US EPA Environmental Technology Verification Report NOx Control Technologies Catalytica Combustion Systems, Inc., Xonon Flameless Combustion System


December 2000

Environmental Technology
Verification Report

NOx Control Technologies

Catalytica Combustion Systems, Inc.
Xonon™ Flameless Combustion
System

Prepared by

MRI&

Midwest Research In stitute	Research Triangle Institute

Under a Cooperative Agreement with

r OA United States
Q • Environmental Protection Agency

ETVETVETV

-------
THE ENVIRONMENTAL TECHNOLOGY VERIFICATION PROGRAM

&EPA

U.S. Environmental Protection Agency

Research Triangle Institute

ETV Joint Verification Statement

TECHNOLOGY TYPE:
APPLICATION:

: NOx AIR POLLUTION CONTROL TECHNOLOGY

A PROCESS-INHERENT NOx EMISSION CONTROL SYSTEM
FOR GAS TURBINE APPLICATIONS

TECHNOLOGY NAME: XONON™ COOL COMBUSTION

COMPANY:
ADDRESS:

430 FERGUSON DRIVE
MOUNTAIN VIEW, CA 94043-5272

http://www.CatalyticaEnergy.com

CATALYTICA ENERGY SYSTEMS, INC.*

PHONE: (650)960-3000
FAX: (650) 960-0127

WEB SITE:

* Catalytica Energy Systems, Inc. is the former Catalytica Combustion Systems, Inc. (CCSI)

The U.S. Environmental Protection Agency (EPA) has created the Environmental Technology Verifica-
tion (ETV) Program to facilitate the deployment of innovative or improved environmental technologies
through performance verification and dissemination of information. The goal of the ETV Program is to
further environmental protection by substantially accelerating the acceptance and use of improved and
cost-effective technologies. ETV seeks to achieve this goal by providing high-quality, peer-reviewed
data on technology performance to those involved in the design, distribution, financing, permitting,
purchase, and use of environmental technologies.

ETV works in partnership with recognized standards and testing organizations; with stakeholder groups
that consist of buyers, vendor organizations, permitters, and other interested parties; and with the full
participation of individual technology developers. The program evaluates the performance of innovative
technologies by developing test plans that are responsive to the needs of stakeholders, conducting field or
laboratory tests (as appropriate), collecting and analyzing data, and preparing peer-reviewed reports. All
evaluations are conducted in accordance with rigorous quality assurance protocols to ensure that data of
known and adequate quality are generated and that the results are defensible.

The Air Pollution Control Technology (APCT) program, one of 12 technology areas under ETV, is
operated by the Research Triangle Institute (RTI) in cooperation with EPA's National Risk Management
Research Laboratory. Midwest Research Institute, on behalf of the APCT program, has evaluated the
performance of a nitrogen oxides (NOx) control technology utilizing flameless catalytic combustion for
stationary gas turbines, Xonon™ Cool Combustion (formally known as Xonon™ flameless combustion.)

VERIFICATION TEST DESCRIPTION

All tests were performed in accordance with general guidance given by the APCT program "Generic
Verification Protocol forNOx Control Technologies for Stationary Combustion Sources" and the specific
technology test plan "Verification Test/QA Plan for Xonon™ flameless combustion system." These
documents include requirements for quality management, quality assurance, auditing of the test
laboratories, and test reporting format.

l

-------
Xonon™

The Xonon™ Cool Combustion system was tested as installed and operating on a Kawasaki M1A-13A
gas-turbine-generator set (1.5 MW) located in Santa Clara, California, on July 18 and 19, 2000. NOx
concentrations were measured using continuous emission monitors (CEMs) following EPA Reference
Method 20 for gas turbines. Other gaseous emissions were monitored using the applicable EPA test
method. Other process variables were monitored using calibrated plant instrumentation.

Tests were conducted to meet the data quality objective of a 95 percent confidence interval with a width
of ±10 percent or less of the mean NOx emission concentration for concentrations above 5 ppmvd,
±25 percent or less below 5 ppmvd and above 2 ppmvd, and ±50 percent or less below 2 ppmvd. In
addition to outlet NOx concentration and the primary process variables, carbon monoxide and unburned
hydrocarbon emission concentrations were also measured using EPA reference methods, and the
installation efforts, site modifications, staffing, maintenance requirements, and similar issues were noted
qualitatively.

A single test run consisted of measuring outlet NOx concentra-
tion and the other parameters over a 32-min steady-state
process condition with the primary variable, ambient tempera-
ture, at either its low point or high point (i.e., early morning or
late afternoon). The test design was a replicated
2x1 factorial using two levels of ambient temperature and
greater than 97 percent of the rated full load. A total of 12 test
runs were conducted over the 2-day field test period. Ambient
temperature variation was small over the test period. Table 1
gives the operating performance envelope over which the
Xonon™ Cool Combustion system was verified.

DESCRIPTION OF XONON™ TECHNOLOGY

This verification statement is applicable to the Xonon™ Cool Combustion system for gas turbine
applications without the air management system. The Xonon™ Cool Combustion system is completely
contained within the combustion chamber of the gas turbine. Xonon™ Cool Combustion completely
combusts fuel to produce a high-temperature mixture, typically about 1300 °C (2400°F). Dilution air is
added to shape the temperature profile required at the turbine inlet.

The Xonon™ Cool Combustion system consists of four sections:

Preburner. The preburner is used to preheat the air before it enters the catalyst module and during
startup for acceleration of the turbine. The preburner tested as part of this verification was a lean,
premixed combustor.

Fuel injection and fuel/air mixing system. This unit injects the fuel and mixes it with the main air
flow to provide a very well mixed, uniform fuel/air mixture to the catalyst.

Xonon™ catalyst module. In the catalyst module, a portion of the fuel is combusted without a
flame to produce a high-temperature gas.

Homogeneous combustion region. Located immediately downstream of the catalyst module, the
homogeneous combustion region is where the remainder of the fuel is combusted, and carbon
monoxide and unburned hydrocarbons are reduced to very low levels (also a flameless combustion
process).

Verification Statement Table 1.
Verification Test
Performance Envelope3

Ambient

Temperature, °C

Low

High

aAt >97 percent of full turbine load.

-------
Xonon™

The overall combustion process in the Xonon™ system is a partial combustion of fuel in the catalyst
module followed by complete combustion downstream of the catalyst in the burnout zone. Partial
combustion within the catalyst produces no NOx. Homogeneous combustion downstream of the catalyst
usually produces no NOx, because combustion occurs at a uniformly low temperature. A small amount of
fuel is combusted in the preburner to raise the compressed air temperature to about 470°C (880°F). NOx
in the turbine exhaust is usually from the preburner.

The design of each Xonon™ combustor is customized to the particular turbine model and operating
conditions of the application and would typically be defined through a collaborative effort with the
manufacturer of the turbine to integrate the hardware into the design. Catalytica Energy Systems, Inc.
expects that the Xonon™ Cool Combustion technology incorporated in a Xonon™ combustion system
for a natural-gas-fueled Kawasaki M1A-13A gas turbine is capable of achieving emissions of NOx of less
than 2.5 ppmvd (corrected to 15 percent oxygen [02]) on a 1-hour rolling average basis, and less than 2.0
ppmvd (corrected to 15 percent 02) on a 3-hour rolling average basis. Under the same conditions, the
Xonon™ combustion system is expected to achieve carbon monoxide (CO) emissions of less than
6 ppmvd (corrected to 15 percent 02). The footprint may vary depending on the implementation,
although generically the Xonon™ combustion system would likely be somewhat larger than the
combustor that is typically supplied as standard equipment by the turbine manufacturer. Each unit could
have multiple fuel inputs from separate control valves, and additional instrumentation for control and
monitoring would be integrated into the turbine control system.

This verification statement covers application of the Xonon™ Cool Combustion system to small gas
turbines operated at full load when combusting natural gas within the stated operating condition
envelope. This unit was operated at the test site by the vendor, Catalytica Energy Systems, Inc., for over
4,000 hours before the verification test. Data from this long-term operating period have been submitted
to a number of regulatory authorities for their review and evaluation. While these data and the
instruments used were not verified during this test, within the operating condition envelope the results are
generally consistent with the verification test results. Catalytica Energy Systems, Inc. should be
contacted for these data or other information.

VERIFICATION OF PERFORMANCE

The verified NOx emission results are given in Table 2. The analysis of variance between NOx and
ambient temperature indicated that ambient temperature did not affect NOx emissions over the narrow
range encountered during this verification test.

Verification Statement Table 2. NOy Control Performance

Half-Width of 95%

Ambient

Percent of Full

Mean Outlet NOx

Confidence Interval

Temperature

Turbine Load

Concentration

on Mean Outlet NOx

Range

ppmvd @ 15% Oz

15 to 25°C (59 to 77°F)

98-99%

1.13

0.026

ppmvd = parts per million by volume dry basis.

CO emissions averaged 1.36 ppmvd at 15 percent 02. Unburned hydrocarbon (UHC) emissions averaged
0.16 ppmv (wet basis reported as propane).

ill

-------
Xonon™

The APCT quality assurance (QA) Officer has reviewed the test results and quality control data and has
concluded that data quality objectives given in the NOx Control Technology generic verification protocol
and test/QA plan have been attained. During the verification tests, the EPA and APCT QA staffs
conducted a performance evaluation and a technical system audit at the field test site, which confirm that
the verification test was conducted in accordance with the EPA-approved test/QA plan.

This verification statement verifies the NOx emissions characteristics of the Xonon™ Cool Combustion
system within the range of application tested (see Table 2). Extrapolation outside that range should be
done with an understanding of the scientific principles that control the performance of the Xonon™ Cool
Combustion system. Gas turbine users with NOx control requirements should also consider other
performance parameters such as service life and cost when selecting a NOx control system.

In accordance with the NOx Control Technology generic verification protocol, this verification report is
valid indefinitely for application of the Xonon™ Cool Combustion system within the range of
applicability of the statement.

Original signed by Hugh W. McKinnon 12/15/00
Hugh W. McKinnon Date

Acting Director

National Risk Management Research Laboratory

Office of Research and Development

United States Environmental Protection Agency

Original signed by Jack R. Farmer 12/22/00

Jack R. Farmer Date

Program Manager

Air Pollution Control Technology Program
Research Triangle Institute

NOTICE: ETV verifications are based on an evaluation of technology performance under specific,
predetermined criteria and the appropriate quality assurance procedures. EPA and RTI make no
expressed or implied warranties as to the performance of the technology and do not certify that a
technology will always operate as verified. The end user is solely responsible for complying with any
and all applicable federal, state, and local requirements. Mention of commercial product names does
not imply endorsement.

-------
Environmental Technology
Verification Report

NOx Control Technologies

Catalytica Combustion Systems, Inc.
Xonon™ Flameless Combustion
System

EPA Cooperative Agreement CR 826152-01-2

EPA Project Manager: Theodore G. Brna
Air Pollution Prevention and Control Division
National Risk Management Research Laboratory
Office of Research and Development
Research Triangle Park, NC 27711

December 2000

Prepared by

Craig Clapsaddle
Midwest Research Institute
5520 Dillard Road
Cary, NC 27511-9232

Douglas VanOsdell
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, NC 27709

v

-------
Notice

This document was prepared by Midwest Research Institute (MRI) under a contract with Research
Triangle Institute (RTI) with funding from Cooperative Agreement No. CR826152-01-2 with the U.S.
Environmental Protection Agency (EPA). The document has been subjected to RTI/EPA's peer and
administrative reviews and has been approved for publication. Mention of corporation names, trade
names, or commercial products does not constitute endorsement or recommendation for use of specific
products.

Catalytica Combustion Systems, Inc. becomes Catalytica Energy Systems, Inc.

Catalytica Combustion Systems, Inc. (abbreviated in this report as CCSI), a subsidiary of Catalytica, Inc.,
reorganized into stand-alone, publicly-traded Catalytica Energy Systems, Inc., on December 18, 2000.
The Xonon™ Cool Combustion technology, referred to in this report as Xonon™ flameless combustion,
remains the same, and all references to CCSI should be understood to refer to Catalytica Energy Systems,
Inc. Contact information in the verification statement and report has been updated.

vi

-------
Availability of Verification Statement and Report

Copies of the public Verification Statement and Verification Report are available from

1.	Research Triangle Institute

P.O. Box 12194

Research Triangle Park, NC 27709-2194

Web site: http://etv.rti.org/apct/index.html

or http://www.epa.gov/etv/ (click on partners)

2.	USEPA / APPCD

MD-4

Research Triangle Park, NC 27711

Website: http://www.epa.gov/etv/library.htm (electronic copy)
http://www.epa.gov/ncepihom/

vii

-------
Acknowledgments

The authors acknowledge the support of all of those who helped plan and conduct the verification
activities. In particular, we would like to thank Ted Brna, EPA's Project Manager, and Paul Groff,
EPA's Quality Assurance Manager, both of EPA's National Risk Management Research Laboratory in
Research Triangle Park, NC. We would also like to acknowledge the assistance and participation of all
the Catalytica Combustion Systems, Inc., personnel who supported the test effort.

For more information on NOx Control Technologies Verification Testing, contact

Douglas VanOsdell
RTI

P.O. Box 12194

Research Triangle Park, NC 27709
(919) 541-6785, dwv@rti.org

For information on Xonon™ Cool Combustion, contact
Chuck Solt

Catalytica Energy Systems, Inc.

430 Ferguson Drive
Mountain View, CA 94043-5272
(916) 729-5004

csolt@,CatalvticaEnergy.com or catalytica@csolt.net
For commercial issues, contact
Joseph Cussen

Catalytica Energy Systems, Inc.

430 Ferguson Drive
Mountain View, CA 94043-5272
(650) 940-6393

j cussen@CatalyticaEnergy.com

viii

-------
Abstract

Nitrogen oxides (NOx) air pollution control technologies (APCTs) are among the technologies
evaluated by the APCT Environmental Technology Verification (ETV) Program. The APCT
program developed the Generic Verification Protocol for NOx Control Technologies for
Stationary Combustion Sources to provide guidance on the verification of specific technologies.
The critical performance factor for this verification is the NOx emission concentration within the
performance envelope of the test. This protocol was developed by RTI and MRI, reviewed and
discussed by a technical panel of experts, and approved by EPA. The protocol states the critical
data quality objectives for aNOx control technology verification, as well as noncritical but still
important measurements of other performance parameters.

The Catalytica Combustion Systems, Inc., Xonon™ flameless combustion system was submitted
to the APCT ETV program for verification. A test/quality assurance (QA) plan, prepared in
accordance with the generic verification protocol, addressed the site specific issues associated
with the verification test. The verification was conducted the week of July 17, 2000, at the
Xonon™ installation on a 1.5-MW gas turbine in Santa Clara, CA. The mean outlet NOx
concentration during the verification was determined to be 1.13 ppmvd at 15% 02. The
measured NOx concentration was well within the stated data quality objective for the NOx
measurement. Other important performance and operating parameters were also measured.

-------
Contents

Section	Page

ETV Joint Verification Statement	i

Notice	vi

Acknowledgments	viii

Abstract 	ix

Figures	 xii

Tables	 xii

Acronyms/Abbreviations	xiii

1.0 Introduction 	1

2.0 Description and Identification of Xonon™ Flameless Combustion System	2

3.0 Procedures and Methods Used in Testing	4

3.1	Test Design	4

3.2	Sampling Methods 	5

3.2.1	Sampling Locations	5

3.2.2	NOx, CO, UHC, and 02/C02 Sampling Procedures 	8

3.2.3	Sampling Methods Requirements	11

3.2.4	Process Data Collection 	15

3.2.5	Ambient Conditions Sampling	17

3.3	Data Acquisition and Data Management 	17

4.0 Statement of Operating Range of Test	19

5.0 Summary and Discussion of Results	21

5.1	Results Supporting Verification Statement	21

5.1.1	Statistical Analysis of Variance 	21

5.1.2	Variability of NOx Emissions	22

5.2	Discussion of QA/QC and QA Statement	22

5.2.1	NOx Measurement DQO	22

5.2.2	Reference Method QC 	23

5.2.3	Audits 	25

5.3	Deviations from Test Plan 	26

6.0 References 	28

x

-------
Contents (continued)

Appendixes

A	QA/QC Activities and Results

A.l Pre- and Post-test Calibration Results	30

A.2 Reference Method Performance Audit Results	58

A.3	Letter Summarizing Results of Technical System Audit and
Performance Evaluation	80

B	Raw Test Data

B.l	Raw Concentration Printouts from Labtech Notebook	86

B.2 Raw Data - Ambient Conditions	116

B.3 Emission Concentration Summaries 	121

B.4	Turbine Process Data	134

C	Equipment Calibration Results

C.l	Calibration Gas Certifications	150

C.2 Calibration Results of Ambient Measurement Equipment 	158

XI

-------
1

2

3

4

5

iM

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

. 7

. 7

. 8

. 9

18

ige

. 4

. 6

11

12

13

15

20

21

23

23

24

24

24

25

26

Figures

Turbine exhaust sampling location 	

1.5-MW gas turbine	

Ambient conditions sampling location 	

Extractive sampling system	

Method 20 N0x/02 gas turbine emissions measurement flowchart

Tables

Verification Test Design (Target Values) 	

Summary of Measurements	

Reference Analyzers and Measurement Ranges 	

Gas Analyzers Interference Test Gas Concentrations 	

Calibration Gas Concentrations	

Method 20 Traverse Points 	

Operating Parameter Ranges 	

Pollutant Emission Concentrations for Xonon™ Verification Test 	

Reference Method QC Criteria 	

Method 205 Summary Data Verification of Mass Flow Controllers 1 and 2
Method 205 Summary Data Verification of Mass Flow Controllers 1 and 3

Analyzer Interference Results 	

Response Times (seconds)	

Method 20 Calibration Error and Drift Results	

NOx Analyzer Performance Evaluation Audit	

Xll

-------
Acronyms/Abbreviations

ADQ

Audit of data quality

ANSI

American National Standards Institute

APCT

Air Pollution Control Technology

ASME

American Society of Mechanical Engineers

CCSI

Catalytica Combustion Systems, Inc., renamed Catalytica Energy Systems, Inc

cfm

Cubic feet per minute

CO

Carbon monoxide

cv

Coefficient of variance

DQO

Data quality objective

EED

MRI's Environmental Engineering Division

EPA

Environmental Protection Agency

ETV

Environmental Technology Verification

fpm

Feet per minute

GVP

Generic Verification Protocol

HMI

Human/machine interface

IR

Infrared

ISO

International Standards Organization

MFC

Mass flow controller

MRI

Midwest Research Institute

NESHAP

National Emission Standard for Hazardous Air Pollutants

NIST

National Institute of Standards and Technology

NOx

Nitrogen oxides

OD

Outside diameter

PE

Performance evaluation

ppmv

Part per million by volume

ppmvd

Part per million by volume dry basis

ppmvw

Part per million by volume wet basis

QA

Quality assurance

QAO

Quality assurance officer

QC

Quality control

QMP

Quality management plan

QSM

Quality system manual

RH

Relative humidity

RTI

Research Triangle Institute

SOP

Standard operating procedure

ss

Stainless steel

TEI

Thermo Enviromental Instruments, Inc. (sometimes identified as TECO)

TSA

Technical systems audit

UHCs

Unburned hydrocarbons (same as total hydrocarbons)

Xlll

-------
Section 1.0

Introduction

The objective of the Air Pollution Control Technology (APTC) Environmental Technology
Verification (ETV) Program is to verify, with high data quality, the performance of air pollution
control technologies. A subset of air pollution control technologies is nitrogen oxides (NOx)
emission control technologies. One of these NOx emission control technologies is the flameless
combustion system known as Xonon™, developed by Catalytica Combustion Systems, Inc.
(CCSI) of Mountain View, California. The Xonon™ flameless combustion system is an
advanced combustion process designed for gas turbines that is capable of producing NOx
emissions below the current level of 9 to 25 parts per million by volume on a dry basis (ppmvd)
at 15 percent oxygen (02) obtainable with dry, low-NOx combustion techniques.

Control of NOx emissions is of increasing interest, particularly related to the National Ambient
Air Quality Standard for ozone. The Environmental Protection Agency (EPA) recently
completed a rulemaking to reduce more than 1 million tons of NOx each ozone season and
offered to develop and administer a multistate NOx trading program to assist the affected states.
Additionally, many state and local permitting agencies are requiring unprecedentedly low NOx
emission levels.

To evaluate the performance of the Xonon™ flameless combustion system, a field test program
was designed by Research Triangle Institute (RTI) and Midwest Research Institute (MRI) with
assistance from CCSI. A site visit to the host facility was completed on April 17, 2000, and a
test/QA plan was developed and approved by EPA on June 28, 2000. The verification field test
was conducted on July 18 and 19, 2000.

The host facility was the Silicon Valley Power Gianera generating station located at 4948
Centennial Drive in Santa Clara, California. The Xonon™ flameless combustion system was
installed on a 1,500-kW gas-turbine-generator set manufactured by Kawasaki (Model
M1A-13A).

The verification statement for the Xonon™ flameless combustion system verification test is
presented in the preceding section. A detailed description of the Xonon™ flameless combustion
system is presented in Section 2. The procedures and methods used for the verification test are
discussed in Section 3. The operating range over which the verification test was conducted is
presented in Section 4. The results of the verification test are summarized and discussed in
Section 5.

Appendices describing QA/QC activities and results (Appendix A), raw test data (Appendix B),
and equipment calibration results (Appendix C) are attached.

-------
Section 2.0

Description and Identification of Xonon™ Flameless Combustion System

The Xonon™ flameless combustion system is completely contained within the combustion
chamber of the gas turbine. The Xonon™ system completely combusts fuel to produce a high-
temperature gaseous mixture, typically over 1300 °C (2400°F). Dilution air is added to shape the
temperature profile required at the turbine inlet.

The Xonon™ combustor consists of four sections:

5. Preburner. The preburner is used for startup preheat of air before it enters the catalyst
module and acceleration of the turbine. The preburner could be a conventional, diffusion
flame burner or could be a dry, low-NOx type (lean, premixed) burner. For this Kawasaki
turbine, the preburner was a lean premix burner.

6. Fuel injection and fuel/air mixing system. This system injects the fuel and mixes it with
the main air flow to provide a very well-mixed, uniform fuel/air mixture to the catalyst.

7. Xonon™ catalyst module. In the catalyst module, a portion of the fuel is combusted
without a flame to produce a high-temperature gas.

8. Homogeneous combustion region. Located immediately downstream of the catalyst
module, the homogeneous combustion region is where the remainder of the fuel is
combusted, and carbon monoxide and unburned hydrocarbons are reduced to very low levels
(also a flameless combustion process).

The overall combustion process in the Xonon™ system is a partial combustion of fuel in the
catalyst module followed by complete combustion downstream of the catalyst in the burnout
zone. Partial combustion within the catalyst produces no NOx. Homogeneous combustion
downstream of the catalyst usually produces no NOx, because combustion occurs at a uniformly
low temperature. A small amount of fuel is combusted in the preburner to raise the compressed
air temperature to about 470°C (880°F). NOx in the turbine exhaust is usually from the
preburner.

-------
When a Xonon™ combustion system is installed, initial startup and shakedown are supervised by
CCSI personnel, and the requisite training to operate and service the equipment is provided at
that time. Maintenance procedures and spare parts requirements are identified during design of
the combustor for the specific turbine model, and this information is provided upon delivery of
the equipment. CCSI indicates the elapsed time between installation and commissioning to be
less than 1 month.

After initial commissioning, the Xonon™ combustion system is expected to require minimal
ongoing service. CCSI expects the catalyst module to have a useful life of approximately 8,000
operating hours, requiring a replacement of the module at this interval.

This verification report covers application of the Xonon™ flameless combustion system to small
gas turbines operated at full load when combusting natural gas within the stated operating
condition envelope. The same pilot unit was operated at the test site by the vendor, CCSI, for
over 4,000 hours before the verification test. Data from this long-term operating period have
been submitted to a number of regulatory authorities for their review and evaluation. While
these data and the instruments used were not verified during this test, within the operating
condition envelope the results are generally consistent with the verification test results. CCSI
should be contacted for these long-term data or other information.

CCSI Xonon™ Product Performance Expectations

CCSI expects that the Xonon™ flameless combustion technology incorporated in a Xonon™
combustion system for a natural-gas-fueled Kawasaki Ml A-13A gas turbine is capable of
achieving emissions of NOx of less than 2.5 ppmvd (corrected to 15 percent oxygen [02]) on a
1-hour rolling average basis, and less than 2.0 ppmvd (corrected to 15 percent 02) on a 3-hour
rolling average basis. Under the same conditions, this Xonon™ combustion system is also
expected to achieve carbon monoxide (CO) emissions of less than 6 ppmvd (corrected to
15 percent 02).

-------
Section 3.0

Procedures and Methods Used in Testing

A generic verification protocol (GVP) for testing NOx control technologies was prepared and
approved by the NOx Control Technology Technical Panel (RTI, 2000a). The GVP established
the guidelines for the verification test design, the data quality objective (DQO) for the primary
verification parameter (for this verification test, NOx concentration corrected to 15 percent 02),
and the test methods to be used. A test/QA Plan (RTI, 2000b) was written to apply the GVP to
the Xonon™ verification. This section details the test design and the test methods used for the
verification test of the Xonon™ flameless combustion system.

3.1 Test Design

The GVP for NOx Control Technologies provides extended discussions on the experimental
design approach for NOx control technologies verification testing. The specific design for this
test is described below.

Table 1. Verification Test
Design (Target Values)3

The critical measurement for the Xonon™ flameless combustion system verification was the
level of NOx emitted in ppmvd at 15 percent 02. This verification test was designed to measure
the outlet NOx emission concentration under targeted field test conditions with the Xonon™
flameless combustion system operating at a specified high load and the encountered low and high
ambient temperature for the test days. Historical ambient
temperature data suggested that its effect might be
detectable by conducting sets of tests at dawn (cold) and in
the afternoon (hot). Associated emissions concentrations
were also measured using EPA reference methods, but the
test was not designed around acquisition of these data.

Ambient temperature was an important measurement for
establishing the bounds of the verification test design.

A 2 x 1 factorial experimental design was used with each of
the parameters. Two replications of the factorial design (six
test runs in each replication) was used for a total of 12 test
runs. Table 1 gives the factorial design with the target
values for each parameter. As required by the DQO, the
product of this test design was the verified mean NOx
emission concentration(s) and the achieved 95 percent
confidence interval of the mean for the specified operating
range.

The factorial design allowed for statistical significance tests
to determine whether the outlet NOx concentration varied
significantly with ambient temperature. Further, since two
replicates were done, the significance of interactions
between ambient temperature and outlet NOx concentration

Test
Run

Ambient
Temperature
(time of day)

1

Low (dawn)

2

Low (dawn)

3

Low (dawn)

4

High (afternoon)

5

High (afternoon)

6

High (afternoon)

7

Low (dawn)

8

Low (dawn)

9

Low (dawn)

10

High (afternoon)

11

High (afternoon)

12

High (afternoon)

Turbine load >95% maximum.

4

-------
could also be tested. If the outlet NOx concentration did not change significantly with ambient
temperature, the results are valid for the range of ambient temperature covered by the test. If the
outlet NOx concentration did vary significantly with ambient temperature, the results need to
include information indicating the dependence of outlet NOx concentration on ambient
temperature. The results of the statistical significance tests are presented in Section 5.1.1.

Because the turbine was operated at constant full load (>97%) during the entire testing period,
the process was assumed to be at equilibrium during all testing.

3.2 Sampling Methods

Table 2 lists all the measurement parameters for this verification test. They are categorized in the
table as performance factors (e.g., direct emission measurements), associated impacts (e.g., CO
and UHC emissions), and test conditions that were documented. Included in Table 2 are the
factors to be verified, parameters to be measured for each factor, the measurement method for
each parameter, and explanatory comments. The facility contact provided data for process
condition parameters collected from the turbine human/machine interface (HMI) computer.
Measurement methods and procedures are described in Sections 3.2.2 through 3.2.5.

3.2.1 Sampling Locations

Sample locations were chosen so that they met the minimum specified sample location criteria of
the sample methods used or yielded a representative sample. The pollutant emission sampling
location, process operating condition measurement locations, and ambient conditions measure-
ment location are presented in Sections 3.2.1.1 through 3.2.1.3, respectively.

3.2.1.1 Pollutant Emission Sampling Location—

The NOx, CO, UHC, 02, and C02 concentrations were measured in the turbine exhaust stack (see
Figure 1). Two sets of sampling ports were available, but neither met Method 20 criteria. As
noted in the test/QA plan, the top set of sampling ports were judged as the most likely to yield a
representative sample; therefore, the top sampling ports were used.

3.2.1.2 Process Conditions Measurement Locations—

Several parameters related to the operating conditions of the gas turbine during the verification
test runs were recorded. These include electric power output, fuel flow rate, inlet temperature to
the compressor, compressor discharge pressure, compressor discharge temperature, temperature
into the catalyst, temperature out of the catalyst, and the exhaust gas temperature. The
measurement locations for process and turbine parameters are identified in Figure 2 and are in
relation to where the measurements are taken in the gas turbine.

-------
Table 2. Summary of Measurements

Factors to be
Verified

Parameter to be
Measured

Measurement Method

Comments

Performance factors

NOx emissions

Outlet NOx conc.,
ppmv

EPA Ref. Method 20
(40 CFR 60 App. A)

MRI provided and operated
analyzer

Associated impacts

CO emissions

Outlet CO conc.,
ppmv

EPA Ref. Method 10
(40 CFR 60 App. A)

MRI provided and operated
analyzer

UHC emissions

Outlet THC conc.,
ppmvw

EPA Ref. Method 25A
(40 CFR 60 App. A)

MRI provided and operated
analyzer

02/C02 emissions

Outlet 02/C02
conc.,%

EPA Ref. Method 20
(40 CFR 60 App. A)

MRI provided and operated
analyzer

Test conditions documentation

Percent of turbine's
rated capacity

Electrical power
turbine rating

Real power sensor

MRI collected data from
facility contact

Fuel type

—

—

Natural gas

Fuel flow

Fuel flow rate

Coriolis-type flowmeter

Facility contact provided data
from turbine HMI computer

Fuel sample results

Natural gas
composition

Chromatographic analysis

From fuel sample results
obtained from CCSI

Ambient conditions

Air temperature

Thermocouple or
Thermohygrometer following
EPA Quality Assurance
Handbook for Air Pollution
Measurement Systems,
Volume IV: Meteorological
Measurements

MRI conducted temperature,
pressure, and humidity
measurements concurrently

Air pressure

ASTM D3631-95: aneroid
barometer or equivalent

Air humidity

Thermohygrometer
equivalent to ASTM E337-
84(1996)e1

Compressor
parameters

Inlet temperature

Array of thermocouples on
turbine

Facility contact provided data
from turbine HMI computer

Discharge
temperature

Array of thermocouples on
turbine

Facility contact provided data
from turbine HMI computer

Discharge pressure

Pressure gauge

Facility contact provided data
from turbine HMI computer

Catalyst inlet
condition

Temperature at
catalyst inlet

Array of thermocouples on
turbine

Facility contact provided data
from turbine HMI computer

Catalyst outlet
condition

Temperature out of
the catalyst

Array of thermocouples on
turbine

Facility contact provided data
from turbine HMI computer

Catalyst hours of
operation

Hours of operation
since catalyst
installed

Clock counter

Information provided by CCSI
facility contact

Exhaust temperature

Exhaust gas
temperature

Array of thermocouples on
turbine

Facility contact provides data
from turbine HMI computer

Compressor/turbine
status

—

Pressure ratio compared to
rated value

Information provided by CCSI
facility contact

6

-------
Figure 1. Turbine exhaust sampling location.

Total
fuel
flow

Compressor
discharge
pressure

Figure 2. 1.5-MW gas turbine.

7

-------
3.2.1.3 Ambient Conditions Measurement Location—

Parameters related to the ambient conditions during the verification test runs include the ambient
air temperature, ambient air pressure, and ambient relative humidity. The measurement location
for the ambient conditions is shown in Figure 3. The temperature (T), pressure (P), and relative
humidity (RH) measurement devices were placed on the platform just below the gas turbine air
inlet filters. In this location, the measurements are representative of the inlet air conditions (as
recommended in Section 4.3.4 of EPA Quality Assurance Handbook for Air Pollution Measure-
ment Systems, Volume IV: Meteorological Measurements, Templeman, 1995). An aspirated
radiation shield was used to prevent biases caused by direct sunlight exposure.

Figure 3. Ambient conditions sampling location.

3.2.2 NOx, CO, UHC, and 02/C02 Sampling Procedures

Turbine exhaust gas was sampled for NOx, CO, UHC, and 02/C02 using EPA reference methods.
All sampling followed the requirements of the specific test method being used unless otherwise
stated in this document or approved by RTI before the verification test. The analytical systems
were calibrated before and after each 32-min test run following the procedures in each applicable
EPA Reference Method (40 CFR 60 App. A).

3.2.2.1 Sampling System—

A diagram of the extractive gaseous measurement system used for the testing is shown in
Figure 4. Two independent sampling systems were used, one for CO, 02, C02, and NOx and

-------
Figure 4. Extractive sampling system

9

-------
another for UHC. All analyzers, calibration gases, and the sampling manifold were housed in an
environmentally controlled trailer. The sampling system components were stainless steel (SS),
Teflon, or glass. These materials have been proven to be inert for the gases of interest.

The sampling system for measurement of CO, 02, C02, and NOx consisted of

Unheated stainless steel probe; 1.27 cm (0.5 in.) outside diameter (OD) (since the stack gas
temperature was ~ 510°C [950°F], the probe was not heated);

Heated (~121°C [250°F]) glass-fiber filter to remove particles with a diameter >1 |im;

•	Heated (~121°C [250°F]) Teflon sample line (~3 m [10 ft] long and 0.95 cm [0.38 in.] OD)
to transport the sample gas to the moisture removal condenser; temperature of the sample line
was regulated with a thermostatic heat controller;

Chiller condenser system submerged in an ice bath to condense and remove moisture in the
sample gas; the condenser is a two-pass system to condense moisture while minimizing the
liquid/air interface; a peristaltic pump was used to continually remove condensed water
vapor; the water vapor dewpoint after the chiller was estimated to be ~ 3.5°C (38°F);

Unheated Teflon sample line (~ 2.3 m [75 ft] long and 0.95 cm [0.38 in.] OD) to transport the
sample gas from the chiller (located on the scaffold platform near the sample ports) to the
sample manifold; just upstream of the sample extraction pump was a second glass-fiber filter;

Teflon-lined sample pump to extract sample gas from the stack; sampling rate was
-10 L/min; and

Individual rotameters regulated the sample flow to each analyzer and excess sample gas was
dumped through the bypass.

The sampling system for measurement of UHCs consisted of

Unheated SS probe; 1.27 cm (0.5 in.) OD;

Heated (~121°C [250°F]) glass-fiber filter to remove particles with a diameter >1 |im;

•	Heated (~121°C [250°F]) Teflon sample line (-23 m [75 ft] long and 0.63 cm [0.25 in.] OD)
to transport the sample gas directly to the hydrocarbon analyzer; temperature of the sample
line was regulated with a thermostatic heat controller; and

Sample gas was extracted by a heated pump contained within the hydrocarbon analyzer.

The sampling system was calibrated by directing each calibration gas to the probe through an
unheated Teflon tube. The probe was "flooded" with calibration gas, and the sample pump
pulled as much of the calibration gas as needed to the system manifold. Excess calibration gas
was dumped out the probe. This process of calibrating the system does not pressurize the
sampling system and mask any leaks (see Section 3.2.3.5.2 for description of CO analyzer
calibration).

Calibration gases were generated from a single, high-concentration EPA protocol gas with an
Environics Model 2020 gas dilution system. The Environics system consists of four electronic

10

-------
mass flow controllers (MFCs). MFC 1 was used for the nitrogen dilution gas. MFC 2 (0 to
10 L/min) and MFC 3 (0 to 1 L/min) are used in combination with MFC 1 to generate the
specified calibration gas concentration by diluting a high concentration standard gas. MFC 4
(0 to 0.1 L/min) was not used. The Environics system was calibrated at the factory on July 11,
2000. Also, the calibration of the combined MFCs that were used for this test (e.g., 1+2 and
1+3) was checked in accordance with EPA Method 205 the day before the field test began. The
Method 205 data are summarized in Section 5.

3.2.2.2 Reference Analyzers—

The reference analyzers used for quantifying the gaseous concentrations are listed in Table 3.
The table also includes a description of the analyzer and the measurement ranges used for this
test. Measured pollutant concentrations were extremely low relative to the measurement ranges.
Most notably, the UHC concentrations were about 0.1 to 0.2 part per million by volume on a wet
basis (ppmvw) as measured on a 0- to 100-ppmvw range. Method 25A specifies a measurement
range of 1.5 times the expected concentration, which is unfeasible at extremely low
concentrations.

Table 3. Reference Analyzers and Measurement Ranges

Pollutant

Reference Analyzer

Measurement
Range

Description

NOx

Thermo Environmental
Instruments (TEI) 42H

0-20 ppmv

Uses the principle of chemiluminescence to
measure the concentration of NOx in the sample
stream. The instrument uses a heated can NOz
converter.

Thermo Environmental
Instruments (TEI) 48

0-50 ppmv

Uses the principle of gas filter correlation and non-
dispersive infrared (GFC-NDIR) to measure the
concentration of CO in the sample stream.

UHC

J.U.M VE 7

0-100 ppmvw

Uses the principle of flame ionization detection
(FID) to measure the concentration of hydrocarbons
in the sample stream.

o2/co2

Servomex 1440C

0-25% / 0-20%

The Oz detector uses the principle of paramag-
netics, and the COz detector uses a single- beam,

dual-wavelength IR technique.

3.2.3 Sampling Methods Requirements

Each of the sampling methods has different criteria to ensure the quality of the sample and the
data collected. Each of these requirements is presented in the following sections.

3.2.3.1 Analyzer Interference Test—

An initial interference check was completed on the NOx, CO, 02, and C02 analyzers before their
first use. For the interference test, the gases listed in Table 4 were injected into each analyzer.
For acceptable analyzer performance, the sum of the interference responses to all of the
interference gases must be <2 percent of the analyzer span value. The interference test results
are presented in Section 5.

-------
3.2.3.2 N02 Converter Efficiency
Test—

The N02 converter efficiency is tested
as part of routine analyzer QC Method
20. The test relies on the oxidation
reaction of NO in the presence of
oxygen. NO reacts to form N02 in
equilibrium with NO. For the test, a
clean, leak-free Tedlar bag was filled
half full with the mid-level NO
calibration gas. The bag was then
filled with 20.9 percent 02 gas. The
bag was attached directly to the NOx
analyzer sample inlet. After
approximately a 2-min stabilization
period, 30 1-min average NOx
analyzer readings were recorded. For
an acceptable converter, the 1-min average response at the end of 30 min is required to not
decrease more than 2 percent of the highest peak 1-min value. That is, the analyzer should be
capable of converting all the NO to N02. The results of the N02 converter efficiency check are
presented in Section 5.

3.2.3.3 Response Time Test—

3.2.3.3.1 Method 20 Response Time - NOx and 02/C02. To determine the response
time according to Method 20 procedures, the zero gas (i.e., N2) was injected into the sampling
system at the probe. When the analyzer's readings were stable, the zero gas was turned off so the
effluent could be sampled. When a stable reading was obtained, the upscale response time was
determined as the time required for the computer readout to record a 95 percent step change from
the zero reading to the stable effluent concentration. Then the high-level calibration gas for each
analyzer was injected into the sampling system at the probe. When the analyzer's readings were
stable, the high-level gas was turned off so that the effluent could be sampled. When a stable
reading was obtained, the downscale response time was determined as the time required for the
computer readout to record a 95 percent step change from the calibration gas reading to the stable
effluent concentration. This procedure was repeated until three upscale and three downscale
response times were completed. The longest of all the upscale and downscale response times
was reported as the system response time for that analyzer. For Method 20, the response time
must be 30 s or less. The response times are presented in Section 5, Table 13.

3.2.3.3.2 Method 25A Response Time - UHC. For EPA Method 25A, only an upscale
response time test is required. To determine the upscale response time, the zero gas was injected
into the sampling system at the probe. Then, the high-level calibration gas was injected into the
sampling system. The upscale response time was determined as the time required for the
computer readout to reach 95 percent of the high-level calibration gas reading. This procedure

Table 4. Gas Analyzers Interference Test
Gas Concentrations

CO 55 CO^ oT

NOx Analyzer Interference Gases

498 ppmv 201 ppmv 9.98% 20.9%

CO Analyzer Interference Gases
NA NA 9.98% NA

02 Analyzer Interference Gases
498 ppmv 197 ppmv 9.98% NA

COz Analyzer Interference Gases

498 ppmv 197 ppmv NA 20.9%

NA = Not applicable.

-------
was repeated three times, and the average was reported as the response time. The response time
is presented in Section 5, Table 13.

3.2.3.4 Preliminary 02 Traverse—

Method 20 requires a preliminary 02 traverse to be conducted at multiple sample points across
the stack's cross-sectional area. The preliminary 02 traverse determines the eight lowest 02
concentration sampling points from an array of multiple points. These eight low 02 points are
used as the traverse points for the individual test runs. However, since this stack had a cross-
sectional area of 0.66 m2 (7.1 ft2), only eight traverse points would be used for the preliminary 02
traverse. Therefore, a preliminary 02 traverse was not necessary and was not done, and eight
traverse points for the test runs were selected in accordance with EPA Method 1.

3.2.3.5 Calibrations—

Table 5 lists the calibration gas concentrations used for the reference method testing. EPA
protocol gas was used to calibrate the analyzers. Each of the reference methods has different
calibration procedures. The individual method calibration procedures are described in
Sections 3.2.3.5.1 through 3.2.3.5.3. The gaseous pollutant measurement system was calibrated
before and after each test run. Also, no test run started more than 2 hours after a pretest
calibration, and all post-test calibrations were completed within 1 hour of the end of a test run.

Table 5. Calibration Gas Concentrations

Calibration point

o
o

NOx

UHC

Zero

Pure N2

Low-level

5.02 ppmv

15.0 ppmv

29.9 ppmv

Mid-level

11.99%

3.01%

10.03 ppmv

29.9 ppmv

49.9 ppmv

High-level

20.9%

9.98%

17.04 ppmv

44.9 ppmv

84.9 ppmv

3.2.3.5.1 Method 20 Calibration Procedures. The NOx calibration gas was
201.85 ppmv NO in a balance of N2. The 02 calibration gas was 38.4 percent 02 in a balance of
N2. The C02 calibration gas was 40.05 percent C02 in a balance of N2. Copies of the calibration
gas certifications are attached in Appendix A. As noted earlier, a gas dilution system was used to
make the targeted gas concentration levels shown in Table 5 from the single, high-concentration
EPA protocol gas.

For calibration error checks of both the NOx and diluent analyzers, the zero gas and mid-level gas
were introduced separately into the sampling system at the probe. Each analyzer's response was
adjusted to the appropriate level. Then the remainder of the calibration gases were introduced
into the sampling system, one at a time. The acceptable response of the analyzer to each calibra-
tion gas must be within ±2 percent of span.

At the conclusion of a test run, the zero and mid-level calibration gases for each analyzer were
introduced separately into the sampling system. Both the zero drift and calibration drift,

-------
calculated in accordance with Equation 1, must be within ±2 percent of span. If a drift was
greater than 2 percent of span, the test run would have been considered invalid and the measure-
ment system would have been repaired to satisfy drift tolerances before additional test runs were
conducted. Method 20 calibration results are summarized in Section 5. Individual pre- and post-
test run calibrations are presented in Appendix B.

Percent drift = (Final response - Initial response) / Span value x 100 (1)

3.2.3.5.2 Method 10 Calibration Procedures. The CO calibration gas was 199.8 ppmv
CO in a balance of N2. The calibration gas certification is shown in Appendix A. The gas
dilution system was used to make the targeted gas concentration levels from the single, high-
concentration EPA protocol gas.

CO analyzer calibration error checks were conducted before the start of each day's testing. The
calibration error check was conducted (after final calibration adjustments were made) by
separately injecting each of the four calibration gases (zero, low-, mid-, and high-level) directly
into the analyzer and recording the response. If the calibration error was greater than 2 percent,
the analyzer would have been repaired or replaced and recalibrated to an acceptable calibration
error limit before proceeding.

Zero and upscale sampling system calibration checks were performed both before and after each
test run to quantify the reference measurement system calibration drift and the sampling system
bias. Upscale calibration checks were performed using the mid-level gas. During these checks,
the calibration gases were introduced into the sampling system at the probe so that they were
sampled and analyzed in the same manner as the sample gas. Drift means the difference between
the pre- and post-test run system calibration check responses. Sampling system bias means the
difference between the system calibration check response and the initial calibration error
response (direct analyzer calibration) at the zero and upscale calibration gas levels. Method 10
calibration results are summarized in Section 5. Individual pre- and post-test run calibrations are
presented in Appendix B.

3.2.3.5.3 Method 25A Calibration Procedures. The UHC calibration gas was
190.6 ppmv propane in a balance of nitrogen. Copies of the calibration gas certification are
located in Appendix B. The gas dilution system was used to make the targeted gas concentration
levels shown in Table 5 from the single, high-concentration EPA protocol gas.

For calibration error checks, the zero gas and high-level gas were introduced separately into the
sampling system at the probe. The UHC analyzer's response was adjusted to the appropriate
level. Then the low- and mid-level calibration gases were introduced into the sampling system,
one at a time. The acceptable response of the analyzer to each calibration gas must be within
±5 percent of the calibration gas value.

-------
At the conclusion of a test run, the zero and mid-level calibration gases were introduced
separately into the sampling system. Both the zero drift and calibration drift, calculated in
accordance with Equation 1, must be within ±3 percent of span. If a drift was greater than
3 percent of span, the test run would have been considered invalid, and the measurement system
would have been repaired before additional test runs were conducted. Method 25 A calibration
results are summarized in Section 5. Individual pre- and post-test run calibrations are presented
in Appendix B.

3.2.3.6 C02 Trap—

Method 10 requires that C02 be removed from the sample gas that is sent to the CO analyzer.
The C02 is removed because the commonly used, nondispersive infrared technology instrument
for measurement of CO exhibits an interference from C02. However, the TEI Model 48
incorporates the technique of gas filter correlation to eliminate the C02 interference from the
measurement of CO. Since the TEI Model 48 does not have a C02 interference (see the
interference test results in Section 5), the C02 trap was not used.

3.2.3.7 Sample Location by Method 20 and Traverse Point Selection by Method 1—

Two sets of sampling ports were available on the
turbine exhaust stack. One set was located
immediately after a long 90° horizontal-to-
vertical upward bend in the stack. The second
set was located approximately 4.6 m (5 duct
diameters) downstream of the 90° horizontal-to-
vertical upward bend and 0.5 m (0.5 duct dia-
meters) upstream of the stack exit. Neither of
these port locations is ideal; however, the top
ports were used (see Figure 1). Only one of the
top ports was used for the Method 20 traverse
because the scaffold was only set up on one side
of the circular stack. Therefore, MRI did not
have safe access to the second port for the
Method 20 traverse. Because the gas concentration was not stratified across the one available
diagonal traverse, all parties agreed that double traversing across the single port was acceptable.
Table 6 shows the point locations.

3.2.4 Process Data Collection

Process data were collected from the turbine control's HMI computer to document the test
conditions. The CCSI facility contact provided the data from the HMI computer. Table 2
identifies the parameters that were measured and the party responsible. The test condition
documentation parameters taken from the HMI computer were retrieved at 1-min intervals for
each test run. Process data, at 1-min intervals for each test run, are presented in Appendix C.
The process data measurements are summarized in Sections 3.2.4.1 through 3.2.4.6.

Table 6. Method 20 Traverse Points

Distance

Percent of from

Stack Stack Wall

Point Diameter (cm)

1 6.7 (0.9)

2 25.0 (3.5)

3 75.0 (10.6)

4 93.3 (13.2)

-------
3.2.4.1 Electrical Power Generation by Turbine —

To determine the operating rate of the turbine during the verification test, the electrical power
production from the electrical generator was recorded. This measurement was taken with a Real
Power Sensor that determines the electrical power supplied at the generator terminals. At this
writing, CCSI is not aware of any calibration of this device since commissioning of the site in
October 1998. The output has been noted by CCSI to be consistent with the City of Santa Clara
meter on several occasions.

3.2.4.2 Fuel Flow Rate—

The fuel flow rate into the combustion system was measured with a Coriolis-mass flowmeter.
The flowmeter was calibrated for natural gas at the factory and was recalibrated on June 28,
2000. (The flowmeter is periodically compared to the City of Santa Clara's main turbine
flowmeter.)

3.2.4.3 Compressor Inlet Temperature—

Compressor inlet temperature (also referred to as "ambient temperature" by the facility) was
measured with two 1/8-in. diameter sheathed K-type thermocouples located in the inlet air duct.
These devices are calibrated on a semiannual basis using a calibrated thermowell device.

3.2.4.4 Compressor Discharge Pressure—

Compressor discharge pressure was measured using two pressure taps and two absolute pressure
transducers. The transducers were originally calibrated at the factory and are periodically re-
calibrated by CCSI personnel using specially maintained and calibrated pressure-sensing devices.
The absolute pressure transducers were last calibrated in March 2000.

3.2.4.5 Catalyst Inlet/Catalyst Outlet Temperatures—

The air temperature just upstream of the catalyst and the gas temperature just downstream of the
catalyst were measured by separate thermocouple arrays. The catalyst outlet temperature was
measured with a series of four to eight thermocouples installed at the exit from the catalyst bed.
The thermocouples are calibrated by CCSI personnel whenever the thermocouple hardware is
changed.

3.2.4.6 Turbine Exhaust Temperature—

The turbine outlet temperature was measured by four 1/8-in. diameter sheathed K-type
thermocouples installed at the exit of the turbine, just upstream of the stack's silencer. These
thermocouples were factory calibrated, were recalibrated by CCSI personnel upon receipt, and
were recalibrated upon installation in the spring of 2000.

-------
3.2.5 Ambient Conditions Sampling

Three ambient air conditions were measured three times during each test run: temperature,
pressure, and relative humidity. Temperature and humidity were measured using an equivalent
technique to ASTM E337-84(1996)el. ASTM E337-84(1996)el uses an aspirated wet-bulb and
dry-bulb device to determine relative humidity, but MRI used a thermohygrometer to obtain the
relative humidity and ambient temperature. Pressure was measured using the ASTM D3631-95
method. Ambient pressure was measured with an aneroid barometer. The thermohygrometer
and aneroid barometer were placed in a mechanically aspirated, grey steel box. The accuracy of
the thermohygrometer measurements are ±3 percent for relative humidity and ±0.7°F for ambient
temperature based on the manufacturer's performance specifications. The relative humidity is
detected using the principle of changes in the capacitance of the sensor as its thin polymer film
absorbs water molecules. Temperature is measured with a negative temperature coefficient
thermistor. Results of the ambient conditions measurements are shown in Appendix C.

The equipment used to make the ambient conditions measurements is carefully maintained by
MRI's Field Measurements section. The instrumentation was calibrated according to MRI
standard operating procedures (SOPs): MRI-0721 - Calibration of Thermocouple Probes,
Thermocouple Indicators and Digital Thermometers, MRI-0722 - Calibration of Pressure
Gauges, and MRI-0729 - Qualification and Calibration of Hygrometers at MRI's laboratory
before being transported to the field measurement site. Results of those calibrations are
presented in Appendix C.

3.3 Data Acquisition and Data Management

Data to document the process operating conditions of the turbine and Xonon™ system were
recorded by the turbine's HMI computer. These data were provided by the facility contact to
MRI's Field Team Leader in electronic format after each three-run test series. Process data are
shown in Appendix B. Data to document the ambient conditions were recorded manually on the
sheets shown in Appendix B. A Labtech Notebook was used to record the concentration signals
from the individual analyzers. The Labtech Notebook recorded the analyzer output at 1-s
intervals and averaged those signals into 1-min averages. At the conclusion of a test run, the pre-
and post-test calibration results were manually transcribed into a Microsoft Excel spreadsheet to
calculate drift and system bias. After a series of test runs, the test run values were electronically
transferred from the Labtech Notebook into a Microsoft Excel spreadsheet for data calculations
and averaging. The calculations done by Microsoft Excel used the default rounding convention.
The raw data printouts from the Labtech Notebook and the test run averages are shown in
Appendix B.

For Method 20, the first 1-min average, after moving to a new traverse point, is typically
disregarded as not representing the concentration at that traverse point. However, for this test
program, since gaseous stratification was not present and test runs were 32 min in length, all data
were used in the NOx and 02 concentration averages.

-------
For this test program, the data measurement and collection activities for the Method 20
measurements shown in Figure 5 were used. This flow chart includes all data activities from the
initial pretest QA steps to the passing of the data to the Task Leader. These steps were followed
in the field. Data for other methods used during this verification test were collected and handled
in the same manner as the Method 20 data.

Figure 5. Method 20 N0x/02 gas turbine emissions measurement

flowchart.

18

-------
Section 4.0

Statement of Operating Range of Test

For this verification test of the Xonon™ flameless combustion system, the CCSI representative
indicated that the emissions performance of the technology was guaranteed to be less than
2.5 ppmvd NOx at 15 percent 02 and to be less than 6 ppmvd CO at 15 percent 02. Without the
air management system, this emission guarantee was valid only at full turbine load conditions.
Therefore, the verification test was done at full turbine load.

During consultation with the CCSI representative, the only parameter identified that could
possibly have an effect on emissions was ambient temperature. In general, lower ambient
temperatures result in slightly higher NOx emissions for Xonon™ -equipped turbines because
slightly more fuel must be used in the pre-burner to achieve the desired inlet temperature to the
catalyst. To evaluate the effect of ambient temperature on NOx emissions, the verification test
was conducted during and after sunrise (to achieve the lowest ambient temperature of the day)
and during the afternoon (to achieve the highest ambient temperature of the day). The ambient
temperature range experienced during the 12 test runs was from 15.1 to 25.3°C (58.8 to 77.2°F).

Data to document the process operating conditions of the turbine and Xonon™ flameless
combustion system were recorded by the turbine's HMI computer. The operating conditions
during the 12 test runs are presented in Table 7. The bottom two rows of Table 7 show the
minimum and maximum values for each parameter. These minimum and maximum values form
the operating range over which this verification test was conducted.

The natural gas collected during the test showed that the fuel had a dry gas higher heating value
of 3.778 x 107 gross J/m3 (1012.9 Btu/ft3). The gas analysis is attached in Appendix B.

-------
Table 7. Operating Parameter Ranges

Run

Ambient
Temp.

Turbine Load

Fuel Flow
Rate

Compressor
Inlet Temp.

Compressor
Discharge
Pressure

Compressor
Discharge
Temp.

Temp, at
Catalyst Inlet

Temp. Out
of Catalyst

Exhaust Gas
Temp.

°C

°F

(MW)

(%)a

kg/h

Ib/h

°C

°F

kPa

psig

°C

°F

°C

°F

°C

°F

°C

°F

1

15.2

59

1.39

97.9

428

944

17

62

910

132

356

672

480

895

847

1557

524

980

2

16.3

61

1.38

98.1

425

937

17

63

903

131

356

673

480

895

847

1557

527

981

3

17.4

63

1.37

97.9

423

932

19

65

896

130

357

675

480

896

848

1558

529

983

4

25.2

77

1.25

98.9

406

894

26

78

869

126

364

686

484

903

850

1562

535

994

5

24.1

75

1.27

98.6

405

893

25

77

869

126

364

686

484

903

851

1564

534

993

6

21.3

70

1.29

98.4

411

907

22

72

876

127

361

682

481

898

849

1560

532

989

7

14.6

58

1.36

98.2

428

944

16

60

910

132

355

670

480

896

846

1555

526

979

8

16.3

61

1.35

98.3

425

938

17

63

910

132

356

673

480

896

847

1556

527

981

9

17.4

63

1.33

98.2

423

932

19

65

896

130

357

674

480

896

847

1556

529

983

10

20.7

69

1.29

98.4

412

908

22

72

882

128

361

681

481

898

846

1555

531

988

11

21.9

71

1.25

99.0

408

899

25

76

876

127

362

684

481

898

847

1556

533

991

12

23.5

74

1.23

98.4

407

898

25

76

869

126

362

684

482

900

849

1559

533

991

Minimum

15.2

59

1.23

97.9

405

893

16

60

869

126

355

670

480

895

846

1555

526

979

Maximum

25.2

77

1.39

99.0

425

944

26

78

910

132

364

686

484

903

851

1564

535

994

aNote: Turbine load (%) is the percent of turbine capability at the prevailing ambient conditions.

-------
Section 5.0

Summary and Discussion of Results

A verification test of the Xonon™ flameless combustion system was conducted on July 18
and 19, 2000, in Santa Clara, California. The purpose of the verification test was to evaluate the
NOx emission performance for the Xonon™ flameless combustion system as claimed by CCSI.
The test was conducted according to a test/QA plan that was approved by EPA on June 28, 2000.

The results of the verification test are summarized in Section 5.1. An important part of the
verification test was the extensive QA applied to this field test. The results of all the QA and
quality control (QC) checks performed during this verification test are summarized in
Section 5.2. A few minor deviations from the test plan were encountered, and those are
discussed in Section 5.3.

5.1 Results Supporting Verification Statement

The pollutant emission concentrations measured for the 12 test runs are presented in Table 8. As
can be seen, the NOx emission concentration was below the 2.5 ppmvd at 15 percent 02
performance claim offered by CCSI. Also, the CO emission concentration is well below
6 ppmvd at 15 percent 02. In addition, the unburned hydrocarbons concentrations were very low
and virtually undetectable during the 12 test runs.

Table 8. Pollutant Emission Concentrations for Xonon™ Verification Test

Run

Ambient
Temp. (°F)

NOx

(ppmvd @ 15% Oz)

UHC (as propane)
(ppmvw)

1.15

1.19

0.17

1.14

1.71

0.16

1.08

1.50

0.17

1.06

1.10

0.15

1.11

1.03

0.17

1.13

1.22

0.15

1.22

1.10

0.18

1.17

1.02

0.13

1.13

1.19

0.20

1.14

1.91

0.12

1.12

1.88

0.18

1.13

1.46

0.19

5.1.1 Statistical Analysis of Variance

This section describes the statistical analysis of the verification test data. As discussed in
Section 3.1, detection of ambient temperature effects required wide swings in daily temperature,

-------
which did not occur during the test period. The measured values from the verification test are
compared to the performance capability range specified by CCSI. The first step in the statistical
analysis was to perform the analysis of variance of NOx concentration on ambient temperature.
This step determines if ambient temperature has a significant effect on NOx emissions at the
95 percent confidence level.

The analysis of variance produced a P-value of 0.1647. Only when the P-value is less than 0.05
would the ambient temperature have a significant effect on NOx at the 95 percent confidence
level. Therefore, the turbine's NOx emissions were not affected by ambient temperature over the
range of 58°F to 77°F.

5.1.2 Variability of NOx Emissions

Because NOx emissions were not a function of ambient temperature, the 95 percent confidence
interval was calculated for the entire 12-run data set. The 95 percent confidence interval was
found to be ±0.026 ppmvd at 15 percent 02. Therefore, the NOx emission concentration for this
verification test can be stated as follows:

1.13 ± 0.026 ppmvd at 15 percent 02 at the 95 percent confidence level.

5.2 Discussion of QA/QC and QA Statement

Extensive QA/QC was applied to this verification test, much more than is typically applied to an
emissions test. The following QA and QC activities were part of this test:

A DQO for the NOx concentration measurement,

Reference method QC checks,

A technical system audit to evaluate all components of the data gathering and data
management system,

A performance evaluation sample to check the operation of the NOx measurement system,
and

A data audit of 30 percent of the critical measurement (NOx concentration) and 10 percent of
the noncritical measurement.

The results of each of these QA and QC checks are presented in Sections 5.2.1 through 5.2.3.

5.2.1 NOx Measurement DQO

The DQO for the NOx emission concentration measurement was stated in the test/QA plan as
follows:

For the NOx emission concentration measurements, the overall NOx emission
must be within ±10 percent of the mean emission concentration above 5 ppmvd,
±25 percent below 5 and above 2 ppmvd, and ±50 percent below 2 ppmvd.

-------
The DQO was computed as the half-width of the 95 percent confidence interval of the mean
divided by the mean. Since ambient temperature was not significant, all 12 test runs were
included in the DQO assessment.

As presented in Section 5.1.2, the half-width of the 95 percent confidence interval was
0.026 ppmvd at 15 percent 02 and the mean NOx concentration was 1.13 ppmvd at 15 percent 02.
Therefore, the DQO for NOx equates to 2.3 percent, well within the DQO limit of 50 percent
below 2 ppmvd.

5.2.2 Reference Method QC

The reference methods
used to measure emission
concentrations of NOx,
02/C02, CO, and UHCs
have specific QC criteria
that must be met. The QC
criteria ensure the accuracy
and stability of the
measurement system and
are summarized in Table 9.
The results of the reference
method QC checks are
summarized in
Sections 5.2.2.1 through
5.2.2.7. The raw data for
the QC checks are in
Appendix A.

Table 9. Reference Method QC Criteria

Method

Check

Criteria

Method 205

Dilution error

± 2% of reference
value

Method 20

Interference

N02 converter efficiency
Response time
Calibration error
Drift

< 2% of span
98%
< 30 s
± 2% of span
± 2% of span

Method 10

Calibration error
System bias
Drift

± 2% of span
± 5% of span
± 3% of span

Method 25A

Calibration error
Drift

± 5% of gas value
± 3% of span

Method 1

Traverse point

± 1 inch

5.2.2.1 Method 205 Dilution System Verification—

A gas dilution system was used to Table 10. Method 205 Summary Data

generate the targeted calibration gas Verification of Mass Flow Controllers 1 and 2

concentrations from single, high-
concentration EPA protocol gases
specific to each analyzer. This
dilution system must be verified in
the field before each test program
according to EPA Method 205
procedures. The dilution system
verification was done with the NOx
analyzer on a 0- to 50-ppmv
measurement range. The results of

the verification of MFCs 1 and 2 and 1 and 3 are presented in Tables 10 and 11, respectively.
For acceptable performance, the three-injection average at the low and high dilution points and

Standard
Calibration
Points

Reference
Value
Concentration
(ppmv)

Average
Analyzer
Reading
(ppmv)

Error

(%)

Low dilution

24.90

24.64

1.03

Mid-level supply

25.59

25.46

0.52

Upper dilution

44.90

45.08

-0.40

-------
the mid-level supply gas must be within
±2 percent of the reference value. As
indicated in Tables 10 and 11, all dilution
points were within the required ±2 percent.

5.2.2.2 Interference Test—

Before an analyzer is used, it must be
demonstrated that other gases in the
effluent do not interfere with the measure-
ment technique. This test was done for
the NOx, CO, 02, and C02 analyzers as
required by the reference method. For acceptable performance the total interference from all the
gases injected must be ±2 percent or less. The interference results are presented in Table 12.
Those results show that none of the analyzers exhibited unacceptable interference.

5.2.2.3 N02 Converter Efficiency Test—

Before each test program, the NOx analyzer must demonstrate
that the N02 converter is at least 98 percent efficient. The
performance criteria state that, during the 30-min N02
converter efficiency test, the last NOx analyzer reading must
not decrease by more than 2 percent from the highest reading.

The N02 converter showed a 0.2 percent decrease (5.02 ppmv
was the highest reading and 5.01 ppmv was the last reading),
well within the criteria for an acceptable converter. During
the entire N02 converter efficiency test, the readings ranged
from 4.97 to 5.02 ppmv.

5.2.2.4 Response Time Test—

A response time test was done for NOx, 02, C02,
and UHCs. Method 20 requires a response time of
30 s or less. The results of the response time tests
are summarized in Table 13.

5.2.2.5 Method 20 Calibrations—

For Method 20, the two calibration criteria are calibration error ( ±2 percent of span) and drift
(±2 percent of span). The largest calibration error and drift for the NOx, 02, and C02 analyzers
are presented in Table 14. See Appendix A, Pre- and Post-test Calibration Results. As shown in
Table 14, all calibration criteria were met.

Table 11. Method 205 Summary Data
Verification of Mass Flow Controllers 1 and 3

Standard
Calibration
Points

Reference
Value
Concentration
(ppmv)

Average
Analyzer
Reading
(ppmv)

Error
(%)

Low dilution

25.07

24.64

1.73

Mid-level supply

25.59

25.29

1.18

Upper dilution

45.10

44.87

0.51

Table 12. Analyzer
Interference Results

Interference

Analyzer

(% span)

NOx

-0.25

-1.80

0.80

o
o

1.50

Table 13. Response Times (seconds)

NOx

o
o

UHCs

-------
5.2.2.6 Method 10 Calibrations—

For Method 10 as performed for this test, the three
calibration criteria are calibration error (±2 percent
of span), system bias (±5 percent of span), and
drift (±3 percent of span). The largest absolute
calibration error was 0.46 percent, the largest
system bias was -1.28 percent, and the largest drift
was -0.44 percent. See Appendix A, Pre- and
Post-Test Calibration Results. All calibration
criteria were met.

5.2.2.7 Method 25A Calibrations—

For Method 25 A, the two calibration criteria are calibration error (±5 percent of the gas value)
and drift (±3 percent of span). The largest calibration error was -0.39 percent and the largest drift
was -0.42 percent. See Appendix A, Pre- and Post-test Calibration Results. All calibration
criteria were met.

5.2.3 Audits

Independent systematic checks to determine the quality of the data were performed throughout
this project. These checks consisted of a technical system audit, a performance evaluation audit,
and a data audit as described in Sections 5.2.3.1 through 5.2.3.3. The combination of these three
audits and the evaluation of the method's QC data allowed the assessment of the overall quality
of the data for this project. MRI's Task Leader managed the collection of and reviewed the field
data as detailed in Sections B10.1, Cl.l, and CI.2 of the test/QA plan.

5.2.3.1 Technical System Audit—

The technical system audit (TSA) was conducted by Robert Wright, RTI Quality Manager, and
Michael Tufts of ARCADIS Geraghty and Miller, an EPA contractor. This audit evaluated all
components of the data gathering and management system to determine if these systems had been
properly designed to meet the QA objectives for this study. The TSA included a careful review
of the experimental design, the test plan, and procedures. This review included personnel
qualifications, adequacy and safety of the facilities and equipment, standard operating procedures
(SOPs), and the data management system.

The TSA began with the review of study requirements, procedures, and experimental design to
ensure that they met the data quality objectives for the study. During the system audit, the Task
QA Officer inspected the analytical activities and determined their adherence to the SOPs and the
test/QA plan.

The draft summary of Wright's TSA is provided in Appendix A. In general, the TSA found that
the test program, as conducted, met all the data quality objectives for the study.

Table 14. Method 20 Calibration
Error and Drift Results

Largest
Absolute
Calibration
Error (%)

Largest
Drift (%)

NOx

-0.55

1.15

0.46

0.32

o
o

0.43

1.20

-------
5.2.3.2 Performance Evaluation Audit—

A performance evaluation (PE) audit was conducted by
Mike Tufts for EPA. For the PE audit, a performance
evaluation sample (PES) was supplied to check the
operation of the NOx analytical system. The PES was
measured for 6 continuous minutes on two occasions
for a total of 12 measurements. The NOx measurement
systems read the 1.00 ppmv NOx PES as 0.991 ± 0.012
ppmv at the 95 percent confidence level. A summary
of the performance evaluation audit is presented in
Table 15.

The method performance also was assessed using the
method QC samples described in Sections 5.2.2.1
through 5.2.2.7.

5.2.3.3 Data Audit—

The data audit, an important component of a total
system audit, was completed to determine if systematic
errors were introduced. The data audit was performed
by Jack Balsinger, the MRI task QA officer, by
randomly selecting approximately 30 percent of the
NOx data and 10 percent of the remaining data and
following them through the calculations. The scope of
the data audit was to verify that the data-handling
system was correct and to assess the quality of the data generated. The data review and data
audit were conducted in accordance with MRI standard procedures.

In addition to the data audit, a data check was performed by James Surman of MRI. The data
check was conducted to find errors in transposing data from the raw data printouts to the
calculation sheets in the Microsoft Excel spreadsheets. Data were reviewed for completeness,
and the method QC results were checked for acceptability. The Microsoft Excel spreadsheets
were checked for accuracy relative to the reference method requirements, and simulated data
were used to check the accuracy of the computations. Three minor errors were found and
corrected. Two errors were typographical, and one error was a spreadsheet format error.

5.3 Deviations from Test Plan

One deviation from the test plan was experienced during the field test, and one corrective action
was taken.

The test/QA plan indicated that an eight-point traverse—four points on one diagonal traverse and
four points on another diagonal traverse—was to be done during the Method 20 sampling.

Table 15. NOx Analyzer
Performance Evaluation Audit

Time

NOx System
Readings (ppmv)