CORMIX1 AN EXPERT SYSTEM FOR MIXING ZONE ANALYSIS OF
CONVENTIONAL AND TOXIC SINGLE PORT AQUATIC DISCHARGES
                         by
       Robert L. Doneker and Gerhard H. Jirka
            DeFrees Hydraulics Laboratory
       Department of Environmental Engineering
     Cornell University, Ithaca, New York  14853
       Cooperative Agreement No. CR813093-01-0
                   Project Officer

               Thomas 0. Barnwell, Jr.
                  Assessment Branch
          Environmental Research Laboratory
               Athens, Georgia  30613
          ENVIRONMENTAL RESEARCH LABORATORY
         OFFICE OF RESEARCH AND DEVELOPMENT
        U.S. ENVIRONMENTAL PROTECTION AGENCY
               ATHENS, GEORGIA  30613

-------
                        FOREWORD


      The information in this document has been funded wholly
or in part by the United States Environmental Protection
Agency under Cooperative Agreement number CR813093 to Cornell
University.  It has been subjected to the Agency's peer and
administrative review, and it has been approved for publication
as an EPA document.
                            11

-------
                      Table of Contents

Abstract	iv

List of Tables	vii

List of Figures	   viii

Acknowledgements   	  x

Chapter I Legal Background and Introduction  	    1
         l.A History of the Clean Water Act	    1
              I.A.I  The  Federal  Water Pollution Control
              Act  of 1972	    3
              l.A. 2 The Clean Water Act of 1977	    4
         l.B Development of Mixing Zone Concept  	  8
         l.C Special Mixing Zone Requirements for Toxics    12
         l.D The NPDES Permit System	   12
         l.E Need  for Regulatory Assessment Tools ....   13
         l.F Justification for Expert  Systems Approach   .   14
         l.G CORMIX1 Summary	   17

Chapter II Hydrodynamic Background on  Mixing Processes   .   18
         2.A Analysis of Subsurface Flow Regions  ....   18
              2.A.I  Description  of   Turbulent  Jets  and
              Plumes	   20
              2.A.2  Elements  of  Dimensional Analysis of
              Buoyant Jets	   20
                   2.A.2.1.   Simple   Jet   in   Stagnant
                   Environment	   21
                   2.A.2.2. Simple Plume in
                   Stagnant Environment 	   23
                   2.A.2.3    Generalizations:   Jet/Plume
                   Interactions and Effects of Crossflow    25
              2.A.3 Length Scales for  Buoyant Jets With or
              Without Crossflow 	   25
                   2.A.3.1 Discharge Length Scale ....   25
                   2.A.3.2 Jet/Crossflow Length  Scale .  .   27
                   2.A.3.3 Plume/Crossflow Length Scale  .   27
                   2.A.3.4 Jet/Plume Length Scale ....   27
              2.A.4 Typical Regimes of Buoyant Jets ...   28
                   2.A.4.1   Weakly    Deflected   Jet   in
                   Crossflow (mdnf)	   28
                   2.A.4.2   Strongly   Deflected  Jet  in
                   Crossflow (mdff) 	   29
                   2.A.4.3  Weakly   Deflected   Plume  in
                   Crossflow (bdnf) 	   30
                   2.A.4.4  Strongly   Deflected  Plume  in
                   Crossflow (bdff) 	   31
                   2.A.4.5 General  Behavior  in Unbounded

-------
                   Crossflow	   31
         2.B Flow Interaction With Free Surface	   38
              2.B.I Flow Classification of Near-Field
              Regions	   38
              2.B.2 Analysis of Surface Interaction
              Processes	   38
                   2.B.2.1 Near Horizontal Surface
                   Approach	   44
                   2.B.2.2 Near-Vertical Surface
                   Impingement With Buoyant Upstream
                   Spreading	   44
                   2.B.2.3 Near-Vertical Surface
                   Impingement With Full Vertical Mixing    48
                   2.B.2.4 Near-Vertical Surface
                   Impingement With Unstable
                   Recirculation, Buoyant
                   Restratification, and Upstream
                   Spreading	   48
         2.C Analysis of Far-Field Mixing Process ....   50
              2.C.I Buoyant Surface Spreading 	   50
              2.C.2 Passive Diffusion 	   52
              2.C.3 General Behavior in the Far-Field .   .   54

Chapter III CORMIX1 Program Structure 	   56
         3.A Discussion of Logic/Ml Elements  	   58
              3.A.I DATIN	   59
              3.C.2 CLASS	   63
              3.C.3 SUM	   64
         3.C Discussion of Hydrodynamic/Fortran Elements    65
              3.C.I PARAM	'"	   65
              3.C.3 HYDRO	   65

Chapter IV Data Comparison and Validation 	   69
         4.A Near Field Flows (sub-surface regions)  ...   69
         4.B Near Vertical Surface Impingement With
         Buoyant Upstream Spreading 	 72
         4.3 Unstable Surface Impingement With Buoyant
         Upstream Spreading .	79
         4.4 Summary	79

Chapter V Design Case Studies	81
         5.A Case 1: AB CHEMICAL CORP., WEST VIRGINIA ... 81
              5.A.I The Problem Statement	81
              5. A. 2 CORMIX1 Analysis	82
         5.B Case 2: SAN ONOFRE UNIT 1	85
              5.B.I The Problem Statement	85
              5.B.2 CORMIX1 Analysis  	 87

Chapter VI Conclusions and Recommendations  	 89
References
91
                              VI

-------
APPENDIX A	(Vol.  II)    95
         A.I DATIN	(Vol.  II)    96
         A.2 CLASS	(Vol.  II)   138
         A.3 SUM	(Vol.  II)   178
         A. 4 PARAM	(Vol.  II)   200
         A. 5 HYDRO	(Vol.  II)   216

APPENDIX B	(Vol.  II)   304
         B.I AB Chemical Co. Analysis	(Vol.  II)   305
         B.2 San Onofre Analysis	(Vol.  II)   329
                              VII

-------
                     -  List of Tables

Table 1.1 Key Federal Water Pollution Control Laws

Table 1.2 Examples of Conventional, Nonconventional,
and Toxic Pollutants  	
Table 1.3 Examples of Technology-Based Effluent
Limitations Under The Clean Water Act of 1977 	    7

Table 1.4 State Legal Mixing Zones  	   10

Table 2.1 Trajectory and Dilution Relations 	   32

Table 2.2 Trajectory and Dilution Constants 	   33

Table 2.3 Flow Transition Rules	   37

Table 3.1 Hydrodynamic Simulation Modules   	   67

Table 3.2 Hydrodynamic-Simulation Protocols 	   68
                             Vlll

-------
                       List of Figures

Figure 2.1 Illustrative Near Field and Far Field of Submerged
Buoyant Discharge 	   19

Figure 2.2 Pure Jet in Stagnant Environment	   22

Figure 2.3 Simple Plume in Stagnant Environment 	   24

Figure 2.4 Examples of Combined Effects of Momentum Flux,
Buoyancy Flux and Crossflow on Flow Behavior	   26

Figure 2.5 General Behavior for Buoyant Jets in Unconfined
Crossflow	   35

Figure 2.6 Flow Classification System 	   39

Figure 2.7 Four Major Conditions of Flow Interaction with
Water Surface	   40

Figure 2.8 Buoyant Surface Spreading  	   51

Figure 2.9 Passive Diffusion Process  	   53

Figure 3.1 System Elements of CORMIX1 	   57

Figure 3.2 Schematization of Discharge Configuration  .   .   60

Figure 4.1 Fan's Buoyant Jet Trajectory, Expr.  20-12, R = 1270

Figure 4.2 Fan's Buoyant Jet Trajectory, Expr.  40-4, R = 4  71

Figure 4.3 Fan's Buoyant Jet Dilution, Expr. 20-12, R = 12  73

Figure 4.4 Fan's Buoyant Jet Dilution, Expr. 40-4, R = 4  .74

Figure 4.5 Wright's Buoyant Jet Trajectory, Expr. 2-2 ... 75

Figure 4.6 Flatten and Keffer, 9o = 60.0"	76

Figure 4.7a Simulation of Stable Surface Impingement/Upstream
Spreading	77

Figure 4.7b Plan View of Surface Impingement/Upstream
Spreading	78

Figure 4.8 Plan View Comparison of San Onofre Prediction and
Field Data	80

Figure 5.1 Schematization of Cross-section  	 83
                              IX

-------
Figure 5.2 Plot of CORMIX1 AB Chemical Co. Predictions  .  .  86




Figure 5.3 San Onofre Longitudinal Cross-Section  	  87

-------
                      Acknowledgements

    This  study   was  conducted   at  Cornell  University  in
cooperation  with  the   Environmental   Research  Laboratory,
Athens,   Georgia.      The   authors  want  to  extend  their
appreciation to Dr. Thomas O. Barnwell,  Jr., project officer,
who  provided   the  initial  stimulus  and  further  guidance
throughout the study.

    The authors also acknowledge  the assistance  given by Dr.
Anil  Nerode,  Chairman,  Department  of  Mathematics, Cornell
University, in the development of expert  system structure and
logical elements.   Dr.  Joseph H.-W. Lee, University of Hong-
Kong, helped in the formulation  of  trajectory  laws  for the
submerged  jet  phases  during  his sabbatical stay at Cornell
University -

    The work was carried out using the computer  facilities of
the  DeFrees  Hydraulics  Laboratory.    Mr. Cameron Willkens,
electronic  technician,  assisted  in  computer  hardware  and
software problem solutions.
                              XI

-------
                          Chapter I
              Legal Background and Introduction
l.A History of the Clean Water Act
    Prior to  1948, States,  local, and regional agencies were
primarily responsible for controlling water pollution.   After
the realization  in the mid-1800's of the role of contaminated
water in the transmission of disease,  State Boards  of Health
were  formed  to  administer water pollution control programs.
Most early pollution control  programs focused  on water-borne
infectious diseases like typhoid and cholera (Ortolano,1S84).

    Table  1.1  outlines  key  federal water pollution control
legislation since 1943.  The 1948 Water Pollution  Control Act
was designed  to provide  technical services  to the States  to
strengthen their water pollution  control programs.   The 1948
Act focused  on the primacy of the State role in water quality
management.   Federal action  against polluters  could only  be
taken with  the consent  of the State from which the pollution
was alleged to originate.

    The Federal  Water Pollution  Control Act  (FWPCA)  of 1956
expanded the federal role in controlling water pollution.  The
Act provided  a program  of subsidies  for municipal treatment
plant construction, strengthened powers of enforcement against
polluters,  increased funding for State water pollution control
efforts, and provided new support for research and teaching.

    Each  of   these  programs   were  included  in  the  many
amendments  to the Act in the 1960's and 1970's.

    The Water Quality Ace of  1965  set  new  requirements for
States  to   establish  ambient  water  quality  standards  and
increased  the  level  of  federal  funding.    Water  quality
standards  were  designed  to  protect  designated  water uses
within a stretch  of  river.    The  Act  required  that State
agencies set  water quality  criteria to meet these standards.
Criteria established the suitability  of  water  for different
activities. If the uses of water within a stretch of river and
the  criteria  designed  to  protect  those  uses  were known,
ambient water quality standards could be set.

I.A.I The Federal Water Pollution Control Act of 1972
    Prior  to  the  1972  Federal  Water Pollution Control Act
(FWPCA) only States had power to develop ambient water quality

-------
Table 1.1 Key Federal Water Pollution Control Laws
Source:  Ortolano, 1984
Year
Title
Selected New  Elements

of Federal Strategy*
1948
Water Pollution
Control Act
1956
Federal Water
Pollution Control
Act  (FWPCA)
1965
Water Quality Act
1972
FWPCA Amendments
1977
Clean Water Act
1981
Municipal Waste
Treatment
Construction
Grants Amendments
Funds for State water
 pollution control
 agencies
Technical assistance to
 States
Limited provisions for
 legal action  against
 polluters

Funds for water pollution
 research and training
Construction grants to
 municipalities
Three stage enforcement
 process

States set water quality
 standards
States prepare
 implementation plans

Zero discharge of
 pollutants as a goal
BPT and BAT effluent
 limitations
NPDES permits
Enforcement based on
 permit violations

BAT requirements for
 toxic substances
BCT requirements for
 conventional pollutants

Reduced federal share in
  construction grants
  program
* The table entries include only significant new
changes established  by the law.

-------
standards applicable to interstate or navigable waters.  Water
quality standards depended  upon  intended  use -agricultural,
industrial,  or recreational.

    Enforcement of  water quality  standards was only possible
if  water  quality  fell  below  standards.     This  hampered
enforcement efforts  because proof  of causation was difficult
in waters receiving  wastes  from  various  polluters.   State
water quality  standards could  be lowered to attract industry
away from States with more stringent water quality standards.

    Congress decided to take rigorous action in 1972  with the
FWPCA  amendments.    The  Act established a uniform system of
water  quality  standards,  permits,  and  enforcement.    The
"goals" of the legislation were to produce fishable, swimmable
water by 1983 and  a total  elimination of  water pollution by
1985  (Findley and Farber, 1983) .

    Major changes  in the  FWPCA of  1972  included i) national
water   quality   goals,    ii)    technology-based   effluent
limitations, iii)  a national  discharge permit system, and iv)
federal court  action against   sources in  violation of permit
conditions  (Ortolano, 1984).

    Congressional intent  in passing the FWPCA was to rule out
arguments  of  assimilative  capacity   of  receiving  waters.
Congress  wanted  uniformity   of   standards  and  enforcement.
Ambient water quality   standards   were  intended  to  be  "more
stringent"   than  effluent  standards.     The  aim of  the  1972
amendments   was   to  restore    and maintain   "the  chemical,
physical,  and  biological  integrity  of  the nation's waters"
 (Weyerhauser Co.  v. Costle  590 F.2d 1001).

    The 1972  amendments  gave broad  powers  to  the federal
Environmental  Protection   Agency  (USEPA)  administrator   to
define pollutants and   to   determine  and  promulgate  effluent
limitations.    Effluent  limitations  were  set  according  to
industry through  the National   Pollution Discharge Elimination
System  (NPDES) permit system.   These discharge limits were set
independent  of the particular   context in  which  the pollution
discharge occurs.  Dischargers in  violation of NPDES pollution
limits were  subject to  enforcement action.

    The Act  contained   ambient water  quality  standards  that
supplemented  federal   discharge   standards for point  sources.
Point sources were defined  as  "any discernable,  confined, and
discrete conveyance  .... from which pollutants are,  or may  be
discharged."

    The 1972 FWPCA  required   that industry  dischargers  meet
"best  practicable  control  technology  currently   available"

-------
 (BPT)  standards   by  1977  and  "best  available  technology
economically achievable" (BAT) standards by 1983.

    The  Act  required  public  sources  of  pollution  to use
secondary treatment  by 1977  and use  "best practicable waste
treatment over life of the works" by 1983.

    Specific sections of the Act include:

    Section 301 of the Act  set  standards  for  point sources
that  were  not  publicly  owned  treatment  works  (POTW).   It
requires  dischargers   to   reduce   emissions   using  "best
practicable control  technology currently  available" (BPT) by
1977 and  "best available  technology economically achievable"
 (BAT) by 1983.

    Section  302   of  the   Act  set  ambient  water  quality
standards.  Ambient water  quality  standards  were  to comply
with  State  or  federal  law, whichever was more stringent to
achieve ambient water quality goals.

    Section 306 of the Act  pertains  to  new  sources.    This
section required  such facilities to meet standards equivalent
to 1983 BAT standards.

    Section 307  covers toxic  water pollutants.    It requires
that standards  be developed  for toxic water pollutants based
on public health and welfare and not technical feasibility.

    Section 402 of the Act empowers the federal  government to
create  a  National  Pollution  Discharge  Elimination  System
(NPDES).   This pollution permit system  empowers  the  USEPA to
set national  effluent standards and grants States, with USEPA
approval,  the  responsibility  of  administering   the program.
NPDES  applies  to  any  discharge to receiving waters.   NPDES
permits  had  to  incorporate   applicable  limitations  under
sections  301,   302,  306,   and  307  of  the  Act,  including
enforcement to meet 1977 and 1983 deadlines.

    Section 505 of provides  the  right  of  citizen  suits to
enforce  provisions  of  the  Act.    States  have the primary
responsibility to enforce the provisions of  the Act,  but the
Federal government  has the  right to  step in and enforce any
provision of the Act.


1.A.2 The Clean Water Act of 1977
    In  1977  the  FWPCA  was  amended  by  congress.    These
amendments  are  known  as  the  Clean  Water Act  (CWA).   Five
general categories of pollutants covered in the Act are;
i) conventional, ii) nonconventional, iii)  toxics,  iv) heat.

-------
and v)   dredge and  fill spoil.  The Act distinguishes between
new and existing source for setting effluent standards.  Table
1.2 lists examples of the first three pollutant catagories.

    Pollutants  designated  as  "conventional"  would  be  "as
defined by the administrator in  compliance  with  the  Act as
amended,  generally   those  pollutants   that  are  naturally
occurring,  biodegradable,  oxygen  demanding   materials  and
solids.   In addition, compounds which are not toxic and which
are  similar  in   characteristic   to   naturally  occurring,
biodegradable substances  are to be designated as conventional
pollutants for the purposes  of the  provision"  (Congressional
Research  Service-, 1977) .

    Pollutants designated as "nonconventional" would be "those
which are not toxic or conventional."   (Congressional Research
Service,  1977).   Table  1.3 illustrates the kinds of effluent
standards set by USEPA under the 1977 amendments.

    A   new   class    of   effluent   standards   called  "best
conventional pollution  control technology"  (BCT) were created
for conventional  pollutants.    Cost  consideration  could be
taken   into  account  by  USEPA  in  determining  BCT effluent
regulations  for   conventional   pollutants,   but   not  for
nonconventional pollutants or toxics.

    Congress modified  BAT standards in the Clean Water Act of
1977.   This action was in  response to  criticism the original
BAT  effluent  limitations  required  too  high  a  percentage
removal of pollutants  and  the  cost  of  reduction  in these
residuals  would  be  much  greater  than  the  benefits.  BAT
standards apply to unconventional and toxic pollutants.

    A variance provision for BAT standards for nonconventional
pollutants  is  contained  in  section  301(g) of the Act.   It
allows the USEPA along with State approval to  modify effluent
standards  for  nonconventional  pollutants  if  this  did not
interfere with water quality standards or public health.

    All river segments within  States are  classified as water
quality limited  or effluent  limited under  section 303 (e) of
the Act.   Effluent  limited  segments  are  defined  as those
stream reaches  for which  ambient water quality standards  can
be met in 1977  by  application  of  best  practicable control
technology currently available (BPT) to industry and secondary
treatment to publicly  owned  treatment  works  (POTW).   When
ambient  water  quality  standards  can  not be met by BPT  for
industry and secondary treatment  for POTW,   these reaches  are
classified as water quality limited.

-------
Table 1.2 Examples of Conventional,  Nonconventional,
and Toxic Pollutants
Source: Technical Guidance Manual For The Regulations
Promulgated Pursuant to Section 301(g) of the CWA 1977
Conventional
Nonconventional
biochemical
 oxygen demand
 (BOD)

PH

total suspended
 solids(TSS)

fecal coliform
 bacteria
chemical
 oxygen demand
 (COD)

chlorine

aluminum


barium
zinc

cyanides

toluene

benzene


asbestos
oil and grease
ammonia
                    copper

-------
Table 1.3 Examples of Technology-Based Effluent
Limitations Under The Clean Water Act of 1977
Source: Ortolano, 1984.
Publicly Owned Treatment Works:

Requirements for 85% BOD removal, with possible case-
-by-case variances that allow lower removal
percentages for marine discharges.
Industrial Discharges  (bases for effluent limitations)

Toxic pollutants - BAT

Conventional pollutants  - BCT; in determining required
control technology, USEPA is directed to consider  ''the
reasonableness of  the relationship between the costs of
attaining a reduction  in effluent and the effluent
reduction benefits derived."
Nonconventional pollutants - BAT, but with possible
case-by-case variances that allow for lower degrees  of
treatment.

-------
l.B Development of Mixing Zone Concept
    The concept  of a  mixing zone is defined as an "allocated
impact zone"  where  numeric  water  quality  criteria  can be
exceeded as long as acutely toxic conditions are prevented.  A
mixing zone can be thought of  as  a  limited  area  or volume
where  the  initial  dilution  of  a  discharge  occurs (Water
Quality  Standards  Handbook,1982).    Water  quality criteria
apply  to  the  boundary  of  the  mixing zone, not within the
mixing zone itself.  EPA  and  its  predecessor  agencies have
published numerous  documents giving  guidance for determining
mixing zones such as the  National  Academy  of  Science Water
Quality Criteria  1968  (Green  Book),  EPA publications Quality
Criteria for Water 1976  (Red Book),  and Guidelines  for State
and  Area  Wide  Water  Quality  Management Program.  Guidance
published by EPA in  Water  Quality  Standards  Handbook (Oct.
1982) supersedes these sources.

    In setting  requirements for  mixing zones, USEPA requires
that the size be "the area or volume of an  individual zone or
group of  zones be  limited to  an area  or volume as small as
practicable that will not  interfere with  the designated uses
or  with  the  established  community  of  aquatic life in the
segment for which the uses are designated" and the shape be "a
simple configuration  that is  easy to  locate in  the body of
water  and   avoids  impingement   on  biologically  important
areas.... shore hugging  plumes should be avoided."  Within the
mixing zone USEPA requires "any mixing zone should  be free of
point or nonpoint source related:

    (a)  Material  in concentrations that will cause acute
    toxicity to aquatic life;

    (b)  Materials in concentrations that settle to form
    objectionable deposits;

    (c)  Floating  debris, oil scum and other matter in
    concentrations that form nuisances;

    (d)  Substances in concentrations that produce
    objectionable color, odor, taste or turbidity; and

    (e)  Substances in concentrations which produce
    undesirable aquatic  life or result in a dominance of
    nuisance species."

(USEPA,  Water Quality Standards Handbook, 1982) .

    The proposed rules for  mixing zones  recognizes the State
has discretion  whether or  not to  adopt a mixing zone and to
specify its dimensions.  USEPA allows the use of a mixing zone
in permit applications except where one is prohibited in State

-------
regulations.    State  standards  require  that  water quality
criteria be met at the edge of the  regulatory mixing  zone 1)
to provide  a continuous zone of free passage that meets water
quality criteria for free-swimming and drifting  organisms and
2) to  prevent impairment  of critical  resource areas (USEPA,
Technical  Support  Document  for  Water  Quality-based Toxics
Control,  1985).    A  review  of individual State mixing zone
policies shows that 48 out of 50 States  make use  of a mixing
zone in some form and can be seen in Table 1.4

    For  discharges  into  streams,   17  of the 31 States that
propose a mixing zone specify that  the mixing  zone shall not
exceed 1/4  of the  cross-sectional  area  and/or volume of the
stream flow, and the remaining  3/4   of  the  stream  shall be
maintained  as  a  zone  of  passage for swimming and drifting
organisms.

    The remaining States  have  varying  requirements allowing
dimensions  of  the  mixing  zone  to  be as low as 1/5 of the
cross-sectional area (Ohio) to  as much  as 3/4  of the cross-
sectional  area  (South  Dakota).    West Virginia is the only
State that specifies a length dimension for mixing zones.  The
length  of  the  mixing  zone  must   be less than 10 times the
average width of the stream or less  than  5 times  the average
width of  the stream  for warm  water and  cold water streams,
respectively.

    In  States  which  specify   a  mixing   zone  for  lakes,
dimensions for  the mixing  zone vary  from 10% of the surface
area  of  the  lake  to  300-1000'  radial  limits  around the
discharge point.

    Pennsylvania and  Arizona are  the two  States that do not
make reference to a mixing zone.  Therefore the USEPA does not
recognize any  mixing zone  for these States and water quality
criteria must be met at  the  point  of  discharge  unless the
applicant and  the State  develop a  mixing zone  on a case by
case basis.

    With the exception of West  Virginia,  the  length  of the
mixing zone is not specified.  Usually, the size of the mixing
zone is determined on a case-by-case basis taking into account
the critical resource areas that need to be protected.  Mixing
zones should be used  and evaluated  in cases  where mixing is
not  complete  within  a  short  distance of the outfall.  EPA
recommends careful  evaluation of  mixing to  prevent zones of
chronic toxicity  that extend  for miles downstream because of
poor mixing.   (USEPA,  Technical  Support  Document  for Water
Quality-based Toxics Control, 1985).

-------
Table 1.4 State Legal Mixing Zones

Source: Draft Technical Guidance Manual for the
Regulations Promulgated Pursuant To Section 301(g]
State
Alabama
Alaska

Arizona
Arkansas
California
Colorado
Connecticut
Delaware

D.C
Georgia
Florida




Hawaii
Idaho
Illinois

Indiana
Iowa
Kansas
Kentucky
Louisiana
Maine
Maryland
Massachusetts
Michigan
Lake
Minnesota
Mississippi
Missouri
Montana
Nebraska
New Jersey
New Hampshire
New Mexico
New York
Water Body
0
river , streams
lakes
NR
large streams
0
0
streams
streams
lakes
estuary
0
streams , rivers

lakes, estuaries


0
0
all
streams
streams
streams
streams
streams
streams
streams
0
0
streams
Michigan
streams
0
streams
0
0
streams
streams
streams
streams
Dimensions
0
<= 1/3 CS
<= 10% SA
NR
<=l/4 CS
0
0
<=l/4 CS
<=l/3 CS
<=10% SA
<=10% SA
0
<=800 meters
<=10% total length
<=125,600 m**2
(600' radius)
<=10% SA
0
0
<=600' radius
<=l/4 CS
< = l/4 CS
<=l/4 CS
<=l/4 CS
<=l/3 CS
<=l/4 CS
<=l/4 CS
0
0
<=l/4 CS
<=1000' radius
<=l/4 CS
0
<= 1/4 CS
0
0
<=l/4 CS (thermal)
<=l/4 CS
<=l/4 CS
<=l/2 CS (thermal)
                          10

-------
Table 1.4 (Continued;
Nevada
North Carolina
North Dakota
Ohio
Oklahoma
Oregon
Pennsylvania
Rhode Island
South Carolina
S. Dakota

Tennessee
Texas
Utah
Vermont
Virginia
Washington
W. Virginia
               streams
                    0
               streams
               receiving
                watercourse
               mouth of
                receiving
               streams
                    0
                    NR
               streams
                    0
               streams

                    0
               streams
                    0
               streams
                    0
                    0
               warm water
               fish streams

               cold water
               fish streams
Wisconsin
Wyoming
Guam
Puerto Rico
               lakes
               streams
                    0
                    0
               streams
               IMZ
               FMZ
Virgin Islands streams

Where:
     <=l/3 CS
          0
     <=l/4 CS

     <=l/3 CS

     <=l/5 CS
     <=l/4 CS
          0
          NR
     <=l/4 CS (thermal)
          0
     <=3/4 CS or
100 yds of streams width
          0
     <=l/4 CS
          0
     <=l/4 CS
          0
          0

     <=33% CS,   length
         <=10*width

     <=20% CS,  length
         <=5*width

     <=300' any direction
     <=l/4 CS
          0
          0
     <=l/4 CS
     <= 400 '
     <= 4000 '
     < = 1/4 CS
CS = cross-sectional area
NR = no reference
IMZ = initial mixing zone
                              SA = surface area
                              0 = not listed
                              FMZ = Final mixing zone
                          11

-------
l.C  Special Mixing Zone Requirements for Toxics
    In order to prevent lethal concentrations of toxics in the
regulatory  mixing  zone,   regulations   can  prohibit  lethal
concentrations  in   the  pipe   itself,  or  require  that  a
concentration known  as  the  criterion  maximum concentration
(CMC) be  met within  a short  distance from the outfall.  The
CMC  is  a  concentration  that  prevents  lethality  or acute
affects in tested species.  If dilution of the toxic discharge
in the ambient environment is allowed, this requirement, which
will be  defined here  as a toxic dilution zone  (TDZ), is more
restrictive than  the legal  mixing zone  for conventional and
nonconventional  pollutants.    In  order to provide turbulent
mixing that will  minimize  organism  exposure   time  to toxic
material,  the  outfall  structure  must  meet   the  following
requirements for a TDZ  (USEPA, Toxics, 1985):

     -The CMC must be met within 10% of  the distance  from the
     edge  of   the  outfall  structure  to  the  edge  of  the
     regulatory mixing zone  in any spatial direction.

     -The CMC must be  met within  a distance  of 50  times the
     discharge  length scale  in any spatial direction.  The
     discharge  length  scale is  defined as  the  square-root of
     the  cross-sectional   area of  any discharge  outlet.  This
     restriction will  ensure a  dilution factor  of  at least 10
     within  this  distance   under  all   possible  circumstances,
     including  situations   of  severe   bottom    and    surface
     interaction.

     -The CMC must be met  within  a  distance  of  5  times the
     local water  depth  in any horizontal direction.  The local
     water   depth  is  defined  as   the   natural  water  depth
     (existing  prior  to  the  installation  of  the discharge
     outlet)  prevailing   under  mixing   zone  design  condition
     (e.g. low  flow  for  rivers).  This restriction  will  prevent
     locating the discharge  in  very  shallow   environments or
     very  close  to  shore,  which  would result  in significant
     surface  and bottom  concentrations.
 l.D  The  NPDES  Permit  System
     Any  discharge  into  a   navigable   watercourse   must   have  a
 National   Pollution    Discharge   Elimination   System   (NPDES)
 permit.  The permit is  designed   to  insure   that  the discharge
 meets  all   applicable standards.   The permit is granted  either
 by USEPA,  or,  if the  State has a   USEPA  approved   program,  the
 State.   The applicant  must  supply   the  reviewing agency  with
 all  data needed to  grant  the  permit.    Data  required   in the
 application include:

       Name  and exact  location of  facility

                              12

-------
    - Nature of business engaged at a facility, including what
    is or what will be manufactured

    - The manufacturing process and maximum production levels

    - Schematic of water flow through the facility

    - Exact  location,  'flow  rates,    flow  frequencies,  and
    chemical composition of each facility discharge

    - The  waste-water  treatment  currently or to be employed
    for each waste stream

    - Pollutant test data
l.E Need for Regulatory Assessment Tools
    Implementation of the  mixing  zone  policy  in  the NPDES
permitting process requires that the applicants and regulators
predict  the  initial  dilution  of  the   discharge  and  the
characteristics  of  the  mixing  zone.    If the discharge is
toxic, the CMC value must be determined for  the discharge and
special requirements  for a  TDZ must be met within the mixing
zone.    Given   the   large   number   of   possible  ambient
environments,   discharge   configurations,  and  mixing  zone
definitions,  the  analyst  needs  considerable  training  and
experience  to  conduct  accurate and reliable effluent mixing
analysis.

    Dilution of the effluent in the receiving water  is caused
by different  mechanisms along  its path.  In the "near field"
of the  source,  dilution  is  caused  mainly  by  jet induced
entrainment.   Further away,  in the so-called "far field" the
jet  velocity  decreases  and  ambient  diffusion  becomes the
primary mechanism of effluent dilution.

    The most direct way of determining pollutant concentration
downstream is by physical measurement.   Non-polluting tracers
can also be injected to give indications of  effluent dilution.
Such  field studies require  considerable time  and effort, and
field personnel  need specialized  training  to perform studies
reliably.   Field studies are  in  many  cases  impractical and
expensive.  For example, if in situ observations are used  they
must  represent conditions  that  are  present  during critical
dilutions,  not  merely  a  typical  dilution  (Draft 301  (g)).
Field studies for analysis of dilution for toxic discharges  is
patently unacceptable, so simulation must  be used  to determine
dilutions.
                              13

-------
    Because of  the complexity of the physical mixing process,
permit writers are increasingly relying on mathematical models
to analyze  the fate and transport of pollutants  (Tait, 1984).
The difficulty with many present models  is that  they tend to
become  specialized  and  give  accurate  results  only  for  a
particular type of outfall.  The user must be careful to use  a
model  that   was  intended   to  make  prediction  under  the
conditions with which he is concerned  (Draft 301  (g)).

    USEPA has developed a  number  of  models  to  predict the
initial  dilution  of  discharges.    A  few  these are PLUME,
OUTPLM, DHKPLM, MERGE, and LINE.  Applicants are  not required
to use  these models  in analysis,  but must  be  able to prove
that  the  methodology  chosen  gives  reasonable  estimates of
initial dilution.
l.F Justification for Expert Systems Approach
    In determining the characteristics of the mixing zone, the
analyst, either the NPDES  applicant or  regulatory authority,
may  choose  from  a  wide  variety of predictive models.  The
models range in complexity from simple  analytical formulae to
highly   intricate   numerical   solutions   to   differential
equations.  Although the USEPA has prepared assessment manuals
and actually  endorsed certain  models in specific situations,
the average user has little reliable  guidance on  which model
is  appropriate  for  a  particular  situation,  or  which  is
actually best (Mullenhoff, et. al., 1985).  Examples of "model
abuse" are ubiquitous.  Often unnecessarily complicated models
are employed, creating a  needless burden  for both regulators
and dischargers.

    Even when  a particular  model is  appropriate for a given
discharge,  the model may not give reliable results over a wide
range  of   conditions.      Model  developers  often  fail  to
explicitly specify limits of applicability,  or model users may
simply overlook important restrictions to model applicability.
An example of a frequent error in the application of the USEPA
plume  models  is  the  violation  of  the  assumption  of the
infinite receiving environment.    In  reality,   the  plume may
attach  to  the  bottom  or may become vertically fully mixed,
possibilities which may occur  due to  changes in  the ambient
environment  such  as  low  flow  conditions.    Consequently,
analysts have reported model "predictions" in  which the plume
diameter exceeds actual water depth!

    Once the  correct choice  of model is assured, the analyst
often faces the considerable  task of  assembling the required
design data  base.   This can  be a frustrating and cumbersome
task for the unexperienced analyst who has little  guidance on
what design  base to  choose, where to obtain data, which data

                              14

-------
is  crucial  to  the  analysis,   and  which data may simply be
estimated.   Because of these difficulties,   a large investment
in time  is required  for the analyst to become fully familiar
and proficient with the  use of  at least  one model,  or more
likely, a group of models.  The analyst in reality must become
highly skilled or an "expert" in the use and interpretation of
number  of  simulation  models.    Such expertise in model use
requires expensive training and is rare.   This  is the reason
for the development of expert system tools for the analyst.

    In  essence  expert  systems  mimic  the  way an expert or
highly experienced person would  solve a  problem.   An expert
system is  a structured  computer program  that uses knowledge
and inference procedures obtained  from experts  for solving a
particular type  or class  of problem called a "domain".   This
knowledge is encoded into  a  "knowledge  base"  which enables
inexperienced personnel to solve complex problems by using the
same basic reasoning process that an expert would  apply.  The
knowledge  base  includes  a  set  of  "objective"  or  widely
accepted facts about a  general problem  area.   This includes
the set of parameters or data an expert would seek in order to
characterize a specific problem.  The inference procedures are
"subjective"  rules  of  judgement  which the expert might use
when analyzing the problem.  The  inference procedures provide
the rules for selecting an appropriate solution to the problem
from the knowledge base.  The  inference procedures  allow the
expert  system  user  to  search  rapidly  and  systematically
through the knowledge base to obtain  a solution  to the given
problem. This  element uses structured search techniques based
upon mathematical logic.

    The  development  of  an  expert  system  for  mixing zone
analysis   promises   significant   advantages  compared  with
existing   conventional   simulation   techniques   for  water
pollution control and management:

    -it  assures  the  proper  choice  of  model  for  a given
    physical situation.

    -it assures that the chosen model  is applied methodically
    without skipping essential elements.

    -it  guides  the  acquisition  or  estimation  of data for
    proper model prediction.

    -it allows a flexible application of design strategies for
    a  given  point source,   screening of  alternatives,  and if
    necessary,  switching  to different  predictive models thus
    avoiding rigid adherence to a single model.
                              15

-------
    -it  flags   borderline  cases   for which  no  predictive model
    exists,  suggesting  either  avoidance  of   such  designs  or
    caution  by  assigning  a degree of uncertainty.

    -it  allows   a  continuous   update of   the knowledge  base  as
    improved predictive  models,  experimental   data,  and field
    experience  with  particular designs become  available.

    -it  provides  a  documented  analysis  listing the  knowledge
    and  decision logic  that have  lead to  the problem  solution.
    Thus,  unlike  conventional programs  or  computer algorithms
    an expert  system is not a  "black box".

    -it  provides a common  framework  whereby   both regulators
    (Federal  or   State), applicants,   and   the  scientific
    community  can arrive  at  a consensus  on  the state-of-the-
    art  hydrodynamic mixing and  pollution control.

    -finally,   and  perhaps  most  importantly,   it provides a
    teaching environment  whereby  the  initially inexperienced
    analyst   through  repeated  interactive use gains physical
    insight  and understanding about initial mixing processes.

    Expert  systems  are  a   technology  that   has  enormous
potential  utility   to  solving   problems  in  environmental
science.  At the present time,  several  preconditions  must be
satisfied  before this technology can be applied successfully,
such as (Barnwell et al. ,  1986):

    -The problem domain must  be  narrow,  e.g.  restricted to
    a particular well defined problem area.

    -Expertise  to   solve  the   problem  must  exist;   Expert
    Systems  cannot create expertise,  they  can  only document,
    disseminate, and enhance  it.

    -Formalization of concepts involved must be available in  a
    format compatible with the tool used.

    If these preconditions can  be  satisfied,  Expert systems
appear to  be  a  powerful addition to the analyst's repertoire
of solution  faculties.  The  analysis  and   simulation  of the
discharge  of    pollutants  into   water  courses  meet  these
preconditions.   Experts in hydrodynamic mixing exist.  It is  a
well defined  problem area,  the problem being too complex for
glib analysis,  but  not  too   large  to  codify   the knowledge
needed to  solve the  problem nor intractable to  the novice if
expertise is available.  Mixing  zone  analysis appears ideally
suited for exploiting Expert  Systems technology.
                              16

-------
l.G CORMIX1 Summary
    The  problem  addressed  is  to  develop  a  tool  for the
analysis  and  design  of   submerged  single-port  continuous
buoyant discharges  into a non-stratified aqueous environment.
The expert system will be labeled CORMIX1,  for  Cornell Mixing
Zone Expert  System,  Subsystem  :L.    CORMIX1 is a subsystem of
CORMIX,  to  be  developed,  which  will   include  stratified
environments,   negatively-buoyant   discharges,   and  bottom
attachments.   CORMIX1 is  primarily intended  for applications
to flowing ambients(such as rivers or estuaries), although the
limiting  cases  of   non-buoyant   discharges   and  stagnant
environments  are included.  The emphasis of CORMIX1 will be on
discharge geometry and characteristics  of   legal  mixing zone
(LMZ) requirements,  including the  toxic dilution zone (TDZ).
CORMIX1  will   summarize  dilution   characteristics  of  the
proposed  design,  flag  undesirable  designs,  give  dilution
characteristics at legally important regions if specified, and
will have  the capability  to recommend  design alterations to
improve dilution characteristics.

    The  subsequent  chapters  in  Volume  I  are  Chapter  II
Hydrodynamic   Background  on  Mixing  Processes,  Chapter  III
CORMIX1 Program  Structure,  Chapter  IV  Data  Comparison and
Validation,  Chapter  V  Design  Case  Studies, and Chapter VI
Conclusions  and  Recommendations.      Chapter   II  contains
discussions of  the hydrodynamic analysis used to simulate the
jet mixing  process.    Chapter  III  details  CORMIX1 program
structure,   logic   programming,   and   FORTRAN   simulation
programming.     Chapter  IV  presents  a  validation  for  the
hydrodynamic   simulations   by   comparison  with  field  and
laboratory  data.    Chapter  V  contains  two  typical design
examples  demonstrating  the  use  of  CORMIX1 for mixing zone
analysis.    Chapter  VI   outlines  conclusions   of  CORMIX1
performance and recommendations  for future improvements.

    Volume II  contains the  CORMIX1 source   code  listings and
the  output  from  an  interactive  session   using  the design
examples presented in Chapter V  of this volume.
                              17

-------
                         Chapter II
         Hydrodynamic Background on Mixing Processes
    The  hydrodynamics  of  a  buoyant  effluent   continuously
discharging  into  a   receiving   body   of   water   can  be
conceptualized as  a mixing  process occurring in  two  separate
regions. In the first region, the  initial jet characteristics
of  momentum   flux,  buoyancy   flux,  and  outfall   geometry
influence the jet trajectory and mixing.  This  region will be
referred  to  as  the  "near  field",  and encompasses the jet
subsurface flow and  surface  impingement.    In   this region,
designers  of  the  outfall  can  affect  the  initial mixing
characteristics  through  appropriate  manipulation  of design
variables.

    As  the  turbulent  plume  travels  further  away  from the
source the characteristics of  the issuing  source become less
important.     Conditions  existing  in the ambient environment
will control  trajectory and  dilution of  the turbulent plume
through buoyant surface spreading and passive diffusion.  This
region will be referred to here as the "far field".

    The hydrodynamic  analysis treats  the near  field and far
field regions  separately.   An illustration of the near field
and the far field of a subsurface plume rising to  the surface
and traveling downstream appears in Figure 2.1.

    The strategy for analysis will be to first present the
mechanics  of subsurface near field properties for pure jets
and pure plumes in stagnant ambient environments,  extend these
results to  ambient crossflows,   and finally  to generalize the
analysis  to  flows  containing  momentum  and  buoyancy.   The
general  results  will  be  extended  to  include non-vertical
trajectories  and  transitions  between  flow  regions.    The
effects of   flow surface  interaction will  then be discussed,
followed by a discussion of the far field mixing process.

2.A Analysis of Subsurface Flow Regions
    A release with no buoyancy is referred to as a "nonbuoyant
jet" or "pure jet" .  A  release of  buoyancy only  (no initial
momentum)  is called a "pure plume".  A release containing both
momentum and  buoyancy  is  designated  a    "buoyant  jet" or
"forced plume".   For  simplicity, a  region within the actual
pure jet,  pure plume, or  buoyant  jet  or  forced  plume will
often be referred to as a "flow".

-------
  Plan  View
   u.
  Side  View
                  Near Fied
Far  Field
Figure 2.1  Illustrative Near Field and Far Field of Submerged Single Port
            Discharge
                                   19

-------
    For a  buoyant jet in a stagnant unstratified environment,
List and Imberger  (1973)  propose  three  flow  regions where
buoyant jet  behavior is  determined by different effects.  In
the first region, near the issuing source, the geometry of the
discharge  is  important.    In  the  second  region,  initial
kinematic momentum flux of the discharge predominates.  In the
third and  ultimate region,  yet further away from the source,
buoyancy flux  of  the  initial  discharge  becomes important.
Characterizing  the   flow  by   the  predominating  mechanism
controlling  the  flow  within  a  region  is  the  essence of
"asymptotic analysis" which will be pursued herein.

    The  effects   of  momentum   and  buoyancy  thus  can  be
considered separately  to  reduce  the  number  of independent
variables under  consideration.  For example, the solution for
a  pure  jet in  a crossflow  can be  applied as  an approximate
solution  to  that  portion  of . buoyant  jet  flow  where jet
momentum dominates the flow.  Likewise the results  for a pure
plume can be applied  to  the buoyancy-dominated regions for the
buoyant plume.

    Additional factors,  such as crossflow velocities, can also
be treated  within   the   framework  of asymptotic analysis as
shown by Wright  (1977).

2.A.I Description  of  Turbulent  Jecs and Plumes
    Most people  are  familiar  with  the  sight of   smoke rising
from a   smokestack into  the atmosphere.   The  smoke plume  rises
and spreads  narrowly;  rising   near  vertically   at   first and
eventually bending   over as   it is  carried away by the ambient
wind.   The smoke plume  is  an  example of a turbulent  plume, the
discharge  contains   both momentum   and buoyancy.   The buoyancy
is produced  by   the   lower density  of   the  heated  air with
respect to the  cooler ambient  air.

    The discharge  of a   fluid  such  as sewage   into  the  ocean
behaves in a similar   fashion.    The sewage   has  momentum from
being  injected   through  the discharge  orifice.   The  sewage has
the density  of  freshwater and  thus  is  buoyant   with  respect  to
the greater  density  of  the ambient  saltwater.

    Turbulent   jets    are  characterized   by   a   long  narrow
turbulent  zone.   Following release  from  a nozzle,  the jet flow
becomes unstable  at  its boundary   and breaks  down into  the
turbulent  motion.  Typically,  the  size  of the  turbulent eddies
increases  with    increasing   distance   along   the  trajectory
 (Holley and  Jirka, 1986) .

2. A.2  Elements  of  Dimensional  Analysis  of Buoyant Jets
     Several  assumptions  are made  in  order to reduce  the
independent   variables    under   consideration.     Only   fully

-------
turbulent jets  are considered so the effects of viscosity  can
be  neglected.    The  Boussinesq  approximation   is   assumed;
density   differences   between   the   jet  and   the   ambient
environment are small and are important  only in   terms  of  the
buoyancy force.

    The   three   variables   used  to  describe   buoyant   jet
characteristics in crossflow are the kinematic  fluxes  of mass,
Qo =   (n/4)D2u0 , momentum  Mo =  Uo Qo , and buoyancy Jo  = gi Qo ,
where D is  the  diameter  of  the  orifice,  uo   is   the exit
velocity, ua   is the crossflow velocity, and gi  is the reduced
gravitational  acceleration caused  by  the  density difference
between the jet and ambient  environment.  This  term is defined
as  g'0  =  g (pa  -  PO)/PJ   where  g   is  the gravitational
acceleration  and  pa and  po are the ambient and jet  discharge
densities, respectively.

    For the case of  a buoyant   jet discharged   into  a flowing
environment,   dimensional  analysis  proceeds   as  follows.   A
general dependent  variable, ij>,  such  as   local centerline  jet
velocity,  can  then be   expressed  as a function of the various
independent variables:

          <|> =   f (Qo ,Mo , Jo , ua , s)                         (1)

where  s  is  the distance  along  the  jet  trajectory.   A function
is  then   empirically   created   by  grouping   the   independent
variables  on  the right  hand side   of  Eq.   (1)  together.   The
created  function   has  to  be dimensionally consistent with  the
desired  dependent  variable-

    First,  the following  paragraphs   present   the  details of
dimensional analysis   for the   simple  case of  a pure  jet and a
pure  plume in a stagnant environment.   Then,   the  general  case
of  a  buoyant  jet   in   crossflow  is presented.   The jet  and
ambient  flow  variables can  be   combined   into  various length
scales  that   measure   the   relative   forces   affecting a  flow
within a  particular  trajectory  distance.

    The  asymptotic  approach  will  provide  solutions   that  are
valid  only  within   certain specified regions and furthermore
require    experimentally   determined    coefficients.      The
individual solutions,   however,   can  be  linked by appropriate
transition conditions  to provide  an overall   prediction which
can be  considered  a  first order approximation to the complete
problem.

2.A.2.1.  Simple Jet  in Stagnant Environment
    Consider  a pure  jet  in a  stagnant ambient  fluid   (Figure
2.2).    Initially  as   the flow  exits  the orifice the velocity
profile  is near uniform.   After a  short  distance  s along the

                               21

-------
                        a.  Instantaneous appearance
                       ENTRAPMENT
                       VELOCITY
     Uo, po • pa, c0
                                                       CONCENTRA TION
                                                       PROFILE
                                                     AMBIENT DENSITY pa
                          • ZONE OF
                          FLOW ESTABLISHMENT
                       b.  Time-averaged  conditions
Figure 2.2   Pure  Jet  in Stagnant Environment  (Ref.  Holley and
Jirka,  1986)

                                   22

-------
jet  trajectory,  the  velocity  distribution  is  assumed  to  be
gaussian.    The  region  where  this   velocity   distribution
transformation occurs is called the zone  of  flow  establishment
(zofe).   The details of the zofe will not be considered in any
of the  following analysis;  i.e. the   jet is   assumed to come
from a point source.

     The  maximum  velocity  uc  occurs   along  the  trajectory
centerline  and  a  similarity  profile   is  assumed  for the
velocity distribution.    Similar  conditions   pertain  to the
centerline  concentration  cc   of  pollutant  mass.    The jet
centerline velocity uc  decreases  with   distance  s  from the
orifice  as  the  jet  entrains  the  stagnant  ambient fluid.
However,  the  momentum  flux  M  at  any  section   along the
trajectory is  conserved.   The magnitude and variation of the
jet centerline  velocity  depend  primarily  upon  the initial
kinematic momentum flux and the distance  along  the trajectory,
Uc = (Mo , s) .  Using techniques  of dimensional   analysis, the
result implies that ucs/Mo1/2  = constant

         Uc  = cMo ! / 2 s- i                                (2)

where c  is a constant.

    The  width  b of  the jet at trajectory distance s can also
be expressed as b = (Mo,s).  The  only possible dimensionally
consistent equation is b/s = constant

         b = bis                                      (3)

where bi is a constant.

    The  dilution  S  at  any  cross-section  along  the jet  is
defined by S =  Co/cc , where  Co is  the  concentration  at the
exit  nozzle.    From  the  mass  conservation  equation CoQo  =
ccUcb2 ,  and dilution S as a function of s

         S = siMo i /2 s/Qo                              (4)

where si is a  constant.

2.A.2.2. Simple Plume in Stagnant Environment
    A pure plume  rises  vertically  and  has   an increase  in
vertical  momentum  flux  with  distance  z  above   the source
(Figure  2.3).  The  buoyancy flux  is constant  for  any cross-
section  of  the plume  as it  rises.    For the  pure  plume,  the
centerline velocity is a function  of   the   buoyancy flux  and
distance z,  uc = 4>(Jo,z).  The centerline velocity  of the  jet
can be  obtained from  dimensional reasoning

         Uc    c(Joz)1/3                                (5)

                               23

-------
                          a.  Instantaneous appearance
                   £NTRAINUENT ^
                   VELOCITY
                                                    AMBIENT
                                                    DENSITY pt
                                        CONCENTRATION AND
                                        BUOYANCY PROFILE
                                        VELOCITY PROFILE
                          b.  Time-averaged conditions
Figure 2.3  Simple  Plume  in  Stagnant  Environment  (Ref. Hollev
and  Jirka, 1986)
                                    24

-------
    The  width  b  of  the  plume  at trajectory distance z is
expressed as b = (Jo >z) ,  leading to

         b = b3 z                                       (6)

where ba  is a constant.

    Dilution S for the pure plume is given  by mass continuity
equation as  was shown  in Eq.   (4) for  the pure jet.  Noting
that  buoyancy  flux  is  conserved  in  the  pure  plume, the
dilution  for  the  pure  plume  can  expressed by the buoyant
acceleration g' which decreases with distance  z as   the plume
rises and  becomes diluted by the ambient fluid.  The decrease
in g' is directly proportional to the amount  of ambient fluid
entrained in  the plume,  so  S  = gd/g'.  Using the continuity
equation  for  buoyancy   flux,  dilution  can  expressed  S  =
4>(Jo,Qo,z) as g' z5 / 3 /Jo 2 ' 3 =  constant

         S = S3 Jo1 /3 z5 /3 /Qo                           (7)
2.A.2.3 Generalizations: Jet/Plume  Interactions  and  Effects  of
Crossflow
    If additional parameters  influence  the  flow   field,  then a
general asymptotic   solutions  for the whole  flow field  can not
be  found.  However,   there  may  be  individual   regions  where
specific  asymptotic   solutions  of  the  type  developed  in the
preceding sections may still  apply.  The  next  section presents
the analysis  for the  discharge of the simple jet or  plume into
a uniform ambient crossflow ua.  As  the jet  or  plume rises  it
will be deflected by  the ambient current  as  illustrated  in the
several examples of  Figure 2.4.

    As can be seen in  Figure  2.4   there  are  still specific
regions  where  the   flow  exhibits  certain  simple behavior.
These regions are separated by the  transition  zones  which are
described by  length  scales  (labeled  IM, 1m,  lb in Figure  2.4).
The development of these important  length scales which  specify
the spatial   distribution of  the asymptotic  regimes  of  general
buoyant jets  in crossflow is presented  in the  next section.

2.A.3 Length  Scales  for Buoyant Jets With or Without Crossflow
    In general, functional relationships  in  the  form of  length
scales are  sought which  describe  the  relative importance  of
discharge flux to momentum  flux,  momentum  flux to buoyancy
flux,  and discharge  to  crossflow in  controlling  flow behavior.
The length scales will  describe the  distance over  which the
dynamic quantities in  Eq. (1) control the flow.

2.A.3.1 Discharge Length Scale
    Initially as the  jet exits the port in the zone  of  flow

                              25

-------
                         Transition
                Discharge
           ua=0
                                               Plume-like
                                       like
                     a)  Buoyant Jet in Stagnant Environment
                                \Transition
           Ur
                                                        Jet
                                     Deflected  Jet

                     b)  Pure  Jet in  Crossflow
           u,
                            //  )   Strongly Deflected Plume

                               /Weakly Deflected Plume
                     c) Pure  Plume  in  Crossflow
Figure   2.4   Examples   of  Combined  Effects of  Momentum Flux
Buoyancy Flux and Crossflow on  Flow Behavior
                                  26

-------
establishment, port  geometry controls the flow.  The distance
over  which  the  port   has  effect   on  the   flow  can  be
characterized as  a discharge length scale.  Momentum controls
the flow initially for the buoyant jet.   The discharge length
scale  IQ  relates  the  mass  flux to momentum flux, and from
dimensional reasoning
= Qo /Mo1 /2
                                                       (8)
which is proportional to the diameter D of  the orifice  for a
round jet.  For distances less than IQ the flow will be in the
zone of flow establishment.  Thus for s/le « O(l)  the source
geometry will  have a significant effect on the flow behavior,
but for S/!Q »  O(l) the  effect of  the initial  geometry is
lost to  jet momentum  or buoyancy which will control the flow
behavior.

2.A.3.2 Jet/Crossflow Length Scale
    The presence  of a  crossflow ua  will deflect  the jet as
shown  in  Figure  2.4b.    The  behavior  of  the pure jet in
crossflow depends on the  relative magnitude  of the crossflow
to jet  momentum.   The distance to the position where the jet
becomes strongly affected (i.e.  deflected in  the case  of an
oblique  discharge)  by  the  ambient  crossflow is given by a
jet/crossflow length scale 1m

         1m = Mo1 /2 /Ua .                                  (9)

    Thus for  s/lm  «   O(l)   the  initial  jet  momentum will
dominate  and  crossflow  is  of secondary importance,  and for
s/lm » O(l) ambient velocity  will  have  the  most important
influence on jet behavior.
2.A.3.3 Plume/Crossflow Length Scale
    Arguments  presented  for  the  effect of crossflow on the
pure plume flow are in analogy for the jet in  crossflow.  The
plume/crossflow length  scale It   for the pure plume rising to
be deflected by the  crossflow  as  shown  in  Figure  2.4c is
determined through dimensional reasoning
         lb
= Jo /Ua3
do:
    Thus  for  z/lb  «  0(1)  the  initial  jet buoyancy will
dominate and crossflow is  of  secondary  importance,  and for
z/lm  »  0(1)  ambient  velocity will have the most important
influence on plume behavior.

2.A.3.4 Jet/Plume Length Scale
    Most flows contain both momentum and buoyancy.  Because of
the  buoyant  acceleration,  any  jet containing buoyancy will

                              27

-------
have a  momentum flux  continually added   to  it.    When  a  flow
contains both momentum  and  buoyancy,  momentum   controls  the
flow  initially  until  the buoyant acceleration  overcomes  the
effect of initial momentum  and ultimately  controls  the flow.
The  distance  from  momentum  dominated to buoyancy  dominated
flow  for  a  buoyant  jet  in   a  stagnant   environment   is
characterized  by  a  jet/plume  length  scale  IM  (See  Figure
2.4a).     Dimensional   analysis   suggests   the  functional
relationship

         IM - Mo3/4/Jo1/2                              (11)

    So for  Z/!M «  0(1) flow  behavior will be  controlled  by
momentum and for z/ln » 0(1) flow behavior will  be controlled
by buoyancy,  i.e. approach that of a vertically  rising-plume.

    In the rare case that IM < IQ,   there will  be  no momentum
dominated flow and the flow will be entirely plume-like expect
for the region very near the issuing source.

    The ratio of IQ/!M  is proportional  to the   reciprocal  of
the usual discharge densimetric Froude number Fo   = uo/(g6D)1/2
which relates  the  momentum  forces  to  buoyancy  within the
plume. The  theoretical pure plume has a Froude number of 0(1)
and a pure jet Fo  -> ».

2.A.4 Typical Regimes of Buoyant Jets
    This section  presents  analytical  results   for  jets and
plumes   issued   vertically   upward  from  a  point  source,
perpendicular to the crossflow.

2.A.4.1 Weakly Deflected Jet in Crossflow  (mdnf)
    For a relatively weak crossflow,  the jet would  behave the
same as  if it  were in a stagnant environment, except that  it
is slightly advected by  the  ambient  current  (Figure 2.4b).
This region is defined for z/lm « O(l).

    Considering a  jet issued  perpendicular to the crossflow,
after the region of flow establishment the vertical relation
would be similar to Eq.  (2) and the kinematic relationship
for a jet moving horizontally with the crossflow  velocity

         dx/Ua = dz/Uc                                 (12)

Substitution for the vertical  velocity given  in Eq.  (2)  and
integrating gives the trajectory relationship for the weakly
deflected jet  flow  (Wright's   (1977)  "momentum-dominated  near
                               28

-------
 field",  or  mdnf1)  expressed  in  terms of the jet/crossf low
 length scale

          Z/lm = tl (X/lrc)1 /2                             (13)

 where ti  is a trajectory constant.

     This relation holds  in the region z/lm  « 0(1).   Eq .  (13)
 is valid  for small  values of  lg /lm .   In the rare  case  that
 lc /1m is large, the  effect of  geometry is  important  and Eq .
 (13) can not assumed to  be valid.

     Jet width  b  is  similar to   the  jet  issued in a stagnant
 environment and is given by Eq.  (3) .

     The dilution  S is similar to  Eq .  (4), and is  expressed  in
 terms of IQ
           S = si ( z/le )                                 (14)

 where si is the dilution constant for  the mdnf  flow.
 2.A.4.2 Strongly Deflected Jet in Crossflow  (mdff)
     For z/lm  » O(l) the ambient flow will  have  a more direct
 effect on the flow pattern.  For a strongly  deflected  jet  the
 vertical velocity  has decayed  to less  than the  value for  the
 ambient   crossflow;   thus   ambient    crossflow  will  have
 significantly deflected the jet as shown in  Figure 2.4b.

     The behavior of  the bent-over jet is assumed  to be roughly
 equivalent to that of  a cylindrical  line impulse  located at
 the  same  vertical  rise.    Scorer  (1954)  describes a line
 impulse as an instantaneous release of   nonbuoyant fluid along
 a horizontal  line source.   The  characteristic  variables  are
 the momentum impulse,  M1(defined  as  the   kinematic momentum
 flux  per  unit  length  for an infinitesimal period of time),
 vertical rise, and time.  Applying dimensional analysis

          M't/z3  = constant                             (15)
     1 In  the  following  the  abbreviated  descriptions  for
subsurface flows  (mdnf, mdff,bdnf  and bdff) as suggested by
Wright (1977) will be used  for  convenience  since  they are
frequently used in the literature.  Care must be exercised in
their interpretation so as to avoid  confusing them  with the
designation  "near-field"  and  "far-field"  as  used in this
study (See introductory comments  at  the  beginning  of this
Chapter).

                               29

-------
    To  apply   this  analogy   to  the  pure   jet,   Mo/ua   is
substituted for M' and x/ua  replaces  t  in  Eq.   (15).    The
trajectory relation  for the strongly deflected jet  flow  (i.e.
Wright's (1977) "momentum-dominated far field", mdff)   is then
expressed in terms of the  jet/crossflow length  scale

         Z/ln, = t2 (Z/lm )1 /3                            (16)

where t2 is a trajectory constant.

    Similar to Eq.  (3) jet width b is proportional  to position
z
    b = b2 z                                            (17)

where b2 is a constant for the mdff flow.

    The continuity equation  is used to  determine  the dilution
at any  position  z,   CoQo=cb2ua . In terms of  the  jet/crossflow
length  scale the  dilution

          S  = s2 (z2 /1m IQ )                               (18)

where 32  is  a  dilution  constant  for the mdff  flow.

2.A.4.3 Weakly  Deflected Plume  in  Crossflow (bdnf)
    For a relatively  weak  crossflow,   the  pure   plume  would
behave  the  same  as  if   it   were in  a  stagnant  environment,
except  that  it  is  advected  with   the  ambient  current (Figure
2.4c) .

    For values   of z/lc  « O(l),  the  flow will  behave as  plume
in a ' stagnant   environment   but   will   be   advected  with  the
crossflow.   Proceeding in  analogy to   the mdnf flow (Section
2.A.4.1)  the   trajectory  equation for   the   weakly deflected
plume   flow  (i.e.  Wright's  (1977)   "buoyancy-dominated near
field", bdnf).    The  relationship in   terms  of   the buoyant
length  scale

          z/lb  - t3 (x/lb )3/4                             (19)

    Plume width b is similar to  the plume issued in a stagnant
environment and is given by  Eq.  (6).

    The dilution S is similar  to  Eq.  (7),  and is  expressed in
terms of  lb, IQ,  and 1m

          S  =  S3 (Ib^'z3 '3 ) /{lei- )                      (20)

where 33  is  the dilution constant  for  the bdnf  flow.
                               30

-------
2.A.4.4 Strongly Deflected Plume  in  Crossflow  (bdff)
    For z/lb  » O(l)  the ambient flow will have  a  pronounced
effect on the flow pattern.  When strongly  defected,  theplume
vertical velocity  has decayed  to less than the value for  the
ambient   crossflow;   so   ambient    crossflow   will   have
significantly deflected the plume as  shown  in  Figure 2.4c.

    For z/lb » 0(1), the plume should behave  as a thermal,  an
instantaneous release of buoyancy-driven  fluid  along  a line
source.  Again,  note  the  analogy   to the mdff flow  (Section
2.A.4.2)  caused  by  the  line   impulse  M1.    The important
variables are J',  the buoyant weight  per unit  length,  vertical
rise, and time.  Dimensional reasoning implies

         J't2/z3 = constant                            (21)

Substituting x/ua  for t and replacing J1 by Jo/ua in  Eq. (21)
yields the  trajectory relationship   for the strongly  defected
plume flow  (Wright's  (1977)  "buoyancy-dominated  far field",
bdff) expressed in terms of length scales

    z/lb = t4 (x/lb)2'3                                 (22)

where t4 is a constant.

    Plume width b is analogous to Eq. (6)

    b = b4 z                                            (23)

where b4 is a constant for the bdff flow.

    The continuity  equation is used  to determine the  dilution
at any position s,  coQo=cb2ua. In  terms of  the jet/crossflow
length scale the dilution

         S = s< z2 / (lc1m )                               (24)

where Si is a dilution constant for the bdff flow.

    The   various   trajectory    and   dilution  relationships
presented in the previous  sections   are  summarized  in Table
2.1;  Table   2.2  lists   tentative  constants  proposed   for
simulation studies  for  CORMIX1.     These  values   are  to be
further determined in future work.

2.A.4.5 General Behavior in Unbounded Crossflow
    The general case of the trajectory and  dilution  for  a flow
containing both momentum and buoyancy is  considered next.  The
correct choice  of flow regions depends on  relative  importance
of the various  length  scales  associated  with   a  particular
discharge  (Figures 2.4b and 2.4c).

                              31

-------
  Table 2.1 Trajectory and Dilution Relations  for
  Submerged Flows
  Source: Wright, -(1977)

Flow Region     Trajectory                Dilution

  mdnf      z/lm=ti (x/lm ) l ' 2         S = si(z/lg)

  mdff      Z/l»=t2 (X/lm )i /3         S = S2 (Z2 /Imlo )

  bdnf      z/lb=t3 (x/lb )3/4    S = s3 (It,1/3 z5/3 ) / do 1m )


  bdff      Z/lb=t4 (X/lb )2/3         S - S4Z2/(lQlra)



  Where :*

  mdnf = weakly deflected jet
        "momentum dominated near field"

  mdff = strongly deflected jet
        "momentum dominated far field"

  bdnf =  weakly deflected plume
        "buoyancy dominated near field"

  bdff = strongly deflected plume
       "buoyancy dominated far field"

  * Designations in quotes according to Wright,  (1977)
                            32

-------
Table 2.2 Trajectory and Dilution  Constants
mdnf :      ti =2 . 65




mdff:      t2=1.44




bdnf:      t3=2.36




bdff:      t«=1.15
si =0.42




S2=0.38




s3=0.42




S4 =0.41
bi =0.34




b2 =0.34




b2 =0.34




b2 =0.34
                           33

-------
    The possibility  of flow  attachment to  the bottom  is not
included in the analysis.  Future work on CORMIX1 will include
the effect of bottom attachments.

a) Possible Transitions
    If  the  buoyant  jet  is  discharged   into   an  unbounded
cross-flow, the ratio of lm/Ib will indicate which of the
regions   (i.e.  mdnf,  mdff,  bdnf,  and  bdff)   occur  for   a
particular  flow.     Provided   that   lm   and   lb  are  both
substantially larger than IQ  (generally true   in  practice) two
possible  transitions can occur  (See Figure  2.5).

    Case  I)   For  lm/lb »  0(1)  the  buoyancy in the plume  is
relatively weak compared to momentum,  and a large distance  is
required  for  the buoyancy  to generate additional momentum  to
control flow characteristics.  Therefore the  flow will  develop
as: mdnf  -> mdff ->  bdff.

    Case  2)    If   lm/lb   << O(l), the buoyancy  force  is much
stronger  and the flow will  be   a  weakly   deflected  jet when
buoyancy  forces  begin to  dominate.    Therefore  the flow will
develop as: mdnf ->  bdnf -> bdff.
b) Coordinate Systems for Oblique Discharges:
    In CORMIX1 a global cartesian coordinate  system  (x,y,z)  is
placed at the bottom of the water body with the  origin  (0,0,0)
directly below the  center  of  the  discharge   orifice.   The
height of  the discharge  orifice above  the bottom is ho .  The
positive x-axis is located at the  bottom and  directed  in the
downstream direction following  the ambient flow.  The positive
y-axis is located at the bottom and points to the left,  normal
to the  ambient flow  direction  (x-axis).  The positive  z-axis
points vertically upward.   The  angle   between  the discharge
axis y*  and its  projection on the horizontal plane  (i.e. the
discharge angle  above  horizontal)  is  60 .     The discharge-
crossflow angle  ao is  the angle between the projection of  y*
on the x-y plane and the  x-axis   (ao  =  0.0°   for co-flowing
discharges,  ao  = 180" for counter-flowing discharges).

    A primed  coordinate position,   (x1,y',z'),  within a given
flow region is specified  with respect   to the   virtual  source
for that  flow region.    A  virtual source  is needed for each
flow region because the previous asymptotic  analyses assume a
point discharge  source, which is physically unrealistic.  The
primed coordinate system is related to   the global coordinated
by
         (x,y,z)  = (x',y',z')  + (xv,yv,Zv)               (25)

where  (xv,yv,zv)  is the global position  of the virtual source
for that flow region.  The position of the virtual source

                              34

-------
                           Transition)
        2 = 0
           a)  Em» Eb ;  Momentum  Dominates
          Transition
         z=0
b)
                    ",  Buoyancy Dominates
Figure 2.5   General   Behavior for  Buoyant Jets in Unconfined
Crossflow  (Assuming Near-Vertical Discharge)

                               35

-------
(Xv,yv,zv)  is  computed by  taking the  known flow solution at
the transition,  as given  from the  previous flow  region, and
back calculating  the source  position using the equations for
the given flow region.

    In general,  the analysis  is extended  to non-vertical 3-D
trajectories  within  the  ambient crossflow.  A supplementary
transverse coordinate   45.0-                     (26)

    xt = C(l) IM         ro  <  45.0°                     (27)

where  C(l)  is a  constant of   the  order   of 1,   and  ro   is the
angle  defined by  the  discharge  axis y*  and  the crossflow (x-
axis).

    The  dilution and  size of  the  plume will be constant at the
transitions   while  a  slight  discontinuity  in  the  centerline
velocity profile will occur.

    A  complete list of all  transition rules  used in CORMIX1,
included those  for bounded   flows,  appears in Table 2.3.   All
constants  of  the   order   1,   C{1),   are   assumed  to be  1 in
CORMIX1.     Future  work  plans  include  further refinement of
these  constants with  experimental  data.
                               36

-------
Table 2.3 Flow Transition Rules
Trans . Current Next
Rule Flow Flow
Module Module
0 01 11
21
34
1 11 21

2 11 22

' Relation

Zf =


ro > 45°
ro < 45°
ro > 45°
ro < 45°

ho


ti=C
x/=C
ff=c
X/-C




(1) IM
(DIM
(Dim
(l)l,n
3
4
5
6
7
Module
01
11
16
22
31
32
33
34
41
61
21 22 Zf1
16 22 £j=C(
22 31
16 31 Zf
11 33
21 32 Zf
41 62 xr =xi +(2/3) (
( [ (S.Olbhi )
=C(l)lb
=H
=0.8H+0.2ho
bl a/ 2 /lbi 72 )
/(Si flmlQ ) ]3/2-l|
= zone of flow establishment (zofe)
= weakly deflected jet (mdnf)
= strongly deflected jet (mdff)
= strongly defected plume (bdff)
= near-horizontal surface approach
= near-vertical surface impingement with
buoyant upstream spreading
= near-vertical surface impingement with
vertical mixing
= near-vertical surface impingement, upstream
spreading, vertical mixing, buoyant
restratification
= buoyant surface spreading
= passive diffusion
                          37

-------
2.B  Flow Interaction With Free Surface
    The  preceding  section  assumed  a turbulent plume rising
with a  crossflow  in  an  infinitely  deep  water  body.   No
boundary  effects   of  the  flow  meeting  the  surface  were
considered.  This section analyzes the  boundary effects  of a
flow interacting with the water surface within the near field.

    The interaction of the flow with the water surface will be
characterized by the local ambient water  depth H.   The ratio
of  the  water  depth  length  scale  H  to  the length scales
discussed  previously  is  used   to  characterize   the  flow
interaction with the surface.
2.B.I Flow Classification  of Near-Field Regions
    The  flow  classification system  appears  in Figure  2.6.  The
flow  classification   system   uses    the  ratio    of   1m/H   to
characterize  the   discharge as "deep  water"  or  "shallow  water"
based on the  momentum  of  the flow as  it contacts   the  surface.
A deep   water discharge   will  have  relatively weak momentum  as
the flow contacts  the  surface, while  a  shallow  water  discharge
will have relatively  strong   momentum  as the flow impinges  on
the surface.

    Ratios of lb/H and IM/H are used to  further  classify the
properties of the  flow as  it contacts the surface  in  both deep
and shallow  water.

    Discharges can be  classified as   "stable"   or "unstable".
Flows with strong  vertical momentum at  surface  contact tend  to
be unstable.   In this  case the jet  is deflected   downward  by
the surface   and an unstable  recirculation zone occurs  around
the jet  as it entrains the fluid deflected  down from
the surface.   In a stable discharge,  buoyancy tends to  have a
stabilizing   effect on  the   flow   as  contacts   the surface,
causing  the  flow to form  a stratified layer on  the surface.


2.B.2 Analysis of  Surface Interaction Processes
    Four major  possible  flow   regions  exist  (Figure 2.7) for
the  flow interaction with   the   surface;  i)  near-horizontal
surface  approach,   ii)  near-vertical   surface impingement with
upstream spreading,   iii)  near-vertical   surface impingement
with  full   vertical   mixing,   and   iv)  near-vertical surface
impingement      with      unstable     recirculation,    buoyant
restratification,  and  upstream spreading.   As   shown  in  Figure
2.6,  each   of  these   four  possible  flows can be defined  by
combinations  of  the ratio of the length scales  !„, , lb , and IM ,
with the local water depth H.

    A control volume approach  is used for the following

                               38

-------
  90*<#0<45*

(Near) Vertical
Deep
Water
with
Weak
Momentum
      lb-
      the flow is minf-^>  mdff->  bdff), are included in CORMIX1 but for simplicity
      are omitted in  Figure 2.6.

-------
                Cross-section
                                             f (rectangular)

                                'i (round)      width bh
                                 width b
               a) Surface Approach (Near-Horizontal)
Figure   2.7     Four  Major Conditions  of  Flow Interaction with
Water Surface (i indicates inflow values  in control  volume and
f outflow values)

                                40

-------
       Side  View
    u,
       Plan  View
        Stagnation
        Point
           Inclined Front
                      y
     b) Surface  Impingement  with  Buoyant Upstream  Spreading
Figure 2.7 (continued)
                               41

-------
            Side View
         c) Surface  Impingement  with Full  Vertical  Mixing
Figure 2.7 (continued)
                              42

-------
        d) Surface  Impingement  with  Buoyant  Upstream  Spreading,
           Full Vertical  Mixing, and  Buoyant  Restratification
Figure  2.7 (continued]
                                 43

-------
sections.   When the plume contacts the surface, bv and bh are
defined as the vertical  depth  and  horizontal  width  of the
subsequent  flow,  respectively.  The  variable subscripts "i"
(initial) and "f"  (final)  (e.g. bi ,  Sf ) denote  control volume
inflow and outflow quantities, respectively.


2.B.2.1 Near Horizontal  Surface Approach
    In this surface approach  condition, the bent over flow
approaches the  water surface near horizontally at  impingement
angle 81  < 45°  (Figure 2. la).  The  flow  is   advected with the
ambient velocity   field  at a  rate equal to ua .  This situation
occurs for weakly  buoyant  and deep water  cases, hence the flow
will be strongly deflected when it contacts the surface.  This
type of surface interaction occurs for flow classifications VI
and V2 which are defined as deep water with weakly  buoyant
discharge  (for  1m /H < O(l) and lb /H <  O(l), respectively).

    Experimental evidence   (Jirka and  Harleman, 1973) suggests
that within a short  distance  after   surface  impingement the
concentration   distribution   for  a  2-D  flow changes from the
assumed   gaussian  distribution  to    a   top-hat    or  uniform
distribution  (Figure  2. la).   Using a control volume approach
the initial centerline dilution is related  to the  final bulk
dilution,  and  a  bulk  mixing  process  is  assumed with Sf =
21/2Si   (Fan 1967).   An  equivalent  cross-section aspect ratio
for the   outflow section  of  2:1  is assumed.   The continuity
equation  for the control volume in Figure 2.7a

          Uabi 2 TT =  Ua 2bh f * bv f                          (28)

where bi  is the  initial dilution  and half-width,  bv f  is the
final  flow   vertical   width,  and  bh f   is  the   final  flow
horizontal    half-width.        This     is    evaluated    as
bvf =bh f =bi
2.B.2.2   Near-Vertical   Surface   Impingement  With  Buoyant
Upstream Spreading
     In this surface approach condition, the  weakly bent flow
impinges on  the surface  at a  near-vertical angle 61  (Figure
2.7b),  where  81 >  45°.   After impingement  the flow spreads
more or  less radially  along the  water surface  as a density
current.    In  particular,  the  flow  spreads  some distance
upstream against  the ambient  flow, and  laterally across the
ambient flow.   This  spreading  is  dominated  by  the strong
buoyancy of the discharge.

    The stabilizing  effects of  buoyancy are greater than the
destabilizing forces of jet  momentum and  the crossflow, thus
this is classified as a stable discharge condition.  This type

                              44

-------
of surface  interaction occurs in flow classifications V3, V5,
H3 and H5 and is defined for strongly buoyant discharges  (lb/H
> 0(1))   in both  deep water   (lm/H <  0(1)) and shallow water
(lm/H >  0(1)).   For deep  water discharges,  lm <•  lb implies
IM/H  <   1,  so  the  flow  will  be weakly deflected plume at
surface approach.  For shallow water cases, to insure that the
flow is  a weakly  deflected plume, the additional requirement
will be that 1M/H < 1.

    The  lateral  spreading  of   the  flow   in  the  surface
impingement  region  is  driven  by both  the flow momentum and
buoyancy force.  Of interest is the upstream  intrusion length
Is ,  dilution  S, horizontal width bh, and vertical depth bv of
the density current at surface impingement.

    The analysis of this flow  region follows results presented
by Lee  and Jirka,  (1981), and Jones, et  al.,  (1983).  Lee and
Jirka analyze the properties of a buoyant subsurface discharge
in stagnant  water including   the effects of recirculation and
buoyant restratification.  Jones, et al.  presents a methodolgy
to  predict   the  upstream  spreading  of  a  buoyant  radial
discharge  in crossflow.  The strategy for solution  will be to
use the results  of Lee and Jirka to provide initial conditions
for the methodology of  Jones, et  al. for  predicting buoyant
upstream spreading and horizontal width.

    A length scale IN representing the turbulent mixing action
of the horizontal momentum  flux  versus  stability  effect of
buoyancy force is given by

    IN = (defected horizontal momentum flux)3/4/Jo1' 2 (29)

For the  weakly  deflected plume, Holley and Jirka, (1986) give
an expression for the vertical momentum of a plume

    M s 0.  85Jo2 /3s4/3                                 (30)

Where, s   (= H)   is the  distance along   the plume trajectory.
Substituting  appropriate  values  into   Eq.   (29), the length
scale for  a weakly deflected plume at impingement

    IN = 0.367H(l-cos6i)                               (31)

where  the  factor  (l-cos8i)   accounts  for   the  deflected
horizontal momentum  flux, in  analogy to the vane equation in
classical  fluid mechanics.

    Jones,  et al. defines an   intrusion   length  scale  li, by
relating   the  interaction  of  the  buoyancy  force  with the
crossflow  force
                               45

-------
    li = Jo / 2rrCD ua 3                                     (32)

where CD  is a drag  coefficient  of  O(l).

    Thus, for  a weakly   deflected plume  at surface approach,
the ratio of length scales  obtained from Eqs .  (31)  and (32)

    li /1N = 0.54(lb /H) (I/ (l-cos0i ) )                     (33)

which  describes   the  relative  importance  of    buoyancy  to
momentum forces at  surface  impingement.

    The upstream   intrusion length  Ls  is found  from observing
the Figure 5-14 in  Jones, et al.,

    Ls/li - 4.2(li/!N)-2/3         for li /!N  <  3.3      (34a)

    Ls /li = 1.9                         li /!N  > 3.3      (34b)

Noting that li  =   lb/5 with  CD  =   1 and  since  the  flow is a
weakly  deflected   plume  at  surface  approach,   the upstream
intrusion length Ls in Eq.  (34a)  can be expressed

    Ls = 1.261b ( (l-cos(8i ) )/(It/H) )2/3                  (35a)

for lb/H <  6.11(1-cos (9i )) , and in Eq.  (34b)

    Ls = 1.91b                                          (35b)

for lb/H >  6.11(1-cos (61 ).

    To calculate  the  dilution in this region,  first  note  that
Jones, et   al.  in   Figure 7-8  gives the dilution for  a radial
surface  discharge

    S/Fs =  1.6(li /!N J1 /3                                (36)

where Fs is  a  radial surface  spreading Froude  number.   This
Froude number is  defined

    Fs = ur / (g1 lo )1/2                                  (37)

where 10 is  a characteristic length scale defined

    lo = (2Trnhi M/2                                    (38)

where n and  hi  are the  radius and depth of the  buoyant radial
surface  spreading flow,  respectively.

    Now  the results of Lee  and Jirka are used to  evaluate the
surface  spreading Froude number Fs , so the dilution  S  from Eq.

                               46

-------
 (36)  can  be  found.     In   this  analysis,  the  initial  radial
surface spreading region  uses   a simplified   control  volume  to
relate  the  properties   of   the vertically  buoyant  jet  at  the
entrance   of   the   surface   impingement    region   to   the
characteristic  parameters    of  the   horizontal   axisymmetric
buoyant surface jet at the exit of  this  region.
    Lee  and
impingement
          Jirka  define  the  Froude
number  at  surface
Fi  =
           / (g' hi
where ur  is the  radial surface  spreading velocity  and hi  is
the depth  after impingement.   For   large values   of H/D,  the
value Fi = 4.62 and the value hi/H =  0.0775.

    The radius  of the flow n  is n  =  eH, where e  =  0.11.   By
substituting  these  asymptotic  values   into  Eq.   (38),   the
characteristic length scale 10  from Jones, et  al. becomes
    lu = 0.23H
                                                   ;4o:
which when combined with  asymptotic  values  for  Eq.  (39)  gives

    Fs - Fi (hi /lo )> /2   =2.65                          (41)

indicating that  the flow  in  this  region  is  jet-like-   Finally,
note  that the radial  surface  spreading Froude number  Fs  can  be
expressed  as  in terms   of   the  discharge  flux  variables  as
Q/ (Mo3/ 4 /Jo 1 /2 ) -

    This result  can be  then  used  to determine   a bulk dilution
at the  end of   the region  St .    From Eqs.   (36)  and (41)  the
expression  for   final   dilution   in  the  surface impingement
region
     Sf  -  3.49Si (lb/H)1 ' 3 (1-cos(81
                                     1 / 3
                                                    42!
     The  geometry   of  the surface flow as revealed by Jones,  et
 al.  will be  used  to determine the width  and depth  within the
 region.    From Jones,   et al .   Figure 7.1,   the width,  bi, t ,  at
 impingement  is assumed to be 2.6 times larger that L3 :
     bh
    =  2.6Ls
 The  typical  depth of the flow in the upstream intrusion region
 hs ,  is   found using  the vertical  length scale from Jones, et
 al.  where
        = CD ua 2 /g'
                                                    44:
                               47

-------
with CD  = 0.8.   Noting  that  g'=  go/Si,  where  Sf  is  the total
bulk  dilution,  using  the  continuity   equation   ua2Qo/Jo   =
ua3 Qo Mo * ' 2 / (Jo Ua Mo 1 i 2 } =  Iglm/lb the  stagnation flow  thickness
hs

         hs  = 0 . 8Sf 1m IQ /lb                             (45)

The final  depth bvf   (at x=0)   is  found   using the continuity
equation, bvf (x=0) = QoSf/(bvfua),  and the  previous  equation

         bv f  =  Ibhs / (0. 8bh f )                           (46)
2.B.2.3 Near-Vertical  Surface  Impingement  With  Full Vertical
Mixing
    In this case, the weakly bent   flow  impinges   on the water
surface at  a near-vertical angle  (Figure 2.7c).   This  case  is
defined by shallow water  ( 1/H   >  0(1)   ) and  weak buoyancy (
lb/H <  O(l) ).  Given the shallow ambient water  depth  and the
weak buoyancy of  the  discharge,   the   flow   becomes unstable
after impingement.  This  occurs  in flow  classifications V4 and
H4,   and  results  in   a recirculation  region immediately
downstream  that  extends over   the  full  water depth.  The
recirculation region causes the  flow  to   entrain  ambient fluid
from  the  flow  itself   causing  dilution within the plume  to
decrease.  Because of unstable  recirculation,  the centerplane
dilution decreases  to Sf  = St/R,  where R is a  recirculation
factor.  Experimental data indicate R ranges  from 1.0  to  2.0,
and an  average  value  of 1.5 will be assumed. The  final  flow
width, bhf , is  found from the continuity equation

         bh f =   Sf 1m lo / (2H)                            (47)

and final  outflow location Xf is approximated as

         Xf = xi +  2H                                  (48)

where xi is the  plume position  at the beginning  of the  region.
The  distance    2H  accounts  for  the   typical   length  of   a
recirculating  zone  (Holley and  Jirka, 1985).


2.B.2.4  Near-Vertical   Surface   Impingement   With  Unstable
Recirculation,   Buoyant	Restratification,    and    Upstream
Spreading
    In  this   surface   approach   region,   the   flow  rises  near
vertically and   impinges  on   the water   surface   (Figure 2.7d).
After impingement the mixed  flow recirculates over the  limited
water depth   and becomes  partially re-entrained  into the  flow.
This  surface  interaction  occurs in flow  classification V6 and
is defined by  shallow   water  (lm/H > O(l)  ),  strongly  buoyant

                              48

-------
(Ib/H >  O(l) ), as well as momentum  dominated  (Ix/h  >  0(1)  ).
Both momentum and buoyancy are  strong,   but  momentum  dominates
in the  near field.  The degree of  recirculation  and  hence  the
overall   mixing   in   this    region    is     controlled    by
restratification  of  the  flow   at  edge  of the  recirculating
region.    The  restratified  flow  spreads   along  the  water
surface.    In  particular  the   flow  spreads  some  distance
upstream against the ambient current, and  laterally across  the
ambient flow.

    The analysis  of the  flow  is based  on the  work of  Lee  and
Jirka,  (1981)  which  was  originally  developed  for stagnant
conditions.  The final dilution Sf  is given  by

         Sf = 0.76H3/3/ (lM2/3Ic )                  (49)

The  dilution  is  controlled  by   buoyancy   forces   only,  the
effects of momentum become dissipated by   the turbulent mixing
action.  This result can be understood  if  the values  of IM
and IQ  are replaced  in Eq.  (49)  by  their  respective efflux
definitions, the results show that  dilution   is a  function  of
the buoyancy flux only.

    The  surface   buoyant  spreading  properties  after   the
unstable recirculation are analyzed similar  to  the development
for  the   near-vertical  surface   impingement  with  buoyant
upstream  spreading  presented  in    section    2.B.2.2.      In
particular, for  the unstable case, the  limit of li/lu -> »  is
of interest, so that Eq. (34b) applies.

    Since the near-field momentum is  dissipated, the  length  of
the upstream  intrusion Ls  is found from Jones, et al.  (1984),
Figure 5-14.  This figure  plots  Ls/li  vs.  li/!N    and  is  a
constant line  of Ls/li  = 1.9  for  li/!N ->  ».  Recalling li  =
lb/5 from Eq. (32) for shallow  ambient  conditions  and noting
that, the relationship for upstream intrusion length  Ls

    Ls  = 0.151b                                         (50)

The upstream  intrusion thickness   hs is  found  in analogy from
Section 2.B.2 . 2

         hs - 2St lmlc/lb                               (51)

The   final  half-width  bhf ,  is   found  from  the  continuity
equation

         bv f  =  Ibhs / (2bbf )                             (52)

The  region  is assumed  to extend downstream a distance equal to
2H  as  in Eq.  (48).

                               49

-------
2.C Analysis of Far-Field Mixing Process
    After  the  flow  interacts  with  the  water  surface  as
described by  the  previous  sections,  the  far  field mixing
begins   (Figure  2.1).      This region consists of one or two
regions,  depending  on  discharge  characteristics.    In the
general case,  the flow contains sufficient buoyancy and there
will be a  buoyant  surface  spreading  region  followed  by a
passive  diffusion  region.    The  surface spreading region is
characterized  by  dynamic  horizontal  spreading  and gradual
vertical  thinning  of  the  flow   after  interacting with the
surface  (Roberts,  1979,  Koh  and  Brooks,  1975).   Boundary
interaction may  occur, and  in bounded   sections the flow may
become   laterally  fully  mixed.    In  the  passive diffusion
region,  the  dilution  is  controlled by the turbulent mixing
action  of the flowing ambient  water  body.     Again, boundary
interaction may  occur, and the flow  may  become both laterally
and vertically fully mixed  in this  region.    If the  flow is
non-buoyant   or  weakly  buoyant   there   is no  buoyant surface
spreading region, only a passive diffusion region.
2.C.I Buoyant  Surface  Spreading
    In  this  region  the buoyant surface  plume  spreads  laterally
along   the   water   surface   while   it  is  being  advected  by  the
ambient  current  (Figure 2.8).  The  plume  behaves   as  a density
current  and  entrains  some   ambient fluid in  the  "head region"
of  the  current.  The mixing  rate is usually  relatively  small.
Furthermore, the surface plume may  interact with  a nearby bank
or  shoreline.  The  plume depth may  decrease during this  phase.
The analysis of  this region  is based on arguments  presented by
Jones,  et. al.  (1985).

    The  continuity  equation  for  the density current

    uaah/ax  +  a(vh)/ay = we                            (53)

where we  is  the  net velocity across the   interface.  Benjamin
 (1967)  has derived  an  equation for  the  spreading  velocity VB

          ve2/(g'h)  - C                                (54)

where   h   is  the  density   current  thickness   and  C  is  a
coefficient  that depends on  the  relative   depth h/H  and is of
order   O (1   to   2).     Combining  the   Eqs.   (53)  and (54)  and
integrating  gives

         Uad(hb)/dx =  qe (x)                            (55)

where qe (x)  is the  localized head entrainment.


                              50

-------
                                                Front
    Plan  View
    u,
      —  I
      -*-  l^-Initial          A
              Condition
     Cross-section A-A
Frontal Zone
    ••
Figure 2.8   Buoyant Surface Spreading


                                51

-------
    The localized head entrainment  is  qe (x)  =  PVB h  where  p  is
a constant O(0.15 to 0.25)  (Simpson and  Bitter,  1979;  Jirka
and Arita, 1987).  The constant  p will be  set  equal to 0  25 in
CORMIX1.

    The  flow  half-with  b   is  obtained   for  any downstream
distance x  by using the boundary condition  for  the streamline
(vB=Uadb/dx) and  integrating  Eq.   (53)   (for  the unattached
case )
         bh =  (bi3/3 +  3/21b1/2(x  - xi))2/3            (56)

where xi   is the  downstream distance   at  the beginning  of  the
buoyant  spreading  region.    The  2/3  power   law   of   plume
spreading is in agreement with the previous work of Larsen  and
Sorensen,  (1968) .

    The vertical plume  width bv for  any bh  in  the   region  is
given by integrating Eq .  (54) to obtain
         bv = bv i  (bh /bh i )                              (56)

where   p   represents   the  additional  dilution  caused  by
entrainment at the head of the density current.

    The bulk dilution  S given by co/c, is equivalent   to go' /g'
as is  Eq.  (7).   Buoyancy conservation in  the density current
(analogous to momentum conservation)  can be  expressed

    uaa(g'h)/3x + a(g'veh)/ay = 0                      (57)

Integrating Eq. (57) and  noting   that  PVB h=pua h (db/dx) gives
the expression for dilution  S

         S = Si (bv /bv i  ) P                               (58)

2.C.2 Passive Diffusion
    In this  region the  background  turbulence  in  the ambient
shear flow becomes  the  dominating   mixing  mechanism (Figure
2.9).    The  mixed  surface flow  is growing  in depth and in
width.  The flow  may interact  with  the  channel  bottom and/or
banks .

    The  analysis   of  this  region follows  classical  diffusion
theory  (e.g. Fischer,  et al. 1979).   The   standard  deviation  o
of a  diffusing plume  in  crossflow  can be  written  in terms of
the transverse  turbulent diffusivity E

         a2 =  2Ex/ua                                    (59)
                               52

-------
     Plan  View
Possible Bonk Interaction
                                                       Plume
                                                       Centerline
               •Initial Conditions
     Side View
                                                  Possible Bottom Interaction
Figure  2.9   Passive  Diffusion  Process

                                       53

-------
In open channel flow the  eddy  diffusivity can be related to
the friction velocity u*  and the  channel  depth H

         Ez = 0.2u*H                                   (60a)

for vertical diffusivity,  and

         Ey = 0.6u*H                                   (60b)

for  horizontal  diffusivity-     Due   to  some anisoptropy in a
typical  channel   flow,   the   diffusivity  in  the  horizontal
transverse  direction   is  larger  than  the diffusivity in the
vertical direction.  The   friction  velocity  is given  by u*  =
(f/8)ua where f is  the  Darcy-Weisbach friction factor.

    Solution of  Eq.  (59)   gives  the   global coordinate system
expressions for flow width bv  and  depth  bi,   at any downstream
distance x

    bv = ( (2Ey /ua ) (x-Xi )  + bv i 2  M/2                   (61)

    bh - ( (2EZ/ua ) (x-xi )  + bhi2)1/2                    (62)

where xi,  bvi, and bti  are the  distance,  width,  and depth of
the  flow,  respectively,   at   the  beginning  of   the  surface
spreading  region.    This assumes that  the above lengths  are
expressed in terms  equal  to one standard  deviation,  or  bv  = oy
and bh = az .

    The continuity  equation applied   to  a   plume  in crossflow
ua , oy o2 =SQ0  yields the dilution  S

         S = ay a2 / (lu,lQ )                               (63a)

In the case that when the  plume is vertically fully  mixed

         S = oy H/ (Imlg )                                (64b)

2.C.3 General Behavior in  the  Far-Field

a) Boundary Interaction
    If buoyant  surface   spreading  occurs,   the   process will
continue  until  either   the   transition  to  passive  turbulent
diffusion occurs (as described in the  next  section  2.C.3 part
b))  or  the flow  attaches  to  both banks in  bounded  sections,
whichever occurs first.

    The end of the  passive turbulent  diffusion region occurs
when  the  plume  becomes   both vertically  and laterally fully
mixed  and  no  change  in dilution   occurs   with  increasing
downstream direction.   In unbounded  sections,  the  plume  will

                               54

-------
never  become   laterally  fully   mixed,   so   the   simulation
terminates  at  some  large  downstream   distance    preset   by
CORMIX1.

b) Transition Between Surface  Spreading  and Passive  Diffusion
    A flux Richardson number,  Rf ,  defined locally  as the  ratio
of buoyant energy flux to  shear  energy production  will  be used
as a  stability criterion  for the transition  from  the  buoyant
surface spreading  to the  passive diffusion  flow.  Written in
terms of  the eddy diffusivity convention (Tennekes  an  Lumley,
1972)

         Rf= -gKa (dp/dz) / (pKM (du/dy)2 )                 (63)

in which KM , KH = eddy   diffusivity   for  momentum  and  for a
scalar (heat ),  (KB =KM )  respectively;  p(y) = local  density;  and
u(y)  is the  local velocity.

     A critical  value  of  Rf c  =  0.10 to 0.20 has  been suggested
by Monin  and Yaglom,   (1971)  and  Turner, (1973) .   Above this
value,  turbulence   is  damped   and  a  stable   profile   can be
maintained  and  the  density  current  flow continues;  below this
value,  turbulence   erodes  the   stable   density  profile  and
turbulent diffusion  then controls the flow.

     For   the  buoyant  surface  spreading  region  Jirka  (1979)
suggests  the  appropriate Richardson  number is  of the form

          Rf  =  K2 (go'H/u*2 ) (1/S) (h/H)                     (64)
 where  K  is  von Karman constant  (S0.4).    The ratio  of h/H in
 Eq .  (64)   is  a  length scale representing depth of the density
 current  to  overall channel depth.

     Noting  that Si ,  bv i ,   and  xi   are  the  dilution, initial
 depth,   and  downstream  distance   at the start of the buoyant
 surface  spreading flow,  the distance to  transition, xt , to the
 passive  spreading flow

 xt = Xo + (2/3) (bi 3/2 /Ib1 / 2 ) { [ (Slbho ) / (fSi IO.IQ ) ]3/2 -1)   (65)

 If the   plume  is  attached to one  boundary Ib  is replaced with
 21b .
                               55

-------
                         Chapter III
                  CORMIX1 Program Structure
    The  Cornell   Mixing  Zone   Expert  System  Subsystem  1.
(CORMIX1)  is a series of software  subsystems or  elements for
the analysis  and design  of conventional or toxic single port
submerged  buoyant  or  nonbuoyant  pollutant  discharges into
unstratified watercourses,  with emphasis  on the geometry and
dilution characteristics of the  initial mixing  zone.   It is
designed as an analysis tool for regulators, dischargers, (and
students of hydraulics).

    The  user  supplies  CORMIX1  with  information  about the
discharge   and   ambient   environment.      CORMIX1  returns
information detailing the hydrodynamic  mechanisms controlling
the flow,  dilution, geometric information concerning the shape
of the pollutant plume or flow in the ambient water  body, and
design  recommendations  allowing  the  user  to  improve  the
dilution characteristics of the flow.    If  specified  by the
user, CORMIX1 also presents information about the legal mixing
zone dimensions and dilution, toxic mixing  zone requirements,
and  zone  of  interest  characteristics  for  the  flow.  The
minimum hardware configuration  for  CORMIX1  is  an IBM-PC/XT
with a printer for hardcopy output.

    The  purpose   of  CORMIX1  is  to  obviate   for the novice
analyst the need for  detailed hydrodynamic  understanding and
experience.    A   general environmental science  or engineering
background at the  BS  level  appears  to be  minimum educational
requirement  needed   to   compile  and  supply  relevant  data,
interpret  the system  information,  and  ultimately   learn and
become   knowledgeable    about    hydrodynamic  mixing  through
repeated interactive  use.   Two working days appears   to be the
minimum  time  needed  for  a  first time user to gain initial
facility   with    system    requirements,    limitations,   and
interpretation of  results.

    Figure 3.1  shows the   system elements  of CORMIX1.   During
system use the elements are loaded sequentially  by the user.
CORMIX1 is implemented  in   the  programming  language Fortran,
and M.I  (Teknowledge, Inc.), an expert systems "shell".

    M.I  is  an  expert   systems  programming language, or more
precisely, a shell.   A  shell  is  a  self-contained  inference
engine that does not  contain the  knowledge  base,  but  has

                              56

-------
                      M.I
           DAT IN
          User  Input
                   Iteration
                   Alternatives
                   Corrections
               file.  CXD
                   Fortran
           PARAM
          Parameter
          Computation
               file. CXP
                     M
          CLASS
          Flow
          Classification
              file.  CXI
file.
CXC
           Fortran
    HYDRO
Prediction/Simulation
     Program
                                        file. CXO
                                     SUM     MJ
                                Summary
                                Evaluation
                                Recommendations
                                (Legal/Engineering)
Figure 3.1  System Elements  of CORMIX1
                            57

-------
facilities for  both forward and backward reasoning, debugging
aids,  consistency  checking,  input  and  output  menus,  and
explanation facilities.

    Two  programming  languages  are  used  to  exploit  their
respective   strengths   while   avoiding   their   respective
weaknesses.    M.I,  as  a  knowledge  base   language, is very
efficient in knowledge representation  and symbolic reasoning;
however  it  is  relatively  weak  in  numerical computational
ability.  On the other hand, Fortran is  ideal for computation
of   mathematical   functions    (Fortran  stands  for  formula
translator) but is poorly suited for the tasks associated with
symbolic  reasoning.    Thus  M.I  is employed to implement the
knowledge acquisition, model selection,  and  analysis  of the
hydrodynamic  simulation   portions  of   the  expert  -'system.
Fortran is used for the computation  of various  length scales
and  in the hydrodynamic flow simulation models.

     It is  interesting to  note that  the entire system could
have been programmed in  a language  such as  Fortran, or even
assembly  language;  the  real  issue  is  one  of programming
efficiency.   For  instance,  a routine  written in  5  lines of
Fortran  code  might  take  100 lines of assembly level source
code.   Since  M.I  was  developed  to  encode  and manipulate
symbolic logic, it does so with great efficiency, allowing the
programmer to write in  5 lines  of code  what might   take 100
lines in  Fortran  or  1000 lines   of assembly.  In essence the
selection of M.I as  the language  for the  symbolic reasoning
tasks gives the programmer significant leverage.
3.A Discussion of Logic/Ml Elements
    The  M.I  elements  of  CORMIX1 are DATIN, CLASS, and SUM.
M.I  is  very  similar  in  structure  to  PROLOG  (PROgramming
LOGic).    PROLOG was  developed in  Europe and  is designed to
manipulate logical  expressions(Clocksin  and  Mellish, 1984).
An M.I  program is  built from statements containing facts and
if-then rules about facts.  This is called the knowledge base.
The knowledge  base is supplied by the user corresponding to a
problem  domain,   in  this  case  buoyant  submerged  jets and
hydrodynamic mixing processes.

    M.I   programs  are  driven  by  a "goal" which the program
tries to validate by searching the knowledge base to construct
a "proof"  by using  the facts and rules in the knowledge base
needed  to  deduce  the  goal  as  a  valid  hypothesis.   The
following  section  gives  a  more detailed explanation of how
this is  accomplished, using the  CORMIX1  module  DATIN  as an
illustrative example.
                              58

-------
3.A.I DATIN
    DATIN is an M.I program for the entry of relevant data and
for the initialization of the  other  program  elements.   The
user executes DATIN by typing "DATIN".  DATIN then prompts the
user for needed information.

    CORMIX1  deals   with   submerged   buoyant   single  port
discharges  into  an  unstratified  water  body.    The system
assumes a schematic rectangular  cross-section bounded  by two
banks -  or by  one bank  only for  coastal or other laterally
unlimited situations.  The user receives detailed instructions
on how  to approximate actual cross-sections that may be quite
irregular  to   fit   the   rectangular   schematization.  The
representative  schematization  with all relevant hydrodynamic
variables that DATIN gathers, appears in Figure 3.2.

    Even in this simple  schematized  ambient  geometry, there
remains a  tantalizing amount of geometric and dynamic detail:
the discharge location   in  relation  to  the  bottom  and the
shoreline;  the  discharge  orientation  may be with the flow,
against the flow, or vertically upward across the  flow, or at
some arbitrary  angle, the water depth may be deep or shallow;
the ambient flow may be  stagnant or fast and highly diffusive;
and the  discharge  flow  may be non-buoyant or highly-buoyant;
with high or low efflux  velocity.

    The purpose of  DATIN is to specify completely the physical
environment of  the discharge,  as well  as legal or regulatory
specifications.  The following data groups need to be entered:
general   site   and   case   identifer  information,  ambient
conditions  (geometry   and   hydrography),  discharge conditions
 (geometry and  fluxes), and  information desired  including legal
mixing  zone  definitions and  toxic  dilution  zone criteria.
DATIN  provides  consistency checks, and  gives advice for input
parameter selection.

     DATIN tries to  satisfy  the goal creating a  valid parameter
input  file   for the other  CORMIX1  elements.   The goal is the
statement that drives  the execution of DATIN.   This is written
in  M.I  as

     goal =  param_input_file.                           [1]

Here   the   goal  is to  satisfy   or  find a valuation  for the
expression  "param_input_file".

    All rules  in  M.I  are  stated  as:  if   (expression(s) or
clauses called  the "premise"  or  "head"   of the rule) -  then
(an expression or clause called the "conclusion" or  "tail" of
the  rulel  statements.     The  premise  of  a  rule in M.I can
contain more than one  expression connected by and/or

                              59

-------
                    PLAN  VIEW
      W
 CROSS-SECTION
                                                       42
H
                                                   D.
                   ///////////// / s /
               Nearest bank
   Flux  quantities:    00= discharge
                     M0= U0Q0= momentum  flux
                     J0 = ( A/OO//DQ) g 00 = buoyancy flux

                                                       -M
Figure 3.2  Schematization of  Discharge Configuration
                , C,
                               60

-------
statements but the conclusion of the rule can contain only one
expression.  M.I  will  try  to  satisfy  the  goal   (here the
expression "param_input_file")   by searching for a rule in the
knowledge  base  whose  conclusion   contains  the  expression
"param_input_file = (valuation)".

    A -rule in DATIN that has "param_input_file = known" in its
conclusion is (a separate  line  for  each  expression  in the
premise of the rule^ has been added to improve readability) :

    if site_description = found and
    ambient_conditions = found and
    discharge_parameters = found and
    zone = found
then param_input_file = known.                          [2]

Here,   in   the   conclusion   of  the  rule  the  expression
param_input_file is assigned the valuation "known".

    First, a explanation is given on how M.I  uses information
contained  within  if  -  then  rules  to assign valuations to
expressions.  M.I always tries to  satisfy a  valuation in the
conclusion of rule by proving its premise.  Thus, M.I tries to
satisfy all  expressions in the premise of the  rule, beginning
in statement  [2] with the first expression "site_description =
found".  If  the valuation of the variable  in first  clause is
satisfied,   i.e.  the  expression site_description does indeed
have the valuation ''found",  then  M.I  tries  to  satisfy the
second  expression,   "ambient_conditions  =  found".   If this
valuation  is satisfied, M.I will try to satisfy  the remaining
expressions  in  the  premise  of  the  rule.  Whenever in the
premise the  valuations for all expressions  are  satisfied, the
rule succeeds or ''fires".  When  the rule fires,  the expression
in the conclusion the rule can be given a valuation  and  added
to the facts known in the knowledge base.

     So  how  does  M.I  know the expression "site_description"
has  the valuation  "found"?  Because  there is  another rule in
the  knowledge base which is

     if site_name = SN and
     discharger_name = DN and
     pollutant_name =  PN and
     design_case = DC  and
     grid_interval = NSTEP
then site_description = found.                         [3]

This statement  is invoked  by statement   [2] in DATIN when it
tries  to  find  a   valuation   for   the   first  expression
"site_description".   Since there  is no present valuation for
the  expression  "site_descriptibn", M.I  locates statement  [3]

                              61

-------
with the  expression "site_description"  in  its  conclusion.  If
all  expressions  in  the  premise  of   statement   [3]  can be
assigned valuations,  then the  expression  site_description is
assigned the valuation "found".    All   capitalized  values in
statement [3]   (SN, DN,  PN, DC,  and NSTEP)  are variables in
M.I.  Variables in M.I  are  only  assigned  a  valuation when
considered within  a rule.   M.I  will try  to find a valuation
for site_name, the first expression in statement  [3].   Within
the DATIN there is another rule

    question(site_name)  ="Enter  a  descriptive  name for the
    discharge location".                               [4]

This rule is a treated as a "fact",  and M.I  prompts the user
for  a  valuation  of  "site_name" with  the message within the
quotes of statement [4].   The  user enters  a value  which is
bound to  the variable  SN, giving  the  expression "site_name"
the valuation of SN.  M.I continues to find valuations for the
remainder  of  the  expressions  in statement [3]  in a similar
manner.  When all expression in  the premise  of statement [3]
are assigned  a valuation,  the conclusion  "site_description =
found" is added as a fact  to the  knowledge base,  which will
allow M.I  to return  to statement [2] to seek a valuation for
the expression "ambient_conditions".

    Thus,  as  was  shown  with  the   previous  example,   the
knowledge  base  DATIN  is  built  from  rules  which  contain
expressions  that  force  M.I  to  seek   valuations  from other
rules.    The  process  of  seeking  valuations of expressions
continues either until all the  valuations  are  found  or the
rule base  is exhausted without finding  a valuation.  M.I will
never  assign a valuation  which is  in contradiction  within a
rule,  so  we are  assured whatever  valuations we conclude are
taken  from a rule  within the  knowledge base.   Care  must be
taken  in  program structure however, since  the search strategy
of  M.I may not  consider all rules needed to find  a valuation
for a  given expression.   In general, the  rule base should be
programmed in a  "tree"  structure, with  the most  general and
independent  rules  at  the  beginning of  the program, and rules
which  depend on  valuations from other  rules following  in the
program.     The  most  dependent and nested rules should occur
last in  the  knowledge  base.

    When a valuation for  a clause  in  the premise  of  a rule is
found  not   agree  with   the  valuation  given  for  that clause
within the rule, e.g.   the expression  discharge_parameters in
statement  [2]   is  found   to have the valuation  "unknown",  then
the rule fails,  no valuation  can be assigned to the  expression
"param_input_file"  from   that rule.   M.I   will stop trying to
satisfy  the  remaining  expressions  in  the premise  of  that rule.
M.I    will    continue   to    try   to  satisfy   the   expression

                              62

-------
"param_input_file"  by   looking  for   another  rule  in  the
knowledge base with "param_input_file" in the  conclusion of a
rule.

    The actual  rule corresponding  to statement   [2] in DATIN
contains additional clauses that  control the  manner in which
intermediate  conclusions  are  stored in memory, messages are
displayed on  the monitor,  and other  statements which create
and  manipulate  external  files  for  use  in  other  CORMIX1
modules.

    When the rule  in  DATIN  corresponding  to  statement [2]
fires,   the  "cache"  of  DATIN  is written to an external DOS
file.   The cache  is a  list of  all expressions  within DATIN
that have  been assigned a valuation.   This cache file is read
by the  next sequential element in CORMIX1,  the Fortran program
PARAM.   At its termination, DATIN directs the user to exit the
M.I environment and execute PARAM.

    A complete program listing  of  the  DATIN  knowledge base
appears in Appendix A.I.  Appendix B contains the output of an
interactive session using DATIN  for  the  design  examples in
Chapter V.
3.C.2 CLASS
    CLASS  is   an  M.I  program  that  classifies  the  given
discharge into one of  the many  possible flow configurations,
e.g. a simple jet or plume, an unstable vertically mixed case,
or mixing controlled by ambient flow.  The user executes CLASS
by typing "CLASS".

    Only  one  flow  classification  can  be  selected  for  a
particular discharge configuration from the many possible flow
classifications explained  in  Section 2.B.I and shown in Figure
2.6.

    The  goal  of  CLASS   is  to  find  a  valuation  for  the
expression   "flow_class"   from  the flow classification scheme
appearing in  Figure   2.6.    Each  of  the   17  possible flow
classification  has  an  alphanumeric  label(eg.  VI,  V2, H3,
etc.). CLASS inputs a  cache created by PARAM  that contains the
length  scales  and  other dynamic  variables needed for flow
classification, and  uses  the knowledge base rules to assign
the  appropriate   classification  to   the   flow.     A  rule
corresponding to  flow  case V2 would appear in simplified form
for illustration  purposes  as

    if vertical_angle  = THETAO and
    THETAO > 45.0 and
    THETAO <= 90.0 and

                              63

-------
    1m/H  <  1  and
    'Ib/H  <1 and
    IM/H  >  1
 then flow_class = V2.                                   [5j

 When  the   appropriate   flow   classification  rule   fires,   a
 detailed hydrodynamic  description  of   the  flow   is provided  to
 the user.   This detailed output includes  a  description  of  the
 significant near field mixing   processes,  or the hydrodynamic
 mixing zone (HMZ).  The  HMZ  is  the region  where the  particular
 design of the outfall  can have effect  on   initial dilution.
 The HMZ is  defined  to  give additional  information as an  aid  to
 understanding mixing   processes and   to   distinguish it from
 purely legal  mixing zone definitions.    CLASS also creates  a
 cache output  file that supplies the next CORMIX1  element, the
 Fortran   hydrodynamic    simulation   program   HYDRO,   with
 instructions  for running  the  appropriate  simulation.    At its
 termination CLASS directs the user to  exit the  M.I environment
 and execute HYDRO.

    The complete knowledge base for CLASS  appears in Appendix
 A. 2.   The interactive output from CLASS appears in  Appendix  B
 as discussed in the design examples Chapter  V.


 3.C.3 SUM
    SUM is an M.I  program  that   summarizes  the hydrodynamic
 simulation results   for the  case  under consideration.  SUM is
 executed by the user   by  typing  "SUM".    SUM comments  on the
mixing   characteristics,   evaluates   how  applicable  legal
requirements  are  satisfied,   and  suggests  possible  design
alternatives to improve dilution.   Thus,  SUM may be used as an
interactive loop to guide the   user  back  to   DATIN  to alter
design variables.

    The  output  of  SUM  is  arranged  in   four  groups; site
summary,  hydrodynamic  simulation  summary,    data analysis,  and
design  recommendations.     The  site  summary  gives the site
identifier  information,   discharge  and   ambient  environment
data,    and   discharge   length   scales.    The  hydrodynamic
simulation summary  will list  conditions   at  the  end  of the
hydrodynamic mixing  zone, legal mixing zone conditions,  toxic
dilution  zone   conditions,   region   of  interest  criteria,
upstream intrusion  information,  bank attachment locations,  and
a passive diffusion mixing summary, depending if the preceding
conditions are  specified or occur.  The data analysis section
gives further  details  on  toxic dilution   zone  criteria,  legal
mixing    zone    criteria,   stagnant   ambient   environment
information, and region of  interest  criteria.     Finally the
design  recommendations   section   gives design  suggestions for
improving  initial    dilution.     Factors   effecting  initial

                              64

-------
dilution  are  the  discharge  momentum flux, discharge angle,
outfall location,  discharge  buoyancy,  bank  attachment, and
ambient environment conditions.

    The listing  of the SUM knowledge base appears in Appendix
A.3.  The interactive listing of SUM appears in Appendix  B and
is discussed in the design examples Chapter V.
3.C Discussion of Hydrodynamic/Fortran Elements
    The FORTRAN elements of CORMIX1 are PARAM and HYDRO. These
element are programmed in  FORTRAN because  of the  limitations
of  M.I  for  computing  mathematical  expressions.   PARAM and
HYDRO are executed after  the user  has successfully  completed
DATIN and CLASS, respectively.
3.C.I PARAM
    PARAM is a Fortran program  that  computes relevant physical
parameters for the given  discharge  situation.   This includes
the various  length  scales;   IQ ,  1» ,   lm ,  It. , fluxes and other
values  needed  by   the  other  CORMIX1  elements.    PARAM is
executed by the user by typing  "PARAM".

    PARAM also  computes the  maximum  value  for each specified
mixing or interest zone for each  of  the  possible hydrodynamic
simulation  termination   criteria,  i.e.  maximum  downstream
distance,  plume  area,  or   plume   cross-section.     At  the
termination of  PARAM the user  is directed to execute the next
CORMIX1 element, CLASS.

    A  complete  listing  of  the  program   PARAM  appears  in
Appendix A.4.   Appendix  B provides the interactive output of
PARAM using the design examples presented  in Chapter V.
3.C.3 HYDRO
    HYDRO is a  Fortran  program  that  runs  the hydrodynamic
simulation  program   for  the   flow  classification  program
specified in CLASS.  The simulation program elements are based
on the similarity theory presented in Chapter II.

    HYDRO consists of control programs or "protocols" for each
hydrodynamic flow classification (V1,V2,H3,  etc.) as specified
by CLASS.   Each  protocol executes a series of subroutines or
"modules" corresponding to the  flow phenomena   (e.g. Wirght's
(1977)  mdnf,   mdff,  bdnf,  bdff; surface interaction modules,
etc.) which may occur in that flow classification.  Thus HYDRO
assembles  the  appropriate  simulation  from  the modules are
which are arranged like "pigeon holes".
                              65

-------
    Table 3.1  lists all modules called by the protocols for a
hydrodynamic simulation.  The "type" column  indicates whether
the  analysis  was  based  on  a  continuous or control volume
approach.  Table 3.1 lists the flow classification as given in
CLASS and  the protocols for assembling the simulation modules
corresponding to that flow classification.  In Table  3.2, the
protocols list  the simulation  modules to be called in order,
from left to right, to complete the hydrodynamic simulation.

    HYDRO creates a  tabular  output  file  of  the simulation
containing information  on geometry   (trajectory, width, etc.)
and mixing  (dilution, concentration).  The user executes HYDRO
by typing "HYDRO".

    After  HYDRO  has   executed,  the user may view the tabula
output file, giving detailed information  on the trajectory and
dilution of  the hydrodynamic flow simulation.  The program
listing  of   HYDRO  appears  in Appendix A.5 and the interactive
output appears  in  Appendix  B  as   discussed  in  the design
examples chapter  V.
                               66

-------
Table 3.1 Hydrodynamic Simulation Modules
MODULE
(MOD)
NAME
TYPE
01

11

16

21

22

31

32


33


34


41

61
zof e

mdnf

mdf f

bdnf

bdff

surface approach

surface impingement
upstream spreading

surface impingement
full vertical mixing

surface impingement
unstable near-field

buoyant spreading

passive diffusion
Control Volume

Continuous

Continuous

Continuous

Continuous

Control Volume

Control Volume


Control Volume


Control Volume


Continuous

Continuous
                           67

-------
Table 3.2 Hydrodynamic Simulation Protocols
Flow Classification: VI, HI
Modules:       01   11   21   22   31   41   61
Transition Rule:  013507
                                      T
                                   End of HMZ
Flow Classification: VIP,HIP
Modules:       01   11   16   22   31   41   61
Transition Rule:  024507
                                      T
                                   End of HMZ

Flow Classification: V2,H2
Modules:       01   11   16   31   41   61
Transition Rule:  02507
                                 T
                              End of HMZ

Flow Classification: V3,  H3
Modules:       01   11   21   32   41   61
Transition Rule:  01607
                                 T
                              End of HMZ

Flow Classification: V4,  H4
Modules:       01   11   33   61
Transition Rule:  050
                              T
                         End of HMZ

Flow Classification: V5,  H5
Modules:       01   11   21   22   31   41   61
Transition Rule:  013507
                                      T
                                   End of HMZ

Flow Classification: V6,  H6
Modules:       01   34   41   61
Transition Rule:  007
                       T
                    End of HMZ
                          68

-------
                         Chapter IV
               Data Comparison and Validation
    In  this  chapter  the  predictions   of   CORMIX1  will  be
compared with laboratory and field data.  This   chapter  is not
meant to be an extensive validation of CORMIX1 predictions for
all possible flow classifications,  but rather   a  limited test
of  key  CORMIX1  elements.    Future  work   plans  for CORMIX1
include  additional  calibration  and   validation  with  both
available laboratory and field data.

    The comparison  of CORMIX1  predictions will   focus  on the
near field flows, the  unstable  surface  impingement process,
and  the  near  vertical  impingement  with   buoyant  upstream
spreading module and buoyant surface spreading.
4.A Near Field Flows  (sub-surface regions)
    To validate the near field flows, CORMIX1 predictions were
compared with  laboratory data from Fan  (1967), Wright  (1977),
and Flatten and Keffer  (1971) .

    Figures  4.1  and   4.2  show  two  cases  of  Fan's  (1967)
trajectory  data   plotted  with  CORMIX1  projections.    Fan
released a dyed salt solution into uniform ambient flow within
a 40.Oxl.10xO.61  m flume.  Fan did not include the effects of
surface impingement in  his  analysis.    Photographs recorded
trajectory and  concentrations were measured with conductivity
probes.

    Figure 4.1 shows an experiment 20-12  with a  jet injected
perpendicularly (8  = 90.0° ) into crossflow (velocity ratio R
= uo/ua= 12)  and a Froude number Fo  = 20 representing a weakly
deflected forced  plume.  In this experiment, CORMIX1 predicts
a VI classification with a mdnf -> mdff -> bdff flow.  CORMIX1
appears to slightly under predict Fan's observed trajectory by
about 20%.  This discrepancy could  be accounted  for by Fan's
method for  defining the  plume centerline as the intersection
of  lines  of  equal  concentration  within  the  plume cross-
section, and  not location of maximum concentration within the
plume.  CORMIX1 assumes a maximum concentration along the flow
centerline.

    Figure 4.2 illustrates experiment 40-4 where the effect of
a larger crossflow with  (R = 4.0) a less buoyant jet  (Fo =

                              69

-------
            0.40  -
          o
          LjJ
          X
n 0.20 H
            0.00
                      • = Fan. 1967
                         Uo = 0.174 m/s  UO = 0.20B m/s
                         Froude No. = 20.0
                         Run No. 20-12
                                        - = CORMIX1
                   I I I I I I I I I | I I I I I I I I I | I I I I I I I I I | M I I I I I I I | I I I I I I I I I |

               0.00      0.20      0.40     0.60      0.80      1.00
                                     DISTANCE X (m)
Figure 4.1   Fan's  Buoyant  Jet Trajectory,  Expr.  20-12,  R =  12

                                    70

-------
              0.40  -
             n 0.20  H
               0.00
                            Fan, 1967
                            Uo = 0.408 m/s UO = 1.63 m/s
                            Froude No. =  40.0
                            Run No. 40-4
                                      CORMIX1
0.00      0.20      0.40      0.60      0.80
                      DISTANCE X (rn)
                                                                   1.00
Figure  4.2   Fan's  Buoyant  Jet  Trajectory,  Expr.  40-1,  R =  4

                                     71

-------
40.0)   where  the  flow  becomes  significantly deflected soon
after emerging from the discharge nozzle.  In this experiment,
CORMIX1 predicts  a VI  classification with  a mdnf -> mdff ->
bdff flow.  CORMIX1  trajectory  predictions  appear  to agree
well with Fan's data.

    Dilution predictions from CORMIX1 that correspond with the
experiments in Figures 4.1  and 4.2  are presented  in Figures
4.3  and   4.4,  respectively.     In   both  cases,  dilution
predictions are in close agreement  with  Fan's  results.   It
should be  noted that  all CORMIX1  near field predictions are
continuous, the apparent "kink" in Figure 4.4 (and Figures 4.5
and  4.6)  is  due  to  the  chosen  step  size  and  could be
eliminated by a higher resolution.

    Figure  4.5  shows  CORMIX1   trajectory  predictions  for
laboratory experiment 2-2 by Wright, (1977).   Wright conducted
his experiments in a 8.7x0.61x0.61 m towing tank.    Figure 4.5
shows an  experiment with R - 37 and Fo  =67.  As  can be seen,
CORMIX1 predictions  appear  to  be  in  strong  agrement with
Wright's data.   Unfortunately, no dilution data was available
for this experiment.

    Figure  4.6  shows  the  effect  on  a  nonbuoyant  jet of
changing the  angle between  the axis  of the discharge to the
horizontal plane from 80 = 90°,  as in  the previous examples,
to  9o  =  60°.    This  experiment  by Flatten and Keffer was
conducted with air in a 2.44  x 1.22  m wind  tunnel.  CORMIX1
predictions appear  to be  in good  agreement near the source,
but tend to deflect  more  strongly.    No  dilution  data was
available for this experiment.

4.B  Near  Vertical  Surface Impingement With Buoyant Upstream
Spreading
    Field  data  for  a   deep  wastewater   outfall  off  the
California coast  was obtained by the Allan Hancock Foundation
 (1964);  (see  also Chen,  1980).     In  this  case,  a strongly
buoyant  discharge with gd = 0.225   (Ap/p = 0.025)  was released
16.8 m below  the surface through  a  2.0 m diameter outfall into
the  ocean  with  an  ambient  current  of 0.175 m/s.  CORMIX1
predicts  a flow  classification   of  V5,  indicating  a  stable
discharge  configuration  with  buoyant upstream  spreading after
surface  impingement.   The  results of  the  CORMIX1 simulation
and the   actual data  appear  in Figures 4.7a and 4.7b.   Figure
4.7a  shows a   side  view  of   the   discharge  as  predicted by
CORMIX1.   Unfortunately, no field data is available to  compare
with  these subsurface predictions.  Field  data is  limited to
photographic  observation  of  the  surface plume.    Figure  4.7b
shows  a  plan  view   of the  CORMIX1  prediction   with field  data
available  from  this   remote  sensing   evidence  (Chen,  1980).
CORMIX1  predicts an upstream  intrusion of  91 m  with a  flow

                               72

-------
     CM

      o-
    z:
    o
                                      Fan, 1967    - =  CORMIX1
                                      Ua = 0.174   UO =  2.08 (m/s)
                                      Froude No. = 20.0
                                      Run No. 20 -  12
             i	1—i i  i i 111	1	1—r i i i  111	1	1—i  i i i 111

      10 ~2            10  ~1             1               10
                    DISTANCE ALONG TRAJECTORY s1 (m)
Figure  4.3  Fan's Buoyant Jet  Dilution,  Expr.  20-12, R  =  12


                                  73

-------
        o-
                                        Fan, 1967    -  = CORMIX1
                                        Ua = 0.408   UO  = 1.63 (m/s)
                                        Froude No. = 40.0
                                        Run No. 40-4
                         10 -1
                     DISTANCE ALONG TRAJECTORY s1  (rn)
Figure  4.4   Fan's Buoyant Jet  Dilution,  Expr.  40-4, R  = 4


                                  74

-------
           0.40  -J
        £

        M
        LJ
        X
          0.00
               0.00
                                     Wright,  1977
                                     Ua  = 0.046 m/s UO
                                     Froude  No. = 67
                                     Run  2-2
                      CORMIX1
                      1.720 m/s
 i  |  i  I  i I  I  i  I  l  l  |  r~~]i  i  i
0.40              0.80
  DISTANCE X (m)
Figure  4.5  Wright's  Buoyant Jet  Trajectory, Expr.  2-2  R =  37

                                75

-------
     o

     6
     LO  -
   o O
   UJ
   X
     O  -
Flatten  and Keffer,  1971
THETAO  = 60.0 deg SIGMAO 0.0 deg
Ua = 1.58 m/s  UO  = 9.99 m/s
Froude  No. = infinity
                                                     - = CORMIX1
           (I I I (I I I  I j I TTlllIII[1III
                                          I  I I I  I I I I  I I I  I I I I
       0.00
   0.05        0.10        0.15
             DISTANCE X (m)
0.20
Figure 4.6   Flatten and  Keffer,  9o  =  60.0


                                  76

-------
                                    ••20m
                         J0.98m
                                    ^
     u0=O.I75 m/s
              -75
-50
-25
                                    \

                                    \


                                    9*
                                    \\


25
50
75m
Figure 4.7a  Simulation of Stable Surface Impingement/Upstream

Spreading
                              77

-------
                                         Hancock data
                                         CORMIX1
Figure   4.7b
Spreading
Plan   View   of   Surface  Impingement/Upstream
                               78

-------
half width  at impingement of 237 m.  The actual data indicate
an upstream  intrusion of  about 90  m and  flow half-width at
impingement of  about 75  m.   No dilution  or flow depth data
were available.   CORMIX1  has  good  agreement  with upstream
spreading, but  over predicts flow width.  Possible sources of
error in the prediction might  be  a  caused  by  weak density
gradients which  can be expected in such ocean depths.  Future
work  for  CORMIX1  includes  the  consideration   of  density
stratification.   Overall agreement appears to be satisfactory
in this test case.

4.3  Unstable   Surface  Impingement   With  Buoyant  Upstream
Spreading
    Fischer  et  al.  (1979),  presents field data for the San
Onofre nuclear power plant.  The  San Onofre  Unit 1 discharge
is a  thermal discharge  from a 4.3 m diameter outfall located
5.0 ~R
m below the surface  water  off  the  California  coast.    The
ambient current  in the  vicinity of the outfall is about 0.14
m/s and the discharge velocity is  1.45 m/s.    The temperature
difference between  the ambient  current and  the discharge is
11.1'C giving rise to a buoyant  acceleration of  go '  = 0.032.
CORMIXl predicts  a flow classification of V6,  representing an
unstable  shallow   water  discharge   with  buoyant  upstream
spreading.  This case is further discussed as a design example
in Chapter V.

    Figure 4.8 shows the CORMIXl results  compared with actual
field  results  obtained  with  a scanning infrared radiometer
(Figure 5.3 presents the cross-section predicted  by CORMIXl).
CORMIXl predicts  an upstream  intrusion of  37 m  with a flow
half-width of 95 m at surface impingement.   The data   show an
upstream  intrusion  and  half-width at surface impingement of
about 30 m and 85 m,  respectively.    The unstable  near  field
impingement is  a highly complicated hydrodynamic process,  and
the results of CORMIXl  appear to  be in  excellent agreement,
although  the  width  of  the  density current  may be  somewhat
over-predicted with increasing downstream distance.
4.4 Summary
    Overall CORMIXl predictions appear  to  be  in  good first
order agreement  with available  observations in the field and
laboratory.  CORMIXl can  predict buoyant  upstream spreading,
where other  available models   (e.g. PLUME)  would have failed
entirely.

    Although limited data are   available   for  both  field and
laboratory  experiments,  further   efforts  will  be  made  to
compare model  predictions and  adjust parameters  in the flow
classification system.

                              79

-------
                                  CORMIX1  Plume Front  Prediction
                       Grid  Interval =30.5m
Figure 4.8    Plan View Comparison of  San  Onofre Prediction and
Field Data  (Corresponding cross-section in Figure 5.3)

-------
                         Chapter V
                     Design Case Studies
    The  purpose  of  this  chapter  is twofold; i) to give an
overview of the significant  features of  CORMIX1 in discharge
evaluation and  design, and  ii) to illustrate the flexibility
of CORMIX1 in highly divergent design  conditions.   The first
case  presented  represents  a  small  hypothetical industrial
discharge into a slowly flowing river, and the  second case is
an actual large cooling water discharge into the ocean.
5.A Case 1: AB CHEMICAL CORP., WEST VIRGINIA
    This is a hypothetical example of  a discharge in a bounded
section.  The CORMIX1  output  appears  in Appendix  B.I.  This
buoyant discharge  represents  a  complex 3-D trajectory subject
to three legal mixing  criteria;  a toxic dilution zone, a plume
cross-sectional  area   criteria on  a  legal  mixing zone, and a
downstream region  of  interest.   The  analyst seeks pollutant
concentrations at  these locations.
 5.A.I The  Problem  Statement
    AB   Chemical   discharges   an   industrial  effluent  into  the
 Ohio River through a   submerged   pipe   outfall.   The discharge
 flow is  0.2  m**3/s  and  contains 0.5  miCi/m**3  of  Cesium 134.
 Cesium   is considered toxic,  and  has   a   criterion maximum
 concentration (CMC)   value of  0.0005 miCi/m**3  (hypothetical).
 For the  critical  summer   conditions  the  discharge  temperature
 is 35.0'C.

    At the discharge site   the Ohio River is dammed as  a run-
 of-the-river   reservoir.  The  cross-section  is  approximately
 trapezoidal with   a bottom width  of  240 m and bank  slopes of  1
 in 3.  The river  depth is 8.5  m.   Due  to  gate operation  at  the
 dam the  river velocity varies between 0.1 m/s  (near-stagnant)
 and 0.6  m/s.   Typical summer temperatures  are  22.0'C.   The
 river  roughness   conditions   are given   by  a  Manning's n of
 0.024.

    The  outfall is located 50  m   from  the berm  line   near  the
 left bank.  The right bank  is under  the jurisdiction  of  the
 State of Ohio.  The port is pointing  directly  offshore (normal
 to  the  ambient   flow)   and   is   angled   20° upward above  the
                               81

-------
horizontal. The  round port  has a  diameter of  30 cm and its
center lies 0.5 m above the river bottom.

    The mixing zone limitations of the State of  West Virginia
have to  be considered.  These stipulate  that the mixing zone
have a maximum dimension  of 33%  of the  cross-sectional area
and a length equal to 10 times the stream width.
5.A.2 CORMIX1 Analysis
    The first step in the analysis would be to schematize
the bounded cross-section as shown in Figure 5.1.  Stream
cross-sections are usually highly irregular; the trapezoidal
cross-section represents an initial approximation of the
actual stream cross-section.  CORMIX1 assumes an equivalent
rectangular cross-section as shown in Figure 5.1, which the
analyst would approximate.

    Using DATIN,  the site  parameters are  specified.  One of
the advantages of logic  programming is  the "transparency" of
the  rule  base  to  the  user.    For  example,  DATIN  seeks
information concerning  the distance  from the  outfall to the
nearest shore.  During the session, DATIN prompts:

    What is the distance from the nearest bank or shore to the
effluent discharge point  (m)?                               [1]
If the user wants  to  know why  this information is needed,
he could determine the  rule  in  the knowledge  base that DATIN
is  trying  to   evaluate  simply  typing  "why", to which DATIN
responds:

    M.I  is trying  to  determine whether  the  following  rule is
    applicable  in  this  consultation:

    kb-33:
         if bounded_section  =  1 and
         distance_bank  = YB  and
         stream_width = BS and
         BS*0.5 >= YB and
         location_bank  = NBANK
    then nearest_bank = found

    The  following  entries are  also under  consideration:

         kb-29  (a  rule)
         kb-4   (an initialdata)                       [2]

The user   can  conclude  that the question in expression  [1] is
asked because  DATIN is  seeking a valuation for  the  expression

                               82

-------
         Cross-section View
            50m
  W. Vo.
                       Ohio
         Plan View
                              + 0.5
m
                                           8.5m
                              240m
262.75m

37.5m
t

uo =
t t t

O.I m/s
t t

f
                             D = 0.3m
                                     -r'20'
                                    0.5m
                     Detail of  Discharge
Figure 5.1  Schematization of Cross-section

                              83

-------
"nearest_bank"  in  "kb-33".    The  flag  "kb-33"  is a label
created by DATIN for this rule.

    If the user  is  interested  in  why  DATIN  is  seeking a
valuation  for  "nearest_bank"  in  kb-33, expression  [2] also
lists that kb-29 and kb-4 are also under  consideration.  Thus
the user  can conclude that kb-33 was invoked by another rule,
kb-29; and kb-29 was invoked by kb-4 an  "initialdata"  (similar
to  a  "goal")  statement.    The user can see kb-29 by  typing
"list kb-29"  to which  DATIN  responds   (simplified  here, for
full printout, see appendix B.I)

    kb-29:if  depth_at_discharge = H and
              nearest_bank = found and
              ambient_velocity_field = found and
          then ambient_conditions = found               [3]

to which  the user  can  conclude that "ambient_conditions" for
the discharge were being sought, when  the rule  in expression
 [2] for  "nearest_bank" was  invoked.

    Another   advantage    to  logic  programming  is  in  error
handling.     It   is    simple   to   write  rules    that   reject
contradictory data.     For  example,  when  schematized as a
rectangular  cross-section,   the stream   width  is   262.75 m and
the distance  to  the   nearest  bank  (W. Va.)  is 37.5 m.   If the
user  made an  error  and responded to expression  [1] by  entering
the   distance to   the  Ohio   shore  of   225.25 m, DATIN would
respond:

    The  distance  to nearest  bank is in error.
    The   value  must   be  less  than  half   the  stream  width.
    Recheck  and re-enter a  value less than or  equal
    to  131.375  (m).                                    [4]

and   the  user  is  given  another chance to enter the  correct
value  of  37.5 m.    This   result can  be  explained  as  follows.
The rule  kb-33 in  expression [2]  would fail when evaluating
the "BS*0.5  >= YB"  clause (stream width  = BS =  226.75)  in the
premise.     Since   kb-33  failed  when   DATIN  was seeking   a
valuation   for    the    expression    "nearest_bank"    it will
automatically seek   another   rule  in the knowledge base with
"nearest_bank" in   its conclusion.   Error   handling advice  is
placed   in   the   next  rule   that  has   "nearest_bank"   in its
conclusion.   Thus  the  message  shown  in expression   [4] comes
from  the  next rule   in  the  knowledge base with "nearest_bank"
in its conclusion.

    After completing  DATIN,   the   analyst  executes   PARAM,
followed  by  CLASS.   In CLASS  the analyst is advised  of the
intermediate  conclusions reached; i.e. the discharge is  a near

                               84

-------
horizontal "H"  case, 1m/h  = 0.885 indicating deep water with
weak momentum, ID/H = 0.923 showing weak buoyancy, and finally
ln/h  =0.864  concluding  that  buoyancy  dominates at surface
impingement.   CLASS  assigns  a  flow  classification  of HI,
specifying the flow simulation will consist of mdnf -> bdnf ->
bdff ->  surface  approach  ->  buoyant  surface  spreading ->
passive diffusion.

    Viewing the output of HYDRO shown in Figure 5.2, the plume
attaches to the left bank at x = 650 m downstream.   The plume
meets the  legal criteria of 33% of the stream cross-sectional
area at x =  850 m  downstream of  the orifice,  S =  395, c =
0.0014 miCi/m**3.   At  x = 1200 m the effluent is fully mixed
within the cross section with c = 0.00053 miCi/m**3.

    In SUM,  the analyst is alerted that the  assumed criterion
maximum concentration  (CMC) value  for the toxic discharge is
not met within the legal restrictions.  The user is advised to
improve dilution by changing the exit velocity, decreasing the
discharge angle Go ,  locating the port  in a  deeper section of
the cross-section,    or  by orienting 'the discharge so it does
not attach to the left bank first.
5.B Case 2: SAN ONOFRE UNIT 1
    While the previous case  represented a  hypothetical small
stable deepwater  discharge, San  Onofre is  an existing large
cooling water outfall  located  on  the  coast  of California.
Large   cooling   water   outfalls   are  typically  unstable,
characterized by strong momentum and relatively weak buoyancy.
The  CORMIXl  session  appears  in  Appendix  B.2.  The output
includes all available help  a user  could request  during the
consultation.
5.B.I The Problem  Statement
    The  San  Onofre  Nuclear  Generating  Station, Unit 1, is
operated by Southern California Edison Company. It has a power
output of  450 MW. The  cooling water  from unit 1 is discharged
about 1000 m offshore   at  a  local water  depth of  5.0 m. The
bathymetry   is    sloping    approximately  linearly  from  the
shoreline.

    The discharge  port  is  round with  a diameter  of 4.3  m and
extends about  0.2m above the surrounding bottom. The cooling
water is discharged vertically at  a flowrate  of 21 m**3/s. The
design  ambient    temperature  is  24.0°C  and  the  condenser
temperature rise  is 11.1'C (20°F).

    The site is characterized by weak currents along the shore


                              85

-------
      Cross-section
                     AB  Chemical Co.



      a) Close-Up Side  View of Near Field
                                  W.Va.
               Plan View
                           Ohio River
                             Ohio






                        b) Overall Plan View




Figure  5.2 Plot  of CORMIX1 AB Chemical Co. Predictions



                                 86

-------
of 6 cm/s in the southerly direction. The bottom is smooth and
sandy with  an  estimated  Darcy-Weisbach  friction  factor of
0.015 (Fischer et al.,  1979, Lee and Jirka, 1981).


5.B.2 CORMIX1 Analysis
     The representative cross-section in this case would place
the discharge 500 m from shore in 5.0 m of water.

    Class assigns a flow  classification of  V6, indicating an
unstable  discharge,  characterized  by  strong  momentum  and
relatively weak buoyancy.   The interactive  session of CORMIX1
appears in Appendix B.2.

    Figure   5.3   plots  the   cross-section  of   the  CORMIX1
predictions.   The plan  view of  this simulation  is shown in
Figure  4.8.

    SUM alerts  the analyst to the upstream buoyant intrusion,
and the possible  plume bank attachment.  Again   the analyst is
advised  to   improve  dilution   by   some   of  the  same options
outlined  in  section 5.A.2..
                               87

-------
           -30    -20   -10
                               5.0m
10    20    30   40m   x
Figure  5.3    San Onofre Longitudinal Cross-Section  (Vertical
Distortion = 6,  See Corresponding Plan View in  Figure  4.8)
                              88

-------
                         Chapter VI
               Conclusions and Recommendations
    U.S. water  quality policy  allows for  a mixing zone as a
limited area or volume of water where the  initial dilution of
a discharge occurs.  Water quality standards apply at the edge
and  outside  of  the  mixing  zone.    Toxic  discharges have
additional  regulatory  restrictions, which require additional
dilution analysis.  The implementation of  this policy  in the
National   Pollution   Discharge  Elimination  System  (NPDES)
permitting process places the burden of  prediction of initial
dilution on  both regulators  and dischargers.  Given a myriad
of possible  discharge  configurations,  ambient environments,
and  mixing  zone  definitions, the analyst needs considerable
training and expertise to conduct accurate and reliable mixing
zone analysis.  An expert system, CORMIX1,  was developed as an
analysis tool for regulators and dischargers.

    CORMIX1 predicts the dilution  and trajectory  of a single
buoyant discharge into a unstratified ambient environment with
or without  crossflow.   CORMIX1 uses  knowledge and inference
rules  obtained   from  hydrodynamic  experts  to  classify and
predict buoyant jet mixing.    CORMIX1  gathers  the necessary
data, checks  for data consistency, assembles and executes the
appropriate  hydrodynamic  simulation  models,  interprets the
results of  the simulation  in terms of the legal requirements
including  toxic  discharge  criteria,   and  suggests  design
alternatives to improve dilution characteristics.

    The results   of the hydrodynamic simulation are in good to
excellent  agreement  with  field   and  laboratory  data.   In
particular,   CORMIX1   correctly   predicts   highly  complex
discharge  situations  involving  boundary    interactions  and
buoyant intrusions,  a result  not predicted by other currently
available  initial mixing models.

    However, simulation models of   complex phenomena,  such as
the  hydrodynamics  of  jet  and plumes described here, should
always be  given with a caveat.   The  analyst  is   forced to make
many   assumptions   when   modeling   the  complex  mechanisms
controlling buoyant jet flow.

    In reality, many physical  processes  which   occur  in the
environment  are  difficult  to  simulate  or control  for in
laboratory experiments. If  the   basic   assumptions  within the

                               89

-------
methodology are  violated in  the analysis of a discharge, the
resulting analysis will be  tenuous  at  best.    For example,
CORMIX1 assumes  a uniform  velocity field in a uniform cross-
section.  This is  a good  first order  approximation for many
simple  discharges.    But  the  extreme case represented by a
discharge into  a cross-section  that is  highly variable with
downstream   distance,   characterized   by   strong  velocity
fluctuations within  the flowfield   (such as  a discharge into
stream  section  with  rapids),  would  be beyond the scope of
CORMIX1.

    What has  been  attempted  here  is  to  place  a modestly
complex   hydrodynamic   simulation   methodology  within  the
framework of a rule based expert system.   Many of  the common
pitfalls  to  model  use  -  incomplete or contradictory data,
choice   of   appropriate   simulation   model,   and   faulty
interpretation of  results - appear to be mitigated within the
context of an expert system methodology.

    CORMIX1 educates the user  to  the  important hydrodynamic
processes controlling  the flow.   It  will give 3-D discharge
trajectory and dilution.  It  will  alert  the  user  to where
significant  legal  criteria  apply  to  the  discharge.    It
predicts buoyant upstream spreading, which presently  no other
model  can  simulate.    It  allows  for a rapid evaluation of
design  alternatives,  and  gives  the  user  suggestions  for
improving dilution  characteristics of the discharge.   Overall
CORMIX1 appears to be an  excellent  first  cut  tool   for the
analyst.

    As stated  in Chapter  IV, further  work should be done to
determine the constants in the flow  classification system and
all other constants within the model.  More research should be
devoted to the concepts involved in the design recommendations
in  SUM.    The  existing  data base is limited for conducting
rigorous validation studies indicating  a need  for additional
field and laboratory data.

    The  problem  domain  of  CORMIX1,   single port positively
buoyant discharges into a  uniform  density  field,   should be
extended   to   include   near   field  boundary  attachments,
negatively   buoyant   discharges,    and   density  stratified
environments.   The application of computer generated graphics
to plot simulation results would enhance user understanding of
simulation results.

    Future work  plans for CORMIX1 are: further calibration of
model  constants,   consideration  of   dynamic  jet  boundary
attachment, analysis of negatively buoyant discharges, and the
effects of density stratified environments.


                              90

-------
                         Reterences


Allan  Hancock  Foundation,  University of Southern California
(1964) ,  "Final Report  on   an   Investigation  on  the  Fate of
Organic  and   Inorganic  Wastes  Discharge   into  the  Marine
Environment and  Their  Effects   of   Biological Productivity",
September 15.

Benjamin,   T.   B.    (1968)   "Gravity  Currents	and	Related
Phenomena", J. Fluid Mechanics, Vol.  31,  pt.  2.

Barnwell,  T.   O.,Brown  L.    C.,   and   Marek,  W.,   (1985).
"Development  of  a  prototype   expert  advisor  for the enhanced
stream  water   quality   model   QUAL2E",    Internal  Report,
Environmental  Research  Laboratory,  Office  of  Research and
Development,  U.S.E.P.A., Athens,  Georgia, September, 1985.

Briggs,  G.   A.   (1969).    Plume  Rise,  U.S.  Atomic  Energy
Commission,   Division  of   Technical  Information,  Oak Ridge,
Tennessee.

Congressional  Research Service,   (1977).  "Legislative History
of the  Clean  Water Act 1977",  Congressional  Research Service,
Library of Congress, October  1978,  No.  95-14  p. 330.

Chen, J. C.,   (1980).    "Studies  on  Gravitational Spreading
Currents",  Rep.  KH-R-40,  W.M.  Keck Laboratory of Hydraulics
and   Water  Resources,  California   Institute of  Technology,
Pasadena, Calif.

Clocksin, W.   F. and  Hellish,  C.  S.,  (1984).  Programming in
Prolog, 2nd ed., Springer-Verlag.

Fan   L. N.  (1967) .  "Turbulent  Buoyant  Jets into Stratified or
Flowing  Ambient  Fluids,"  Rep.  KH-R-18, W.M. Keck Laboratory
for Hydraulics and  Water  Resources,  California  Institute of
Technology, Pasadena, Calif.

Fay   J  A , Escudier, M. and  Hoult,  D.  P.  (1970).  "^Correla-
tion  of Field  Observation of   Plume__Rise." ,   J.  Air Pollution
Control Association, vol. 20.,  No.33, pp. 6818-6835.

Findley and  Farber (1983),   Environmental Law	in a Nutshell,
West  Publishing Co.
                               91

-------
Fischer, H.  B. et  al .  (1979).   Mixing in Inland and Coastal
Wat_ers ,  Academic Press,  New York.

Gaschnig, J., Reboh, and J. Reiter,  (1981).   "Development of a
Knowledge-Based Expert   System for  Water Resources Problems",
Palo Alto, Calif.                                         ~~

Holley,  E. R., and  Jirka, G.  H.  (1986).  "Mixing in Rivers",
Technical  Report  E-86-11,  U.S.  Army  Corps  of  Engineers,
Washington, B.C.

Jirka, G. H. and D. R. F. Harleman  (1973).    "The Mechanics of
Submerged  Multiport   Diffusers  for  Buoyant  Discharges  in
Shallow Water" , M.I.T.,  Ralph M. Parsons  Laboratory for Water
Resources and Hydrodynamics, Tech. Rep No. 169.

Jirka,  G.  H.,  and   Joseph  H . -W .  Lee,  (1987).   Hydraulic
Structures  Design  Manual,  E.   Naudascher,  Ed.,   Vol  10.
International Association of Hydraulic Research,  (in print).

Jirka,  G.  H.,  J.  M.  Jones,  and  F.  E.   Sargent,  (1980).
"Theoretical and  Experimental  intermediate   Field Dynamics of
Ocean Energy  Conversion Plants",   Progress Report, 1978-1979,
School  of   Civil  and   Environmental  Engineering,  Cornell
University .

Jones,  J.  M.,  G.  H.  Jirka,  and  D.  A.   Caughey,  (1985).
"Numerical  Techniques  for  Steady Two-Dimensional  Transcritical
Stratified    Flow   Problems,    with  an  Application   to   the
Intermediate  Field  Dynamics  o f  Ocean Thermal  Energy Conversion
Plants" ,    Report   No.    ANL/EES-TM-27T ,   Argonne  National
Laboratroy .

Koh,  R.  C.,  and Brooks,  N.   H.   (1975).   "Fluid   Mechanics of
Waste   Water   Disposal  in   the Ocean" ,  Annual review  of  Fluid
Mechanics,  Vol  7,  pp.  187-211.

Larsen,  J.  and  Sorensen,   T.   (1968) .     "Buoyancy   Spread of
Wastewater  in  Coastal  Regions",  Eleventh Conference  on Coastal
Engineering,  London,  Vol.  2.

Lee,  J.  H.  W-,  and   Gehard  Jirka,   (1981).     "Vertical  Round
Buoyant  Jet   In  Shallow   Water",   Journal   of the  Hydraulics
Division,  American  Society of  Civil Engineers,   Vol.  107,  No.
HY12, December.

List  E.   J.  (1982).   "Mechanics of Turbulent Buoyant  Jets  and
        '  Chapter in Turbulent Buoyant  Jets  and Plum^s_,  (1982).
                               92
 W.  Rodi ,  ed.,  Permagon Press.

-------
List, E.  J-, and Imberger J.,  (1973).   "Turbulent  Entrainment
in Buoyant Jets",  Proc.  ASCE,   J.   HydraulicsDivision,99,
1461-1474.

Morton,  R.  B.   (1959).     "Forced   Plumes,"  Journal  of  Fluid
Mechanics, vol 5, pp  151-163.
Monin,_A. S., and Yaglom, A.  M.,  (1971).      Statistical  Fluid
Mechanics:  Mechanics   of   Turbulence,   Vol.  1,   MIT Press,
Cambridge, Mass, p.485.

Morton,  B.  R.,  Taylor,  G.I.   and   Turner,  J.   S.   (1956).
"Turbulent   Gravitational    Convection   form  Maintained  and
Instantaneous Sources,"   Proceedings  Royal  Society  of London,
A234, pp. 1 -23.

Muellenhoff,   W.   P.,   et.   al . ,   (1985).   "Initial Mixing
Characteristics  of  Municipal   Ocean   Discharges   (Vol. 1&2)",
U.S.E.P.A,  Environmental  Research   Laboratory,  Narragansett,
R.I. .

Naib, S. K. A.  (1974).  "Deflection of   a Submerged  Round Jet
to Increase  Lateral  Spreading",   La  Houille Blanche,  Vol. 29,
No .   6 .

National Academcy of  Science,  (1968).  "Water   Quality  Criteria
(1968) " ,  (Green  book),  Report   of   the   National  Technical
Advisory Committee to the  Secretary  of   the Interior,  Federal
Water  Pollution  Control  Administration,  U.S.  Dept. of the
Interior, Washington, B.C..

Ortolano, Leonard, (1984). Environmental  Planning and  Decision
Making, John Wiley &  Sons.

Flatten,  J.  L.  and  Keffer,  J.  F.,   (1971).    "Deflected
Turbulent Jet Flows", Transactions of  the   American Society of
Mechanical Engineers, Vol 38,  Dec. 1971,  pp. 756-758.

Roberts, P.  J.  W.   (1977).   "Dispersion  of Buoyant Wastewater
Discharge From Outfall Diffusers  of Finite  Length", Report No.
KH-R-35,  W.  M.  Keck  Laboratory  of   Hyydraulics  and Water
Resources,  Division  of  Engineering and  Applied  Sciences,
California Institute  of Technology, Pasadena,  Cal.

Rajaratnam, N.  (1976). Turbulent -Jets, Elsevier.

Scorer,  R.  S.,   (1954).     "The Behay_ip_r__of_ Chimney  Plumes,"
International Journal of  Air Pollution~Vol 1, pp 198-220.

Tennekes, H. and Lumley,  J.L.,   (1972).     A   First	Course  in
Turbulence, MIT  Press, Cambridge, Mass,  p.  99.
                               93

-------
Roberts,  P.   J.   W.,   (1979).    "Line Plume and Occ-an  Outfall
Dispersion" , Journal  of  the Hydraulics Division,   Proc.  of  the
Amer . Soc.  of  Civil   Engineers, Vol  105, Mo. HY4 ,  April,  pp.
313-331.

Rodi, W . ed,   (1932).     Turbulent  Buoyant  Jets   and  Plumes ,
Pergamon Press,  New York.

Shwartz,  J.   and  Tulin,  M.P.   (1972).     "Chimney Plumes in
Neutral  and   Stable   Surroundings" ,  Atmospheric  Environment^
vol. 6  pp.  19-36.

Tait,   H.   (1984) .     "Section  402  Guidelines   for Reviewing
National Pollutant  Discharge  Elimination   Permits" , National
Coastal Ecosystems  Team, Rept.  No. FWS/OBS-8 4/0 5 ,  U."S. Dept.
of  the  Interior,  Washington, B.C.,  September.

USEPA  (1972).   "Water  Quality Criteria __ 1.972.",  EPA-R3-7 3-00 3 ,
Environmental   Studies   Board,  Committee   on   Water  Quality
Criteria, Washington,  B.C..

USEPA  (1976).   "Quality Criteria for  Water  1976"   (Red Book),
Guidelines  for  State and  Area Wide  Water  Quality Management
Program, Washington,  B.C.  (Chapter  5).

USEPA,  (1982).   Water Quality Standards   Handbook.    Office of
Water Regulations and Standards, Washington, D C.
USEPA ,  (1984).   "Technical Guidance Manual^ for  the _Re_g_u l.a.IA°DLS
Promulgated  Pursuant to Se_ctig_ n  3011 _.|gj_ ____ p_f_._t he _CI^ a n  Water
Act of  1977  (Braft)  ",  Washington B.C., August.
USEPA,  (1985) .   "Tejc^m_:lc_al__Sjjpj2p_^
based Toxics   Control",   Office  of  Water,  Washington, B.C.,
September .

Wright   S.  J.   (1977).     " Ej£f_e_cJ:_s ___ o r __Ajmb_i._e_n_t ___ Crossflows and
Bensity  Stratif icaticm_gn the Characterise i c _ B e h ay i o r q f R o u n d
Turbulent Buoyant _Je_t_s" ,  Rep. KH-R-36, W.M. Keck Laboratory of
Hydraulics  and  Water  Resources,    California  Institute  of
Technology, Pasadena,  Calif.
                               94

-------
                         APPENDIX A

    Appendix  A  contains  the  code  listing for all programs
within CORMIX1.   The M.I programs are presented first followed
by  the  Fortran  programs.    Section A.I has the listing for
DATIN, the M.I data entry module.  Section  A.2 shows  the M.I
listing for  CLASS, the  flow classification program.  Section
A.3  lists  SUM,  the   M.I   program   that   summarizes  the
hydrodynamic simulation  output.   Section A.4  is the Fortran
program PARAM which computes the  length  scales  used  in the
flow classification  system.  Section A.5 contains the Fortran
listing for the hydrodynamic simulation program HYDRO.
                              95

-------