/day) refinery. This paper covers the tests
of fugitive emissions from 5,800 valves, 12,000 flanges, 115 pumps, 5 com-
pressors, 3 cooling towers and 3 oil/water separation pools.
VALVES AND FLANGES
Test Method
Fugitive emissions from valves and flanges were measured by spraying
the components with soap solution and characterizing the leak rates by the
rate of bubble formation (Figure 1). Each component found to be leaking was
rated and tagged designating it as a small, medium or large leaker. By measur-
ing the leak rates for a number of small, medium and large leakers, a charac-
teristic leak rate was determined for each of these leak sizes.
The technique for measuring large leak rates was simply to enclose
the component in a polybag and let the gas escape from the bag through a meter
-------
Pi
O*
K)
T
NT
s~\
Soap
Bubble
Leak
Detector
Pump Can
Valve, Flange,Pump Seal,Compressor Seal
'— Bubbles indicating source of leak
Figure 1. Soap bubble detection and temperature evaluation of hydrocarbon fittings.
-------
H. Taback
as shown in Figure 2. For medium and small leaks a dilution method shown in
Figure 3 was used. A pump drew a steady stream of filtered air through the
polybag and meter. The sample stream was continuously sampled. The test
began when the hydrocarbon analyzer indicated a steady state concentration.
The leak rate was then determined as the product of the air flow times the
hydrocarbon concentration. Samples for GC/MS analysis were taken as shown
in Figures 2 and 3. Table 1 is a typical analysis report generated from the
GC/MS analysis.
Results
A breakdown of valve and flange test results is presented in Table 2.
The number of flanges tested was approximately twice the number of valves
since most installations of valves in pipelines involve two flanges. In spray-
ing fittings with soap solution all of the accessible valves and flanges in
a process area were inventoried and tested. The valve type breakdown shown
was based on the testing at the large refinery while those listed as un-
classified were measured at the small refinery. The concept of recording valve
type did not begin until after the small refinery tests. Instead, the emphasis
was to assess emissions by process unit as presented below.
A surprising result in Table 2 was the proportionately larger number
of leaks of all types found in the plug valves. Plug valves were believed to
represent "improved technology" over gate valves. However, plug valves re-
quire periodic lubrication to prevent leaks. In nearly every case of a leak-
ing plug valve, the leak could be stopped by application of sealing grease.
The fact that these leaks were found was an indication that the refineries
were in a normal maintenance condition when the tests were conducted. Another
probable explanation for the greater leakage from plug valves is that they arc
used primarily for gas service which is a more difficult application from a
leakage standpoint.
A breakdown of valve and flange emissions by pipeline size and fluid
content at the large refinery is presented in Table 3. Ethane and propane
lines contained gaseous product while all of the other products were liquids.
Table 4 presents a breakdown of the emissions from the small refinery
by processing units. For each component at each unit the number of components
-------
cr
to
4 mil Polyethylene Bag
T °F
Seal
("snoop" tested)
Valve
Pump
Compressor
Open Top
Surge Bottle
•"f
\"
O O
C amr-k 1 £*
Total Hydrocarbon Bottle
Analvzer
«•
Figure 2. Leak rate and concentration measurement of ambient temperature
fittings. High leak rates.
-------
U1
4 mil Polyethylene Bag
T °F
Background Filter
Activated
Charcoal
Small Pump
(DC/Manual)
Screw
Clamp
Valve
Pump
Compressor
Water
Bubbler
Pressure
Control
Seal
("snoop" tested)
Total Hydrocarbon
Analyzer Tedlarj
Bag or I
Bottle'
X
H
Figure 3. Leak rate by dilution sweep and sampling of ambient hydrocarbon fitting.
Low leak rates.
-------
H. Taback
Table 1. TYPICAL GC/MS ANALYSIS REPORT
SAMPLE NUMBER:
BOTTLE 10308 A (DILUTED)*
ARB
CODE
1
1
3
2
3
2
2
3
EPA
NUMBER
43201
43202
43203
43204
43205
43212
43214
43215
CHEMICAL NAME
METHANE
ETHANE
ETHYLENE
PROPANE
PROPYLENE
N-BUTANE
ISO-BUTANE
ISO-BUTYLENE
MOL.
WT
UG/L
%WT
PPM
%VOL
16
30
28
44
42
58
58
56
34300.
157000.
81700.
231000.
968000.
4810.
11300.
7260.
2.3
10.5
5.5
15.5
64.8
0.3
0.8
0.5
52200.
127000.
71000.
128000.
561000.
2020.
4720.
3160.
5.5
13.4
7.5
13.5
59.1
0.2
0.5
0.3
TOTALS 1490000. 100.
949000. 100.
TOTAL PPM FROM GC AS HEXANE
TOTAL PPM FROM TOC AS HEXANE
423000.
455000.
2 COMPOUNDS OF ARB CLASS I
3 COMPOUNDS OF ARB CLASS II
3 COMPOUNDS OF ARB CLASS III
# DENOTES COMPOS WITH EPA NO. ASSIGNED BY ARLI
( ) DENOTES NO. OF ISOMERS, IF MORE THAN ONE
* RESULTS COMPUTED TO REFLECT THE TRUE CONCENTRATION IN THE ORIGINAL SAMPLE.
-------
H. Taback
Table 2. KEFINEKY EMISSION SUMMARY, LEAKING VALVES BY VALVE TXFE
Valve Type
Plug
Gate
Control
Unclassified
Total
Flanges
Number
Tested
1300
3100
75
1300
5800
12000
Leaks
Measured
15
5
2
3
25
0
Leakers Identified
Small
76
47
9
25
157
38
Medium
21
6
0
35
62
20
Large
24
4
3
2
33
7
-------
H. Taback
Table 3. REFINERY EMISSION SUMMARY (LARGE REFINERY)
VALVE AND FLANGE LEAKS BY SIZE AND FLUID SERVICE
Propane
Small
Medium
Large
Light Gasoline
Small
Medium
Large
Gasoline
Small
Medium
Large
Naphtha
Small
Medium
Large
Gas Oil
Small
Medium
Large
Fuel Oil
Small
Medium
Large
Crude
Small
Medium
Large
Residual Oil
Small
Medium
Large
Ethane
Small
Medium
Large
Preon
Small
Medium
Large
Sour Hater
Small
Medium
Large
Valves less
than 2 in.
928
56
10
8
137
0
0
0
538
5
1
1
56
3
0
0
227
0
0
0
327
4
2
0
96
0
0
0
62
0
0
0
52
1
1
0
37
0
0
0
47
0
0
0
Valves 2 in.
and Greater
596
39
12
16
88
0
0
1
358
13
0
0
60
1
0
0
352
1
0
0
220
1
0
0
126
4
1
0
29
0
0
0
56
4
0
5
30
0
0
0
50
0
0
0
Fittings &
Flanges
Less than
2 in.
1180
13
0
0
146
0
0
0
551
1
0
1
230
0
0
0
4
0
0
0
765
0
0
0
367
0
0
0
70
0
0
0
73
1
1
0
37
0
0
0
—
0
0
0
Fittings &
Flanges
2 in. and
Greater
1583
3
0
0
249
0
0
0
1007
0
0
0
176
0
0
0
1004
1
0
0
655
0
0
0
357
0
0
0
80
0
0
0
152
0
0
0
75
0
0
0
~
0
0
0
-------
H. Taback
Table 4. SMALL REFINERY EMISSIONS, VALVE, FLANGE AND PUMP INVENTORY
No. Tested
(% of Total)
Valves
Large leaks
Medium leaks
Small leaks
Flanges
Large leaks
Medium leaks
Small leaks
Pumps
Large leaks
Medium leaks
Small leaks
Reformer Unit
500 (100%)
0
26
13
852 (70%)
0
13
7
12 (100%)
0
1
1
Naphtha Unit
318 (100%)
0
0
5
889 (70%)
0
0
0
7 (100%)
0
0
0
Crude Unit
475 (80%)
2
9
7
1,319 (80%)
0
7
11
30 (100%)
1
0
2
tested is indicated along with the percentage of those components that were
tested in that unit. For example, on the crude unit 80% of the valves were
tested. The 20% of the valves not tested were not readily accessible without
special apparatus.
Table 5 summarizes the leak rate measurements and calibration of
visual leak rating. The leak rates were measured as described above. The
"large," "medium," "small" designations were assigned in the field prior to
measuring leak rate. Thus a "large" gas line peak ranged from 7 to 38 Ib/day
with an average of 18 Ib/day.
-------
H. Taback
Table 5. REFINERY EMISSION SUMMARY, VALVE LEAK RATE MEASUREMENTS
Location
Code
Fluid
Leak Rate
Ib/day
Average Leak Rate
(calibration of
visual leak rating)
Large Leakers
L
L
L
L
L
L
L
S
S
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Gas
Liquid
25
7
19
13
14
15
38
22
12
tmmmtm
18
Large Gas Valves
18 Ib/day
Large liquid valves
18 Ib/day
Medium Leakers
L
L
L
L
L
L
Gas
Gas
Gas
Liquid
Liquid
Liquid
2
6
2
2
3
5
Medium Gas Valves
3 Ib/day
Medium Liquid Valves
3 Ib/day
Small Leakers
L
L
S
L
S
L
L
L
L
Gas
Gas
Gas
Gas
Gas
Liquid
Liquid
Liquid
Liquid
0.02
0.02
0.02
0.0002
Small Gas Valves
0.3 Ib/day
Small Liquid Valves
0.02 Ib/day
10
-------
H. Taback
The computation of emission factors for refinery valves is presented
in Table 6. For valves in gas service the leak rate per valve is 0.4 Ib/day
and for liquid service is 0.003 Ib/day. To compare this to AP-42 which makes
no distinction between gas and liquid service refer to Table 6. The total
emissions for gas and liquid service is 620 + 49 = 669 Ib/day divided by the
total valves (1700 + 2800 » 4500, 669/4500) equal 0.15 Ib/day/valves which is
identical to the value given in AP-42.
PUMP SEALS
The pump seals were tested by first sniffing for leaks using a Bacha-
rach TLV total hydrocarbon analyzer then tenting the seal using the measure-
ment technique shown in either Figures 2 or 3.
Refinery pump test results are shown in Tables 7 and 8. While Table
7 indicates that mechanical seals and packed seals have the same proportionate
number of leakers (approximately 50%). Table 8 shows that mechanical seals
have generally lower leak rates and especially for gas service. Table 6 pre-
sents the emission factors. The mechanical and packed seals show nearly the
same emission factor for liquid service (< 26 RVP) while for gas service
the mechanical seal emission factor was one-sixth that for the packed seal.
The leak rate data for the small refinery (Table 8) agree with those for the
large refinery amazingly well. During testing the type seals and fluid RVP
were not recorded at the small refinery. Later it was established that all
their pump seals were mechanical and generally the fluid RVP was below 26
psi. These assumptions were made in developing emission factors.
Based on the above data an overall pump emission factor was calculated
(refer to Table 6) . Total emissions = 25 + 140 + 5 + 170 = 340 Ib/day divided
by seal tested (93 + 19 + 12 + 4 = 128), 340/128 = 3 Ib/day/seal or 0.5 tons/
year/seal. This agrees with the AP-42 value of 3 to 5 Ib/day/seal.
11
-------
Table 6. KVB FUGITIVE EMISSION DATA FOR PETROLEUM REFINING OPERATIONS*
Device Type
Total
Inventory
Product NO.
No. of Leakers Identified
(Average Leak Rate,
Ib/day-leak)**
Large Medium Small
Total
Emissions
(Ib/day)
Emission
Factor
(Ib/day device)
Valves s Fittings
Valves Gas 1700 29(18) 23(3) 100(0.3) 620 0.4
Flanges Gas 3100 0 1(3) 17(0.3) 8 0.003
Valves Liquid 2800 2(18) 4(3) 32(0.02) 49 0.02
Flanges Liquid 5700 1(18) 0 2(0.02) 18 0.003
H
8-
p>
Pump Seals
Mechanical
Mechanical
Packed
Packed
< 26 RVP
> 26 RVP
< 26 RVP
> 26 RVP
93
19
12
4
3(7)
2(70)
1(4)
1(170)
* Based on test data from Source #7 only.
**See Table 5 for determination of average leak rate.
4(1)
0
0
0
17(0.01)
8(0.06)
5(0.06)
0
25
140
5
170
0.3
7
0.4
40
-------
Table 7. REFINERY EMISSIONS, PUMP SEALS
U)
t-3
(U
Reid
Seal Type
Mechanical
(Large Refinery)
Total
Packed
(Lajrge Refinery)
Total
Unclassified
Vapor Pressure
psi
> 26
< 26
> 26
< 26
No. Tested
19
44
63
4
12
16
49
Small
8
14
22
0
5
5
3
Leakers Identified
Medium
0
3
3
0
0
0
1
(B
O
Large
2
2
4
1
1
2
1
(Small Refinery)
-------
H. Taback
Table 8. REFINERY EMISSION SUMMARY, PUMP SEAL LEAK RATE MEASUREMENTS
Reid Vapor
Pressure
psi
LARGE REFINERY
Large Leakers
Medium Leakers
Small Leakers
SMALL REFINERY
Large Leakers
< 26
< 26
< 26
> 26
> 26
> 26
< 26
< 26
< 26
< 26
< 26
< 26
< 26
< 26
< 26
< 26
< 26
< 26
> 26
> 26
< 26
< 26
< 26
< 26
Seal Type
Mechanical
Mechanical
Packed
Mechanical
Mechanical
Packed
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Packed
Packed
Packed
Packed
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Mechanical
Average Leak
Rate , Ib/day
Measured (Calibrations
Leak Rate of Visual Leak
Ib/day Rating)
10 >
4
4 4
5 } __
•* 70
130
170 170
2
1
0.002
0.05
0.01
0.002
0.002
0.005
0.2
0.05
0.0005
0.0005
^
0.01
OOfi
• W \J
°'°7 > 0 04
0.007
24
5
2
2
7
Medium Leakers
< 26
Mechanical
14
-------
H. Taback
COMPRESSOR SEALS
Only five refinery compressors were located and tested. One tiny leak
of 0.0003 Ib/day was located. However, in some related tests conducted in oil
production fields an emission factor of 28 Ib/day/seal was determined .
OIL/WATER SEPARATOR
Open separators were found at both refineries visited. The largest
separator found is shown in Figure 4. Samples were taken from pools in each
refinery. The sampled oil ..was taken to the KVB laboratory where the oil was
separated from the water and the oil was placed in a dish for evaporation
tests at the recorded pool temperatures. The measured evaporation rates were:
2
Refinery Evaporation Rate (Ib/hr'ft )
Large 0.01
Small 0.0002
Small 0.004
The large separator in Figure 4 had a surface area of 14,000 ft . At the
2
rate of 0.01 Ib/hr-ft the emissions from that separator would be 140 Ib/hr.
The operators of the refinery estimated that the water flow through the sepa-
rator was 3000 gal/min. or 180,000 gal/hr- At these rates the emission factor
for the separator would be 140/180 or 1 lb/1000 gal waste water. If the KVB
measurement was correct, the emission from that separator would be 600 tons/
year. The local air pollution control agency rated the emissions from this
separator at 1.5 ton/year. Clearly, more work is required in this area.
COOLING TOWERS
Of the three cooling towers tested, valid data were obtained on only
one. The unit tested is illustrated in Figure 5. It was a large tower
15
-------
H. Taback
Figure 4. Oil/water separator tested.
16
-------
Water 120 °F Water
Cascades Cascades
\ fim f
Reservoir
95 °F
FORCED DRAFT TOWER
42,600 gal/min
(per refinery data)
1460 U gm/liter (lab analyses)
Sampling
Points
I
42 \i gm/liter
(lab analyses)
PROCESS COOLING
Figure 5. Forced-draft cooling tower schematic.
-------
H. Taback
serving an FCC unit, the gas plant for that FCC and a reformer unit. The
water circulation was 42,500 gal/min. Cooling water circulated through the
various processes and returned to the tower where the water was evaporatively
cooled by forced air circulation.
Water samples were taken at the inlet and outlet of the tower as
shown and analyzed (xylene extraction and GC analysis) on organic content.
The organic content was identified as 100% isopentane with concentrations
indicated on the schematic. The emissions were determined to be the dif-
ference in organic concentration times the flow rate as follows:
Change in concentration = 1460 - 42 = 1418 y gin/liter
„ . . „ . IAIO y gm 3.785 liter gm Ib 42,600 gal
Emission Rate = 1418 frf — x - - - x r-rr2 - x ~r=-. - x — — - - * — =
liter gal 10b y gm 454 gm min
= 7QO _
.
min hr hr day day yr 2000
= 130^2.
yr
To relate this to AP-42 the emissions must be put into lb/10 gal.
Other emission factors are in lb/day/1000 GPM
0.5 Ib/min x min/42,600 gal x 10& = 12 lb/106 gal
700 Ib/day * 42,600 GPM/1000 = 16 Ib/day/lQ3 GPM
A comparison of these results with results calculated from published emis-
sion factors is shown in Table 9. EPA publication AP-42 lists an emission
factor of 6 lb/10 gal of cooling water. The API document, referenced on the
table, discusses the emissions and indicates that a 1957 study (probably
the Los Angeles joint project) specified an emission factor of 3 to 5.3
lb/day/1000 GPM while a "more realistic average figure used by some refineries
is 8 to 10 lb/day/1000 GPM." From these emissions factors an emission rate
was calculated based on the 43,0001 GPM water circulation rate. These emis-
sions are shown in Table 9 compared to the KVB measured emissions from -which
emission factors were calculated as indicated in the table.
18
-------
Table 9. COOLING TOWER EMISSIONS AND EMISSION FACTORS
43,000 GPM Water Flow
Data
Source
AP-42
API 931*
1957 Study
Best Estimate
KVB Tests
Note : Underlined
Emission
Ib/day
400
200
400
700
figures were given and
Rates Emission Factors
ton/yr lb/10 gal lb/day/1000 GPM
70 £
40 3-5
70 8-10
130 12 17
other figures were calculated.
*API Publication 931, Manual on Disposal of Refinery Wastes, Volume on Atmospheric
Emissions, Chap. 7, Pages 7 - 12, Hydrocarbon Emissions, API Refining Dept.,
Washington, D.C., February 1976.
cr
to
n
7?
-------
H. Taback
The higher emission factor determined by the KVE test can be explained
by the fact that the hydrocarbon emitted by the cooling tower test was iso-
pentane, a fairly volatile material. Since the AP-42 and API "realistic"
estimate agree, KVB feels that there are valid test results to confirm the
AP-42 emission factor.
EMISSION PROFILES
In developing the emission profiles for refineries, the following facts
and assumptions were used:
1. The typical refinery process schematic is as shown in
Figure 6.
2. The primary leaks were from fuel gas lines. (We esti-
mate that approximately 25% of fuel used in the refinery
was commercial natural gas and the balance was re-
finery fuel).
3. Losses from lines carrying heavy liquid were negligible.
4. Approximately 50% of the crude oil processed in a re-
finery ends up as blended gasoline and 30% becomes
other distillate products.
Valves
As discussed above, the leak rate for valves in gaseous products was
0.4 Ib/valve-day whereas the comparable leak rate for valves in liquid service
was 0.2 Ib/valve-day. Estimates of the total valves in gaseous and liquid
service in refineries located in the Basin were 25% and 75% respectively. By
applying these percentages to the associated leak rates, it was determined
that approximately 85% of leaks from valves result from those handling gaseous
products.
Using the above information, a "typical" fugitive emission profile
for valves was generated. Table 10 gives a summary of the calculation pro-
cedure. Listed are the analyzed emission profiles for leaks in various pro-
duct streams along with the estimated contribution from each to the total.
For the purpose of this analysis it was assumed that the percentage of valves
in liquid service are as follows:
20
-------
Refinery Fuel
Separated
Crude Oil
from Pro-
duction
t
Gas
Crude
Fraction-
ation
i
Straight
Run
Gasoline
1
Naptha^
— . L
Re for
i
Gas
' 4
Reformed
Stock
_.
Gas-Oil ^
I
\
Gas
s
^
Cracking
Unit
J
Gas j *" Gas (4%)
PlantsJ ^- LPG (3»)
, .. _ Blended
1»- uicnaing m> Gasollne
Cracked
Gasoline
s Distillates
' *~ (30»)
to
er
fa
o
7?
Bottoms
Bottoms
I
Heavy
*" Products
Figure 6. Refining process schematic.
-------
10
Table 10. COMPOSITE PROFILE FOR REFINERY FUGITIVE EMISSIONS FROM VALVES
Refinery Gas
Fraction of Enissions 0.64
Organic Compound!
(wt. percent)
Methane 16.4
Ethane 5.1
Propane 15.9
Propylene
H- Butane 26.8
I-Butane 11.3
Butcne
N-Pontane 9.4
I-Pentane 10.7
H-lloxane 2.8
I-Kexane 1.3
N-lleptan« 0.1
I-Heptane 0.2
N-Octane
I-Octane
N-Nonane
I-llonane
M-Decane
I-Decane
C-7 Cycloparaf fins
C-8 Cycloparaf fins
C-9 Cycloparaf fins
Toluene
Xylene
Benzene
Cyclohexane
Straight Run
Natural Gas Gasoline
0.21 0.015
84.5
11.0
3.6 0.2
0.4
0.5 0.7
0.1
1.3
3.2
1.1
1.4
3.4
58.1
4.7
2.8
8.5
2.6
1.0
0.2
2.6
3.2
2.6
1.4
0.9
Naphtha
0.015
0.6
0.6
1.3
O.4
6.2
6.6
1.3
6.4
5.6
8.3
20.6
16.6
2.4
0.6
4.8
6.2
7.0
2.5
1.5
___ . i. i . in,- • -— — — •*
Reformats Gas-Oil
Stock Stock
0.015 0.03
10.0
0.9 3.8
13.9 3.3
24.7 6.5
21.5 7.4
3.8
19.3 11.8
6.0
10.0 8.3
16.0
9.7 7.9
4.4
3.5
7.3
Cracked
Gasoline
0.03
1.1
1.1
19.7
21.4
15.4
12.8
14.9
4.1
0.2
3.9
0.3
3.0
1.6
0.5
Distillate
O.C45
4.2
1.0
5.3
12.7
2.5
11.0
5.0
9.0
3.0
8.6
2.0
9.9
3.0
6.5
0.5
14.8(1)
1.0
H
£
P>
. o
JT
Composite
28.6
5.8
11.5
0.1
18.3
7.4
—
7.7
7.9
3.4
1.6
1.4
0.8
1.8
0.4
0.6
0.5
0.8
0.3
0.2
—
0.1
0.5
0.2
0.1
0.1
100
100
100
1OO
100
100
100
1OO
1OO
'" <
-------
H. Taback
Straight Run Gasoline 10%
Naphtha 10%
Reformate Stock 10%
Gas-Oil Stock 20%
Cracked Gasoline 20%
Distillate 25%
For this analysis, the distribution was not critical since the leaks from
valves in liquid service constitute only 15% of the total emissions.
Because 85% of the emissions are from valves in gas service, the
composition of the natural gas and refinery fuel gas had a very significant
influence on the composite emission profile. A much more detailed analysis
would have also incorporated leaks from other gaseous lines within the re-
finery; however, data on these internal gaseous stock transfer were not
available.
Tests were also conducted on 80 pumps to characterize emission rates
from these fugitive sources. For the purposes of this analysis it has been
assumed that leaks from pumps occur only from the liquid product lines as
previously described. A summary of the calculation procedure is given in
Table 11. As with valves, a much more complex analysis was possible; however,
existing data are only sufficient to make a cursory estimate.
CONCLUSIONS
The conclusions reached as the result of these tests were:
1. The emission factors for valves, flanges and pumps in
AP-42 are reasonable.
2. In valves and flanges the large leakers (averaging
18 Ib/day/valve) accounted for over 80% of the fugi-
tive emissions. Medium leaks (averaging 3 Ibs/day/
valve) accounted for over 10%. Together they ac-
counted for approximately 95%.
3. Refinery fugitive emissions are primarily paraffins
with low photochemical reactivity.
23
-------
TaJale 11. COMPOSITE PROFILE FOR REFINERY FUGITIVE EMISSIONS FROM POMPS
to
Straight Run Reforraate
Gasoline Kaphtha Stock
Fraction of Emissions
Organic Compounds
(wt. percent)
Methane-
Ethane
Propane
N-Butane
X-Butane
Butene
N-Pent«ne
I-Pentajie
N-Hexane
I-Hexane
N-Heptane
X-Kcptane
N -Octane
I-Octano
N-Nonane
Z-!«onane
N-Decane
I-Decane
C-7 Cycloparaffins
C-B Cycloparaffins
C-9 Cycloparaffins
Toluene
Xylene
Benzene
Cyclohexane
0.10
0.2
0.7
0.1
1.3
3.2
1.1
1.4
3.4
58.1
4.7
2.8
8.5
2.6
1.0
0.2
2.6
3.2
2.6
1.4
0.9
100
0.10 0.10
0.9
0.6 13.9
0.6 24.7
1.3 21.5
0.4
6.2 19.3
6.6
1.8 10.0
6.4 9.7
5.6
8.3
20.6
16.6
2.4
0.6
4.8
6.2
7.0
2.5
1.5
100 100
Gas-Oil
Stock
0.20
10.0
3.8
3.3
6.5
7.4
3.8
11.8
6.0
8.3
16.0
7.9
4.4
3.5
7.3
100
Cracked
Gasoline
0.20
1.1
1.1
19.7
21.4
15.4
12.8
14.9
4.1
0.2
3.9
0.3
3.0
1.6
0.5
100
Distillate
0.30
4.2
1.0
5.3
12.7
2.5
11.0
5.0
9.0
3.0
8.6
2.0
9.9
3.0
6.5
0.5
14.8(1)
1.0
100
Composite _
3.3
1.2
3.7
7.9
0.8
0.2
11.1
6.6
11.0
5.5
8.5
4.1
12.0
2.8
3.9
3.1
5.1
1.9
1.1
0.1
0.8
3.0
1.3
0.5
0.5
100
fu
cr
(U
n
111
-------
H. Taback
4. More work is required to develop experimental techniques
for increasing evaporative emissions from oil/water
separators, cooling towers, etc. and to derive emission
factors from these sources.
ACKNOWLEDGEMENTS
This work was sponsored by the California State Air Resources Board;
Dr. John R. Holmes, Research Division Chief and Mr. Jack .Paskind, Program
Manager. The refinery fugitive emission data reported in this paper were
the efforts of the KVB, Inc. project team: H. J. Taback, Program Manager,
T. W. Sonnichsen, Project Engineer, N. Brunetz, Test Director, A. R. Brienza
and J. F. Macko, Test Engineers, and G. L. Anderson, Technician. GC/MS
analyses were performed by Analytical Research Labs, Inc. of Monrovia, California.
REFERENCES
1. "Compilation of Air Pollutant Emission Factors," AP-42, Third Ed.
Plus Supplements, EPA, OAQPS, Research Triangle Park, NC.
2. Taback, H. J- et al., "Control of Hydrocarbon Emissions From Sta-
tionary Sources in the California South Coast Air Basin," Final
Report Vols. I and II, KVB, Inc. 1978. (A preliminary draft was
submitted December 1977. The final draft should be released in
July or August 1978.)
3. Sonnichsen, T. W. et al., "Hydrocarbon Emissions from Petroleum
Production Operations in California's South Coast Air Basin,"
paper to be presented at the APCA Meeting, Houston, Texas,
June 25-30, 1978.
25
-------
FUGITIVE HYDROCARBON EMISSIONS -
MEASUREMENT AND DATA ANALYSIS METHODS
Karen M. Hanzevack
Exxon Research and Engineering Company
Florham Park, New Jersey
ABSTRACT
A statistical data analysis method for determining fugitive source emission
factors from field leak rate measurements is presented, along with a discussion
of general measurement techniques. Data on valve emissions obtained in a
measurement program conducted at an Exxon chemical plant are highlighted, both
as an example of the data analysis method, as well as in the discussion of the
results of this study. A comparison is made among valve emission factors from
product line areas receiving varying degrees of maintenance. This comparison
shows the potential for emission reduction through improved maintenance at a
given facility, although further studies would be needed to accurately quantify
the expected reduction for the general case.
RESUME
Karen has a Master's Degree in Chemical Engineering which she received
from the University of Michigan and she is now employed by Exxon Research and
Engineering. She has been there since 1974 working in the general areas of
measurement and control of hydrocarbon emissions from both marine vessels and
inplant sources. Currently she is head of the Hydrocarbon Emissions Control
Group within the Environmental Control and Safety Division of Exxon Research
and Engineering.
-------
Karen M. Hanzevack
FUGITIVE HYDROCARBON EMISSIONS -
MEASUREMENT AND DATA ANALYSIS METHODS
Information on hydrocarbon emissions from fugitive sources has
recently been obtained at an Exxon chemical plant. A measurement program
was conducted which included monitoring emissions from fugitive sources,
including valves, pumps, compressors, and safety valves. Although the
specific purpose of the study was to provide accurate fugitive emissions
estimates for the particular plant under study, the measurement techniques
and data analysis procedures developed as part of the program are of general
interest in evaluating or preparing for other current or future test programs
in this area.
The test program which was conducted can be broken down into three
phases: screening, sampling, and data analysis. The screening phase
included determining the total number of components of each source in the
plant, as well as developing a sampling plan. The sampling involved
"bagging" specific sources identified in the screening phase, and develop-
ing hydrocarbon leak rate data for each source. Finally, a statistical data
analysis method was identified which was judged to result in the best
estimates of emission factors for leaking components for each source type.
These values, combined with data on the percentage of components which leak
and the total number of components in the plant, can be used to estimate
total fugitive emissions.
The screening and sampling phases of this program will be briefly
outlined. The main emphasis in this presentation will be on the methods
used for analysis of the data developed in this study. Specifically, the
work done on valves will be highlighted, both as an example of the data
analysis methods and in the discussion of the results of this study.
SCREENING AND SAMPLING PLAN DEVELOPMENT
The first objective of the screening phase was to determine the
total numbers of plant-wide components for each source type. These values
were tabulated from chemical plant flow sheets for each of the three product
line areas within the plant. The total number of valves was over 20,000,
while the number of other components ranged from approximately 200 to 600.
Based on these numbers, it was decided that random sampling of
approximately 10% of the pumps, compressors, and safety valves could be
expected to result in representative samples without requiring an unreason-
ably large measurement effort. The components to be sampled were randomly
selected from the various process units throughout each of the three product
line areas in the plant. The data obtained in sampling these components
were used both to determine the percentage of each source which leaked and to
determine the actual leak rates of the leaking components.
—1—
-------
Karen M. Hanzevack
However, due to the extremely large number of valves in the plant,
a more detailed sampling program was developed to set the number of samples
at a reasonable level, while still providing a data base which would be an
adequate representation of the entire plant valve population. The approach
taken was to categorize the valves by service (liquid or vapor), size
(£ 4 inches <), and line pressure (_< 50 psi <) such that eight classes of
valves were established.
Approximately 10% of the valves in each product line were screened
with soap solution to -detect leaking components. Specific valves were
selected for screening, such that each class of valves was screened in
approximately the same proportion as it occurred in the general population.
All valves found to be leaking were counted and tagged. The advantage of
this sampling approach is that it is likely to give a more accurate
representation of the total population with fewer samples than by using
simply a random sampling procedure. Using this approach, approximately
1% of the total valve population, about 200 valves, were randomly sampled
from the set of valves found to be leaking.
The information obtained during the valve screening provided an
estimate of the percent of the total population of valves which leak.
Therefore, only leaking valves needed to be sampled and analyzed.
FIELD SAMPLING METHODS
A "bagging" technique was used to measure the leak rates of those
valves which were selected during the screening phase of the study. The
"bagging" technique generally consists of enclosing the emission source in a
mylar-type sheeting. Figure 1 shows the sampling equipment used. Plant
instrument air was directed into the mylar enclosure to serve as a carrier
stream for the leaking hydrocarbon vapor. The air was first passed through
a dessicant and charcoal filter, to remove any water vapor and hydrocarbon
present. The flowrate was measured by a calibrated rotameter and held
constant during the sampling; The flowrate was set to maintain a slight
positive pressure in the enclosure. Both temperature and pressure were
monitored. The air/hydrocarbon stream was sampled from an outlet connection
in the enclosure on the opposite side from the air inlet. The dry gas meter
was used to determine that vapor was flowing through the line into the
sample and to verify that a stable rate out of the enclosure had been
obtained prior to collecting the sample.
This method does not require air samples to be taken along with
each source sample, once the efficiency of the air filter has been estab-
lished. Therefore, this method minimizes the number of sample analyses
required, as compared to a method using ambient air as the carrier
stream. Also, this method is not particularly sensitive to leaks out of
the enclosure, since the actual amount of air diluting the hydrocarbon
vapor is measured upstream of any possible enclosure leaks.
-2-
-------
Karen M. Hanzevack
The hydrocarbon/air samples were analyzed by gas chromatography
techniques, giving total hydrocarbon concentrations, as well as a breakdown
by hydrocarbon component. From these data, total hydrocarbon molecular-
weights were determined for each sample, in order to accurately determine^
mass leak rates. The temperature readings were used to accurately determine
flowrate. Replicate samples were generally taken and their results averaged
to get a leak rate for each component sampled. A computer program was
developed to carry out these calculations, resulting in a leak rate, in
oounds of hydrocarbon per hour, for each component sampled.
This measurement method and calculation procedure yielded a set
of hydrocarbon leak rate data for each source. A method for statistically
analyzing these data was then defined, which would allow the leak rate data
from the set of components sampled to be generalized to a representative
emission factor, with confidence limits, for the entire plant-wide population
of leaking components of a given source type.
DATA ANALYSIS
General Data Analysis Technique
By virtue of the methods used for sample selection, it was
concluded that the samples taken were in fact random samples; that is,
every component had an equal chance of being sampled. Therefore, the
samples can be assumed to be representative of the entire population of
components in the plant. This assumption is a necessary condition for the
statistical analysis used to develop estimates of the average emission
rates for the entire population of leaking components based on the leak
rates of the components which were sampled.
The first step in the data analysis is to determine the frequency
distribution of the data for each of the four general source types sampled.
This is done by finding the maximum and minimum leak rate values and
dividing this range into a convenient number of equal cells. The number
of data points which fall into each cell is then plotted against the
mid-point value of each cell to obtain a frequency distribution of the
data set. Such a frequency distribution is shown in Figure 2, using the
valve data as an example.
This frequency distribution provides information about the general
characteristics of the data set. It can be seen from Figure 2 that most of
the valves measured have small leak rates and that very few valves account
for the high measured rates. A similar distribution was obtained for each
of the other source types sampled. These data clearly do not lie in a
common bell-shaped curve which is characteristic of what is known as a
normal frequency distribution.
—3—
-------
Karen M. Hanzevack
Only a relatively small amount of development work has been
carried out on statistical analysis of sampling data from non-normal
distributions. The widely accepted approach to analyzing data from such a
distribution is to seek a transformation of the data set which will result
in a new data set which is normally distributed. This is done because
there is a large body of well-proven and widely accepted theory related
to the analysis of normal distributions. This theory allows for statis-
tics developed from a sample to be projected to the parent population. An
inverse transformation is then carried out to convert the statistics
developed for the transformed (i.e., normal) distribution back to the form
of the original data.
A common transformation to consider in cases such as this one,
where the data are significantly skewed, is to take the natural logarithm of
each data point. A new data set is then obtained made up of the logarithms
of the original data set. Figure 3 shows the frequency distribution for
the data set of logarithms of the original valve data. By observation it
can be seen that this transformed data set is approximately a normal
distribution. Tests for normality exist, and therefore the validity of
the transformation can be checked, and, if need be, correction factors can
be applied to account for deviations from normality.
Standard statistical tests for normality were made, which
included calculating the skewness, or relative spread, and the kurtosis,
or relative height, of the distribution. From these tests it was deter-
mined that the log transformation had resulted in a distribution which
could be assumed to be normal for the purpose of statistical analysis.
This was the case not only for the valve data, but for all source types
sampled. No corrections to account for deviations from normality were
necessary. Therefore, the validity of this analysis technique was verified
for these sets of fugitive emissions data.
By working with the data sets obtained by taking-the natural
logarithm of the original data for each source type, the population mean
values and confidence limits can be determined. This method is judged to
yield a better estimate of the population mean than that obtained by
simply taking the arithmetic average of the original data. This is the
case because, by working with a normal distribution, more information
about the data can be incorporated into the calculation of the population
mean value.
Calculation of Results
Calculations made were based on the general statistical data
analysis techniques discussed above. The computer program was extended to
perform the statistical analysis on the set of leak rate values calculated
from the field data. Table 1 shows the steps in the statistical treatment
of the data. Natural logarithms of the data sets were taken. Using these
values, the mean, standard deviation, skewness, and kurtosis were calculated
for each source type. The skewness and kurtosis values provided a check on
-4-
-------
Karen M. Hanzevack
the assumption of a normal distribution. The mean and standard deviation
obtained for the data set of logarithms were then transformed back into the
form of the original data set. This transformed mean value is the best
estimator for the population mean value. It is this number which is the
emission factor for leaking components.
Confidence limits for the population mean were calculated at the
3a , or 99.75% confidence level. This means that for repeated random
samples of the same size, the calculated mean values of the samples would
lie within the given confidence limits 99-75% of the time. The 3a level
represents a higher degree of conservatism than the 2a (95%) confidence
level normally used in engineering applications. This higher degree of
conservatism was judged to be appropriate for this study mainly because of
the extremely skewed leak rate distributions and the general lack of
experience with this type of data.
Discussion of Results
The valve results are summarized in Table 2. The leaking source
emission factors shown in this table represent the best estimates of leaking
valve emissions within the stated confidence limits at the specific chemical
plant involved in this study. In combining the valve data from each product
line, a factor of 0.031 Ib/hr per valve, with a range of 0.017 to 0.057
Ib/hr, was obtained. Furthermore, 15% of the total population of valves
were found to leak. Therefore, the plant-wide valve emission, factor is
0.0047 Ib/hr. This is equivalent to 0.11 Ib/day per valve, which can be
compared to the valve emission factor from AP-42 of 0.15 Ib/day per valve.
The valve data were also analyzed separately by product line,
in order to observe any systematic differences which might be present. As
shown in Table 2, the product line results show significant variability.
A preliminary analysis of the data suggests that the level of maintenance
in each product line is the main variable that could account for these
observed differences. Being at the same facility, factors such as ambient
conditions and plant age are not significant variables. Also, the screen-
ing data showed that the valve categories of service, line pressure, and
size are represented in approximately the same proportion in each product
line area, although the products being handled do vary.
Maintenance programs are often defined within a given product line
area and, therefore, significantly different levels of maintenance can occur
from one area to another. It has been determined that the three product line
areas, at this facility have three different levels of maintenance. One area
performs what can be referred to as routine maintenance for this facility,
another has a somewhat improved maintenance program in effect, while the
third area has been conducting a special maintenance program to minimize
the loss of high-valued products. The differences in valve emission factors
for each area are significant. The emission factor for the area with the
best maintenance practices is approximately one tenth of the factor from
the area where routine maintenance is performed. The area with an inter-
mediate level of maintenance shows a reduction about half as great.
-5-
-------
Karen M. Hanzevack
Not surprisingly, these results suggest that fugitive emission
factors are a direct function of maintenance. However, it is important
to emphasize that the quantitative results discussed above only apply
specifically to the facility under study. Additional studies at other
facilities would clearly be needed to determine the relative impact of
maintenance on emissions, as compared to other influencing elements which
could vary from one facility to another. Without additional studies,
the generalization of these results to other chemical plants or refineries
would be inappropriate.
CONCLUSIONS
The results of this study suggest that there are two main
areas beyond actual leak rate measurement methods which need to be
thoroughly addressed in any current or future fugitive emissions studies.
Data analysis techniques should be used which provide the best estimate
of total population emission factors for a given source type. Secondly,
information on maintenance procedures should be developed along with
other correlating parameters in order to address the question of the
effect of maintenance on fugitive emission factors.
The data analysis technique identified in this study as pro-
viding the best estimate of total population emission factors is based
on transforming the leak rate data into a normal frequency distribution.
A transformation found to be valid for the data obtained in this study
is a logarithmic transformation. Such a method is recommended over the
use of a direct arithmetic average calculation because it provides a
more representative factor by making use of more of the information
obtainable from the data. Also, it allows for a straightforward calcula-
tion of the confidence levels associated with the emission factors.
Furthermore, it is important that a standard analysis method be adopted,
to avoid the generation and use of different factors based on the same
data set.
Developing information on the levels of maintenance associated
with the specific fugitive emissions data should be included in fugitive
emission measurement programs. Maintenance levels should be defined
so as to permit emission levels to be correlated to maintenance, as well
as to other influencing parameters. This is important not only in
defining emissions, but in the consideration of reasonable means to
reduce these emissions.
-------
Karen M. Hanzevack
Table 1 STATISTICAL TREATMENT OF DATA
Transformation of Data:
x = component leak rate
y - In (x)
Statistics of Transformed Data Set:
Xy - Arithmetic mean of y values
ay = standard deviation of y values
Statistics for Original Data Set:
X = exp ( Xy + 1/2 ay2) = population mean
cr = ®y / *fn » standard deviation of the mean,
where n = sample size
3 a confidence limits = exp [(1L + 3
-------
Karen M. Hanzevack
Table 2 SUMMARY OF VALVE EMISSION FACTORS
Leaking Source Emission Plant Product
Factor (lb/hr/valve) Percentage Line Emission
3a Confidence of Leaking Factor
Mean Value Limits Components (lb/hr/valve)
Source
(Sample Size)
Total Plant Valves 0.031 0.017-0.057 15
(197)
Product Lines:
A-Routine Maintenance 0.032 0.014-0.076 17
(100)
B-Iniproved Maintenance 0.019 0.008-0.047 16
(49)
C-Special Maintenance 0.008 0.002-0.026 8
Program (48)
0.0047
0.0054
0.0030
0.0006
-8-
-------
Instrument
Air
-HXh
Dessicant
Activated
Carbon
Filter For
Carbon Fines
Pressure And
Temperature
Indicators
H
0
s
•
EC
N
P>
O
' XI
Filter Flow/meter
Mylar Enclosure
Over Equipment
Gas Meter
Sample Bomb
Figure 1 - Fugitive Emissions Sampling Train
-------
Karen M. Hanzevack
200-
180-
160-
120-
60-r
404
20-
.172 .400 .629 .858 1.09
LEAK RATE, lij/hr/val'velCEuTMID-POINTS)
Figure 2 - Frequency Distribution Of Valve Leak Rate Data
-10-
-------
Karen M. Hanzevack
-12.2 -9.3 -6.4 -3.5 -0.6
LOGARITHM OF LEAK RATE, In (Ib/hr/vaTve) '(CELL MID-POINTS)"
Figure 3 - Frequency Distribution Of Logarithms Of Valve Leak Rate Data
-11-
-------
Karen M. Hanzevack
QUESTIONS AND ANSWERS
Q. Richard Vincent, GARB - I was wondering if there were any cost
estimates for the three levels of maintenance that you gave results for, in
terms of emissions and emission reductions.
A. - We have not sought that sort of information based on this study.
Q. Bruce Beyaert, Chevron USA - You looked at the effective differences
among three different product lines. Were there any other differences involved
between the product lines, such as gas versus liquid systems, pressure, line
size and other things that affect them?
A. - The categories of valves by service, size and pressure, are repre-
sented in about the same proportion in each of the product line we looked
at.
Q. Charlie Sunwoo, Tosco Corp. - Can you expand on the levels of the
maintenance? What you meant by routine maintenance versus special program?
A. - To some extent, yes. The special maintenance program that was in
effect consists of inspecting for leaks at a frequency of approximaely once
every three to four months. Routine maintenance, I don't want to go into
specifics on this particular program, because this was a chemical plant and
not a refinery. I don't know that what would be referred to as routine in a
refinery would be exactly analogous. I think the main point is that we need
to develop categories of maintenance on a broad scale, not from a given
facility, but in looking at general industry practices, and to use this
information to attempt to correlate the data that is being gathered.
Q. Rex Smith, Xomox Corp. - I think everything is being attributed to
maintenance. Study ought to be given to proper valve selection because, of
course, this can affect the maintenance factor. If the valve is improperly
applied, it can be a high maintenance item. This entire factor, I think, has
been completely disregarded.
A._ - In this particular study, because the different product lines are
at the same facility, the means used for selecting the type of valves would be
approximately the same from one area to another. So, I don't think that valve
selection is a significant variable in this particular study.
Comment - Rex Smith, Xomox - Maybe not a variable, but it might vary the
maintenance features on the particular valve selected. And, possibly the
selection of valves may or may not have been correct.
A_._ - That is certainly true.
Q. Paul Harrison, MRI - We are talking about existing facilities
and the problem of how to correct something as opposed to installing new
technology. I would like to say that your approach, Karen, is whole-heartedly
endorsed. My paper will talk about the same thing, categorization; choosing
-12-
-------
Karen M. Hanzevack
from a selection of leaking components, as opposed to trying to look at the
whole thing as statistical. The selection process is virtually impossible
and the probability of gaining a good statistical distribution of leakers
is highly improbable. You did it right, and I think that's great.
Q. Richard Vincent, GARB - I was just wondering if you or anyone has
any idea of what the relative importance is of fugitive emissions just from
valves, pumps, etc. from chemical plants and other industrial sources, other
than refineries?
A_._ - I would hope that after we see the results of some of the refinery
studies being done, we could go back and compare these results and get a feel
for it. On the basis of one chemical plant study, however, such information
couldn't be developed.
Q. Richard Vincent, GARB - I guess I am looking for a valve count of
all the valves that are relative to how many of them are in refineries?
A. - That information I wouldn't have.
Q. - What chemical plant was this? What type of chemical plant? You
didn't mention that.
A. - That is right, I didn't name the specific facility.
-13-
-------
DETECTION AND CLASSIFICATION OF FUGITIVE COMPONENT LEAKS
P. R. Harrison
Meteorology Research, Inc.
Al'tadena, California
ABSTRACT
After surveying four separate facilities for component leaks, it has
become apparent that less than six percent of the components account for
more than 90 percent of the fugitive emissions from all components. Due
to the large variance of emission rates, rigorous statistical selection from
a representative arithmetic average is highly improbable. Subclasses must
be secured by total facility surveys. Each class will have its mean where
the geometric mean is much closer to the arithmetic mean than it would be
in the total distribution. For similar throughput refineries, the difference
in number of component leakers is a factor of two for newer versus older
facilities which could be inferred to reflect maintenance activities. The
survey technique seems not to be critical except for classifying the component
as to Class I (worst case) to Class IV (insignificant leak). The close in sample
is preferred and helps to negate geometry and wind effects.
Use of transects can characterize the total fugitive emission rates by
use of meteorology and hydrocarbon analysis. Verification by component
leak rate measurements have occurred, with the results showing that "areas
of emissions" such as spills, sewers, etc. which are related to component
leaks are nearly or equally as important as the components themselves.
However, eliminating the component leaks usually eliminates the "area
emissions" (evaporation).
The data for these conclusions are contained within this paper.
RESUME
Paul_Harrison has been with MRI since 1974 where he.-is a Department Manager.
During this time he has done work in the area of fugitive emissions from petroleum
refining. From 1970-1974, Dr. Harrison worked for the City of Chicago where
he was Director of Technical Services in the Dept. of Environmental Control.
Paul has a Ph.D. in Air Pollution Atmospheric Sciences and a Master's Degree in
Meteorology, a Master's Degree in Computer Information and Control Engineering
and a Bachelor's Degree in Physics which is a very impressive background.
-------
P. R. Harrison
DETECTION AND CLASSIFICATION OF FUGITIVE COMPONENT LEAKS
INTRODUCTION
This paper describes some results of work performed on various types of
hydrocarbon treatment and petrochemical facilities. The methods have been
developed to a nearly routine protocol between the years of 1975-1978.
These detailed protocols are beyond the scope of this report. Most of the
impetus for this work was the need to lower hydrocarbon emissions to acceptable
levels and/or "trade-off" requirements in a nonattainment (ozone) area. In
all cases, the method has been accepted by the agencies involved and the
reduction/construction phase in some cases is pending final permits. Much of
the assessment calculations made to date are dependent on emission factors
based on throughput, or component type, or number. Our studies show that
for identical throughputs the percentage of components leaking and total
fugitive emissions are dependent on age, maintenance, total number of components,
and product type. There is less relationship to throughput and size of component.
For some time there have been known inaccuracies in emission inventories
due to the lack of knowledge of the fugitive part of the calculation. Although
a study was conducted in the 1950s, resulting in component factors such as
0.5 pounds/day for valves on, later studies resulted in 5 to 28 pounds/day/
1000 barrels capacity, depending on component type. (The latter number is
for valve emission rates.) Needless to say, there has been a great deal
of controversy surrounding these values and their use as average or state-
of-the-art controls. The argument would have remained academic if not for
the discovery that most of the country was in violation of the primary ozone
standard which made them Class I or nonattainment areas. This secondary
pollutant is heavily linked with nonmethane hydrocarbons (NMHC). Recent
legislation has also mandated trade-offs for new, improved, or expanded
facilities with penalties over one-to-one trades (add 20% in California).
Thus, the fugitives become an abatement area as well as a valuable asset
for trading.
The multiple tasks remaining were to:
1. Identify the total amount of hydrocarbon emitted
a. Isolate the fugitives
b. Determine the nonmethane fraction
2. Identify the component type and size of leak
a. Correlate with throughput, size, product, etc.
b. Compare the number and size of component leaks
with total number and type of component
-1-
-------
P. R. Harrison
c. Resolve the component fraction with the total fugitives
3. Suggest a control strategy
a. Consider economics, manpower, and technology for
maintenance
b. Number needed to be controlled for trade-off needs
c. Amount of control needed for compliance
d. Formulate institutional changes to assure maintenance
e. Consider maintenance of these desired rates
To date, we have accomplished the techniques to accomplish each sub-
task, except 2a and 3a. These tasks are less technical and more management
related but can easily be assessed given the information available from
the other tasks. Based upon experience to this point, the correlations
are not as important as the others and the economics are greatly favorable.
The main objective of a fugitive emission study is to estimate the
rate at which total hydrocarbons are being emitted from the numerous
fugitive sources located in a refinery prior to the Directed Maintenance
Program. To make this estimation with a reasonable degree of accuracy,
it is necessary to determine to what extent, if any, emissions from elevated
sources (flares, cooling towers, etc.) and process variations contribute
to hydrocarbon concentrations used in calculating the fugitive emission
source strength. These factors are a pacing factor in estimating the
absolute accuracy of the abatement program. The repeatability of the results
of the diffusion calculations, especially under similar meteorological and
emission conditions, is very good. By judiciously taking grab samples for
subsequent chromatographic analysis the nonmethane fraction and instrument
calibration is found.
An alternative procedure would be to measure a large number of individual
components and statistically project the reduction. This procedure is very
time consuming and expensive. Even though the statistical procedure may
seem rigorous, it may not represent reality. The explanation for this contra-
diction is the large skewness of the actual distribution toward the few
largest leaking components. The distribution extends over several decades.
Figure 1 is a semi-experiential estimate of the actual distribution of leaking
components found in a "typical" facility. One can see that a few leakers
over the last five decades on the plot account for as much in total emissions
as those in the first five even though the median is around the third (10
g/hr) (i.e., one component leaking at 100 g/hr accounts for 10,000 at 10 2
g/hour). The probability of randomly selecting these two components in
10,000 plus components is very small. The best solution to the probable
errors inherent in the statistical selection is to categorize the components
into significant, small, medium, and large classes by a complete survey and
select a group randomly from each category to find a probable class median.
-2-
-------
u>
O
bt>
4J
-8
n)
W
.9
10
44
a
u
£
O
a,
6
(3
tt
U3
I
10.000 -
I. 000 -
100
10 -
10
10
10 ' I 10
Emission Rate (gm/hr)
100
1,000
10
EC
(U
10
Figure 1. Semi-experiential plot of distribution of leak/emission rate of component in a
"typical" hydrocarbon facility.
-------
P. R. Harrison
Again, this procedure is much more accurate for the component fraction but
does not lend itself to routine measurements and enforcement due to time and/
or expense considerations. Finally, some of the fugitives are area sources
such as spills, liquid component leaks, and sewers. The first choice to
characterize by transects is easily the most rapid and inexpensive method
of obtaining the total fugitive emission rate. It is also probably the most
accurate.
RESULTS
Percent of Component Leaks Related to the Total Number
Of the four facilities, two happen to be moderately large refineries
located in different areas of the United States. Each has virtually the
same capacity and throughput. Table 1 summarizes these results. The older
facility has six percent classified above the insignificant leak rate
(0.01 lib/day) and the newer facility is 1.9 percent. Thus, the rates
differ by a factor of three due to age and possible maintenance practices.
In addition, the older facility produces a higher percentage of gaseous
hydrocarbons (lights). By repairing all the Class I and II leaks they
could possibly obtain average leak rates of less than 0.0014 and 0.002
Ib/day per component (500 and 3200 components, respectively) for the newer
and older facility. This close agreement is an indication of the leverage
of the larger leak rate components on the total rate. In fact, after Class I
maintenance the average component leak rate is substantially reduced by a
factor of nine to tenfold.
TABLE 1. DISTRIBUTION BY PERCENT OF LEAK RATES
FOR TWO FACILITIES OF SIMILAR THROUGHPUT
Class
Total Number
of Components
II
in
IV
Older Facility
Newer Facility
2.0
0.3
2.0
0.7
2.0
0.9
96
98
80,000
50,000
-------
P. R. Harrison
Figure 2 presents a cumulative distribution of components for the
newer facility. These results reinforce the conclusion that a rigorous
one class component selection procedure is not advised.
Detection and Classification Techniques
By use of a classification technique we must operationally define
the ranges and methods in sufficient detail to be quantitative but not
so restrictive as to be unacceptably difficult (expensive). Factors
such as accessability, wind speed and direction, and probe distance
relative to the component are important. In order to investigate the
sensitivity of some of these factors we conducted some simple experiments.
Figure 3 is a sketch of a four-inch valve that was classified as a
Class III leaker (yellow). Table 2 is the result of this test. Figures
4 to 6 present plots of these data. The impetus for these tests is to
equate the MRI method with a proposed EPA method. We see that the
closer one takes the survey the higher the concentration becomes, but
in general, the difference between maxima and minima become less, except
for very brief excursions. This fact, coupled with the ease of classifi-
cation by actually physically touching the probe extension to the valve
stem makes the task more efficient. Each method is acceptable as long
as one has an order of equivalence between the methods.
Use of Transects for Measurement of Total Fugitive Emission Rates
As mentioned earlier, fugitive emissions are mostly related to
component emissions. There are other sources identified by the characteri-
zation techniques that are area sources not easily quantified by individual
measurements. These must be assessed by transect measurements. By use
of classical diffusion analysis one can assess individual areas as well
as the total facility. After securing constituent analysis of the grab
samples one can determine the nonmethane fraction and the hydrocarbon mix
calibration factors. Combining these three data points one can calculate
the NMHC emission rate of the total facility or any subset.
The method has better than 10 percent precision under constant
emission activity and availability of an optimum transect line. Under
dynamic conditions with no control of activities and a facility-edge
transect line the precision can drop to 50 percent, most of which seems
to be caused by changes in activities, not the method. In most cases, the
precision is better than 20 percent. The accuracy is most effected by the
constituent analysis factors. The instrument .calibration holds well over a
length of time to better 10 percent if the temperature is not radically changed.
The calculations use meteorology in such a way that only the vertical
diffusion parameter is assumed from statistical tables. All other parameters
are measured by objective means. An example of a surface transect is shown
in Figure 7. For a more positive plume profile measurement several transects
-5-
-------
P. R. Harrison
99.99
Arithmetic
Mean
Older Facility
e.ei
o o
HHHC Uak Rate (Ib/day)
Figure 2. Cumulative distribution of components with
respect to leak rate.
o o
-6-
-------
P. R. Harrison
wind * 2 mph
probe to analyzer
78-068
Figure 3. Schematic of valve showing sampling positions.
-7-
-------
P. R. Harrison
TABLE 2. VALVE DISTANCE STUDY
Measurements Were Taken Horizontally From Valve Stem
As Noted at Level of Packing Surface Except for *.
Please See Figure 3.
() Values are Minimum and Maximum Values at That Position
Position
Distance
(cm)
5
2.5
0
0*
1
250
(80-1200)
250
(100- 800)
(700-2000)
700
(400-200)
2
150
(90-900)
200
(150-600)
3
20
(15-800)
30
(15-800)
4
15
(10- 60)
15
(10- 30)
100
(40-400)
5
15
(10-500)
25
(15-50)
6
50
(30-100)
30
(15-50)
Ambient = 20-40 ppm
Wind = 2 mph, gusting to 5 mph
*1 cm above packing
-8-
-------
P. R. Harrison
TlOGOppm
8 -
Figure 4. Concentration around a valve stem at 5 cm horizontal
distance from packing.
-9-
-------
P. R. Harrison
-p 1000 pprr
1
" 800
-J- 600
78-070
Figure 5. Concentration around a valve stem at 2.5 cm horizontal
distance from packing.
-10-
-------
P. R. Harrison
-rlOOOppa
78-071
Figure 6. Concentration around a valve stem at the stem and
1.0 cm above packing.
-11-
-------
t-t
H-
W
O
75 m
NJ
I
E
a.
c
rt
li
u
a
a
u
Horizontal Diatance
Figure 7. Average hydrocarbon concentrations obtained downwind of a refinery on
19 July 1977, 2000 to 2200 CDT. Solid lines represent computed Gaussian
plume components.
-------
P. R. Harrison
in height are made. Figure 8 shows the results of such a method. Similar
but inferior results can be obtained by use of fixed point vertical profiles
(see Figure 9).
CONCLUSIONS
The following general conclusions can be made concerning the techniques
and knowledge developed so far for fugitive hydrocarbon emissions from re-
fineries and petrochemical facilities:
• A reasonably accurate assessment of nonmethane
hydrocarbon emission rates for subsets or total
facilities can be made by use of the MRI transect
methods and calculations. Precisions up to eight
percent have been reached.
• Component surveys require a classification scheme
for meaningful statistical selection.
• Initial component surveys require three to six man-
months to achieve. Less time is required after main-
tenance .
• Results of component surveys show that significant leaks
constitute one to six percent of the total components
available, depending on age and maintenance practices.
• Subsequent individual measurement of leak rates of
randomly selected components indicate that the Class
I components (large leakers) constitute about 80 percent
of the total component fugitive emissions. After main-
tenance on Class I leaks the average component leak rate
is less than AP-42 by a factor of ten or more. After
maintenance on combined Class I and II leaks the average
rate is reduced by another factor of ten.
• Until these techniques become Best Available Control
Technology (BACT) the methods can give considerable
trade-offs.
• The remission rate of fixed components is slow, on the
order of months.
-13-
-------
P. R. Harrison
- 20-
«
BO*
I I I I I I I I I I I I I I I
10 -
5 _
15 20
Chart Units
25 27
Figure 8.
Example of vertical cross section, 20 October.
All data are in ppm.
-14-
-------
VIRTUAL SOURCE
OF FUGITIVE EMISSIONS
ELEVATED
POINT SOURCES
i
H-
Ul
\V
H
H-
CO
O
K p
n
ZOOor
o
n
Figure 9. Schematic representation of vertical point profiles (1, 2, 3, 4, 5)
and elevated horizontal transect lines (A, B, C, D).
-------
P. R. Harrison
QUESTIONS AND ANSWERS
Q. Bob Kenson, TRC - The fact that reproducibility was very good on
the horizontal traverse at the fenceline doesn't say anything about
accuracy does it?
A. - How do you measure accuracy when you have nothing to measure
against? So, I take a measured factor but I don't choose to represent it
as any accuracy because I can't calibrate that accuracy. What I can do is
to go through and do this relatively detailed component classification
and measure the leak rate and use this equivalency to see how it compares
with the fenceline measurement. In one case, it was within 75% which is
good considering that there are contributions from non-fugitive sources.
Other cases haven't done that well. In all cases it is always less than
the fenceline measurement which is encouraging. My biggest problem is
finding a laboratory that will give me a reproducible hydrocarbon constituent
analysis.
Q. Bob Kenson, TRC - The second part of my question relates to the
airplane traverse, were they able to show you the term tf , and were they
anything reasonable?
A. - The purpose was not to measure elevated plumes. As far as the
fugitives are concerned, they are mostly a ground-level type source system.
The Turner factors were acceptable within the 20% that we claimed. This is
over flat terrain.
Q. Hal Taback, KVB - During the program that I reported on earlier,
we made an attempt to measure the emissions from a refinery using upwind/
downwind techniques and the results were not in very good agreement with
the technique we used for point source or source type testing. What we did,
in this small refinery, is to have a source test crew in the refinery at the
same time as an ambient crew on the fenceline and they used the techniques
with the sulfur hexachloride for trying to trace the wind patterns. And
with that they predicted the diffusion and actually from the fenceline
measurements, they back calculated the emission factor. The result was
that there was about 10% correlation. There is a final report associated
with that program which should be out soon. We thought that perhaps our
ground-level measurements were missing a lot of emissions that were going
up.
A. - It depends on your release point. If you release in a warm area
it is~possible that you are missing it. That is why you should do at least
one elevated transect, especially when you are using it as a point of
reference. You must do elevated transects in order to get that maximum
value. We have seen no contradictions to the fenceline. Not a single
case in which it could not make sense, where in your case you don't seem
to have enough at the fenceline.
-16-
-------
P. R. Harrison
Comment - H. Taback, KVB - In the_ first chart you _showed,_._yqu had the
AP-42 prediction but the fenceline measurement was substantially lower.
A. - The actual survey did verify the fenceline measurement; in that
case it verified very well, but in a very small facility.
Comment - In that facility you did not have tanks, because when you
are at the fenceline of most facilities you were picking up tanks and other
sources of emissions.
A. - In most cases, we were able to isolate the tanks. The tanks were
downwind of us. One critical thing, of course, is to judiciously choose
your transect line and that is an experiential thing. I would caution you
that, (myself being a meteorologist) many times one will hire meteorologists
to do diffusion analysis. Even within the field of meteorology, diffusion
analysis is a specialized field. A modeler himself may not understand
synoptic conditions and may not be able to direct a field program as well
as he can direct a computer program. Thus, the choosing of the fenceline
transect is critical.
-17-
-------
FUGITIVE EMISSION CONTROL STRATEGIES
FOR PETROLEUM REFINERIES
I. A. Jefcoat, R. G. Wetherold, and W. Leigh Short
Industrial Environmental Research Laboratory
Environmental Protection Agency
Research Triangle Park, North Carolina
ABSTRACT
In a study funded by the Environmental Protection Agency, measurements
of hydrocarbon emissions from fugitive sources are being made at a number of
refineries throughout the country. Sampled sources include valves, flanges,
pump seals, compressor seals, relief valves, process drains, cooling towers,
API separators, and dissolved air flotation units.
In this paper, strategies are presented for the control of fugitive
hydrocarbon emissions from the various refinery sources. Emissions from these
sources in existing refineries are estimated from the results of sampling at
six refineries. The potential for emissions reductions with the proposed
control strategies is presented.
-------
DETECTION OF VOLATILE ORGANIC COMPOUND EMISSIONS
FROM EQUIPMENT LEAKS
K. C. Hustvedt* and R. C. Weber
U. S. Environmental Protection Agency, MD-13
Research Triangle Park, North Carolina
ABSTRACT
The body of information presented in this paper is directed to air
pollution control agencies and refinery and chemical plant operators
concerned with limiting volatile organic compound (VOC) emissions from
equipment leaks. Equipment that may emit VOC as a result of leaks includes
pump seals, compressor seals, valves, flanges, pressure relief devices,
process drains, and cooling towers. Four methods of leak detection are
discussed. The first three utilize a portable VOC detection device (flame
ionization detector) and the fourth a fixed monitoring system. For all of
the methods an elevated VOC concentration is indicative of an equipment leak.
The first detection method consists of using the instrument to sample the
ambient air in close proximity to each potential leak source. In the
second method, the VOC detector is used to traverse a unit upwind and down-
wind of all major ground level sources. Only the unit boundary is surveyed
in the third method. In the fourth method, fixed monitoring systems are
used to detect elevated VOC concentrations or to monitor equipment performance.
After a leak is located, maintenance can be performed and VOC emissions to
the atmosphere reduced.
RESUME
K. C. Hustvedt-is an Environmental Engineer with the EPA in Durham with
the Office of Air Quality Planning and Standards. K. C. is responsible for the
development of guidelines and regulations for the control of volatile organic
compound emissions from the petroleum industry. He received his B.S. degree
in Civil Engineering from Duke University in 1975 and is an Engineer in
Training in North Carolina.
Speaker
-------
K. C. Hustvedt
DETECTION OF VOLATILE ORGANIC COMPOUND EMISSIONS
FROM EQUIPMENT LEAKS
INTRODUCTION
This paper outlines four monitoring methods that have been used to
locate leaks of volatile organic compounds (VOC) from petroleum refinery
and chemical plant equipment. Equipment that is prone to develop leaks
includes pump seals, agitator seals, compressor seals, valves, flanges and
other connections, and pressure relief devices. Process drains and cooling
towers can also emit VOC because of leaks into the water from faulty equipment
or through improper operation. Although the average leak rate is relatively
low, these leaks can be a significant source of VOC emissions to atmosphere
because of the very large number of equipment components. For example, a
15,900 cubic meter per day refinery would have about 27,000 potential leak
sources, and a 22,700 megagram per year maleic anhydride plant would have
about 500, excluding flanges in both cases. Past and present testing of
petroleum refinery and chemical plant equipment has shown that, while leaks
from most equipment are insignificant, a small percentage can have high leak
rates. This situation highlights the importance of developing an effective
leak detection plan. Combinations of the four methods described in this paper
can be used to locate leaking equipment, so that timely repairs can be made
and emissions of VOC to the atmosphere reduced.
DESCRIPTION OF THE EMISSION SOURCES
There are several types of equipment found in petroleum refineries and
chemical plants which can leak and, therefore, emit VOC to the atmosphere.
VOC are emitted directly to atmosphere from vapor leaks and indirectly when
liquid leaks evaporate. Equipment that can leak includes; pump seals,
agitator seals, compressor seals, valves, pressure relief devices, flanges
and other connections. In addition, there can be VOC emissions from cooling
towers as a result of leaks through heat exchangers into the cooling water.
Improperly designed or operated process drains can also be a source of VOC.
-------
K. C. Hustvedt
Finally, are several miscellaneous sources of VOC in such facilities,
generally from wastewater streams contaminated with organic material, however,
these sources are not addressed in this paper.
In pumps, agitators and compressors, the leak occurs between the moving
shaft and the sealing element. These sealing elements include compressed
packings and finely machined surfaces (as in mechanical seals). Leaks in
flanges occur through or past the gasket material. Emissions from process
drains occur when organic materials are allowed to collect before the trap
and thus are exposed to the ambient air. For the most part, organic compounds
in cooling water result from leaks in heat exchange equipment.
Emissions from pressure relief devices can be of two types; an episodic
emission, when a rupture disc pops or when a relief valve opens, and a
continuous emission, when the valve does not reseat properly after the
pressure is relieved. The latter case, in which the leak occurs because of
improper mating of the sealing surfaces, is of more interest here.
Other types of valves can leak in several ways. An in-line valve, such
as a block or control valve, can leak through the valve stem, bonnet, and
flanges. An additional mode of leakage occurs for open ended valves, such as
drain or sample valves. In addition to the sealing failures, leaks can occur
in these valves when they are not completely closed, through either degradation
of the valve or improper operation.
Although leaks from individual sources are relatively small, the total
emissions from these sources can be quite significant because of the very
large number of potential sources. For example, a model typical sized petroleum
refinery of 15,900 cubic meter per day throughput with 10 processing units has
an estimated 250 pump seals, 14 compressor seals, 25,000 valves, 1400 process
drains, and 130 relief valves venting to atmosphere. On the other hand, a
model typical sized maleic anhydride manufacturing facility with a capacity
of 22,700 Mg/yr, has an estimated 15 pump seals and 500 valves.
Not all of the potential sources of VOC emissions will actually leak,
however. Several studies have found that the majority of potential sources
do not leak. Of the remainder, some leak a minor amount, while a few sources
-------
K. C. Hustvedt
account for a large part of the total VOC emitted.3'4'5'6'7'8 For example,
in one of these studies 73 percent of the refinery valves tested had
insignificant VOC emissions to the atmosphere. Twenty-five percent of the
valves tested emitted up to 2.2 kilograms per day (kg/day), and accounted
for one-fourth of the total emissions. The remaining two percent of the
valves each emitted over 2.2 kg/day, and accounted for three-quarters of
the total emissions.9 Data for chemical plants are very limited. One study
found that 34 percent of the pump seals leaked. Since there are many
potential leak sources which have widely varying leak rates, it is necessary
to utilize leak detection methods to effectively locate these leaks. Once
located, the leaking equipment can be repaired and emissions to atmosphere
reduced.
LEAK DETECTION METHODS
Four methods for detection of volatile organic compound emissions from
equipment leaks are presented here. Along with a description of the methods,
the advantages and disadvantages of each approach are discussed, including
relative leak detection effectiveness.
Complete Individual Component Survey
In a complete individual component survey each potential leak source is
screened to determine if the component is leaking. A component is screened
by using a portable VOC detection device to measure the VOC concentration at
the surfaces where leakage could occur. The instrument probe should be moved
along the surface, with emphasis on positioning the probe at the upwind and
downwind side of the component. Some potential sources, such as with process
drains, cooling towers, pressure relief devices, and open-ended pipes, have
open exhaust areas to the atmosphere rather than a seal interface. For these
sources, the probe should be placed at the centroid of this area as well as at
upwind and downwind points. These surveys can be performed by two-man teams,
one operating the instrument and the other recording results.
The major advantage of the complete individual component survey is that
it should locate all significant equipment leaks. By checking each component
individually, all leaks should be noted, and there should be no false indications
-------
K. C. Hustvedt
of high concentrations where no leak is found. The major disadvantage of
this method is that it has a relatively high manpower requirement. In a
complete individual component survey of a refinery similar to the model
15,900 cubic meter per day refinery, two men performing the survey required
a total of almost 1000 manhours.
Unit Area Survey
A unit area survey entails measuring the ambient VOC concentration
within a given distance, for example one meter, of all ground level equipment
within a processing area. These measurements are performed with a portable
VOC detection instrument utilizing a strip chart recorder. The instrument
operator walks on both the upwind and downwind side of the equipment, noting
on the chart record the location in the unit where any elevated VOC con-
centrations are detected. If an elevated VOC concentration is noted, the
components in that area will have to be individually screened to locate
the specific leaking component.
The major advantage of this method is that large leaks from all ground
level sources can be quickly located. The manpower requirements for this
method are much lower than those for the complete component check. Among the
disadvantages of this method are (1) leaks from adjacent units can cause false
leak indications, (2) high or intermittent winds (local meteorological
conditions) can increase dispersion of VOC and cause large leaks to go
undetected, and (3) additional effort is necessary to locate the specific
leaking equipment.
Unit Boundary Survey
In a unit boundary survey the VOC concentration at the edges of each
processing unit is monitored. Again, a portable VOC detection instrument and
strip chart recorder are utilized to perform the measurements. When an elevated
concentration is detected, the operator should note its location and the wind
speed and direction on the strip chart record. As in the unit area survey
the source of the elevated concentration will have to be determined by per-
forming an individual component survey on the equipment in the area of the
elevated concentration.
-------
K. C. Hustvedt
Unit boundary monitoring has the lowest manpower requirement of the
first three leak detection methods. It can be used as a quick way to determine
the presence of large leaks in a unit. However, unit boundary monitoring is
highly sensitive to local meteorology. For example, a large leak may not be
detected because of dispersion. On the other hand, high concentrations that
are recorded may be the result of emissions from non-leak sources, such as
downwash from a process vent. Further, an additional effort must be made to
locate the leaking equipment once an elevated concentration is found.
Fixed Point Monitoring Systems
The basic concept of the fixed monitoring system is that equipment can
be installed at specific sites within the process area to automatically
monitor for leaks.
There are several approaches to leak detection with fixed-point
monitors. These differ in the number and placement of the sample points,
and the manner in which the sample is taken and analyzed. One approach is
to place monitors near specific pieces of equipment, such as process pumps,
compressors, and cooling towers. These monitors can sample either the VOC
concentration or the operation of equipment, i.e. to monitor the pressure
between double mechanical seals to indicate failure of either seal. These
systems will tell the operator the location of the leaks that are detected.
A second approach is to place the sample points in a grid pattern throughout
the process area near the potential leak sources. When an elevated concen-
tration is noted, the operator must perform in individual component survey on
equipment in that area to locate the leaking component. In addition to these
variations in the location of the sampling points, different types of instru-
ments can be used. For example the monitoring system can consist of either
continuous sampling and analysis, or an automatic sequential system where
samples are collected and analyzed at a central location.
One advantage of the fixed-point monitor approach is that the location
and type of sampling and analysis can be tailored to meet the specific
requirements of individual plant sites. An additional advantage is the
ability to sample for specific compounds by GC analysis. Fixed point
-------
K. C. Hustvedt
monitoring systems have the highest capital cost and the lowest monitoring
manpower requirement of the four methods. Further, the approach may still
require the use of a portable VOC detection device to locate the leak,
particularly if process area monitoring is used. This monitoring method
is currently being used in several organic chemical manufacturing facilities,
including ethylene oxide and vinyl chloride production.
SUMMARY
In this paper, four methods of leak detection have been discussed. It
has been shown that, although a complete individual component survey has the
highest manpower requirement, it will specifically locate all leaks. The
other methods detect the presence of large leaks but usually an additional
effort is required to locate the actual leaking equipment. For example, if
an elevated concentration is found at a pump row during a unit area survey,
all of the pump seals, valves, drains, etc., in that immediate vicinity must
be individually sampled to locate the leak. A similar procedure is used to
locate leaks when conducting a unit boundary survey or when using certain
types of fixed point monitoring systems. This is illustrated in Figure 1.
Timely repair of leaks found by these methods will reduce VOC emissions to
the atmosphere.
-------
K. C. Hustvedt
COMPLETE
INDIVIDUAL
COMPONENT
SURVEY
UNIT
AREA
SURVEY
UNIT
BOUNDARY
SURVEY
FIXED-POINT
MONITORS
"ON-CALL"
LOCATION
OF
LEAKS
SPECIFIC
LEAK
FOUND
GRID
PATTERN
SPECIFIC
EQUIPMENT
FIGURE 1. ALTERNATIVE LEAK DETECTION METHODS
-------
K. C. Hustvedt
REFERENCES
1. J. M. Johnson, Exxon Company, U.S.A., personal communication with
Robert T. Walsh, U.S. EPA, July 28, 1977.
2. J. F. Lawson, Maleic Anhydride - Product Report, unpublished draft
report under EPA Contract No. 68-02-2577, March, 1978.
3. J. M. Johnson, op. cit.
4. Los Angeles County Air Pollution Control District,"Joint District,
Federal and State Project for the Evaluation of Refinery Emissions,"
(nine reports), 1957-1958.
5. "Assessment of Environmental Emissions from Oil Refining,"
(unpublished interim reports), Radian Corporation, EPA Contract
No. 68-02-2665, in progress, March, 1976, to March, 1979.
6. K.C. Hustvedt, U.S. EPA, personal communication with H.J. Taback,
KVB, Inc., March 13, 1978.
7. B.F. Ballard, Phillips Petroleum Company, personal communications
with W. Stewart, Texas Air Control Board, September 8, 1977.
8. K.C. Hustvedt, U.S. EPA, personal communication with Paul Harrison,
Meteorology Research, Inc., January 18, 1978.
9. Los Angeles County Air Pollution Control District, op. cit.
10. J. B. Cox, Exxon Chemical Company, U.S.A., personal communication
with J. W. Blackburn, Hydroscience, Inc., February 21, 1978.
11. K. C. Hustvedt, U.S. EPA, personal communication with H. J. Taback,
KVB, Inc., March 13, 1978.
-------
K. C. Hustvedt
QUESTIONS AND ANSWERS
Q. Herb Bruch - Have you looked at the size of the refinery units
and determined costs, like a smaller refinery, say 30,000 bbl/day vs.
one at 100,000 bbl/day?
A. - Costing what? The various methods?
Q. Herb Bruch - Monitoring fittings.
A. - The costs are a function of the number of sources and the
frequency of monitoring. If they had the same degree of fractionation,
a large crude unit would have the same number of sources as a smaller one,
so the cost of monitoring them would be the same. I imagine the small
refineries usually have less fractionation so you have less pumps and
valves. Familiarity of the instrument operator with the area is important;
so that you don't check steam valves, etc.
Q. - J. Daily - Is there a specific make of VOC unit that you use or
is there a broad range?
A. - We use the Century Systems OVA-108.
Q. J. Daily - Is that the only thing you have tried?
A. - We have also tried the Bacharach TLV. When performing this type
of monitoring, we are going from one piece of equipment to another as fast
as possible, finding leak or no leak and moving on. We found the OVA
superior because of the fast response time and a high pumping rate to clear
the sample probe. The Bacharach operates on the hot-filament principle.
Once it gets heavy hydrocarbon on it, it does not clear very quickly. You
have to walk out of the unit. Once you find a large leak, your instrument
is dead for about 10 minutes. With the OVA, a leak can put the fire out,
but one can step back and re-light it very quickly. The OVA is instrinsically
safe in hazardous areas.
£. - Did you actually test this fixed point monitoring system?
A. - I've not performed tests on it myself. There have been some
studies performed on fixed monitoring systems. Vinyl chloride plants in
Region V have been inspected and reports have been issued on them as to
their state of compliance. We are reviewing these reports, and we have
test plans underway.
-------
K. C. Hustvedt
Q_. - I think you had on your evaluation sheet a low labor cost for fixed
point monitoring systems. If you use something of this nature and maintain
it properly, labor is going to be higher than you expect.
A. - This is only monitoring labor that I have referred to on the chart.
Fixed point monitoring systems will have a higher maintenance cost, as will
maintaining the portable instruments. There are problems with these types
of instruments that have been found in chemical plants.
10
-------
STATE REGULATIONS FOR CONTROL
OF REFINERY EMISSIONS
Henry E. Slevers
Texas Air Control Board
Austin, Texas
ABSTRACT
This paper illustrates a state air pollution control philosophy and
summarizes provisions of Texas' Regulations governing control of refinery
emissions. Provisins of TACB Regulation VI "Control of Air Pollution by Permits
for New Construction or Modification" are detailed as Texas' strongest control
mechanism. Regulation V, "Control of Air Pollution by Volatile Carbon Com-
pounds" is reviewed as having greatest impact on control of emissions from
existing petroleum refineries, and as an example of effect current revisions
will have on state policy and control philosophy. This discussion includes
definitional changes, plans for and effect of implementing EPA's Applicable
Control Techniques Guidelines, and difficulties found in designating geograph-
ical areas in which these controls will be applied.
Also presented are anticipated rule changes on fugitive emissions and
current thinking on need for rule changes resulting from fuel conversion
efforts.
RESUME
Hank Sievers is from the Texas Air Control Board. He has a B.S. and M.S.
Degrees in Meteorology from Texas A&M; a B.S. in Electrical Engineering from
the University of Texas and an M.S. in Environmental Health Engineering from
University of Texas. He served in the service with the Air Force. He has
been with the Texas Air Control Board for 5-1/2 years. Currently, he is Chief
of the Control Programs Section, the Control Strategy Division.
-------
Henry E. Slevers
STATE REGULATIONS FOR CONTROL OF REFINERY EMISSIONS
As representatives from other states and territories here can attest, state
air pollution regulatory agencies often find themselves in the role of balancing
conflicting desires of EPA, citizens environmental groups, industry, local gov-
ernments, and the public. This role is not always an easy one. Open communi-
cation and opportunity to share viewpoints helps. As a representative of Texas
Air Control Board, I thank you for this opportunity to present a state viewpoint
on control of refinery emissions.
I'm not here to "talk Texas." However, petroleum refining ranks third among
our nation's industries and Texas' refining capacity is first in the country both
in number of plants, Uo, and barrels produced each day, about three and three-
quarters million. Translated, this figure accounts for more, than a quarter of the
national total. For Texas, with respect to emissions, this production results in
emissions accounting for approximately 21 percent of total point source emissions
in the state, and 13 percent of total point and area sources. Obviously we have
had to develop control strategies as have other states, to impact this significant
percentage. Of the 38 states with oil refineries, half have refinery related con-
trols on emissions of volatile organic compounds. In terms of capacity, these 19
states process about 83 percent of the national total. Development of policy at
the state level, provisions of 1977 Amendments to the Clean Air Act and subsequent
rules, regulations, and interpretative rulings of the Environmental Protection
Agency have resulted in the emergence of issues'—past, pending, and anticipated—
that may be of interest to us all. I plan today to touch briefly on these issues
and hopefully generate a forum for general discussion.
Since its inception in 19&5, the Texas Air Control Board has sought to strike
a reasonable balance among concerns for clean air, new industry, jobs, and a
healthy economy. Our enabling legislation, the Texas Clean Air Act, mandates con-
sideration be given to the character and degree of injury to, or interference
with, health and physical property of the people; social and economic value of
the source; location priority; and the technical practicability of reducing or
eliminating emissions resulting from the source. Through the years we have devel-
oped a control philosophy designed to achieve our clean air goals without sacri-
ficing industrial development or expansion. We have adopted eight regulations
to conduct our air control activities. Seven of them apply to existing sources
and are organized by type of emissions.
Regulation VI, Control of Air Pollution by Issuance of Permits for New Con-
struction or Modification, is our primary control mechanism for new sources.
Permitting new sources is probably our most effective control measure. It both
allows economic application of more stringent controls than feasibly can be
applied to existing sources, and is easier to enforce because permit conditions
are tailored to a specific source. Legal authority for this regulation was given
to the Board in 1971 Amendments to the Texas Clean Air Act. It requires that any
person who plans to construct any new facility or to engage in the modification
of any existing facility which may emit air contaminants must obtain a construc-
tion permit from Texas Air Control Board prior to beginning work on the facility.
If a permit to construct is issued, an operating permit must be applied for within
60 days after the facility has begun operating, in most cases.
In order to be granted a permit to construct and to operate, the owner or
operator of a proposed facility must demonstrate compliance with all Rules and
Regulations of the Texas Air Control Board and the intent of the Texas Clean Air
-------
Henry E. Slevers
Act. He or she must also demonstrate that the proposed facility will not prevent
maintenance or attainment of any applicable ambient air quality standard nor cause
significant deterioration of existing ambient air quality in the area. The appli-
cation also must contain demonstration of proper consideration of land use, utili-
zation of best available control technology, and that technical practicability and
economic reasonableness of reducing or eliminating emissions resulting from the
facility was considered. The proposed facility also must meet, at minimum, re-
quirements of any applicable new source performance standards or emission standards
promulgated by EPA pursuant to authority granted under Sections 111 and 112 of the
federal Clean Air Act, as amended.
In January of this year, the Texas Air Control Board reluctantly voted to
adopt the controversial EPA offset policy, providing that new sources offset pro-
jected emissions. Despite our serious reservations, we have incorporated this
policy into Regulation VI which deals with permits. Essentially we are enforcing
the requirements of EPA's interpretative ruling on UO CPE 51.18. When the pro-
posed facility is in a nonattaintnent area for national ambient air quality stand-
ards, which with one exception in Texas are for photochemical oxidants and TSP,
and if emissions potential for these pollutants is 100 tons per year or greater,
the applicant must attain the lowest achievable emission rate. If allowable
emissions of the proposed facility, determined as the maximum permissible rate
after controls are applied, still are equal to or in excess of 100 tons, then the
applicant must propose offset measures which represent greater than a one-to-one
ratio of decreased over increased emissions. A statement certifying that all
existing facilities of this owner, if any, in the applicable AQCR are in compliance
with all applicable TACB Rules and Regulations or in compliance with an approved
schedule also must be provided. This implementation procedure has revealed some
ambiguities in the Clean Air Act and EPA's interpretative ruling. We have applied
our own interpretation to some of them, making assumptions we believe maintain
the integrity of the intent of both.
One example is that application of offset is limited to sources with potential
of 100 tons of emissions per year. It may be argued that setting this limitation
exempts facilities of a size that, taken together, could represent figures well
over 100 tons and significant percentages of emissions in a given area. Also,
abuse of the intent of providing a net air quality benefit through offset could
incur if industry expansion were designed for growth through increment—through
establishing separate facilities none of which alone would have 100 tons or greater
annual allowable emissions. Such a system abuses both industry and ambient air
quality. For these reasons we apply the 100-ton limitation to combined emissions
allowable for previously permitted facilities and proposed facilities at a given
source location or site. A typical permit falling under the offset policy carries
with it 30-to-kO provisions designed to document and enforce these and other re-
quirements of new source review.
The potential differential between actual and allowable emissions represents
another issue we feel needs clarification. As stated, the interpretative ruling
requires that reductions achieved by offsets result in a net reduction of total
allowable emissions from existing sources. Since actual emissions can be, and
very often are, considerably below allowed emissions, literal application of this
concept could result in manipulation of emissions reductions that literally would
be acceptable, but which could in practice increase actual emissions in the^
affected area. For this reason, our permits section uses actual emissions in cal-
culating and approving offset credits.
-------
Henry E. Sievers
A third issue we believe needs EPA clarification is prohibition of "banking"
offset credits to new sources. This policy discourages voluntary application of
control more stringent than applicable state regulations. It also discourages
cost effective growth since the source owner could, if banking were allowed, ex-
pand when and where it would be to his economic advantage to do so, without losing
offset credits. If banking were allowed on a per company or per site basis, volun-
tary reduction of emissions of sources that are not controlled by state regulations
also would be encouraged. We feel some ambiguity in this policy exists since
states and communities are permitted to commit to emissions reductions when growth
is desired. This procedure could, in effect, be construed as establishment of a
state- or community-operated bank.
Our permits division has been involved in lengthy consultation with EPA over
these issues, and we hope for EPA clarification in the near future.
Thus far I have presented a summary of issues confronting us under permitting
of new sources. Controls on existing sources also have significant impact on
petroleum refineries. Volatile Organic Compounds in Texas are controlled by TACB
Regulation V. I plan to focus on this regulation since petroleum refineries are
a significant source of VOC and represent a potential emission problem by virtue
of the large quantities of petroleum refined and the intricacy of the refining
process. Also, refineries tend to be located in areas where the oxidant national
ambient air quality standards are likely to be exceeded. In Texas, for example,
the petroleum refining industry is concentrated on the gulf coast. Estimates of
nonmethane hydrocarbon emissions in the Houston-Galveston SMSA total about lt-7,1^1
tons annually. This figure represents 23 percent of total nonmethane hydrocarbon
emissions in the area. Ozone measurements there have been recorded at .155 to
.261 ppm in the three nonattainment counties falling in this seven-county standard
metropolitan statistical area.
Regulation V, Control of Air Pollution from Volatile Carbon Compounds, covers
storage of volatile carbon compounds, crude oil, or condensate; loading and un-
loading facilities; filling of gasoline storage vessels; water separation; and
waste gas disposal. Current efforts are directed at substantial revision of this
regulation. Revisions underway are to enhance clarity of existing control mea-
sures, to modify areas regulated, and to incorporate technology outlined in the
first eleven Applicable Control Techniques Guidelines issued by EPA's Office of
Air Quality Planning and Standards.
Regulation V's title will be changed to "Control of Air Pollution from Vola-
tile Organic Compounds" for consistency with EPA nomenclature. Second, to insure
the scope of the regulation is clear, VOC is being defined as any compound of
carbon (excluding carbon monoxide, carbon dioxide, carbonic acid, metallic car-
bides or carbonates, and ammonium carbonate) that has a vapor pressure greater
than 0.1 mm of Hg at standard conditions. Currently we define volatile carbon
compounds as compounds with vapor pressure of 1.5 psia or greater and control them
at those levels. This definitional change will clarify the compounds controlled
and still will apply only to VOC with 1.5 psia or greater. We also are developing
a new Regulation IX for control of carbon monoxide, since it no longer will be
covered by Regulation V.
Applicable Control Techniques Guidelines which pertain to refineries are
Control of Refinery Vacuum Producing Systems, Wastewater Separators, and Process
Unit Turnarounds, issued in October, 1977; and Control of Volatile Organic Emis-
sions from Storage of Petroleum Liquids in Fixed-Roof Tanks, issued in December,
1977.
-------
Henry E. Slevers
With slight differences in emphasis and wording, our existing Regulation V
covers technologies outlined in these documents except for process unit turnarounds.
This process currently is being incorporated. Application of technologies in Regu-
lation V have accounted for fairly significant reductions of uncontrolled refinery
emissions in the state, about 30 percent, with 18 percent attributed to controls
on wastewater separators, 6 percent from vacuum jets, and 6 percent from fluid
catalytic cracking units (FCCU). Estimates are that, in the Houston area, these
controls have resulted in a 50 percent reduction over a l;-year period. Application
of RACT for process unit turnaround is expected to account for an additional 10-15
percent reduction of uncontrolled emissions statewide.
As we make revisions to this regulation, and in development of our current
SIP, designation of areas or counties to be covered by these controls is a very
real problem.
Parts of current Regulation V apply to 19 Texas counties which account for 90
percent of total emissions in the state. These counties were selected prior to
state and EPA designation of nonattainment areas for NAAQS. Last July, EPA promul-
gated regulations to expand areas covered for certain controls, including storage
of crude oil. These areas included counties in close proximity to those included
in nonattainment areas for the photochemical oxidant standard. At the state level,
then, we are found with (l) Our original list of counties falling under Regulation
V, (2) a disimiliar EPA-promulgated list for which certain rules apply, and (3)
designated nonattainment areas. Obviously there exists a problem on geographical
application of these controls. Applying Regulation V, as modified, to any county
included in these designations would expand the list of counties covered and gen-
erally impose more stringent rules to a greater number of existing sources.
Petroleum refineries operating in some attainment areas would therefore be forced
to comply with the additional controls required for nonattainment areas. Current
thinking is to include nonattainment counties only, although a firm decision has
not been made.
Thus far I have summarized changes we presently are incorporating. One change
that was considered, but not adopted, warrants some discussion.
As you are aware, last May EPA released a draft document on fugitive emissions,
Control of_ Hydrocarbons from Miscellaneous Refinery Sources along with control tech-
nology outlined in Air Pollution Control Technology Available to. 2.6 Sources of Vola-
tile Organic Compounds.
During negotiations between EPA and TACB on adoption of the offset policy last
fall, TACB applied for a waiver. Provisions of our waiver application included
agreement to modify Regulation V to incorporate RACT as defined by the extremely
stringent provisions of these documents.
In Texas, proposed regulations or proposed regulation changes are discussed
among TACB staff, and with interested or affected parties including regional offices
and groups representing industry, local governments, protection of the environment,
and the public interest. Public hearings are held, comments voiced, and changes
incorporated before adoption.
During these processes, industrial response to the proposed fugitive rule
change was overwhelmingly negative. Exxon Corporation, for example, estimated cost
of incorporating these changes in one refinery at two and one-half million dollars.
When it became evident that TACB would not receive a waiver from offset, adoption
of this entire set of proposed rules was delayed. We do not intend to impose such
-------
Henry E. Sievers
restrictive rules on industry unless publication of final guidelines, expected in
June, requires us to do so. Application of additional industrial controls takes
away opportunities by industry to offset emissions, in turn limiting industrial
growth. Such limits we consider a negative factor in the general welfare of our
state, an inequitable tradeoff since estimates are these controls would result in
a reduction of 3^,000 tons annually, or only 6 percent of total uncontrolled re-
finery emissions.
Several studies on fugitive emission data were conducted in Texas last fall
with the intent of obtaining additional data prior to finalization of EPA guide-
lines. One, instigated by Texas Chemical Council, a chemical industrial trade
association, was conducted by Exxon Chemical. As part of the study, fugitive
sources—valves, pumps, compressors, flanges, etcetera—were measured for losses.
Preliminary findings indicated that fugitive losses through valves may be as much
as 70 percent less than EPA estimates indicate. Therefore, emissions factors may
be much lower than EPA figures, even though several vents were found to have
losses greater than expected.
Another study was conducted at Amoco/Texas City Refinery by TACB staff and
EPA to cover fugitive emissions from standpoints of sampling, leak detection,
and control technology.
Informal estimates were that proposed fugitive emission controls would require
addition of an estimated U5 persons, an increase of 7 to 8 percent of total oper-
ators, resulting in increase of about l.U million dollars for the 23 process units
involved. Final results of both studies are pending. Joint committees have been
set up with Texas Air Control Board and the Texas Chemical Council to review final
results of these studies, when released, and to coordinate activities in implement-
ing EPA guidelines for fugitive emissions, when published.
Issues I have outlined thus far occupy us today. In terms of future develop-
ments, probably the most far reaching change facing Texas, especially the Texas
Gulf Coast, is boiler fuel conversion. We expect most petroleum refineries to
convert to fuel oil as natural gas resources dwindle. The primary impact will be
increase in emissions of sulfur dioxide and particulate matter, which may require
modification of TACB Regulations I and II. Current thinking is to control the
sulfur content in allowable fuel oils, although limited availability and cost pose
problems for such an approach. At present we are researching probable effects of
fuel conversion by applying dispersion modeling techniques.
In summary, I think it is fair to say Texas Air Control Board has taken, and
continues to take all necessary steps to protect public health and welfare accord-
ing to the legislative mandate of the Texas Clean Air Act. In establishing regu-
lations and in proposing rule changes, we take into account the very real con-
siderations of the comprehensive impact our actions will have in Texas. We believe
that a strong industrial and economic base is as essential to general welfare as
is clean air, and that pollution control agencies must take this fact into account.
We also constantly review our regulations in an ongoing process dependent on newest
conclusions derived from research and analysis of chemical reactions and state-
of-the-art technology. We cannot, as it were, operate in a vacuum. To this end
we must all continue to work—EPA in consideration of technical feasibility and
economic reasonableness; state air pollution agencies in enforcing air control
regulations with an eye toward balance; and local governments and industry in pur-
suing clean air goals with good faith efforts.
Thank you for your attention. I welcome and solicit comments or questions.
5
-------
Henry E. Slevers
QUESTIONS AND ANSWERS
£. - Can you tell us a little bit more about how you expect the SIP
revision process to go, specifically, do you believe that Texas will complete
its revision by the end of this year so as to avoid the construction ban in
the middle of 1979?
A. - Yes indeed. We are currently revising Regulation V to incorporate
these RACT. We hope to have this done shortly and to go to public hearings
with it—probably in June. We also have been working with our local govern-
ments and planning agencies in developing a transportation planning process
which, as we understand it, is all that is necessary for the SIP revision.
This, we hope, will be done by the end of the year. We are having several
workshops and meetings throughout the state which will take care of the
consultation requirements for the plan. We will also have public hearings
at the end of the year which will incorporate all of our control strategies
and we'll be able to submit a plan by the first of January.
£. - You said you needed a 70% rollback of reductions of overall
hydrocarbons being considered, also organics, is that by rollback or is that
with modeling?
A. - That was using rollback. We also used the EKMA model, and in all
cases the reductions required there are higher than by using the rollback
methods. So we chose, as EPA says we can, the rollback method to determine
the amount of reduction.
£. - Do you think 70% reduction overall is possible?
A. - Yes, it is possible. No, I don't think it is practical. I think
we are hoping that by the time everything is laid before the public and we
show the type of controls that would be necessary to meet this standard, that
possibly the laws will be changed.
£. - How much of that 70% was from stationary sources and how much
from mobile sources?
A. - We can get this anywhere we want to. EPA doesn't say how you
have to get this reduction. What we have done is made a gross estimate of
how much reduction will result from the changes to our Regulation V on exist-
ing sources and from the changes in mobile sources due to the new car
standards, then add in some for permits which have already been issued
before the offset requirement. In any case, it looks like we are going to
get about 11% from fixed sources, probably about 24% from mobile sources in
that area by 1985, which still leaves about 43%.
Q. - Are inspection and maintenance in that 24%?
-------
Henry E. Slevers
A_. - No, we haven't had time to evaluate that. We are probably going to
be forced to have inspection and maintenance. We are not quite sure what
method. The problem is the legislature has to meet to approve it, so we
are not sure of the 24%.
(}. - On your 11% on the fixed, do you have any idea of how much of that
would be from fugitive sources?
A. - Actually that is not including fugitive sources at all, because
this is just from the existing guidelines that we have.
(£. Is TACB continuing the offset policy as part of the SIP revision?
A_. - I don't know that we have any alternative. We've got to allow for
growth somehow and this is one way of doing it. The other way, of course,
is to impose additional restrictions which essentially will allow some growth.
(£. - Would you say that the control plan would be a combination of the
two?
A. - We will, if we can, build in enough decreases to account for some
growth, but you see from these numbers that we aren't quite sure how we
are going to do that.
(£. - From a cost-effectiveness standpoint, is there any consideration
being given to the photochemical reactivity of pollutants or is it blanket
reductions?
A_. - No, we are following EPA guidelines that say essentially everything
except methane and ethane and those two other compounds are photochemically
reactive sooner or later. If not in the immediate area, then sometime
after they are transported to another area. So we are essentially on a
non-methane basis.
Q^. - Do you have any feel for when we will get to see a bound copy of
the revised Regulation V?
A. - Hopefully any day now. Probably it will be a couple weeks and we
will have something ready to send out for another review. We do, of course,
intend to have a fairly complete review of this regulation, and, all of our
other control strategies. So if anyone wants a copy of it, let us know and
we welcome your input and criticism.
-------
INSPECTION AND MONITORING CONCEPTS
FOR REFINERY FUGITIVE EMISSIONS
John H. Nakagama
Director, Environmental Affairs-Western Area
Atlantic Richfield Company
Los Angeles, California
ABSTRACT
A discussion on the concepts to be considered in framing regulations
to control refinery fugitive emissions from valves. Those concepts involve
the inspection, monitoring, reporting, and enforcement segments of any
proposed regulation, and were in part developed through testing at a
refinery in Carson, California.
RESUME
John Nakagama received his Bachelor's degree in Chemical Engineering in 1954
from the University of Utah and joined Atlantic Richfield immediately after
graduation. He spent 20 years in refinery tech service, 17 years as a
supervisor of various process units in a refinery. The last two years
John has been involved in environmental work for ARCO. He is currently
Director of Environmental Affairs for the Western area.
-------
John Nakagama
INSPECTION AND MONITORING CONCEPTS
FOR REFINERY FUGITIVE EMISSIONS
INTRODUCTION
Refinery fugitive emissions as evidenced not only by this workshop but by
recent events in California and Texas have obviously become a major concern
of the regulatory agencies. Since the last concerted effort to address this
subject was made in 1957, this might appear to the critical observer that
the renewed concern is a belated one that has been long overdue. If in
fact hydrocarbons have been polluting the air we breathe, why has not some-
thing been done about it? My response is that this renewed concern is in
fact more indicative of the extent to which regulatory review has come.
Consider that of the volatile organic emissions from stationary sources,
(2)
refineries contribute 2.8% of the total and of that fugitive emissions
represent some portion. Even based on conservative factors the fugitive
emissions are probably less than 1% of the total from stationary sources.
Consider also that by the definition of fugitive emissions we encompass a
wide range of equipment and the very massive number of items which make
practical measurement and control extremely difficult.
Nevertheless, we are not here today to defend any "belated" concern or bemoan
the difficulty of the solution, but in fact to solve the problem. We are as
a company concerned about reducing fugitive emissions and providing input to
regulations that cost effectively accomplishes the same.
A RESTATEMENT OF THE PROBLEM
In this discussion, I would like to limit the problem as one to develop an
inspection and monitoring program for the reduction of fugitive emissions
from valves. There are several reasons for this. One of course is in the
interest of time. But the major reason is that the definition of fugitive
emissions encompass such a wide variety of sources that the quantity of
- 1 -
-------
John Nakagama
emissions from each type are probably totally different. And if not, the
control methods should be. For instance, in the case of flanges per se
the past and more recent tests have shown that emissions from flanges are
essentially negligible and with the sheer magnitude of the number of flanges
there is no practical need for specific regulatory control.
In light of the fact that EPA has now published a guideline series on
vacuum producing systems, wastewater separators and process unit turn-
arounds, this delineation is not totally inconsistent with their own
approach. I would, however, urge that both industry and the regulatory
agencies develop a more reliable data base for both quantity of emission
and cost of control in order that: (1) Industry would have the time and
opportunity to provide some meaningful input on each specific proposal.
(2) The final regulations will be developed to control emissions in the
most cost beneficial mode conserving not only the agency's and industry's
limited resources, but ultimately that of the consumer's.
THE DATA BASE
Regarding the establishment of a more reliable data base, it would be nice
to argue for the completion of the present EPA/Radian study before any
further action is taken. In light of the mandates of the Clean Air Act
Amendments and recent actions by local regulatory agencies, I doubt that
we have the luxury of that option. Regulations, if not adopted, will at
least be proposed and industry will have to respond. With this in mind
we have undertaken two projects at our Watson Refinery at Carson, California.
The first project involved a crude distillation unit with a rated capacity
of 38,000 B/D. The objective of the study was to assess the costs that
would be incurred for the inspection of all valves and fittings for the
detection of any leaks, large or small. In all 11,414 leak sources, of
which 1,415 were valves, were identified and tested. In all 38 leaks were
- 2 -
-------
John Nakagama
discovered, 17 of which were valves and 21 fittings. Of those leaks, 16
valves and 15 fittings were repaired without a unit shutdown. The costs
incurred for this unit inspection was then pro-rated for the entire refi-
nery. If the inspections were conducted on a monthly basis as initially
proposed in the EPA guideline, the cost, exclusive of our present main-
tenance budget, would run $1,500,000 per year. Obviously, if the fre-
quency of inspections are reduced, so are the costs. But the more
pertinent question to be answered is, "what frequency of inspection is
needed?" That question can be partially answered by the fact that when
the California Air Resources Board Inspection Team re-inspected that
unit eight months later, with only the normal maintenance procedures
conducted at that unit in the interim, no leaks were detected.
But in order to provide additional data to answer the question of the
frequency needed, we have initiated a second project. At our LPG,
storage and transfer facility each component will be inspected for leaks.
The leaks will be repaired and each component will be inspected at fixed
intervals of time to determine at what point inspection would be warranted.
If the components once repaired do not leak for a year or more, that's fine,
but I am afraid that any correlation developed will not be in time to influ-
ence any rule making. But then again perhaps the effort will. If is on
this note of lack of data and impending regulations that I offer the follow-
ing concepts.
1. Limit the inspection to high probability areas.
At the outset limit the inspection to those types of units that are
the most probable to have leaks and whose emissions would be the most
reactive and/or toxic. Granted all the data are not available and
valves in any unit probably leak to some degree, but let's make the,
attempt. Units that I would consider include: aromatic, ethylene,
- 3 -
-------
John Nakagama
cracking unit gas plants, alkylation and LPG units. Leave out crude-
vacuum, treating, heavy oil processing, fuel gas systems and even
reforming and hydrocracking units. The latter units operate with a
high hydrogen partial pressure and leak mostly hydrogen that has
its own automatic detection device, auto-ignition, and those leaks
get immediate attention for safety reason.
Whether to include one or more or even all the units depends on the
proposed regulation. If the frequency of inspection proposed is high
then limit the number of units to be controlled to obtain a good data
base and amend the regulation accordingly. If the frequency is practi-
cal and cost effective, then include more units.
2. Inspection Frequency and Method.
Given that the units selected are those high probablity units previously
mentioned, I would suggest that a reasonable frequency for the inspection
for each valve should occur just prior to a scheduled turnaround. During
this inspection, leaks that would have required partial shutdown of the
unit with its resultant emissions can be repaired or replaced during the
shutdown minimizing the overall emissions. Leaks that might have deve-
loped between shutdowns and found and repaired could be taken care of by
a walk-through monitoring inspection.
The method for detecting valve leaks should be left up to the operator.
The major point is to locate and repair, not quantify, the leak. In
some instances, a hydrocarbon detector would be preferred, in others a
soap test or visual inspection would suffice. A hydrocarbon detector
would require calibration and specialized training. Delays and argu-
ments concerning various craft jurisdictions would only hamper the
effort. In those instances, the soap test or visual inspection brings
- 4 -
-------
John Nakagama
the unit operator into the process of effectiving an improvement
in his unit, increasing his awareness and morale and ultimately
increasing the efficiency and effectiveness of the program.
3. Reporting and Enforcement
The most difficult segment of any proposed regulation to control
fugitive emissions is in the reporting and enforcement. If the
reporting is to be a self-enforcement mechanism coupled with severe
penalties, then the costs for reducing these emissions becomes
inordinately high and the whole program is not practical. Consider
that in our Watson Refinery there are more than 120,000 valves.
If it were necessary to number, date, record and file all of the
data on these valves to ensure that no violations occurred, the
cost of the paper work would soon outweigh the cost of locating
and repairing the leak. This sort of effort would be analogous
to requiring the motorist to maintain a file on all possible leak
sources on his automobile. On the other hand, if the effort is too
lax, then emissions reductions may very likely not occur. Somewhere
in the middle ground is the solution.
Since the strict reporting and enforcement approach is intended to
ferret out the indifferent operator, then perhaps another approach
would accomplish that. I would suggest that this could be accomplished
by the enforcement agency reviewing the maintenance policies and prac-
tices that are in fact in force at the facility. Some of things to be
considered should include: (1) procedures for testing of new and re-
serviced valves,(2) selection of the proper lubricant and gasket materials,
(3) on-site inspection of the check lists and tagging system for leaking
valves. (4) maintenance procedures for assignment of repairs, (5) unit
operator training practices. This review coupled with the authority of the
- 5 -
-------
John Nakagama
enforcement inspector to cite the facility for violations under
presently existing regulations should accomplish the goal of reducing
emissions.
SUMMARY
Considering then the probability that industry will soon be faced with
addressing proposed regulations on the control of fugitive emissions, I
would like to re-state some of the concepts that should be considered in
framing those regulations.
1. In light of the inadequate data limit the scope of the regulation
to high probability areas.
2. Inspection of each component should be made just prior to the
scheduled turnaround.
3. The method of inspection should be left up to the facility.
4. The reporting should be limited and supplemented by review of
maintenance policies and practices of the facility.
5. The enforcement can be accomplished through existing regulations.
If in fact these concepts are applied, I am sure that an effective reduction
in fugitive emissions can be accomplished within the framework of an effi-
cient utilization of resources.
- 6 -
-------
John Nakagama
REFERENCES
"Joint District, Federal and State Project for the Evaluation
of Refinery Emissions", Los Angeles Air Pollution Control
District. 1957.
2. Letter from E. P. Crockett, API, to R. T. Walsh, U.S. EPA,
July 18, 1977.
-------
John Nakagama
QUESTIONS AND ANSWERS
Comment - I think you have made a lot of practical suggestions. On
the schedule for inspection valves, I think that this depends very much on
the type of valve, for instance, if you have a lubricated plug valve, every
time you turn the valve it should be greased. Secondly, if you have a packed
valve where the stem is rising and it's a valve which cycles often, it will
require frequent maintenance.
We make some valves for Universal Oil Products for their HF process.
They have a green paint and if there is a leak it turns blue. I think if some
investigation were made of using this paint, valves that are susceptible to
leakage, particularly cold devices, would tend to indicate these leaks auto-
matically; tagging could be avoided. Where you have fittings that are hot,
you are getting a heat leak when you get a vapor leak. If you have a
sensitive heat detection device, such as temp sticks, it would tend to
identify the point where you had leaks. I .particularly think this idea of
paint or some sort of device which changes color to show susceptibility or
chemical reaction with the leakage is a viable method of detecting leaks
rather than using detection devices that are so sophisticated and so refined
that you may pick up background contamination rather than the valve leak
itself.
Comment - I agree with a lot of the concepts and especially the comments
about record keeping. Having been a Control Officer myself, I get sick and
tired of spending hours talking about nothing and stacks of paper and statistics
which nobody believes anyway. But you have got to document them. I cannot
go along with soap bubble tests, however.
P. Harrison - The problem with the detection on a particular interface
is the fact if you don't happen to put soap on that particular point, then
you don't even know it is there, let alone counting bubbles. The other thing
I want to leave with you is the worst leaks, the majority of the worst leaks
you see, are a matter of accessibility. So there are places where people don't
like to go or can't go conveniently and things like that. I know darn well
unless you have an instrument method you won't find them. There are leaks
other places too, but I am talking about some of the real bad ones right now.
Comment - F. G. Mesich. Radian - We disagree with that statement.
J. Nakagama - I am glad you said that.
Atly Jefcoat - I think John has done an admirable job of presenting
some logical approaches. This is the whole purpose of having these symposium
workshops—to have an exchange of ideas. These issues and questions that have
been brought up are ideal for the workshops this afterndon.
-------
API EMISSION MEASUREMENT PROGRAMS
J. G. Zabaga
Mobil Research and Development Corporation
Princeton, New Jersey
ABSTRACT
The organization and responsibilities of the Committee on Evaporation
Loss Measurement (CELM) are described within the framework of the API's
environmental and measurement objectives. The CELM's 12 hydrocarbon evapora-
tion loss bulletins, published during the period 1957-1969, are widely used
by industry and government. However, due to advancements in equipment design
and changes in operating procedures, the bulletins have been shown to generally
overstate emission levels, and must be updated to assure proper use in the
current regulatory climate.
The paper is part of a series of continuing status reports to industry
and government on the CELM's updating efforts.
RESUME
Mr. Joseph Zabaga is an Associate Engineer with Mobil Research and
Development Corporation in Princeton, New Jersey. His duties include
responsibility for the technical aspects of hydrocarbon loss control. He
is currently Chairman of the API Committee on Evaporation Loss Measurements.
He is also a member of the API Advisory Panel to EPA for the refinery study
that we currently have going. He is a member of APCA. CELM is now
revising all of the evaporation loss forces on which hydrocarbon emission
factors are based.
-------
J. G. Zabaga
API EMISSION MEASUREMENT PROGRAMS
INTRODUCTION
In addition to studies of hydrocarbon losses from fugitive
sources, there are a number of other programs underway to
quantify evaporation losses from petroleum handling equipment.
The purpose of this paper is to present an overview of ongoing
studies and the proposed time frame for completion of the general
revision of hydrocarbon emission factor bulletins.
As most of you know, evaporation loss data from API bulletins has
been used by regulatory agencies to develop hydrocarbon emission
factors. However, publication of the bulletins preceded the
Clean Air Act Amendments of 1970, and it has become quite clear
that both industry and regulators need updated emissions data to
properly respond to the Act.
This paper is part of a continuing series of reports to industry
and government on the status of new measurement programs. Much
of the new work is being done by the API's Committee on Evapora-
tion Loss Measurement or CELM, whose activities I would like to
describe.
Figure 1 shows the organizational structure of the API. There
are a number of technically oriented departments, one of which is
Non-Departmental/Industry Affairs. Within that department, as
shown in Figure 2, is the Committee on Petroleum Measurement
(COPM) a free-standing API committee. Titles of the standing
committees briefly describe the scope of the Committee's respon-
sibilities. In general, this is the establishment of standards
for measurements of all types of petroleum products throughout
the industry, including production, refining, distribution and
marketing activities.
The CELM is one of COPM's standing committees. CELM deals only
with the measurement of evaporation loss, which is quite dif-
ferent from the finite measurement responsibilities of other COPM
committees in developing standards for inventory control or
custody transfer.
Figure 3 shows the publications of the CELM that were developed
over a 12-year period from 1957-1969. These bulletins were the
first industry-published data on quantifying evaporation loss of
hydrocarbon liquids, and they have been widely accepted. To this
day they remain, or are the basis for, virtually all data sources
dealing with hydrocarbon evaporation loss. EPA's AP-42 "Compila-
tion of Air Pollutant Control Factors," has emission factors
derived from the bulletins.
-1-
-------
J. G. Zabaga
It is important to realize that all of the bulletins were
published prior to the U.S. Clean Air Act and do not necessarily
comprehend the accuracy levels required by the Act. The publi-
cations were originally developed to enable oil company operating
divisions to prepare cost benefit studies for evaluation of
alternative conservation techniques, when the data and esti-
mating methods became one of the tools used by regulators for
control strategies, it became necessary to review the content of
all of the bulletins for suitability in this new application.
The CELM is the working committee primarily responsible for that
review.
Figure 4 shows the present organization of CELM. Figure 5 is a
pictorial view of the equipment involved within each subcommit-
tee's responsibility. The four groups are staffed by 33 members,
24 from oil companies and 9 from supplier companies. Note that
these activities do not include hydrocarbon loss evaluations from
fugitive or other point sources. Responsibility for fugitive
loss determinations fall to the API environmental committee
concerned broadly with stationary source emissions.
Also omitted from the figure, but implicit throughout this
presentation, is any mention of several recent studies conducted
by the Western Oil and Gas Association (WOGA), and the Chicago
Bridge and Iron/Standard Oil of Ohio (CBI/SOHIO) floating roof
project.
Referring to Figure 5, the petroleum liquids of concern to CELM,
and the basic equipment and handling characteristics are as
follows:
1. Marine vessels — ships and barges — which transport crude
oils and all types of finished product. The depth of the
loaded compartment and its preloading condition are major
factors affecting evaporation loss as described later.
2. opentog floating-roof tanks have been in use since 1880 and
in their present form since 1923. The concept is to curtail
evaporation by floating a disc, the movable roof, in the
liquid. Under certain conditions, some emissions result
from the perimeter sealing ring area. Floating-roof tanks
are usually used for liquids with storage vapor pressures
between about 1.5 to about 11 psia, i.e., motor and aviation
gasolines, and some crude oils.
3. internal floating covers are a logical combination of open
top arid fixed-roof tanks, providing weather protection to
the roof and contents, and evaporation control by use of a
floating deck. Liquids generally stored are the same as in
open top floating-roof tanks.
-2-
-------
J. G. Zabaga
4. Fi xed-r oof tanks have been used for many years to store all
types of petroleum* Recognition by industry that product
could be conserved by use of one of the floating-roof
systems, and later, regulations to limit use to finished
product with less than 1.5 psia storage vapor pressure, have
generally restricted use of fixed-roof tanks to distillates,
residuals and some crude oils.
5. Tank truck and tank car considerations are confined pri-
marily to motor gasolines. A small amount of other finished
product and crude oil are also transported by these means.
The method of loading and the vapor pressure are apparently
the factors most affecting evaporation loss.
MARINE TERMINAL EMISSIONS
In 1976, CELM updated Bulletin 2514A, Hydrocarbon Emissions from
Marine Vessel Loading of Gasolines. A small but important point
in the bulletin title is the use of the term emissions, the
first such use in an API evaporation loss document. This recog-
nizes regulatory agency need for the emission factor format.
Previously CELM was concerned only with an annual determination
of the volume decrease caused by the change of hydrocarbon liquid
to hydrocarbon vapor, which was defined as evaporation loss.
This volume decrease was usually expressed in barrels per year
and has traditionally been used by the industry for inventory
control. Future API bulletins will include both emission and
evaporation loss factors.
Figure 6 shows a summary of emission factors from 83 tests
performed by four companies loading motor gasoline into ships
and barges. These factors average about 40 percent of previously
published data. The format also suggests the parameters that
dominate the evaporation phenomena, depth of the loaded compart-
ment and the hydrocarbon level present in the vessel on arrival
at dockside. For any specific marine terminal the percentage
of vessel compartments that annually arrive clean (i.e. low
preloading concentration) or dirty (high concentration), can
be determined by analysis of ships logs and terminal records.
With this information the factors in Figure 6 can be adjusted
with confidence to predict emission levels when loading motor
gasolines.
Other factors are still required for loading crude oils and for
emissions resulting from water ballasting of vessels after
unloading either gasoline or crude.
-3-
-------
J. G. Zabaga
Late last year the Western Oil and Gas Association (WOGA) com-
pleted a test program on loading California crudes. The program
was specific to one California county and while much data was
acquired, no predictive model was developed. However, eight oil
companies, acting in response to an EPA 114A letter, have under-
taken a coordinated program to develop new data on ballasting
emissions, plus any information necessary to fill gaps in the
previously described API and WOGA studies. Correlation of all of
these efforts should provide the data base necessary for publi-
cation of new, comprehensive marine terminal emission factors
late this year.
FLOATING ROOF TANKS
Emissions from floating-roof tanks have been under intensive
investigation since 1976. While much of the data in the original
API Bulletin (2517) was about 40 years old, the relatively slow
evaporation rate from floating-roof tank seals and the difficulty
in measuring the evaporation loss generally frustrated any new
in-depth studies.
Figure 7 depicts a floating-roof tank and the perimeter seal area
through which losses occur. The amount of evaporation occurring
in a reasonable amount of time, say several months with the tank
dormant, is too small to detect reliably with conventional
measuring techniques.
Two new testing methods have been used in an attempt to overcome
this problem. As shown in the figure, a field tank with product
in it can be taken out of active service and permitted to
weather. 'Separate testing has established that any evaporation
due to tank wall wetting as the roof descends in normal operation
is insignificant. Therefore, any emissions are those that result
from loss of lighter gasoline components through the seal area.
Further, tests have also shown that natural convective mixing
maintains a homogeneous liquid in the tank. This establishes the
format for one test approach. Periodic samples are taken and the
density of the bulk liquid remaining is determined with an
extremely precise density meter. Sensitivity is required, in
grams per millilitre, to five decimal places. Establishing a
time rate of density increase (increae due to loss of the lighter
components) permits calculation of decrease in bulk volume.
A test program, using the density change method and involving 13
gasoline tanks, was completed in 1977 by WOGA. Average evapor-
ation loss was about half of previously published values.
-4-
-------
J. G. Zabaga
A second test program, also completed last year, was conducted by
CBI/SOHIO. Here a 20 foot covered pilot tank, designed to
simulate climatic and product variables, was used to capture and
account for all hydrocarbon losses. Being capable of isolating
individual variables, the study program developed a clearer
understanding of the significance of the parameters affecting
evaporation. Wind in particular was determined to play a domi-
nant role in the evaporation process. The test program was
designed for a specific client's application, and within that
framework — the latest in tank and seal technology for storage
of crude oil — indicates emission levels at about 10 to 20
percent of previously published values.
Neither of these programs can be extrapolated to a data base
adequate to provide an emission predicting method for all types
of floating-roof tanks in all geographical and service condi-
tions. Therefore, CELM has developed a program to complement,
correlate and conclude the two efforts just described and to
develop new Bulletins 2517, "Evaporation Loss From Floating-Roof
Tanks," and 2519, "Use of Internal Floating Covers and Covered
Floating Roofs to Reduce Evaporation Loss," applicable to any
open top or covered floating-roof tank. The CELM program com-
bines density change and pilot tank techniques.
Phase I of the program is complete. It includes correlation of
volume change determined by the density meter with vapor measure-
ments of pilot tank losses, plus development of very specific
field and laboratory testing techniques. Phase II of the program
has recently been awarded to Radian Corporation. This work
involves field testing with the density meter to develop scale up
factors for the pilot tank measurements, which can then be used,
with appropriate operating variables, to develop the necessary
data base for a comprehensive understanding of floating-roof
emissions.
Completion of all work necessary to rewrite the API bulletins is
expected in 1979.
FIXED-ROOF TANKS
Current correlations for fixed-roof tanks are confined to
products with a vapor pressure greater than two psia, and due to
a limited data base, to tanks with diameters less than about 150
feet. Therefore, some crude oils, residuals and distillates
stored in larger modern tanks are excluded. It is desirable to
establish emission levels for these products.
-5-
-------
J. G. Zabaga
Last year WOGA completed a 46-tank test program on crude and
distillate liquids that provided some new data on emissions. The
study demonstrated the apparent significance of the way a tank is
operated, e.g., continuous in-out flow, fast filling and slow
emptying, etc., and the effect of vapor pressure on loss, and
generalized that emission levels are about half of API 2518
estimates. This agrees roughly with a recent German study that
indicates that filling losses are 88 percent and breathing losses
11 percent of API 2518 estimates. However, no new emission
formulas were developed by either program.
Two API committees, including CELM, are currently involved in a
joint program, building on these recent studies, to produce a new
prediction method applicable to production field tanks. This
work is to be finished by the end of this year. CELM plans to
expand the production tank study on completion, if necessary, to
develop new correlations suitable for fixed-roof tanks in all
service conditions,
A revised Bulletin 2518, "Evaporation Loss from Fixed-Roof
Tanks," is planned for publication in late 1978 or in 1979.
TRUCK AND TANK CAR EMISSIONS
Emissions from these mobile sources are the final area of CELM's
immediate concern. Advancements in loading methods, e.g., bottom
filling vs. submerged fill pipes or splash loading, and the
effect of vapor recovery units must be evaluated. The effects of
Stage I service station return vapors will also affect evapora-
tion loss. Much information on these items has already been
collected by industry and regulators, and EPA has published
emission factors for gasoline loading in the two Control Tech-
nique Guildlines on terminal and bulk plant operations.
CELM is currently analyzing all of the new loading data, includ-
ing some quite recent truck loading tests intended to provide a
more comprehensive data base. A totally rewritten Bulletin
2514B, "Evaporation Loss Prom Tank Trucks and Tank Cars," is
scheduled for publication late this year.
FUGITIVE EMISSIONS
A related program of interest is API's fugitive emissions study
being sponsored by the Environmental Affairs Department. It is
not a refinery emissions study since API is cooperating with EPA
and Radian in the conduct of their refinery program. The API
project is determining emissions from various equipment components
associated with the production of crude oil and natural gas.
-6-
-------
J. G. Zabaga
Data is being collected at twenty-one key sites which were selected
to provide coverage of products and operations in four geographic
regions. The Air Monitoring Center of Rockwell International is
performing the work which should be completed by the end of this
year. The results should offer interesting comparison with the
Radian findings since some of the devices being examined in the
two studies are similar.
SUMMARY
The CELM organization chart in Figure 8 shows the major areas of
evaporation loss activity of immediate concern to regulators and
industry.
The existing evaporation loss bulletins, originally prepared for
economic comparisons, have been shown to generally overstate
emissions, restricting their proper use to industry and for
control strategies. All of the pertinent bulletins are in some
stage of updating, with publication schedules being expedited to
make the new information available as soon as possible.
-7-
-------
American Petroleum Institute
STAFF ORGANIZATION
1
1 GENERAL COUNSEL
Stark Ritchie
1 1
ASST TO EXECUTIVE ASSOCIATION
VICE PRESIDENT LIAISON
J s Closb L T. Snowdon
1
STATISTICS
F. Mi rphy
SENIOR ADVISOR. 1 °«E"°« Of 1 I VICE PRESIDENT 1
1A80B RE1ATIONS 1 MANAGEMENT 1 1 f—t
_J AND BUDGET 1— 1 C T Siiwyur 1
1 ~ " ~~\ AFFAIRS ""*•
AUMINIblHAIIVt w M Lairil
jtnvictj ^
H J KuiKH MARKETING
1 -i B W C.cnl
CONTROLLER
NnM-DEPAHTMFNTAl
• ' ' - "V w «' ••<
'M"LO¥£E PRODUCT^
RELATIONS _
SERVICES — 1 B R Hall
VV Wik.'isty
| TRANSPORTATION
,- A 1 c V K LeOfMr. 1
i • • • •
( H Binim-i
INDUSTRY AFFAIRS
POLICY DEVELOPMENT t REGULATION
PRESIDENT
Frank N Ikard
I
[SECRETARY |
E E HanimefOeck I
|
ASST
TREASURER
£ C Beletski
1 EXECUTIVE 1
VICE PRESIDENT 1
Charles J DiBona II 1
COMMITTEE CONSERVATION
ON INDUSTRIAL LIAISON
f^B(« ANIMATION
J Atkin R u ™v
1
EXTERNAL LIAISON
M D Garber
•
HVICE PRESIDENT | [ VICE PRESIDENT | | VICE PRESIDENT |
D B Ratnbun 1 1 P G Goufdincj |
ENVIRONMENTAL £ EDITORIAL ">
•« AFFAIRS
A E Gubrud ^ J K'n9 S.
< 1
FEDERAL AGENCIES a CIO S. SPECIAL C
A U MOOIL ^ R Fryklund a
FINANCE AND PUBLIC RELATIONS
1 S P Potter 1
FEDERAL RELATIONS
C E S;indHjr
TAXATION
T A M.trlin
STATE RELATIONS
R H Slewart H H Hardy F J JarnKowil/
C,DC «. CiCCTV
•_ COORDINATION 1
1 ' " """""" 1 CENTRAL REGION
MEO & BIO. SCIENCE E H Sl.r.ims
EASTERN REGION
N C . Hffiwvn
N K Weaver 1
POLICY ANALYSIS ™( ™J' ™
G T Pnlion (o MO vVist,
Cl M.lilS f'rl
D'-l NH f
-------
API COMMITTEE ON PETROLEUM MEASUREMENT
JAN. 1. 1978
COMMITTEE ON
PETROLEUM MEASUREMENT
CHAIRMAN
8 MESSER.JR.
O
t-J
ta
cr
(B
OQ
w
COMMITTEE ON
STATIC
MEASUREMENT
I IIAIHMAN
Ji HIllltUDH
S/C ON NBS
PHYSICAL PROP
DATA PROJECT
Cl IAIRMAN
fl MUhtH
API BOARD COMMITTEE ON
PUBLIC ISSUES
CONTACT
fl THOMAS
STANDARDIZATION
VICE-CHAIRMAN
K. BAILEY
I
COMMITTEE ON
DYNAMIC
MEASUREMENT
CHAIRMAN
E MCALLISTER
API/ISO/OIML
S/CONP.S.5
Cl IAIRMAN
API NON DEPARTMENTAL
INDUSTRY AFFAIRS STAFF
WN SEWARD
JK WALTERS
COMMITTEE ON
NATURAL GAS
FLUIDS
MEASUREMENT
CHAIRMAN
0 KEMP
*
S/C ON AGA 3
ORIFICE METER
PROJECT
CHAIRMAN
E 8UXTOH
COMMITTEE ON
EVAPORATION
LOSS
MEASUREMENT
CHAIRMAN
J. ZABACA
**
S/C ON DATA
FLOATING
ROOF TANKS
CHAIRMAN
G GOOD
-
.
SPECIAL PROJECTS
VICE-CHAIRMAN
AE 8RYSON
I I
COMMITTEE ON COMMITTEE ON
MARINE LOSS PERSONNEL
CONTROL TRAINING
CHAIRMAN CHAIRMAN
R GRIFFITH R. BOYLE
I
ADVISOR TO US
A/C OIML
A.H HALL
I I
COMMITTEE ON
U.S. INT'L. TASK GROUP
TRADE ON LONG RANGE
COMMISSION PLANNING
CHAIRMAN CHAIRMAN
L DOOGION K. BAILEY
I
ADVM™]?AP' COMMITTEE ON
MtlKIO PROGRAM
SEE *•"*—
CHAIRMAN
A.H HALL
*UN«. t A'.Ml II I Hit M AHOi I'HO.U I I MANAiiLMi.NI OHOUP
*MNI i -\','.,ii ii n ni si Aiti M morosAi iinour
Figure 2
-------
o
i
API EVAPORATION LOSS BULLETINS
<-,
•
API BULLETIN 2512: TENTATIVE METHODS OF MEASURING EVAPORATION LOSS ^
FROM PETROLEUM TANKS AND TRANSPORTATION EQUIPMENT (1957) S-
OQ
API BULLETIN 2513: EVAPORATION LOSS IN THE PETROLEUM INDUSTRY—CAUSES »
AND CONTROL (1959)
API BULLETIN 2514: EVAPORATION LOSS FROM TANK CARS, TANK TRUCKS, AND
MARINE VESSELS (1959)
API BULLETIN 2515: USE OF PLASTIC FOAM TO REDUCE EVAPORATION LOSS (1961)
API BULLETIN 2516: EVAPORATION LOSS FROM LOW-PRESSURE TANKS (1962)
API BULLETIN 2517: EVAPORATION LOSS FROM FLOATING-ROOF TANKS (1962)
API BULLETIN 2518: EVAPORATION LOSS FROM FIXED-ROOF TANKS (1962)
API BULLETIN 2519: USE OF INTERNAL FLOATING COVERS FOR FIXED-ROOF TANKS
TO REDUCE EVAPORATION LOSS (1962)
API BULLETIN 2520: USE OF VARIABLE-VAPOR-SPACE SYSTEMS TO REDUCE EVAP-
ORATION LOSS (1964)
API BULLETIN 2521: USE OF PRESSURE-VACUUM VENT VALVES FOR ATMOSPHERIC
PRESSURE TANKS TO REDUCE EVAPORATION LOSS (1966)
API BULLETIN 2522: COMPARATIVE METHODS FOR EVALUATING CONSERVATION
MECHANISMS FOR EVAPORATION LOSS (1967)
API BULLETIN 2523: PETROCHEMICAL EVAPORATION LOSS FROM STORAGE TANKS
(1969)
Figure 3
-------
COMMITTEE ON
EVAPORATION
LOSS
MEASUREMENT
CHAIRMAN
J.G.ZABAGA
IN
P
cr
(U
API STAFF
SPECIAL LIAISON
J.R. ARNOLD
SIC
ON
FLOATING-
ROOF TANKS
2517
G.J. GOOD
SIC ON
INTERNAL
FLOATING
COVERS
2519
R.C. KERN
SIC ON
FIXED-ROOF
TANKS
2518
A.D. KUBAL
SIC OH
TANK CARS,
TANK TRUCKS
251 4B
W.K. BURNS
SIC ON
MARINE
VESSELS
251 4A
T.A. HECHT
Figure 4
-------
j. G. Zabaga
FLOATING ROOF
S/C 2517
te o o o o
INTERNAL DECK
S/C 2519
FIXED ROOF
CLOSED TOP &
NO FLOATING ROOF
TRUCKS OR TANK CARS
I T
SPLASH
LOADING
S/C 2518
S/C 2514
OR
I I
BOTTOM
LOADING
I I
SUBSURFACE
LOADING
BARGES
Figure 5
- 12-
-------
c_
O
API 2514A
MARINE VESSEL LOADING OF GASOLINES
SUMMARY OF AVERAGE HYDROCARBON EMISSION FACTORS
(a
a-
VESSELS
SHIPS
BARGES
ARRIVAL
CONDITIONS
CLEANED
UNCLEANED
CLEANED
UNCLEANED
NO. OF COMPART-
MENTS TESTED
50
21
1
11
EMISSION FACTORS
(lbs./1000
GALLONS LOADED)
1.3
2.5
1.2
3.8
Figure 6
-------
J. G. Zabaga
A. FRT EMISSIONS
FLOATING ROOF
SEALAREA
TANK
SHELL
-EMISSIONS
LIQUID LEVEL
B. TESTING METHODS
1. DENSITY CHANGE
DORMANT FIELD TANKS
DENSITY
GM/ML
6-12 SAMPLES
2-4 WEEKS
TIME
2. PILOT TANK
AIR
HC&CFM
• VARY PRODUCT, TEMPERATURE,
PRESSURE, WIND EFFECT & SEAL
CONDITIONS
• COLLECT & ANALYZE EMISSIONS
Figure 7
-14-
-------
PLANNED PUBLICATION
OF UPDATED BULLETINS
EARLY 1979
S/CON
FLOATING
ROOF TANKS
2517
G.J. GOOD
API STAFF
EARLY 1979
f
S/CON
INTERNAL
FLOATING
COVERS
2519
R.C.KERN
COMMITTEE ON
EVAPORATION
LOSS
MEASUREMENT
CHAIRMAN
J.G. ZABAGA
LATE 1978
C-,
O
S/CON
FIXED-ROOF
TANKS
2518
SPECIAL LIAISON
J.R. ARNOLD
1978
LATE 1978
i
S/CON
TANK CARS,
TANK TRUCKS
2514B
W.K. BURNS
i
S/CON
MARINE
VESSELS
2514 A
T.A. HECHT
Figure 8
-------
J. G. Zabaga
QUESTIONS AND ANSWERS
Q. Frank Mesich, Radian - You said excessively expensive, how expensive?
A_._ - Right now, we are pushing $5,000,000. That doesn't count the
fugitive emission study that you are working on for EPA. I am talking just
about CELM's involvement, and the work that is being done by WOGA, SOHIO and
CB&I, all of which is feeding into this enormous data base. But, it is around
$5,000,000. It may not sound like much for a multi-billion dollar industry,
but when you start breaking it down to individual companies and individual
contractors, it has created some interesting problems and justifications.
-16-
-------
VALVES - A POSSIBLE SOURCE OF FUGITIVE
EMISSIONS IN HYDROCARBON PROCESSES
Alton M. Williamson
Director of Engineering and Research
W-K-M Valve Division
ACF Industries, Inc.
Houston, Texas
ABSTRACT
With recent knowledge of concerns of some for Fugitive Emissions in
Hydrocarbon Processes, the Valve Industry addresses this issue. Valve
types, applications, and potential leak points are discussed. Installation
practices, misapplications, maintenance and purchasing practices are
reviewed. Existing Standards requirements and improvements in packing
materials and seal configurations are studied. Other possible causes
outside the valve manufacturers' control and resulting consequences are
further reviewed.
RESUME
Alton Williamson received his B.S. in Mechanical Engineering from Texas
A&M in 1954. He has taken graduate studies in Business Administration at
the University of Houston and is a Registered Professional Engineer in the
State of Texas. In 1974 he was graduated from Cornell University from their
Executive Program for Advanced Management Training. For the past six years
he has been employed at the W-K-M Valve Division of ACF Industries. He
is now the Director of Engineering and Research. Prior to that, he acted in
the capacity of Chief Engineer for nine years at a major valve company and
has extensive training and experience in the field of materials, non-metallics,
machining, and fabrication processes. Mr. Williamson is an active member of
the Valve Manufacturer's Association serving for the past three years as
Chairman of the Technical Coordinating Committee and representing VMA as Chief
Technical Officer of their organization. He presently is Chairman of the
Nuclear Products Committee of VMA and a member of ASME, NACE, and NMA.
In his various positions, he has been instrumental in the research, design,
and implementation of new valve lines including butterfly, gate, check and
ball valves and, as an inventor, has several patents in the valve-related
field.
-------
A. M. Williamson
VALVES - A POSSIBLE SOURCE OF FUGITIVE EMISSIONS
IN HYDROCARBON PROCESSES
Gentlemen, I consider it a real pleasure and opportunity to
represent the United States Valve Industry and discuss with you
today the role that valves may play in Air Pollution through Fugi-
tive Emissions in Hydrocarbon Processes.
When the VMA was contacted around the first of the year con-
cerning an industry speaker for this Symposium, we were first
alerted of pending concerns and possible studies that were to be
conducted. At our Spring business meeting in Washington, D. C.,
in January of this year, we thoroughly discussed the request for
the speaker and the pending Symposium. We further inquired of all
company members of requests for information from their customers
concerning this subject. Surprisingly, few had been contacted in
any regards about the subject. Now that we have been alerted, we
stand ready to face the issue squarely and seek the truth. Through
our 72 member companies, we represent the majority of the United
States valve production, and manufacture every style of industrial
valve. The VMA has earned the solid reputation of being responsive
to the needs of our industry, including our customers, government
at all levels, suppliers, codes and standards writing groups, and other
regulatory agencies. We join you in investigating this possible pro-
blem and welcome the opportunity to work with you in the future.
Now, let's look at the possible problem and, specifically, at
valves:
I read and studied with interest the paper "Fugitive Emissions -
Current and Projected Studies"^-, by Donald D. Rosebrook and Dr. Robert
Wetherold of the Radian Corporation, Austin, Texas. 'This paper was
just presented at the March, 1978 meeting of the National Petroleum
Refiners Association in San Antonio, Texas. This paper makes many
good points that merit my commenting.
-------
A. Williamson
First, the petroleum refining industry represents a major po-
tential source of hydrocarbon air emissions in the United States.
Second, emissions can be classified as controlled or uncontrolled;
controlled emissions being defined as those that result from stacks and
vents and can be readily identified and quantified by conventional
means. Uncontrolled or fugitive emissions result from leaks which
can occur in virtually any hydrocarbon service. This type of emission
is the one we are discussing in relation to valves.
Third, valves are listed as one of the main potential sources of
fugitive emissions. One major reason is that there are so many valves
in this type of service in each plant. They are also one of the easi-
est pieces of equipment to isolate. By using a special bagging proce-
dure, a fairly accurate leakage rate is said to be measured; within t
11% on methane and JT 10% on propane.
Fourth, we are generally discussing "non-visible and non-audible"
leaks. This is an extremely important point. As an example, valves
leaking at 150 ml/hr or less in hexane service are considered negligi-
ble. This quantity equates to .03 Ib/day. Special "sniffers" used
have found that unseen vapor leaks of 200 p.p.m. equates to 150 ml/hr
or .03 Ib/day. The real problem thus factors down to the very high
number of valves used in refineries and associated industries. Over
10,000 valves per refinery is not uncommon.
SLIDE - This slide is a typical scene of the complex piping system
found in so many refineries or associated hydrocarbon processing
plants. Many, various styles of valves are used in this piping system.
Before we discuss prior or future proposed research into this sub-
ject, let's look at some of the various valve styles and potential leak
points in the valves. A discussion of various valve end connections
will come later.
Most valves have operating stems projecting from at least one
point through the valve body. In addition, there may be body lubrica-
tion, by-pass or drain connections that form other pressure containment
-------
A. Williamson
penetration points to be considered as possible points of leakage.
These body lubrication, by-pass or drain connections are common in
many market areas that utilize ball, gate, swing-check and globe
valves.
SLIDES - The next two slides show Wafer and Swing-check Valves.
Wafer-check valves usually have only one or two possible leak points
excluding the flange faces. These are usually assembly holes that
are used to assemble the main shaft and/or stops in the valve. These
holes are usually tapped with an NPT pipe thread and plugged effect-
ively against fugitive emissions.
Swing-check valves may have screwed, bolted or clamped bonnets, de-
pending upon the design. The clapper shaft may penetrate the pres-
sure containment boundary in the body. Again, body fittings may be
employed as by-pass relief around the clapper, or for other purposes.
The bonnects may have metallic or non-metallic seals, many being sim-
ple flange-type connections.
SLIDE - The Diaphragm or Weir-type ¥alve.
This valve has a screw-type stem with either fixed or adjustable seals
around the stem and static seals at the bonnet.
SLIDES - Plug Valves.
Plug valves can be cylindrical or tapered, top or bottom entry, with
fixed or adjustable seals and packing. This first slide shows a bot-
tom entry valve with a metal seal at the entry opening of the plug.
The t o p of the plug is sealed with a simple dynamic seal that is
both pressure energized and mechanically energized by the plug support
spring.
This second version shows a top entry valve with adjustable packing
gland. An external injectable sealant provides the downstream seal
and assists in sealing externally in both versions. The lubricant in-
ject4-on system is effectively sealed with check valves to prevent emis-
sions backward to the atmosphere.
SLIDES - Globe Valves.
The Globe valve is one of the oldest and most used type valves.
-------
A. M. Williamson
SLIDE - The globe valve also is capable of being furnished with fixed
or adjustable packing around the stem. Its bonnet also can be of var-
ious types of construction.
SLIDE ~ M2111? types of globe valves, in both straight through and
angle-type, are excellent in throttling service and can be a fixed
position handwheel-type valve or an automatic regulating diaphragm-
actuated valve.
SLIDES - Butterfly Valves.
SLIDE - Most Butterfly Valves have stems penetrating both the top and
the bottom of the body. The most popular butterfly valves today, are,
of course, the wafer style. Wafer-style valves may have flange seals
built into the sealing element of the valve, itself. Many butterfly
valves have 0-rings, or other sealing elements between the stem and
body. They also may have primary seals between the disc and the seat
or sealing element in the valve, with additional 0-rings on the stem
or shaft as secondary sealing members.
SLIDE - This body construction is also popular in the wafer-type check
valve, which we looked at earlier. The wafer-style valve comes in sev-
eral configurations, but basically, the true wafer is one that simply
slips between the flanges and is centralized by the bolts in the flanges,
and held in place by the compression that flange bolting exerts on the
valve body.
THREE SLIDES - There are three other thin body styles which are common-
ly referred to as the semi-lug body style, lug-body style, and the
flanged body style. The lug-body style is a wafer-type valve, that has
threaded lugs around the periphery of the body, that allows bolts or
studs to be utilized to hold the valve to a single flange. For low pres-
sures, this style or the flanged style valve may be utilized as a Stop
valve at the end of the line, depending on the individual design.
SLIDE - The Butterfly valve can also be. furnished with adjustable pack-
ing glands around the stem of the valve.
SLIDE - Some designs also have pressure-energized seats, which affect
sealing from either direction in the valve.
-------
A. M. Williamson
SLIDE - This slide shows a common seal for the lower end of a high
performance ANSI Class Butterfly Valve. This particular type valve
is one of the fastest growing valve styles in the industry,
SLIDE - Ball Valves are commonly referred to as spherical plug valves.
Like all valves, they come in assorted styles, shapes, and configura-
tions. They, like the butterfly valve, may have stem seals of simple
0-rings, or other fixed packing, as well as adjustable outside pack-
ing boxes.
SLIDE - Ball valves may be of the floating style, where the weight of
the ball is carried by the sealing elements and line pressure forces
the ball against the downstream seat.
SLIDE - The shaft or operating stem protrudes through the top of the
valve and is sealed at the body. This type valve is generally a down-
stream sealing valve.
SLIDE - There are also trunnion-type ball valves, where the ball is
carried in a relatively fixed position by the shaft extending through
the top and bottom of the valve. This style has seats which are ener-
gized by pressure. The valve is generally an upstream sealing valve.
The ball valve type is one of the most popular styles of valve being
utilized in all types of industry today. Its future in hydrocarbon
processes will continue to be brightened as further progress in tech-
nical advances are made in the development of this style valve.
SLIDES - These slides show various styles of Gate Valves. Basically,
they are variations of the wedge-type gate valve, and a through-conduit
gate valve.
SLIDE - The basic difference between the two being that in the through-
conduit gate valve, the seats are protected in both the full open and
the full close position.
SLIDE - Most gate valves are furnished with an adjustable outside pack-
ing at the stem, with some low pressure styles for less critical ser-
vice being furnished with 0-rings or other fixed type packing.
SLIDE - The packing glands may be conventional stuffing boxes, with a
myriad of various types of materials to choose from. However, they can
also have external, injectable plastic packing sealing capability, as
-------
A. M. Williamson
we will see later. They can also have a way to remove any minute
leakage around the stem, for disposal through a waste-process sys-
tem.
SLIDE - Bonnets, again, vary from clamped, screwed or flanged, with
the most popular by far being the flanged bonnet with either metallic
or non-metallic-type of gasket. This valve is furnished many times
with body fittings, for either purging operations, by-pass around
the gate, or simply a drain or for leak checking or other purposes.
Up to this point, we have discussed, basically, shut-off or
block-type valves. Although a simple on/off block valve can also,
in many cases, be used for throttling, there is a special flow rate
control-type valve. These are commonly referred to as Control Valves.
They may be a globe style, a butterfly, a ball, or some other configu-
ration, but they're designed and sold primarily for flow control. In
addition, we have seen that there are flow direction control valves
which are commonly referred to as check valves. These can be referred
to as a one-direction stop-type valve.
SLIDE - In addition, there is one other type valve, as shown in this
slide, which is commonly referred to as a Pressure-Relief Valve. These
valves are block valves, and are intended to open upon a predetermined
over pressure condition. They may or may not be piped to a waste collec-
tion or process area. If not, their internal sealing means becomes a
source for fugitive emissions investigations.
Now that we've looked at styles of valve construction and the bas-
ic types and some of their general applications, can we draw any conclu-
sion? Basically, the conclusion can be made that although valves have
varying degrees o-f potential leak points, the main point of susceptabil-
ity for a valve in a closed system, that is one that is not being used
on the end of the line or as a safety relief valve, becomes the stem
packing area. Bonnet flanges are a possible source of fugitive emis-
sions also, but generally, not subject to exaggerated movement, except
for process induced vibrations, and consequently, less of a suspect for
leakage.
-------
A. M. Williamson
Let's touch on end connections of valves. Basically, these
may be flanges, threads, welds, or clamp hubs. These connections
are not limited to valves, because there are many more points of
leakage in a refinery or process plant piping system than at the
valves. One of these multitude of potential leak points is flanges.
SLIDES - The next two slides should be readily recognizable to most
of the people here. They are merely design criteria of the various
styles of gasket materials and contact facings, or flanges, as de-
picted in ASME, Section VIII Code. These are readily acceptable in
industry today, when utilized with standard flanges. Many have pro-
ven themselves when tested under extreme high pressure gas and with
helium. It is my suggestion and recommendation to those who propose
to undertake fugitive emissions leakage studies for the future, that
they also isolate this component from other equipment, such as
valves and pumps. I think we ought to know what truly is or is not
causing problems of hydrocarbon emissions. The same general comment
can be made for standard pipe thread connections and any other spec-
ial clamp or threaded connections, where utilized in process piping
for whatever reason,
Now let's look at the possibility that valves may be a major
source for fugitive emissions. What would cause this? Valves manu-
factured from the majority of the U. S. valve manufacturers today will
meet any standard industry recognized test that you care to give the
product at the time of acceptance. That is, the product will meet its
intended use. Therefore, if valves are actually a problem in fugitive
emissions, it must be something that happens after shipment that
causes the problem. I submit that the major problem is age, lack of
proper maintenance and misapplication. There are many installation
practices and purchasing and specification practices that are question-
able. Too many purchasers of valves look upon all types of valves as
simply commodity items, that can be installed with any process, as long
as the combination of pressure and temperature is within the range as
established by the manufacturer. This is simply not true. The results
are often premature failures, caused by the misapplication, and many
-------
A. M. Williamson
times, the valve manufacturer is not even called in to help or even
notified of the problem. This is one of the biggest hindrances in
user/manufacturer relations that stymies further development and
progress in valve improvements. It is imperative that valve manu-
facturers and their customers work closely as a team for the bene-
fit of both in the future.
We are all getting older. As I grow older, this old body be-
gins to wear out. My automobile has to be serviced regularly. Even
parts that are not discussed in the automobile repair and service
manuals, wear out due to age and lack of maintenace. With, in many
cases, over ten thousand valves in a large refinery or process plant,
and with automation coming of age as it has, how do you perform the
maintenance on all of the valves that you have in a plant; much less
set up a preventive maintenance program? From the manufacturer's
standpoint, I recognize that this must be an enormous task. But re-
member, we're discussing emissions that normally can't be seen, nor
heard. Therefore, the problem of maintenance becomes extremely im-
portant.
This brings up other problems that need to be discussed. Most do-
mestic valve companies know their products extremely well. They don't
guess at the recommendations that they make, as their experience in
both laboratory and field testing, along with many years of successful
product applications, give them confidence in their recommendations.
They recommend spare parts for valves. These spare parts are "genuine"
spare parts as made by the American valve manufacturer. They are ex-
pected to and do work in the valve for which they were manufactured.
Outside of the basic valve manufacturing industry, there are many, what
I refer to as, "backyard valve repair shops". Many of them have names
that say "valve manufacturer". They are in the business to service
valves, most of which they do not make. They sell these services on
the basis that they know more about valves and their applications than
the valve manufacturers do. They cannot know the need for exact parts
size, much less tolerances, the required surface finish, or necessary
configurations of the various parts, nor proper maintenance intervals,
procedures or proper materials. Any yet, they are utilized by legiti-
-------
A. M. Williamson
mate, major companies in the U. S. every day. When most valve
manufacturers not only stock and sell original and guaranteed parts
for their products, they generally have a well-educated and trained
valve service department of their own. Many companies have these
service organizations world-wide and can give quick and professional
service.
In addition, there are also companies now selling "on-line"
leak sealing capabilities. They are found selling their services
in refineries, process plants, and yes, even fossil fuel and nuclear
power plants. These service companies come into a plant, by invita-
tion. Without valve manufacturers' knowledge, they will drill holes
into areas of packing glands and stem sealing areas, so that some
kind of a special seal can be pumped into the leaking area to stop a
leak. Since a packing gland feature is available in many domestic
valves as a standard or an option that negates the need for this ser-
vice, why should it be necessary after the valve has been installed,
to ever perform such an emergency service? Was it a misapplication?
Was it a poor choice in the valve for the service? Again, the valve
manufacturer is seldom, if ever, contacted concerning this problem.
Other purchasing practices must be questioned. First consider
the fact that many large, honorable valve user companies have explicit
performance requirements for valves that they approve and list in
their Engineering Standards. It requires extensive test time on the
part of the valve manufacturer to prove that his product will, in
fact, meet the criteria. When it does, he expects to participate in
the business available within the boundaries of the free enterprise
system. I can assure you that often he cannot, because the buying
practices of companies often completely ignore their own Engineering
and Quality Assurance Standards for use of acceptable quality valves.
This is a cause of valve misapplication in process plants. This, also,
brings in an even more important issue - Foreign Produced Valves.
I represent the United States Valve Manufacturers Association. I
know for a fact that foreign manufactured valves are a source of tech-
-------
A. M. Williamson
nical problems in process plants. I will not even address the eco-
nomic problems, even though they also greatly affect the valve
user. I will also agree that there are some excellent quality
foreign-produced valves. But, have they passed equal quality test-
ing programs as required of U.S.-produced valves? Is technical sup-
port, both in the form of qualified people, spare parts and service,
readily available to the user?
A large refinery was persuaded to purchase over $3,500,000
worth of foreign valves. I ask you, what could be the fugitive
emissions problem from the following factual results?
(1) Casting porosity in several valves that worsened as pro-
duction temperatures and pressures were applied.
(2) Valves developed "stretched" or "bowed" yokes.
(3) "T" head assembly sheared off and disengaged from sealing
member.
(4) Valve bodies were worked and stretched to a point where repair
was impossible.
(5) Sealing members seized in position resulting in stems breaking
or stripping of stem bushing, not threads.
We won't even take time to discuss the liability problems of the
prior examples. But to reduce all types of leakage, whether miniscule
fugitive type or other, and to reduce risk of a fire, explosion, or other
disaster in a plant, you must have the best quality valves that you can
get and depend upon.
Let's look now at a few special modifications and some results of
other work done in the valve industry leading toward improved equipment.
SLIDES - These next two slides show various types of seals that can be
utilized in valve bonnets, flanges or any area where two mating parts
must fit together and form a seal. There are so many excellent designs
being utilized today, that it is impossible to list even more than
these few. You will recognize that several of these, such as those
marked with the asterisk, are pressure-energized seals. The basic idea
of utilizing pressure-energized seals is one that should be pursued
10
-------
A. M. Williamson
more and more in the future.
Much work has also been done in improving basic seal materials,
style and configuration. Many excellent variations are available,
with the only real hinderance being one of past specification and
established purchasing habits. Excellent work has been done in lami-
nates, ribbons and filament braid packings.
SLIDE - This slide shows a schematic of a grease or sealant injection
system at the lantern ring, in the packing area around the stem of
any conventional valve. The slide says "Gate or Globe Valve". How-
ever, it's applicable to any type valve.
TWO SLIDES - Here is a slide of an actual valve design, while the
next slide shows a standard packing box, both with and without this
feature. This is a standard feature in many valves, and is probably
not used on more valves for only two reasons: (1) customers decide
they don't need it and (2) it makes the product cost more. Probably,
there are places where it is not needed, but here is one simple
scheme where all you have to do is inject a sealant under pressure,
and you eliminate leakage just due to keeping a higher differential
pressure in the packing than you have on the inside of the valve.
SLIDE - This slide shows a typical wedge gate valve, utilizing a
pressure-energized bonnet seal.
There is another idea that has been used in industry for years,
although you don't see it on many standard valves in plants today.
That idea is to eliminate the packing around the stem on the primary
seal by the use of bellows or other means.
TWO SLIDES - These next two slides show typical straight through
and "Y" pattern globe valves with bellows sealed bonnets. This is
sometimes referred to as a "packingless valve". This does not mean
that there is not a potential leak path at a bonnet.
There are also all welded valve body designs on the market that
have readily been used in general pipeline transmission service.
These designs eliminate body joint connections such as flanges.
11
-------
A. M. Williamson
We have discussed fugitive emissions as a possible problem,
defined minimum suggested leak rates, looked at basic valve de-
signs, possible leak paths, valve user contributed problems, and
a few valve seal improvements. But we must now ask the question,
"Is the need real"? How far are we to take investigations of
fugitive emissions? What price is the United States of America
ready to pay for this type of emission control? Are we going to
eliminate all of our forests and woodlands? Trees give off hydro-
carbons, also. The Blue Ridge Mountains are "Blue" and the
Smoky Mountains "Smoke" for a reason. We can solve any problem
for a price.
Time, today, simply didn't allow an adequate coverage of such
an immense subject from the valve industry point-of-view. Trying
to consolidate into a 20-minute talk, an adequate discussion of
this magnitude, is extremely difficult. However, I can tell you that
our valve members are eauiooed with excellent research facilities,
testine laboratories, and the proper manpower to make use of the .
facilities. I can also assure you that the valve manufacturers of
the United States want to do more than to sell valves. They want to
offer service, as well. We have n2vi?r ducked a challenge. We face
all issues squarely. I assure you that valves being built today are
better than ever before, with new developments every day to make
seals, packings and sealing elements more reliable, longer lasting,
and overall more dependable than ever before. I must, also, leave
with you the honest conveyance .that the valve industry wants to work
for improvement of its products. Your assistance is appreciated.
12
-------
A. M. Williamson
ACKNOWLEDGEMENTS AND REFERENCES
•"•Rosebrook, D_r_. Donald D_. and Wetherold, Dr_. Robert, "Fugitive
Emissions - Current and Projected Studies", from a paper
presented at the 1978 NPRA Annual Meeting, San Antonio,
Texas, March 19-21, 1978.
^Acknowledgement is made to the Officers and Member Companies
of the Valve Manufacturers Association for their assis-
tance in supplying information for this paper.
13
-------
Alton Williamson
do
QUESTIONS AND ANSWERS
Q. - J. Daily, Chevron - On the bellow valve, what percentage
you sell and what is the operating limit?
A. - I can't answer that, I have to speak for a group of 72 that I
represent, and I don't make a bellows seal valve and I would defer to
somebody else.
Q. - A. T. Kott, TAGS - I have a question on the cost of spray packing
versus putting a little chamber in there with a tap like the cooling
chamber you have versus the cost of the bellows. Are the bellows half
again the cost of the valve?
A. - I can't speak for the bellows. I know they exist. I know that they
are not widely used. It is a possibility.
Comment - R. Smith, Xomox - The bellow seal valve is very expensive and
is only used basically in the globe-type of construction. You have to
control the amound of head due to the fact that you could overstress the
bellows. The valve is only used in very exotic circumstances and is
primarily in small size valves. Two inches are relatively large bellow
seal valves. I don't believe the bellow valve is particularly applicable
to the industry.
A. - Your first two parts of the question I can readily answer because
I am very familiar with it. It really doesn't cost that much more to put
the extra feature in there to be able to have plastic injectable packing.
Every oil well valve has got them and I predict that when you measure
emissions out there, you are going to find that the fugitive emissions
around oil wells, whether offshore or onshore, are much samller than
what you have found so far, because of this type of valve. Plus everything
is a metal-to-metal seal otherwise.
Q. - R. Vincent, GARB - In your opinion, would current valve technology
and the expertise of the manufacturers in recommending certain equipment
and certaim maintenance practices suffice to essentially eliminate valve
leakage in refineries and related industries?
A. - I believe it would. If we tested in our plant, tommorrow, just
before a valve left, I believe we'd find that these new valves do not leak.
Even if they are the simple low-pressure 0-ring seal valve.
Q. - R. Vincent - Do you think this would entail any significant
extra capital expense on an on-going basis for refineries? Are we talking
about fancy valves and expensive materials?
A. - It is going to cost them a little more, but not so much from the
valve'standpoint as it is from other areas that they have already addressed.
That is, they are working hard on maintenance, more than just tightening down
nuts. But, when you get to rebuilding valves that you have got to shut down
for, those are real problems.
-14-
-------
PLANS FOR ASSESSMENT OF WATER AND RESIDUALS EMISSIONS
FROM THE PETROLEUM REFINERY
Fred M. Pfeffer
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma
ABSTRACT
The Environmental Protection Agency will prepare a document which assesses
the emissions from a petroleum refinery via wastewater effluent discharge and
residuals from wastewater treatment. The mechanism for preparing this report,
the topics covered, the factors considered in making the assessment, and some
of the information to be utilized in preparing the assessment will be discussed.
RESUME
Our final speaker for the program is Fred Pfeffer from IERL, Ada,
Oklahoma Laboratory. Fred got his B.S. and M.S. Degrees in Chemistry from
the University of Cincinnati and he has over 15 years professional experience
in wastewater treatment with EPA and its predecessor organizations.
-------
Fred M. Pfeffer
PLANS FOR ASSESSMENT OF WATER AND RESIDUALS EMISSIONS
FROM THE PETROLEUM REFINERY
The EPA Office of Research and Development will prepare an assessment of
the emissions from the industrial point source—the petroleum refinery. Two
EPA laboratories are cooperating in this effort and will assemble a final EPA
document: IERL, RTP on air emissions; the Robert S. Kerr Environmental
Research Laboratory on emissions from wastewater effluents and from the
residuals from wastewater treatment. The products of the two laboratories
will be independent parts of the overall assessment; e.g., documents a_, ID, c_
carrying the same EPA report number.
The reports on wastewater and residuals emissions will be based upon
existing information and will contain chapters on the following topics:
a) State-of-the-Art Discussions
1. A characterization of and future trends for the petroleum refining
industry, including the parameters affecting generation of pollutants such as
crude source, process units, operation (refinery classification), upsets,
turnarounds, and wastewater treatment systems in use at the refinery.
2. An evaluation of existing and emerging treatment and control tech-
nology applicable to refining wastewaters.
b) Rationale for Characterizing the Wastewater and Residuals Emissions
Several different approaches have been or are being considered to
assess the pollution potential of refinery waste discharges. In the
case of wastewater, the classical permit parameters such as COD, chromium,
suspended solids, sulfide, cyanides, and oil and grease have been in
common use for such assessments. Use of the more recent "priority
pollutants listing" is another approach. Other concepts are effects-
oriented, using such techniques as fish toxicity, bioassay, biodegradability
(or persistence), carcinogenicity, mutagenicity, and teratogenicity. A
-1-
-------
Fred M. Pfeffer
rationale considering all of these assessment techniques will be presented
in the form of a pollutant "punch" list. In the case of residuals,
leachability of toxic or hazardous substances is an important consideration.
Any pollutant's assessment is method-dependent; therefore, a measurement
procedure will be referenced for each pollutant defined.
c) Emission Data
Available information on the occurance, quantities, sources, and
avenues of emissions of pollutants will be presented. Wherever possible,
this compilation will include accuracy, precision, and variability for
both physical and chemical measurements techniques.
d) Criteria
Criteria will be presented for the current environmental impact of
pollutants, establishing levels in emissions and in the environment. Use
will be made of existing regulations, not necessarily limited to those of
the EPA.
e) Environmental Assessment
A comparison will be made of pollutant emissions to the criteria,
and a statement as to level of treatment and control (T&C) requirements
for obtaining criteria levels will be attempted for each pollutant.
f) T&C Evaluation and Research Needs
Existing and emerging T&C systems will be evaluated relative to the
acceptable emission levels, and T&C research needs will be identified.
The mechanism for producing these reports will be a research grant
utilizing recognized experts for the preparation of each chapter. A
consortium of experts will be assembled to prepare the criteria for
environmental impact and acceptable pollutant levels.
Data available to the grantee is largely composed of wastewater
emissions information (as opposed to residuals) and is the result of the
Robert S. Kerr Environmental Research Laboratory's program of T&C research
in the refining category conducted in cooperation with various API
committees. The major studies are as follows:
a) Raw waste load and final effluent characterization at refineries
representing 85% of domestic refining capacity was completed in 1972.
-2-
-------
Fred M. Pfeffer
•,
The study was specific for 22 common wastewater parameters and included
an investigation of the treatment efficiencies across five activated
sludge systems. The following tables show the classes of refineries
monitored (Table 1), the parameters measured (Table 2), the performance
of activated sludge treatment (Table 3), and final effluent loadings (Tables
4, 5).
b) In December 1976, a study was conducted at SOHIO (Toledo) to determine
the major individual organic compounds entering and leaving each of
three treatment systems in series: full-scale activated sludge; pilot
scale mixed-media filtration; and pilot scale activated carbon. The
study was conducted through an interagency agreement with ERDA (Argonne
Laboratory, Chicago); the draft final report is under review. Prior to
utilization for the assessment, Argonne's analytical results will be
compared to split sample results of EXXON R&E (for the API) and SOHIO.
Figure 1 shows the sampling points in the study.
c) The Robert S. Kerr Environmental Research Laboratory's participation
in a "Priority Pollutants" study for EPA's Effluent Guidelines Division
(EGD) was completed last year. Twelve refineries were sampled to character-
ize and semi-quantitate the priority pollutants entering and leaving the
treatment system. Analytical results from laboratories under contract
to EGD are not as yet complete. Table 6 shows the refineries visited;
Table 7 presents the priority pollutant sampling information. The API
sponsored similar studies at four refineries and has indicated that this
information may be considered a source for the assessment.
d) Also for EGD, the Robert S. Kerr Environmental Research Laboratory
is conducting an indicatory fate study to determine the mechanism whereby
an activated sludge system removes the polynuclear aromatic compounds (PNA's)
found on the priority pollutant list (e.g., via wasted solids; sparging
to the air; or biodegradation). One refinery has been sampled extensively;
the in-house study is in the sample preparation stages prior to final
GC/MS analysis. Figure 2 shows the sampling points; Figure 3 is an engineer-
ing drawing of the air sampling device. The specific organic compounds
studied are found in Figure 4. Two gas chromatograms are reproduced in
Figure 5 for comparative purposes. At the bottom of the figure is a
-3-
-------
Pfeffer
chromatogram of the PNA compounds of interest. The upper tracing is a
chromatogram of a waste sample collected during the study, showing graphically
the analytical problem of determing a few specific materials in a complex
matrix of organics in wastewater. The API has sponsored PNA studies at a
refinery and a publically-owned treatment works and has indicated that this
information will be available for the assessment.
There is very little information available for residuals relative to
that for wastewater. Most of the research concerning residuals has dealt
with oily rather than biological sludges. Two studies have been sponsored
by the API to characterize the sludges found in a refinery: a study by
Jacobs Engineering in 1974, and a much more extensive study involving over
70 refineries conducted in 1976. The API has indicated that these studies
when published will be available for the assessment.
The target date for completion of the Laboratory's portion of the
overall assessment is September 1979. This will not be in time to impact
the revised Development Documents for Petroleum Refining relating to waste-
water and solid waste; however, assessment may prove useful in the mandated
revision process of these documents.
-4-
-------
Table 1. CLASSIFICATION OF REFINERIES MONITORED BY EPA
REFINERY
CLASS
A
B
C
D
E
QUESTIONNAIRES
RECEIVED
14
79
31
12
14
150
REFINERIES MONITORED
NO.
2
8
3
2
2
17
CRUDE CAPACITY
BPSD
13,200
424,400
388,500
397, 300
702,000
1,925,400
i-t
(D
A-
t-h
l-h
m
-------
Fred M. Pfeffer
Table 2. ANALYTICAL PARAMETERS
BOD5
COD
TOG
PHENOL
SUSPENDED SOLIDS
DISSOLVED SOLIDS
pH
ALKALINITY
ACIDITY
OIL a GREASE
T. PHOSPHORUS
AMMONIA N
KJELDAHL N
NITRATE N
CHLORIDE
CYANIDE
SULF1DE
T. CHROMIUM
IRON
COPPER
LEAD
ZINC
-6-
-------
Table 3. PERFORMANCE OF ACTIVATED SLUDGE TREATMENT
rnn
V-> VJ U
OIL a GREASE-
PHFMni
r rtC-IMUL.
TOO
1 OO
ci ii cri np
oULrlUt
PL. I fiiMi'fc?! .
n vuniTs i
AM MOM! A M
AIVIIVlwrJiA N
KJELDAHL N-
T. CHROMIUM—
i pn M
1 re u N
i FT A n
Lli AU
Zl M f*
INC
INFLl
MAX.
228.0
583.0
89.0
16.0
261.0
43.0
11.5
160.0
194.0
2.5
4.25
44.0
0.88
1.73
1.23
JENT (mg/l)
MIN. MEDIAN
27.0
93.0
8.0
0.26
15.0
0.10
6.2
6.0
8.0
0.02
0.03
0.50
0.02
0.00
0.07
85.0
213.0
29.0
3.4
36.0
2.9
8.8
12.0
17.0
0.2
1.43
1.02
0.14
0.10
0.22
EFFLUENT (mg/l)
MAX. MIN. MEDIAN
53.0
201.0
37.0
4.2
83.0
1.7
8.7
124.0
134.0
0.03
1.45
0.50
O.M
1.04
1.84
5.0
46.0
4.0
0.01
5.0
0.0
5.0
0.05
2.0
0.02
0.02
0.18
0.02
0.10
0.04
8.0
70.0
11.0
0.01
25.0
0.3
7.3
II. 0
15.0
O.I
0.26
0.78
0.05
0.11
0.16
% REMOVAL
(MEDIAN OF ALL DATA)
89.2
54.6 I
60.0 *
99.6 S
n>
23.8
90.0
16.2
17.2
30.9
58.2
37.1
57.7
0.0
22.0
-------
Table 4. FINAL CLARIFIER EFFLUENT LOADING (FIVE PLANT COMPARISON)
oo
BOD5
COD
OIL a GREASE
PHENOL
TSS
SULFIDE
AMMONIA N
CYANIDE
T. CHROME
IRON
COPPER
LEAD
ZINC
MEDIAN LOADING (Ib/IOOO bbl )
9973
5.5
42.0
9.2
0.006
17.0
0.10
4.1
0.06
0.025
0.480
0.032
0.397
0.086
2115
3.1
22.0
2.8
0.006
4.0
0.05
17.0
0.01
0.078
0.148
0.005
0.018
0,027
0288
3.9
49.0
7.1
0.006
14.0
0.32
7.5
—
0.136
0.330
0.015
0.064
0.036
6512
4.0
16.0
2.9
0.003
7.9
0.13
O.I
— —
0.085
0.083
0.014
0.030
0.059
6693
1.5
23.0
2.4
0.002
6.4
0.04
5.8
0.210
0.650
0.030
0.046
0.089
Tl
Hi
ft
Hi
Hi
(0
i-i
-------
Table 5. FINAL CLARIFIER EFFLUENT LOADING (SUMMARY OF ALL DATA)
MEDIAN LOADINGS
(Ib/IOOObbl )
BOD5
COD
OIL a GREASE
PHENOL
TSS
SULFIDE
AMMONIA N
KJELDAHL N
CYANIDE
T. CHROME
IRON
COPPER
LEAD
ZINC
3.9
23.0
2.9
0.006
7.9
0.10
5.7
7.5
0.04
0.083
0.330
0.015
0.046
0.059
STANDARD DEVIATION
OF LOADINGS
(% of median)
55
21
41
1100
44
97
31
25
24
29
29
26
20
55
n>
S
Mi
(D
t-h
l-h
-------
WATER
INTAKE
8.
TJ
Mj
CD
t-h
o
Process © AERATION ._^ _/Ov
Ditch API DAF BASIN "V3>
© CARBON (4) MIXSP ®
ADSORPTION * FILTER
F|NAL
EFFLUENT
FIJLLSCALE
PILOTSCALE
Figure 1. Sampling points for 1976 study.
-------
Fred M. Pfeffer
Table 6. REFINERIES VISITED
Refinery
Gulf
Exxon
Hunt
Clark
Texaco
Mobil
Phillips
Shell
Exxon
Exxon
Coastal States
ARCO
Refinery
Location Class
Philadelphia, Pa. C
Baytown, Tex. E
Tuscaloosa, Ala. A
Hartford, 111. B
Lockport, 111. B
Augusta, Kan. B
Sweeney, Tex. C
Anacortes, Wash. B
Benecia, Calif. B
Billings, Mont. C
Corpus Christi, Tex. C
Philadelphia, Pa. B
Treatment System
Trickling Filter
Activated Sludge
Aerated Lagoon
Activated Sludge
Activated Sludge,
Filtration
Activated Sludge
Oxidation Ponds
Aerated Lagoon
Activated Sludge
Activated Sludge
Aerated Lagoon
Activated Sludge
Activated Sludge
-11-
-------
Fred 1'Ceffer
Table 7. PRIORITY POLLUTANT SAMPLING INFORMATION
Parameter
Non Volatile
Metals
Mercury
Cyanide
Phenolics
Asbestos
Volatile
24-hour
Composite
X
X
Grab
Sample
X
X
X
X
X
Sample
Container
1 Gallon Glass
1 Gallon Plastic
1 Quart Plastic
1 Quart Plastic
1 Quart Glass
1 Quart Plastic
40 ml Vial
Preservative
Ice
HN03
HN03
NaOH
H3P04
Ice
Ice
-12-
-------
WASTE
SLUDGE
(0
•x)
l-tl
AIR
SKIMMINGS
API
u>
DAF
®CENTRATE
AERATION
BASIN
RETURN®
SLUDGE
CENTRIFUGE
SAND
FILTER
AEROBIC
DIGESTER
BACK
WASH
NPDES
Figure 2. Sampling points at Sun - Corpus Christ! refinery.
-------
3/8 COPPER TUBING
3 GLASS PIPE
5*
MANOMETER
TAP
INSTRUMENT
AIR FROM
ROTOMETER
1 r
D=
&$$$
Lr
•'"
* * .
»» 6 '
6 " «*
o c* o
*x
,» o
0 "«
_J
Sl^s IKON PIPE
^-DEMISTER AERATION
-*-*^ LEVEl
WATER
METER
*^^ oUcSMtinolo
LEVEL 1 PUMp
(hJ
*^— SPARGER
VENT
H
8.
Hi
(D
l-h
H)
PIPE
.SAMPLE CANISTER
W/XAD-2 RESIN SLURRY
AERATION
BASIN WALL
Figure 3. Air sampling device in the aeration basin.
-------
Fred Pfeffer
^si-
Naphthalene
" Acenaphihylene
CI
Acenaphihene 2-Chlornaphthalene
Anthracene
Fluoranlhene
1,2-Benzanthracene
Pyrene
Benzo(a)Pyrene
11,12-Benzofluoranthene l|12- Benzoperyiene
l,2! 5,6-Dibenzanthracene
Chrysene
3,4-Benzofluoranthene
lndeno(l,2,3-c,d)pyrene
Figure 4. Polynuclear aromatic hydrocarbons studied.
-15-
-------
Fred Pfeffer
Figure 5. Comparative gas chromatograna.
-16-
-------
WORKSHOP DISCUSSION
PART I
Fugitive Emissions Detection and Measurement
Paul R. Harrison
The session, having been held Wednesday afternoon, did not have the
advantage of subsequent sessions; the sesssions and the discussions last
night did touch upon this topic. The topic was, of course, detection and
measurement. As far as the measurement technique is concerned, there seemed
to be a general consensus (although difference in detail as you saw in some
of the papers) that you have to bag fittings and, by some method or other,
extract the material enclosed in the bag or enclosure, measure it, and make
the proper calculations to get its leak rate. What was not discussed, is
the fact that we are talking about the non-methane fraction. Although
AP-42 gives you a lb/day/valve/component it doesn't necessarily talk about
non-methane fractions. Methane is excluded from those leak rates (shown
in paper) . One thing we did not discuss is that you may have to do some
analyses or, at least have analyses available. Most facilities do have
an average analyses of those process streams. You could make it from that
or make your own individual measurements. (Comment - Unknown - Really now
we are talking non-methane and non-ethane under the definition of VOC.) I'm
not completely sure of that because I have heard some people say; "Is that
true in every state?" (That is true in every state.) Then, concerning the
measurement methods for small leakers versus large leakers, most of the he.avy
leakers or what I call Class I leakers, are measured by flow meters or displace-
ment type devices, the smaller leaker requires instrumental measurement
techniques. There is an art to bagging and getting tight seals. Radian
discussed - in one of its papers- not at this meeting, that bagging hot
components is not a difficult task nor is it a mean trick to get done. That
is some technological news and an advancement which some of us did not know
how to do.
Discussion took place concerning the bubble method versus the instru-
mentation OVA or other methods. The bubble method was proposed. I think
the statement was made that we can classify them consistently but there
was no discussion about classification of bubbles, leak rates, and actual
Ib/day. That may be available but was not mentioned, to my knowledge, at
the conference. This is as opposed to the instrumented method which, on
at least two occasions, has projected average leak rates/classification.
The bubble method is found to be an excellent screening method. You know
you have a leak, and you can classify it and then measure it with a bag
and instrumentation. But no correlation with bubble rate or size and so
forth was made.
There were three instrumental techniques mentioned, one was close-in
monitoring. Valves are the most prevalent leakers (using blackboard), the
-1-
-------
packing is around the valve stem. The Radian method for determining whether
it's a leaker or non-leaker is (without any sub-classes at this time) to
put the sensor probe right on the packing at the interface. MRI has used
two methods. The method presently used is to put the probe slightly above
the packing, approximately a centimeter away, and measure the sustained
maximum deflection on the meter. We don't use chart recorders on this
particular survey method. We classify into Class I through IV, Class I
being anything over 10,000 ppm sustained maximum deflection; 1,000-10,000 -
Class II; 100-1,000 - Class III; and anything less than 100 ppm would be
a Class IV or an insignificant leaker. We gave some estimation in the paper,
of the average leak rates of each Class. Class III is 0.01 Ib/day non-
methane. We didn't extract any ethane, but I don't think there was any
there. Obviously the further away you go, the less the deflection you get
on any particular valve. The third method was measurement 5 cm from the
interface. The same thing applies to pump seals. The basic conclusion
is that closer in, you do start eliminating the wind effects, turbulent
effects, around the component. You get higher readings but they seem to
be fairly consistent. We do have a general equivalency of the concentration
at 5 cm versus 1 cm above the stem. The equivalency was approximately
3-1/2 to 4 times higher concentration for maximum sustained deflection which,
as you can imagine, as the wind wafts around the valve varies greatly.
If you can eliminate that problem, then you would like to do just that.
What are you really measuring? We talked about the non-methane fraction
versus the total. For example, around fuel gas lines more than 90% of the
emission is methane and other gases such as hydrogen. Although you get a
large signal there is not much non-methane hydrocarbon. So there is a
problem with the classic classification. The original classification is
based on instrument deflection versus the actual non-methane fraction.
When you report results in that fashion distribution does spread. The
distributions are good on the high end, but because of the correction factor,
there is tailing toward the lower leak rate as for the number of components
versus the leak rate. The technique is getting the large leaks. You may
have to fix large leakers, but they may not be largely non-methane. The
obvious need is to have an explosion-proof and safe instrument. The instru-
mentation that we used, was the OVA. It is well accepted in these
atmospheres, and of course, the soap bubble technique is also explosion-
proof. By the way, there is another technique. It consists of packing a
real stiff grease around the packing gland and pushing a calibrated sample
tube with a squirt of a little snoop in it and watch the bubble go up. I
find it difficult to get a tight seal.
We discussed the types of bags that were used, e.g., Mylar, Tedlar.
If you are looking for very low levels one should use special containers.
We talked about field calibration. If an instrument is used close-in
some condensible, heavy hydrocarbons may be sucked in and your probe will
become contaminated. Then the effective zero point will move up as these
materials accumulate. For large deflections there is no problem because
we are putting them in fairly gross categories. But, one has problems if
you try to calibrate after using the instrumentation near condensible
vapor leaks, especially for low level calibration gases. To make field
-2-
-------
calibrations, especially when you have been around high concentrations
of condensibles, use at least 100-1,000 ppm calibration gases. The
higher the concentration, the better your calibration will be. Don't
try low level zero calibrations unless you have decontaminated the flow
system within the instrument.
The other significant area of discussion was the transect, and how
can you relate that to individual fittings. As far as I know you can't. It
was suggested that measuring the average leak rate of these classes and
multiplying by the number of fittings in the classes you may not reach the
total fenceline fugitive emission rate. The reason for that is that the
fenceline measurement will include spills, low volatility material that
degases slowly, etc. We also concluded that you should never make a
fenceline measurement in low wind speed because of the heat rising effect.
Tedlar seem to be the best for grab samples. They should be analyzed
within 24 hours, especially for some of the heavier hydrocarbons. Hydrocarbons
above GS could be obtained on activated charcoal with carbon disulfide
extraction. However, you lose the Ci-Cs portion.
We talked about in-place hydrocarbon monitors. They seem to give you some
general idea of the performance of the unit, but we didn't come up with any
detailed plan to relate that to individual fittings.
We discussed walk-through monitoring versus fixed monitors. The walk-
through probably is more effective but even that had some problems of
detection. It would give you a little better idea of the performance of
the particular components close to the transect walk line than a fixed
monitor. There is no substitute for individual, complete coverage. But,
there are some compromises which may be desirable.
-3-
-------
WORKSHOP DISCUSSION
PART I
FUGITIVE EMISSIONS CONTROL
Fred Storer
Generally the emphasis of the discussion was on the reduction of leaks
from existing fugitive sources as opposed to what might be done with new
devices to eliminate leaks. The time was spent learning what each other
knew about the different sorts of leaks. With respect to valves, it
was generally agreed that valves were the largest source and most productive
place to emphasize fugitive controls. However, there were no real ideas
offered or general agreement on what were feasible objectives. It was
mentioned that a fugitive emission reduction program would lead to improve-
ments in valve design which would ultimately reduce fugitive emissions in
new sources. It was also emphasized that the results of surveys and
experience with fugitive emission control programs needed to be communicated
to the valve experts within the companies, to the valve manufacturers, and
to the purchasing department. With respect to flanges, it was generally
agreed that flanges were not a productive source for special emphasis.
Relief valves were discussed and their general control techniques: inspec-
tion/maintenance programs, connecting relief valves to flares, and installa-
tion of rupture discs. Both the flare situation in existing plants and the
rupture disc installation were considered in that those items may have some
safety compromises, depending on the individual installation and for that
reason any sort of a general across the board requirement would be very
difficult. Pumps were discussed, and it was agreed that inspection would
tend to reduce emissions from pumps; that the most important sources were
pumps pumping high-vapor pressure materials; that some pumps or maybe even
a majority of pumps that are now equipped with packed seals could be converted
to mechanical seals, although that is not the general rule; and that it was
less probable that an existing pump could be converted to a dual mechanical
seal with a barrier fluid. Compressors were briefly discussed, and the
consensus was that it was very difficult to define an action level or set
what might be reasonable goals for fugitive emission control programs.
Cooling towers were discussed and it was pointed out that many important
leaks in cooling towers are detected, that the leaks result in operating
problems with the cooling tower, problems with the algicides, chlorine
residual, etc. Some people have monitoring programs on cooling towers
both for the detection and for sleuthing out leaks. Also, some people
have good monitors in the plumes from cooling towers to detect leaks of
light hydrocarbon.
-4-
-------
WOEKSHOP DISCUSSIONS
PART I
POINT SOURCE EMISSIONS
Larry Johnson
We had a description from Radian on the point-source sampling that they
are doing as part of the refinery environmental assessment. Most of the
discussion here had been concerned with the valves. They are also looking
as often as they can at three point sources in the plant: those are sulfur
recovery units, CO boilers, and process boilers. Things have gone fairly
smoothly on these. They haven't sampled as many of them as they would
have liked, partly because many of these units do not have sampling ports
or facilities for doing the sampling in a straightforward manner. There
have been a few other minor problems which were discussed. In trying to
use an instrumental method for NO and NO , they had some problems with
getting the sample from the stack to thexinstrument without changes in the
sample, and they have gone to a chemical method to get around that. Something
either the same or very similar to Method 7 which is tricky but works if you
know how to use it. We discussed some of the techniques that were being
used and they seem to be using more or less standard techniques and not
having too much difficulty with them.
The other topic we discussed was discussed in the session on regulations,
but we had a preview of it. This was the problem with offsets and what is
going to happen when it gets difficult or impossible to get an offset. A
little bit of disagreement was generated in the session.
Radian is using source assessment sampling system or SASS train. The
block diagram of this train was developed in the Process Measurements Branch
of IERL/RTP. The actual machinery was put together by Aerotherm/Acurex
This machine is a 5 cfm expanded version of the Method 5 train. In addition
to being larger, it has capability on the front end for size cuts; it has
a module after the filter which can be loaded with sorbent —we use XAD-2
(about 135 grams) and from there on back it looks somewhat like a Method 5
train. In the impingers instead of water we put an acid oxidant which will
catch mercury and similar materials. It is an attempt, and a fairly good
one, to get everything in one shot. The only thing we really don't catch
this way is very, very volatile organics. Anything with a boiling point
below about 100°C, goes on through to the oxidative impingers and gets all
chewed up. The train seems to be, on this project, working about like it is
on others. There are some mechanical sealing problems with the train. The
train can be modified to make sealing better, and future models will have
better seals. It takes three men to run and we usually sample for five hours
with it. This is because many of the sources we look at have very low
loadings. This is also why it's a big train. For a 5-hour sampling run, you
end up with something on the order of an 8-hour clean-up/preparation time. So
you are talking about a 12 or 13-hour day to get a run. But you get organic,
inorganics and sizing.
-5-
-------
WORKSHOP DISCUSSIONS
QUESTIONS AND ANSWERS
D. D. Rosebrook - I would like to add to what Larry said. One of the
things we are trying to do is to determine the efficiency of the control
technology that is now in place on point sources. In order to do this,
we must measure on both sides of any control device. We have tried for
over a year and finally found a place that we could get access on both
sides and we made simultaneous SASS train runs. We had two SASS trains
running simultaneously, so we are finally getting some control efficiency
data.
L. Johnson - If anybody is interested in finding out more about the
train or how it was built or what it does, you can contact me or call Bruce
Harris or Bill Kuykendal and we will send you lots of information.
Jim Daily - What was the control device that he was sampling up and down
stream?
A. - The devices were a scrubber and a CO boiler, both on the same
stack.
L. Johnson - Let me add one thing about the SASS train. Since you are
getting sizing on the front end, you have to have constant flow rate through
it or the cut points will change. Thus, you don't get a true isokinetic
sampling with the SASS train. It has all the controls necessary to do that,
but if you utilize that feature you lose the sizing. So, you can either
take the cyclones out and do true isokinetics or, if your source is even
reasonably stable, you can take your velocity traverse, and set the nozzles
to the right size so that you get isokinetic as long as it doesn't drift
too much. So we have dubbed that as pseudo-isokinetic because as long as
it doesn't shift around it is pretty good.
Jim Daily - Generally speaking, going upstream of the CO boiler, the
piping configurations are such that you never find a spot that you don't have
a turbulent situation. (Where you don't have too few diameters from the
last bend to get true traverse and get a proper sample point.)
A. - Well, we are not too worried about that on environmental assessments
because we are not trying to be as accurate, if you will, as somebody that
is doing compliance testing. Also we don't normally traverse with a SASS
train because it is very big. You can do it. Normally we try to find
some average point and put the train in there and take the 4 or 5 hour sample
without moving it.
Jim Daily - Are you recording particulates?
A. - Particulate sizing and loading, but it is not a compliance test.
You would be surprised how little difference it has made. We have done a
SASS validation run with three different contractors, two SASS trains and
a Method 5 train. The results are very, very close; much closer than any
-6-
-------
of us thought. Now this is in a fairly well behaved source. That report
is not quite available yet, but it will be soon.
Jim Daily - On the critical analysis of the organics in stacks, are
you taking the same type of critical analysis of organics in the ambient
air in the vicinity of the equipment at ground level?
A. - On the Radian program I don't know.
D. D. Rosebrook - No we are only taking the total hydrocarbon concen-
tration at the inlet. We are not doing the complete breakdown.
L. Johnson - In general, on the environmental assessment programs,
the fugitive emissions part of the work is sometimes implemented and some-
times not. We have it in the protocols and we have protocols for doing it.
A truly 100% complete environmental assessment would include that type of
thing. But, budgetary requirements and a number of other things, very often
mean that you trim your program a little bit; and since that is one of the
more expensive things, it is usually one of the first things that are
written out.
Jim Daily - In early gas chromatographic work that was done on stack
sampling from gas-fired furnaces the total organics at ground level in the
vicinity of the furnace were higher than the organics in the stack. Do
you have to take that into consideration?
A. - Yes. That can occur in cooling towers too. That can happen in
a lot of places and isn't unusual. That can happen with particulate too.
Especially if you have a baghouse or something.
(}. - A question on detection of VOC leaks. The EPA's letter to the
states on revising the SIP's this year calls for controlling VOC's as
needed to meet the oxygen standard. The EPA policy statement on volatile
organics published in the Federal Register defines them to exclude methane and
ethane. Is there any practical method from a screening standpoint in
looking for leaks in refineries to detect leaks of VOC as defined by the
EPA so you don't spin your wheels looking for methane and ethane limits?
A. Paul Harrison - There are two methods that occur to me. One, is
knowing what your product line is or what is in your service lines. The
other method involves screening instruments that can be obtained with a
36-inch GC column which is called a "methane column". You can perform a
quick GC analysis on-site, at the sample point, using this particular
device. We have done some of that, but it does take a long time. As you
know our technique is to survey the entire facility as opposed to one or
two units. It takes so long to get GC analysis that we'd rather trust
the product line or service analysis than to actually conduct an individual
GC on every fitting that is leaking.
-7-
-------
K. C. Hustvedt - If you use a charcoal filter on the OVA, take the
total and then put the filter on it and get your methane fraction, and take
the difference between the two to get the non-methane fraction.
£. - What about ethane?
^. - I don't have any comment.
Jim Daily - Refinery fuel gas lines are primarily methane and ethane.
A. - I think by far the best way is to just know what you are sampling.
Comment - Fine if that reflects the regulation you are working under.
But if it is a soap bubble test for total organics, then you are not measuring
VOC.
A. - Good point. Of course, some of the importance behind this is
safety and product savings. So, on one side you have the regulatory problem
and on the other side you have the economics problem.
Comment - There might be some interesting liability problems in having
a record of leaks discovered that you haven't fixed.
J. Daily - There also might be litigation on the required control of all
leaks if the Federal law specifically says methane and ethane are not
important.
D. D. Rosebrook - I have a question for Fred Storer. Could you elaborate
a little more on putting monitors in cooling tower plumes and the value of
this, liability of it?
A. - Someone in the discussion group pointed out that it had been done.
I don't have any knowledge of it and I don't even recall who made the
suggestion. If the person is here maybe they could address that. It is a
fairly simple thing to do, but I don't know how useful the information would
be. I suppose that if you had a problem with reoccurring leaks of light
components that would flash as soon as they would hit the atmosphere, the
monitor might be useful. As a pollution control device, I don't know.
I don't think that was the purpose.
Atly Jefcoat - I have always been under the impression that the
operators or maintenance people in the refinery watch the cooling towers
for the appearance of an oil skim then they would know when a heat exchange
or something was leaking. How widespread is monitoring of cooling towers?
Is this a common practice or is it not?
Fred Storer - I don't know. I don't think it is a common practice. I
think watching the cooling tower and watching things like the chlorine residual
and some of the other parameters in the cooling tower is a common practice
and certainly watching for oil in the cooling tower basin is a common practice.
-8-
-------
Comment - We have one plant that has been doing that and I don't
know what their action level is. It may be that when they have a problem
with the cooling tower, they make a GC analysis of hydrocarbons in the
water to help find the source of a leak.
Comment - That problem was discussed in the Wastewater Treatment
Session also. We've done it and maybe that is the appropriate place to
put a monitor.
Comment - J. Daily - Better refineries do monitor, say weekly, hydro-
carbons in and out through the cooling tower. You can always tell when
you have an instantaneous big leak, it will blow the spray nozzles off the
top of the spray head.
Atly Jefcoat - Some of these things are self-evident.
Comment - Well there are two methods. One is to measure the hydro-
carbon in the water in and out. The other is to put a device in the
effluent side or the air side of the top of the cooling tower and monitor
that for the volatiles.
P.P. Rosebrook - Was there any discussion of flares as a point source?
A. - Larry Johnson - I mentioned it in our session and really there was
no real experience in that group on flares.
-9-
-------
WORKSHOP DISCUSSIONS
PART II
Emissions from Waste Treatment Facilities
Fred Pfeffer
The workshop dealing with Emissions from Waste Treatment Facilities
divided its time between two topics. One was fugitive emissions from the
treatment train, particularly the API separator. The second topic was
solid wastes. We didn't cover anything else from the wastewater treatment
facility.
In the realm of fugitive emissions, we started by discussing the
sampling of non-methane hydrocarbons from an anaerobic pretreatment lagoon.
We went from there to the factors that affect the emissions of hydrocarbons
from physical treatment systems that are ahead of biological systems, such
as effects of sparging with air, the variability in composition of the
influent - particularly the oil content, the presence of emulsions, and
the presence of floating oil.
Then we went on to the problems associated with something like an API
separator. They are not easy to sample. We noted in our discussions that
any attempt to do a mass balance for emissions (what goes in minus what comes
out equals emissions) has not been sucessful. We disucssed how to physically
sample the air emissions including a discussion of "a transectional" system
of measurements. (A system attempting to account for background hydrocarbons,
wind velocity and direction, elevation effects, etc.)
In the solid waste area we discussed the on going studies in solid
wastes concerning leachability, characterization of the leachate and effects
of the leachates. The discussion moved to the very tedious and expensive
analytical problem to prove that a particular solid waste is not "hazardous".
This might entail very extensive biological testing. We discussed the Ames
test, the rodent acute toxicity test, bio-accumulation, effects on algae,
effects on daphne and effects on fish. A very elaborate scheme may be
involved here. Then chemical testing such as complete metals analysis may
be required. And then into organics testing which may, from a specific
organic standpoint, have to account for up to 99% of the organics that are
found in that sample.
We briefly covered the implication of rule making. A problem that may
arise if EPA specifies re-use of the final effluent. If this final effluent
goes back to cooling towers, what happens as far as an increased organic
load getting out from a cooling tower. It was a very enlightening workshop.
-10-
-------
WORKSHOP DISCUSSIONS
PART II
Regulations
Leigh Short
The discussion was summarized very accurately by Lloyd Provost from
Radian who is a statistician who said that we discussed 27 points and
agreed on three of them and he wasn't too sure about the third.
The first item I think worth bringing to your attention was the
distinction between the use of "frequency" in the regulations and frequency
of leaks, versus the way at least it is perceived. The Draft Guideline
Document as written now refers to a concentration measurement. The
discussion concerned whether it would be better to have the regulation
reworded or reworked saying that it will be based on frequency of leaks
of valves. The discussion brought out points concerning how do you define
frequency and how do you define a leak and what technique can be used for
measurement. I think those are obviously three key variables. There also
was a discussion of the soap test versus any instrument test. Frank Mesich
discussed Radian's information to date, which shows that the distribution
of leaks, particularly in valves, is spread across the whole temperature
range they have sampled. There is a very significant fraction of leaking
valves that you can't soap test.
The next part that came through very clearly to me, of the 27 points
discussed, was that each refinery is probably going to be a different
situation and different sampling case. The regulations as they end up at
the state level should provide for flexibility both in how they are admin-
istered and in allowing them to be changed as the process goes on. The
point was made very strongly that there is essentially no data base at this
point to determine what the improvement target should be. The data coming
out of Radian's work, measures leaks in the refineries as they are now
maintained. There is very, very little data in that study, at this point,
about what happens when you go out and try and maintain the leak, how much
you can fix, what percent reduction do you get, and how long does that last.
Another point I want to bring up is that GARB, if I heard them correctly,
has a staff report which is proposing that the regulations there say no
operator shall operate a facility with a leaking valve or a leaking flange.
His definition of a leak was no bubble. Aside from whether you agree with
the regulation or not, it points up the difficulty of defining what is
meant by a leak and what is meant by a frequency. The Radian data, for
example, when they define leaks is at a very different cutoff in terms
of concentration. EPA's document defines a leak as 10,000 ppm. The Radian
leak is defined at a much lower value because they are looking at the
whole statistical range and not just the high leaking range.
-11-
-------
The other point I want to make about flexibility is that the numbers
which Frank Mesich presented and which were assiduously copied down by
everybody there, indicate that the frequency — number of leaking valves —
varies by type of unit within a refinery. The regulation ultimately should
reflect that variation when there is enough data to make a statistically
valid statement. For example, based on Radian's definition of frequency, the
percent of valves leaking ranges from about 20% up to about 50% as you change
units within a refinery.
There was a large discussion about the frequency of checks. How often
should you go in and sample, both looking for leaks and by sampling individual
sources. The usefulness of a walk-through system was discussed in terms of
whether it will tell you what you are looking for and whether the cut-off
which is 100 ppm will directly lead you to the high leaker. Further, will
it lead you to spurious results? A discussion occurred on the definition
of a walk-through, because every refinery is different. If you have one such
as the one I am familiar with, with a huge bank of air coolers, what do you
learn if you walk through that, other than it is very windy.
-12-
-------
WORKSHOP DISCUSSIONS
PART II
Equipment Considerations
Alton Williamson
We started off with a discussion of wastewater separators with HC
control. It has been demonstrated but is not common. These are relatively
new, but tank manufacturers will make them. We discussed in particular
the inert gas separator blanket, internal floating covers, parallel plates,
and CPI's. We should look at the control methods on all of these to determine
whether there are other improvements that are not being used or if some others
are needed.
The subject of vacuum eductors was discussed briefly. What are they?
How are they used? What are some of the problems with them? No one knew
of recent improvements other than the fact that in some of the old systems
they are not being vented to flares and they should be and could be.
The biggest discussion concerned what we heard all week, valves. Valves
are a problem. Having had two or three equipment manufacturers in the session,
we did talk about a lot more than just what's new or what is going on or
what is available that we are not using, some of the problems. We won't
go into all that. The discussion centered around materials, particularly
packing materials and forms of design. If we used more packings with
automatic take-up instead of the common gate and globe valves either with a
fixed or adjustable packing, emissions may be reduced. The problem arises
because refineries are so top heavy with gate and globe valves because of the API
standards and the fact that they are older valves and have proven them-
selves in the past. Non-lubricated plug valves whether they are sleeve type
or ball valve or some other type of valve like that, are lower maintenance
item products, however, they haven't demonstrated that they have the applica-
bility yet. One of the strong recommendations was to tell the valve
manufacturers to get us quarter-turn valyes that are more reliable at higher
temperatures. This led to discussions on metallic seals and other types
of improvement.
We discussed the fact that the standard packing gland can be used in many
different ways. One adaptation that is not used as much as it should be
and could be is purging with nitrogen. Using a higher pressure than your
internal pressure would force the valve to leak inward.
A limitation to the plastic packing injectable system that I had
mentioned yesterday, whether you call it grease or whether it is a special
sealing injection medium, is that the temperatu're must be less- than 500°, so
this would prohibit expanding this feature into higher temperature applica-
tions and maybe we need improvement there.
-13-
-------
We talked about different new packings that have been used. It was
mentioned that some of the very sensitive packings do have very good applic-
ations in packing glands, but they also have some limitations. We discussed
laminous, labryinth type seals, and ribbon type seals, and how they could
be applicable and where we need new improvements in this area.
We discussed a little bit about the necessity of improving fire safe
valves of all kinds. Ball valves have fire safe standards to go by (16 or 18
different standards, depending on whether it is international or domestic).
But, there are not many fire safe standards for other valves. Maybe we need
an improvement there.
How many valve manufacturers really recommend regular maintenance. My
comment was maybe only about 10%. It depends on the type of valve, the
specialty and the project. When you are just buying standard valves off
the shelf for what we call standard API refinery business, I doubt if there
is a maintenance recommendation given except in a repair manual.
We also discussed the need to use more pressure energized metallic seals,
especially in the 600-1200° range which is becoming more common. Also if we
use more packings with sacrificial metals in them, that would help eliminate
corrosion on parts that cause packings to leak prematurely.
We recommended cost-effectiveness studies on both single and double
mechanical seals for pumps and nitrogen purges on the linear rings, I found
it was common that when we require secondary seals, losses in Ib/hr is
reduced drastically. The cost-effectiveness seems to be there and ought
to be examined in more detail.
We discussed secondary seals on open-top and floating roof tanks and
felt that these are cost-effective even when retrofitted and maybe there
ought to be more work done along this line.
In summary, the strongest recommendation was the need for obtaining
better maintenance procedures from the valve manufacturers.
-14-
-------
WORKSHOP DISCUSSIONS
PART II
Fugitive Emissions
James W. Daily
It was brought out that there are 264 refineries in the United States
and they are owned by 156 separate and distinct oil companies. There are
about 8 or 10 representatives at this session, and the decisions that are
being made here are partly bracketed by what these 8 or 10 companies think.
The National Petroleum Refiners Association does represent all 156 companies,
_so we will have_to see what kind of hue and cry comes from the other 146 after
the 10 that are here have^heard and discussed and looked into what is coming out of
this meeting. We spent most of our time on valves, we didn't discuss relief
valves at all. We did talk briefly on pumps, briefly on compressors as to
what kind of pumps and compressors they might be, whether they are recipro-
cating or centrifugal, and whether the leak rate was the same and whether
the solutions were the same. Radian's study shows the leak rates aren't all
that different, but the solutions, of course, are different for obvious
reasons.
The question arose about what else was being studied by Radian. One of
the points that was discussed at our session was the health effects of the
measured emissions from these sources, and what they are contributing to
the overall public health problem from hydrocarbons. The Radian study showed
that the types of contaminants and pollutants that are coming from refineries
are the lighter, more volatile materials, very few were highly olefinic and very
few were heavy aromatics.
There were several suggestions about the appropriate method of control
i.e., whether regulations should be very specific or whether they should
give control limits and let the refineries determine how they would reach
them. The question arose as to whether there is any actual present correlation
between the lower limit hydrocarbon leakage that Radian was measuring and
the lower limit soap for bubble testing.
The cost-benefit question arose. That question is very difficult to
pin down because the benefit, in terms of savings, continues to rise as the
value of the product increases.
The last point that was brought up was that the Clean Air Act Ammendments
of 1977 are written in such a fashion and the procedures are presented in
such a way, that we will reach a zero growth situation very rapidly. In
other words, tradeoffs under the regulations will disappear very rapidly -
any kind of tradeoffs — with the present definitions of non-deterioration
and non-attainment. That covers 80% of the nation's population area. Thus,
we are headed rapidly toward a zero growth. But the saving grace is that the
public recognition of- and public outcry about the situation would be so loud
and hopefully in sufficient time that the law would be changed, the regulations
would be changed.
-15-
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPOR
EPA-600/2-78-199
3. RECIPIENT'S ACCESSION-NO.
4 -TITUS AND SUBTITLE Proceedings: Symposium/Workshop
on Petroleum Refining Emissions (April 1978, Jekyll
Island, GA)
S. REPORT DATE
September 1978
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Susan R. Fernandes, Compiler
8. PERFORMING ORGANIZATION REPORT NO.
78-200-187-24-08
9. PERFORMING ORGANIZATION NAME AND ADDRESS
10. PROGRAM ELEMENT NO.
Radian Corporation
P.O. Box 9948
Austin, Texas 78766
1AB604C
11. CONTRACT/GRANT NO.
68-02-2608, Task 24
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD.COVERED
Proceedings; 11/77-7/78
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES JERL.RTp project officer iS
514-2547.
A. Jefcoat, Mail Drop 62, 919/
IS. ABSTRACT
The proceedings document presentations made during the symposium/work-
shop which brought together knowledgeable and concerned individuals to discuss the
current and future status of emissions from petroleum refineries. The program
included technical presentations from both regulatory and control technology areas of
EPA, from private contract research organizations, from state pollution control
agencies, from refining companies, and from equipment design and manufacturing
companies.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATl Field/Group
Pollution
Petroleum Refining
Refineries
Processing
Leakage
Hydrocarbons
Control Equipment
Petroleum Industry
Valves
Pumps
Sampling
Regulations
Monitors
Inspection
Pollution Control
Stationary Sources
Fugitive Emissions
13B
13H
131
14B
07C
05C
13K
13G
05D
13. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
168
20. SSCURI TY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
------- |