Texas Commission on Environmental Quality Comprehensive Flare Study Project PGA No FY09-04 Tracking No

Transcription

1 August 2010 Revision 0 Texas Commission on Environmental Quality Comprehensive Flare Study Project PGA No FY09-04 Tracking No Quality Assurance Project Plan The University of Texas at Austin Center for Energy and Environmental Resources

2 Texas Commission on Environmental Quality Comprehensive Flare Study Project PGA No FY09-04 Tracking No Quality Assurance Project Plan Prepared by The University of Texas at Austin Revision 0 August 27, 2010

3 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.1 Preface This Quality Assurance Project Plan (QAPP) is submitted in partial fulfillment of the Comprehensive Flare Study Project contract issued by the Texas Commission on Environmental Quality (TCEQ) to The University of Texas at Austin, Center for Energy and Environmental Resources under Grant Activities No FY09-04, Tracking No It has been prepared in accordance with the Environmental Protection Agency QA-R5 document format for National Air Monitoring Stations/State and Local Air Monitoring Stations (NAMS/SLAMS) and Photochemical Assessment Monitoring Stations (PAMS). In this regard, the most current versions (at the time of initial preparation) of the TCEQ NAMS/SLAMS/PAMS QAPPs for air monitoring in Texas have been used as the basis for this document. It is expected that during the life of this project, the requirements of this QAPP will always meet or exceed the TCEQ NAMS/SLAMS/PAMS QAPPs for air monitoring in Texas. Contact: Edward L. Michel Flare Tests Coordinator Comprehensive Flare Study Address: The University of Texas at Austin Center for Energy & Environmental Resources (R7100) Burnet Road, EME (Bldg 133) Austin, TX Revision No. 0 Page 1 of 11 8/10

4 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page Project Manager Air Quality Division Texas Commission on Environmental Quality Danielle Nesvacil Date Project Manager Air Quality Division Texas Commission on Environmental Quality Russ Nettles Date Project Manager Air Quality Division Texas Commission on Environmental Quality Kevin Cauble Date Project Quality Assurance Manager Air Quality Division Texas Commission on Environmental Quality Bryan Foster Date Revision No. 0 Page 2 of 11 8/10

5 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) TCEQ Technical Review Committee Comprehensive Flare Study Project Peter Gogolek, Ph.D. Date TCEQ Technical Review Committee Comprehensive Flare Study Project John Pohl, Sc.D. Date TCEQ Technical Review Committee Comprehensive Flare Study Project Eben Thoma, Ph.D. Date Revision No. 0 Page 3 of 11 8/10

6 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Principal Investigator Comprehensive Flare Study Project The University of Texas at Austin Austin, Texas David T. Allen, Ph.D. Date Project Manager Comprehensive Flare Study Project The University of Texas at Austin Austin, Texas Vincent M. Torres, MSE, PE Date Project Quality Assurance Manager Comprehensive Flare Study Project The University of Texas at Austin Austin, Texas Dave Sullivan, Ph.D Date Flare Tests Coordinator Comprehensive Flare Study Project The University of Texas at Austin Austin, Texas Edward L. Michel Date Revision No. 0 Page 4 of 11 8/10

7 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Project Representative/ Project Director Aerodyne Research, Inc. Billerica, Massachusetts Scott Herndon, Ph.D. Date Project Quality Assurance Officer Aerodyne Research, Inc. Billerica, Massachusetts Charles E. Kolb, Ph.D. Date Revision No. 0 Page 5 of 11 8/10

8 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Project Representative/ Project Director Industrial Monitor and Control Corporation Round Rock, Texas Robert L. Spellicy, Ph.D. Date Project Quality Assurance Officer Industrial Monitor and Control Corporation Round Rock, Texas Curt Laush, Ph.D. Date Revision No. 0 Page 6 of 11 8/10

9 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Project Representative/ Project Director John Zink Company, Inc. Tulsa, Oklahoma Robert Schwartz Date Project Quality Assurance Officer John Zink Company, Inc. Tulsa, Oklahoma Wes Bussman, Ph.D. Date Project Operations Director John Zink Company, Inc. Tulsa, Oklahoma Zachary Kodesh Date Revision No. 0 Page 7 of 11 8/10

10 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Project Representative/ Project Director Leak Surveys, Inc. Early, Texas Bud McCorkle Date Project Quality Assurance Officer Leak Surveys, Inc. Early, Texas Joshua Furry Date Revision No. 0 Page 8 of 11 8/10

11 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Project Representative/ Project Director Telops Quebec City, Quebec Canada Vincent Farley Date Project Quality Assurance Officer Telops Quebec City, Quebec Canada Jean Giroux Date Revision No. 0 Page 9 of 11 8/10

12 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Project Representative/ Project Director TRC Austin, Texas Jim Barufaldi Date Project Quality Assurance Officer TRC Austin, Texas Clayton Elliot Date Revision No. 0 Page 10 of 11 8/10

13 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A1 A1.2 Approval Page (Continued) Project Representative/ Project Director Zephyr Environmental Consulting Austin, Texas Karen Olsen Date Project Quality Assurance Officer Zephyr Environmental Consulting Austin, Texas Maria Gou Date Revision No. 0 Page 11 of 11 8/10

14 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 TABLE OF CONTENTS Section Title Pages Revision Date PROJECT MANAGEMENT A1 Title and Approval Sheets /10 A1.1 Preface A1.2 Approval Page A2 Table of Contents 8 0 8/10 A2.1 List of Figures A2.2 List of Tables A2.3 List of Appendices A3 Distribution List 1 0 8/10 A4 Project/Task Organization /10 A4.1 Project Sponsor A4.2 TCEQ Technical Review Panel A4.3 Principal Investigator and Project Manager A4.4 UT Austin and TCEQ Project Quality Assurance (QA) Officers A4.5 UT Austin Flare Tests Coordinator A4.6 Flare Test Facility A4.7 Direct Flare Stack & Flue Gases Measurements A4.8 Direct Flare Flue Gas Measurements A4.9 Flare Flue Gas Remote Sensing Measurements A4.10 Infrared and Visible Wavelength Video Camera Recordings of Flare Flue Gas A4.11 Flare Test Data Collection, Coordination, Management and Validation A4.12 Statistical Support A4.13 Subcontractors A5 Problem Definition/Background 3 0 8/10 A5.1 Background Revision No. 0 Page 1 of 11 8/10

15 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 TABLE OF CONTENTS Section Title Pages Revision Date A5.2 Conclusions to be Made A5.3 Uses of Data A5.4 Decision Makers A5.5 Principal Customers for the Results A6 Project/Task Description 2 0 8/10 A6.1 Project Overview A6.2 Sampling Measurement Activities A6.3 Standards and Screening Levels A6.4 Assessment Tools A6.4.1 Performance Evaluations A6.5 Project Reports A7 Data Quality Objectives (DQO) for Measurement Data 2 0 8/10 A7.1 General Project Objective A7.2 Measurement Quality Objectives A7.2.1 A7.2.2 A7.2.3 A7.2.4 A7.2.5 A7.2.6 A7.2.7 Detection Limits System Contribution to the Measurement Precision Accuracy Completeness Representativeness Comparability A8 Special Training Requirements/Certification 1 0 8/10 A9 Documentation and Records 2 0 8/10 A9.1 Mechanisms for Documentation of Procedures and Objectives A9.2 Mechanisms for Record Keeping Revision No. 0 Page 2 of 11 8/10

16 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 TABLE OF CONTENTS Section Title Pages Revision Date A9.3 Data Reporting Turnaround Time A9.4 Data Storage MEASUREMENT/DATA ACQUISITION B1 Sampling Process Design (Experimental Design) 2 0 8/10 B1.1 Study Site Design B1.2 Study Site Design Rationale B1.3 Measurement Validation B2 Sampling Methods Requirements 5 0 8/10 B2.1 Continuous Methods B2.1.1 John Zink Company, Inc. B2.1.2 Aerodyne Research, Inc. B2.1.3 Telops B2.1.4 Industrial Monitor and Control Corporation B2.1.5 Leak Surveys, Inc. B2.2 Non-continuous Methods B2.2.1 TRC B2.3 Corrective Actions B3 Sample Handling and Custody 1 0 8/10 B4 Analytical Methods Requirements 1 0 8/10 B4.1 Analytical Procedures B4.1.1 Aerodyne Research, Inc. B4.2 Corrective Actions B5 Quality Control (QC) 2 0 8/10 B5.1 John Zink Company, Inc. B5.2 TRC B5.3 Aerodyne Research, Inc. B5.4 Telops B5.5 Industrial Monitor and Control Corporation B5.6 Leak Surveys, Inc. B6 Instrument/Equipment Testing, Inspection, and Maintenance Requirements 2 0 8/10 B6.1 Instrument Testing/Inspection B6.2 Preventive Maintenance Procedures B6.2.1 John Zink Company, Inc. B6.2.2 TRC B6.2.3 Aerodyne Research, Inc. B6.2.4 Telops B6.2.5 Industrial Monitor and Control Corporation B6.2.6 Leak Surveys, Inc. B6.3 Corrective Maintenance Procedures B6.3.1 John Zink Company, Inc. B6.3.2 TRC B6.3.3 Aerodyne Research, Inc. Revision No. 0 Page 3 of 11 8/10

17 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 TABLE OF CONTENTS Section Title Pages Revision Date B6.3.4 Telops B6.3.5 Industrial Monitor and Control Corporation B6.3.6 Leak Surveys, Inc. B6.4 Availability of Spare Parts B7 Instrument/Equipment Calibration and Frequency 2 0 8/10 B7.1 Calibration B7.1.1 John Zink Company, Inc. B7.1.2 TRC B7.l.3 Aerodyne Research, Inc. B7.1.4 Telops B7.1.5 Industrial Monitor and Control Corporation B7.1.6 Leak Surveys, Inc. B7.2 Traceability B7.2.1 John Zink Company, Inc. B7.2.2 TRC B7.2.3 Aerodyne Research, Inc. B7.2.4 Telops B7.2.5 Industrial Monitor and Control Corporation B7.2.6 Leak Surveys, Inc. B7.3 Documentation B8 Inspection/Acceptance Requirements for Supplies and Consumables 1 0 8/10 B8.1 Sampling Supplies B8.2 Standards B8.3 Spare Parts B9 Data Acquisition Requirements (Non-Direct Measurements) 1 0 8/10 B10 Data Management 2 0 8/10 B10.1 John Zink Company, Inc. B10.2 TRC B10.3 Aerodyne Research, Inc. B10.4 Telops B10.5 Industrial Monitor and Control Corporation B10.6 Leak Surveys, Inc. B10.7 Acceptability of the Hardware/Software Configuration B10.8 Data to Users ASSESSMENT/OVERSIGHT C1 Assessments and Response Actions 3 0 8/10 C1.1 Technical Systems Audit C1.1.1 C1.1.2 Field Technical Systems Audit Field Inspections C1.2 Performance Evaluations Revision No. 0 Page 4 of 11 8/10

18 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 TABLE OF CONTENTS Section Title Pages Revision Date C1.2.1 Field Assessment C1.3 Assessment of Data Quality Indicators C1.3.1 Specific Procedures to Assess Data Quality C C C Data Precision Assessment Data Accuracy Assessment Data Completeness Assessment C1.4 Audits of Data Quality C1.5 Corrective Actions C2 Reports to Management 1 0 8/10 C2.1 Quality Assurance (QA) Audit Reports C2.2 Annual Project QA Report C2.3 Data Reports C2.3.1 C2.3.2 Field Activity Reports Quality Assurance Reports C2.4 Reporting Schedule DATA VALIDATION AND VERIFICATION D1 Data Review, Validation, and Verification 3 0 8/10 D1.1 Data Validation D1.1.1 D1.1.2 D1.1.3 D1.1.4 D1.1.5 D1.1.6 John Zink Company, Inc. TRC Aerodyne Research, Inc. Telops Industrial Monitor and Control Corporation Leak Surveys, Inc. D1.2 Data Custody D1.2.1 John Zink Company, Inc. Revision No. 0 Page 5 of 11 8/10

19 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 TABLE OF CONTENTS Section Title Pages Revision Date D1.2.2 D1.2.3 D1.2.4 D1.2.5 D1.2.6 TRC Aerodyne Research, Inc. Telops Industrial Monitor and Control Corporation Leak Surveys, Inc. D2 Validation and Verification Methods 2 0 8/10 D2.1 John Zink Company, Inc. D2.2 TRC D2.3 Aerodyne Research, Inc. D2.4 Telops D2.5 Industrial Monitor and Control Corporation D2.6 Leak Surveys, Inc. D2.7 Data Review D3 Reconciliation with User Requirements 3 0 8/10 D3.1 Detection Limits D3.2 Precision D3.2.1 D3.2.2 D3.2.3 D3.2.4 D3.2.5 D3.2.6 John Zink Company, Inc. TRC Aerodyne Research, Inc. Telops Industrial Monitor and Control Corporation Leak Surveys, Inc. D3.3 Accuracy D3.3.1 D3.3.2 D3.3.3 D3.3.4 D3.3.5 D3.3.6 John Zink Company, Inc. TRC Aerodyne Research, Inc. Telops Industrial Monitor and Control Corporation Leak Surveys, Inc. Revision No. 0 Page 6 of 11 8/10

20 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 TABLE OF CONTENTS Section Title Pages Revision Date D3.4 Completeness Revision No. 0 Page 7 of 11 8/10

21 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 LIST OF FIGURES Section Title Pages Revision Date A4 Figure A4.A Comprehensive Flare Study Project 1 0 8/10 Organization B2 Figure B2.A LSI Camera Image 1 0 8/10 B5 Figure B5.A Aerodyne Mass Spectrometer 1 0 8/10 Appendix F Revision No. 0 Page 8 of 11 8/10

22 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 LIST OF TABLES Section Title Pages Revision Date A7 Appendix G Reporting Units of Measurements 1 0 8/10 B2 Table B2A Flare Plume Compounds Measured by Aerodyne 1 0 8/10 B2 Table B2.B Flare Plume Compounds Measured by Telops 1 0 8/10 C1 Appendix I Field Inspection Report 2 0 8/10 App. D Draft Tests Comprehensive Flare Study Draft John Zink Tests 1 0 8/10 App. L Appendix L Acronyms 2 0 8/10 Revision No. 0 Page 9 of 11 8/10

23 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 LIST OF APPENDICES Appendix Title Pages Revision Date A Overall Plan View of Flare Operation Facility Area 1 0 8/10 A1 Satellite Photo of Flare Operation Facility Area, Subcontractor Location 1 0 8/10 A2 Satellite Photo of Flare Operation Facility Area, Observer Room, 1 0 8/10 Propylene Storage Location B TRC Modified EPA Method 3A, O 2 and CO /10 B1 TRC Modified EPA Method 18, GC Analysis /10 B2 TRC Modified EPA Method 19, Mass Emission Calculations /10 B3 Heating Value Calculations 1 0 8/10 B4 Aerodyne QC 8 0 8/10 B5 John Zink QC 4 0 8/10 B6 Zephyr Calculations 1 0 8/10 B7 John Zink Calculations 3 0 8/10 B8 Aerodyne DRE 4 0 8/10 B9 IMACC CE 1 0 8/10 B10 IMACC Procedures 8 0 8/10 B11 IMACC Calibrations 6 0 8/10 B12 Telops Methods 2 0 8/10 C1 Schematic of Flare Measurements; Wake Dominated Flare Flue Gas 1 0 8/10 C2 Schematic of Flare Measurements; Buoyancy-Dominated Flare Flue Gas 1 0 8/10 C3 Schematic of Flare Measurements; Remotely Sensing of Flare Flue Gas 1 0 8/10 D Comprehensive Flare Study Proposed Flare Test Plan /10 E Flare Test Plan Modification Process 1 0 8/10 F Quality Assurance Steps for the Aerodyne Measurements 4 0 8/10 G Measurement Data Quality Objectives 4 0 8/10 H Measurement Data Quality Control Activities 5 0 8/10 I Field Inspection Report 2 0 8/10 Revision No. 0 Page 10 of 11 8/10

24 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A2 LIST OF APPENDICES Appendix Title Pages Revision Date J Flare Flue Gas Sampling Device Diagram 1 0 8/10 J1 Flare Flue Gas Sampling Device Drawings, Vertical 1 0 8/10 J2 Flare Flue Gas Sampling Device Drawings, Horizonal 1 0 8/10 K Comprehensive Flare Study Project Schedule of Major Activities for 2 0 8/10 Field Measurements L Steam Flare Tip Drawing 1 0 8/10 L1 Air Flare Tip Drawing 1 0 8/10 M John Zink Waste Gas and Steam Flow Diagram, /10 M1 John Zink Waste Gas and Steam Flow Diagram, /10 N Flare Project Activity and Protocol Documents /10 O Acronyms 3 0 8/10 Revision No. 0 Page 11 of 11 8/10

25 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A3 A3 DISTRIBUTION LIST Texas Commission on Environmental Quality Danielle Nesvacil, Project Manager, Air Quality Division Russ Nettles, Project Manager, Air Quality Division Bryan Foster, Project Quality Assurance Officer, Air Quality Division Texas Commission on Environmental Quality Technical Review Panel Peter E. G. Gogolek, Ph.D., CanmetENERGY John Pohl, Sc.D., Virginia Polytechnic Institute and State University Eben, Thoma, Ph.D., US Environmental Protection Agency Flare Tests Subcontractors Scott Herndon, Ph.D., Aerodyne Research, Inc. Robert L. Spellicy, Ph.D., Industrial Monitor and Control Corporation Bob Schwartz, John Zink Company, LLC Bud McCorkle, Leak Surveys, Inc. Vincent Farley, Telops Jim Barufaldi, TRC Karen Olsen, Zephyr Environmental Corporation The University of Texas at Austin David T. Allen, Ph. D., Principal Investigator Vincent M. Torres, Project Manager Edward L. Michel, Flare Tests Coordinator Dave Sullivan, Ph. D., Project Quality Assurance Officer Revision No. 0 Page 1 of 1 8/10

26 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan A4 PROJECT/TASK ORGANIZATION Section A4 The Comprehensive Flare Study Project is performed by The University of Texas at Austin and its contractors. The overall project organization is shown in Figure A4.A. The project participants involved in the flare testing and their role and responsibilities are presented in Table A4.1. The interrelationships and responsibilities of the participants in this project are listed below. A4.1 Project Sponsor Ms. Danielle Nesvacil, Mr. Russ Nettles Texas Commission on Environmental Quality (TCEQ) Sets the study objectives for the project. Allocates adequate resources to ensure completion of the project in compliance with the stated objectives. Defines the project team and organization Reviews and approves the Quality Assurance Project Plan (QAPP) and any changes. Defines the overall project schedule and deliverables. Determines the ultimate use of the data set developed from the project activities. A4.2 TCEQ Technical Review Panel Dr. Peter Gogolek, Dr. John Pohl, and Dr. Eben Thoma Reviews and comments on the Draft QAPP. Reviews and comments on the Draft Preliminary Flare Measurements Report. Reviews and comments on the Draft Comprehensive Flare Study Report. A4.3 UT Austin Principal Investigator, Project Manager Dr. David Allen and Mr. Vincent M. Torres, The University of Texas at Austin Are the primary contact personnel for the project. Provide project planning, coordination of all project work and preparation of all reports to the Project Sponsor. Provide oversight of subcontractor work and approval of work products. Ensure that all subcontractors are qualified for the operations they will perform and/or the measurements they will be making. Prepare the QAPP for the project for review and approval by the TCEQ. Coordinate the QA activities for the project including QA activities with external agencies and non-agency groups. Coordinate data compilation, oversee and perform data analysis, prepare project draft preliminary flare measurements report and production of project final report. Revision No. 0 Page 1 of 13 8/10

27 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A4 A4.4 UT Austin and TCEQ Project Quality Assurance (QA) Officers Dr. Dave Sullivan, The University of Texas at Austin and Mr. Bryan Foster, Texas Commission on Environmental Quality (TCEQ) Participate in the development, approval, implementation, and maintenance of the project s written quality assurance documents (e.g, QMPs, SOPs, QAPPs). Perform project and laboratory technical systems audits. Participate in the preparation of quality assurance reports. Determine conformance with project quality system requirements. Review and approve proposed corrective actions and verifications. Monitor the implementation of corrective actions. Report on the status of corrective action programs. Assess the effectiveness of the project s quality systems. Coordinate the identification, disposition, and reporting to project management of nonconforming items and activities. A4.5 UT Austin Flare Tests Coordinator Edward L. Michel, The University of Texas at Austin Participate in project planning, participate in coordination of project work and participate in the preparation of all reports to the Project Sponsor. Coordinate the subcontractor s work during the flare tests and review their work products. Coordinate the assurance that all subcontractors are qualified for the operations they will perform and/or the measurements they will be making. Participate in the preparation of the QAPP for the project for review and approval by the TCEQ. Participate in the coordination of the QA activities for the project including QA activities with external agencies and non-agency groups. Coordinate delivery of all subcontractors preliminary report of flare measurements, final report of flare measurements, and delivery of quality assured data in prescribed formats. A4.6 Flare Operations Facility Mr. Robert E. Schwartz, Senior Technical Specialist, John Zink Company, LLC Provide the flare test facility with the capabilities specified in the flare test plan. Provide the qualified technical and test coordination support to operate the flare test facility during the flare tests. Review and certify that all flare test facility instrumentation, sampling equipment, and surveillance cameras meet or exceed the QAPP specifications. Revision No. 0 Page 2 of 13 8/10

28 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Perform quality control checks on all flare test facility instrumentation, sampling equipment and surveillance cameras before and during the flare tests and take corrective action when indicated. Perform scheduled preventive maintenance procedures for all flare test facility instrumentation, sampling equipment, meteorological equipment, and surveillance cameras Record data/information as required in appropriate flare and quality assurance test logs. Calibrate instrumentation. Perform calibration verification checks. Maintain calibration equipment. Participate in the development of updates and revisions to written quality assurance standards (e.g., QMPs, SOPs, QAPPs). A4.7 Direct Flare Stack & Flue Gases Measurements Mr. Jim Barufaldi, TRC Companies, Inc. Provide the stack gas instrumentation, sampling and analysis equipment to perform measurements of stack gases as specified in the flare test plan. Provide the qualified technical support to operate the stack gas instrumentation, sampling and analysis equipment before and as needed during the flare tests. Review and certify that all stack gas instrumentation, sampling and analysis equipment meet or exceed the QAPP specifications. Perform quality control checks on all stack gas instrumentation, sampling and analysis equipment before and during the flare tests and take corrective action when indicated. Perform scheduled preventive maintenance procedures for all stack gas instrumentation, sampling and analysis equipment. Record data/information as required in appropriate flare test and quality assurance logs. Calibrate instrumentation, sampling equipment and meteorological equipment. Perform calibration verification checks. Maintain calibration equipment. Participate in the development of updates and revisions to written quality assurance standards (e.g., QMPs, SOPs, QAPPs). Section A4 Revision No. 0 Page 3 of 13 8/10

29 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan A4.8 Direct Flare Flue Gas Measurements Section A4 Dr. Scott Herndon, Aerodyne Research, Inc. Provide the Aerodyne mobile monitoring lab with all the instrumentation to perform measurements in the flare plume and associated analysis of samples as specified in the flare test plan. Provide the qualified technical support to operate the Aerodyne mobile monitoring lab in advance of, as required, and during the flare tests. Review and certify that all instrumentation, sampling equipment, and meteorological equipment meet or exceed the QAPP specifications. Perform quality control checks on all instrumentation, sampling equipment, and meteorological equipment before and during the flare tests and take corrective action when indicated. Perform scheduled preventive maintenance procedures for all instrumentation, sampling equipment, and meteorological equipment. Record data/information as required in appropriate flare test and quality assurance logs. Calibrate instrumentation, sampling equipment and meteorological equipment. Perform calibration verification checks. Maintain calibration equipment. Participate in the development of updates and revisions to written quality assurance standards (e.g., QMPs, SOPs, QAPPs). A4.9 Flare Flue Gas Remote Sensing Measurements Dr. Robert L. Spellicy, Industrial Monitor and Control Corporation Provide the passive Fourier transform infrared spectrometer with the capabilities specified in the flare test plan. Provide the qualified technical support to operate passive Fourier transform infrared spectrometer before, as required, and during the flare tests. Review and certify that passive Fourier transform infrared spectrometer meets or exceeds the QAPP specifications. Perform quality control checks on passive Fourier transform infrared spectrometer before and during the flare tests and take corrective action when indicated. Perform scheduled preventive maintenance procedures for the passive Fourier transform infrared spectrometer. Record data/information as required in appropriate flare test and quality assurance logs. Calibrate the passive Fourier transform infrared spectrometer as required. Perform calibration verification checks. Maintain calibration equipment. Participate in the development of updates and revisions to written quality assurance standards (e.g., QMPs, SOPs, QAPPs). Revision No. 0 Page 4 of 13 8/10

30 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A4 Mr. Vincent Farley, Telops, Inc Provide an infrared hyper-spectral imaging instrument with the capabilities specified in the flare test plan. Provide the qualified technical support to operate the infrared hyper-spectral imaging instrument before, as required, and during the flare tests. Review and certify that the infrared hyper-spectral imaging instrument meets or exceeds the QAPP specifications. Perform quality control checks on passive the infrared hyper-spectral imaging instrument before and during the flare tests and take corrective action when indicated. Perform scheduled preventive maintenance procedures for the infrared hyper-spectral imaging instrument. Record data/information as required in appropriate flare test and quality assurance logs. Calibrate the passive infrared hyper-spectral imaging instrument as required. Perform calibration verification checks. Maintain calibration equipment. Participate in the development of updates and revisions to written quality assurance standards (e.g., QMPs, SOPs, QAPPs). A4.10 Infrared and Visible Wavelength Video Camera Recordings of Flare Flue Gas Mr. Bud McCorkle, Leak Surveys, Inc. Provide infrared and visible wavelength cameras with the capabilities specified in the flare test plan. Provide videography (60 frames per second) of the flare flue gas during both the steam and air flare tests. Provide the qualified technical support to operate the infrared and visible wavelength cameras before, as required, and during the flare tests. Review and certify that the infrared and visible wavelength cameras meet or exceed the QAPP specifications. Perform quality control checks on the infrared and visible wavelength cameras before and during the flare tests and take corrective action when indicated. Perform scheduled preventive maintenance procedures for the infrared and visible wavelength cameras. Record images from all LSI cameras, data/information as required in appropriate flare test and quality assurance logs. Calibrate the infrared and visible wavelength cameras as required. Perform calibration verification checks. Maintain calibration equipment. Participate in the development of updates and revisions to written quality assurance standards (e.g., QMPs, SOPs, QAPPs). Revision No. 0 Page 5 of 13 8/10

31 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A4 A4.11 Flare Test Data Collection, Coordination, Management and Validation Mr. Edward L. Michel, The University of Texas at Austin Coordinate the participation and data collection of all subcontractors during the flare tests. Review validation of all flare test data. Provide technical support on data management issues that may arise. Document all data management activities. Coordinate delivery of all subcontractors preliminary report of flare measurements, final report of flare measurements, and delivery of quality assured data in prescribed formats. Mr. Robert E. Schwartz, Senior Technical Specialist, John Zink Company, LLC Validate all John Zink Company, LLC, flare test data. Provide technical support on data management issues that may arise. Document all data management activities. Deliver data per schedule in Section C2.4. Submit draft and final John Zink reports to UT Austin flare project manager. Mr. Jim Barufaldi, TRC Companies, Inc. Validate all TRC Companies, Inc. stack test data. Provide technical support on data management issues that may arise. Document all data management activities. Deliver data per schedule in Section C2.4. Submit draft and final TRC reports to UT Austin flare project manager. Dr. Scott Herndon, Aerodyne Research, Inc. Validate all Aerodyne Research, Inc. flare plume test data. Provide technical support on data management issues that may arise. Document all data management activities. Deliver data per schedule in Section C2.4. Submit draft and final Aerodyne reports to UT Austin flare project manager. Dr. Robert L. Spellicy, Industrial Monitor and Control Corporation Validate all Industrial Monitor and Control Corporation flare plume test data. Provide technical support on data management issues that may arise. Document all data management activities. Revision No. 0 Page 6 of 13 8/10

32 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Deliver data per schedule in Section C2.4. Submit draft and final Industrial Monitor and Control Corporation reports to UT Austin flare project manager. Section A4 Mr. Vincent Farley, Telops, Inc. Validate all Telops, Inc. flare plume test data. Provide technical support on data management issues that may arise. Document all data management activities. Deliver data per schedule in Section C2.4. Submit draft and final Telops reports to UT Austin flare project manager. Ms. Karen Olsen, Zephyr Environmental Corporation Validate all flare plume test data. Provide technical support on data management issues that may arise. Document all data management activities. Submit draft and final Zephyr reports to UT Austin flare project manager. Mr. Bud McCorkle, Leak Surveys, Inc. Validate all Leak Surveys, Inc. flare plume tests images and associated date/time data. Provide technical support on data management issues that may arise. Document all data management activities. Deliver data per schedule in Section C2.4. Submit draft and final Leak Surveys, Inc. reports to UT Austin flare project manager. A4.12 Statistical Support Mr. Robert E. Schwartz, Senior Technical Specialist, John Zink Company, LLC Provide statistical evaluation of John Zink Company, LLC, flare test data to assist in achieving the study objectives. Provide statistical evaluation of these data to quality assure data. Mr. Jim Barufaldi, TRC Companies, Inc. Provide statistical evaluation of all TRC Companies, Inc. stack test data to assist in achieving the study objectives. Provide statistical evaluation of these data to quality assure data. Dr. Scott Herndon, Aerodyne Research, Inc. Provide statistical evaluation of Aerodyne Research, Inc. flare plume test data to assist in achieving the study objectives. Provide statistical evaluation of these data to quality assure data. Revision No. 0 Page 7 of 13 8/10

33 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A4 Dr. Robert L. Spellicy, Industrial Monitor and Control Corporation Provide statistical evaluation of Industrial Monitor and Control Corporation flare plume test data to assist in achieving the study objectives. Provide statistical evaluation of these data to quality assure data. Mr. Vincent Farley, Telops, Inc. Provide statistical evaluation of Telops, Inc. flare plume test data to assist in achieving the study objectives. Provide statistical evaluation of these data to quality assure data. Ms. Karen Olsen, Zephyr Environmental Corporation Provide statistical evaluation of all flare plume test data to assist in achieving the study objectives. Provide statistical evaluation of these data to quality assure data. Mr. Bud McCorkle, Leak Surveys, Inc. Provide statistical evaluation of Leak Surveys, Inc. flare plume test data to assist in achieving the study objectives. Provide statistical evaluation of these data to quality assure data. Revision No. 0 Page 8 of 13 8/10

34 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan A4.13 Subcontractors Mr. Robert Schwartz, Project Representative John Zink Company, LLC East Apache Tulsa, Oklahoma Section A4 Mr. Jim Barufaldi, Project Representative TRC 9225 US Highway 183 South Austin, Texas Dr. Scott Herndon, Project Representative Aerodyne, Research, Inc. 45 Manning Road Billerica, Massachusetts Mr. Bud McCorkle, Project Representative Leak Surveys, Inc Early Blvd. Early, Texas Dr. Robert L. Spellicy, Project Representative Industrial Monitor and Control Corp. 800 Paloma, Suite 100 Round Rock, Texas Mr. Vincent Farley, Project Representative TELOPS St-Jean-Baptiste Avenue Quebec City, Quebec Canada G2E 6J5 Ms. Karen Olsen, Project Representative Zephyr Environmental 2600 Via Fortuna, Suite 450 Austin, Texas According to terms of the contract, responsibilities include but are not limited to: Provide and operate test, instrumentation and/or sampling equipment according to this approved QAPP. Perform quality control checks on test, instrumentation, and/or sampling equipment as specified in this approved QAPP. Calibrate and maintain all equipment as required for use during the flare tests period. Perform data validation. According to terms of the contract, contractor s communications responsibilities include, but are not limited to: Maintain an open line of communication between The University of Texas Project Representatives, TCEQ Personnel, and other subcontractors. Attend and/or provide information for, if requested, any meetings that may be requested by The University of Texas Project Representatives and TCEQ Personnel. The types and frequency of communications may include: Revision No. 0 Page 9 of 13 8/10

35 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A4 Cell phone, land-line, and exchanges several times per week among The University of Texas Project Representatives, TCEQ Personnel, and subcontractors. Daily on-site meetings at the John Zink flare test facility in Tulsa, Oklahoma between the subcontractors, TCEQ Personnel and The University of Texas Project Representatives. Intermittent meetings among The University of Texas Project Representatives, TCEQ Personnel, and subcontractors. Written and electronic versions of reports of test data in the format specified by the The University of Texas Project Manager per the contract deliverables schedule. Section C2.4. The University of Texas Project Manager and Project Quality Assurance Officers monitor the subcontractors through these communications. Revision No. 0 Page 10 of 13 8/10

36 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A4 Bryan Foster, TCEQ Project QA Officer TCEQ Project Sponsor Ms. Danielle Nesvacil, Project Manager Mr. Russ Nettles, Project Manager TCEQ Technical Review Panel Peter Gogolek John Pohl Eben Thoma The University of Texas at Austin Center for Energy & Environmental Resources (CEER) David T. Allen, Principal Investigator Dave Sullivan, UT Austin Project QA Officer Project Support Personnel Vincent M. Torres, Project Manager Edward L. Michel, Flare Tests Coordinator Zephyr Environmental Corporation Karen Olsen John Zink Company, LLC Bob Schwartz Aerodyne Research, Inc. Scott Herndon TRC Companies, Inc. Jim Barufaldi Leak Surveys, Inc. Bud McCorkle Industrial Monitor and Control Corporation Bob Spellicy Telops, Inc. Vincent Farley Figure A4.A Comprehensive Flare Study Project Organization Revision No. 0 Page 11 of 13 8/10

37 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Table A4.1 Responsibilities of Project Participants During Flare Tests Section A4 Company/Entity and Role Texas Commission on Environmental Quality (TCEQ) TCEQ Technical Review Panel (TRP) The University of Texas at Austin (UT Austin) Principal Investigator, Project Manager and Flare Tests Coordinator John Zink Company, LLC (Zink) Flare Operations Facility Aerodyne Research Inc. (ARI) Direct Flare Flue Gas Measurements Primary Responsibility/Measurement to be Performed During Flare Tests Be on site and ensure flare tests are addressing project objectives and rule on any changes proposed to the test plan. Witness tests, offer comments on technical implementation of test procedures and provide advice on any proposed changes to the test plan to TCEQ Ensure all flare test equipment and direct and remote sensing measurement instrumentation is ready before the tests begin each day and ensure/oversee that subcontractors pretest QC checks are performed. Conduct daily briefings with all project participants to review tests to be run that day. During the tests, direct the John Zink personnel when to begin flare tests and what operational conditions are to be run during each test. Also communicate and coordinate with all project participants during each test to ensure that each knows what test is being conducted, when it begins, when it ends and when measurements should be made of test points in between. Coordinate with ARI to ensure proper placement of the plume sampling device. Conduct post test briefings with all project participants to learn of and address problems encountered during the day. Propose changes to the test plan. Allow personnel on site and ensure utility infrastructure support is operating properly. Operate flare equipment during each test at parameters specified by UT Austin, measure and record operating parameters, and ensure flare is operating as expected. Direct crane operator to position plume sampling device at location specified by UT Austin and ARI. Ensure that all instrumentation is ready to conduct direct measurement of flue gas constituents before each flare test begins. During each test, specify to UT where the plume sampling device should be positioned, and make, record, and send to control room display direct measurements of flue gas per Appendix G. What and Frequency Be On-site Continuously each day. Rule on Proposed Test Plan Changes As needed. Witness Tests and Offer Comments and Advice - At daily briefings. Instrumentation Readiness & Briefing - At the beginning of the day and as needed at other times. Flare Tests Activities Continuously during each test. Post test briefing At the end of days briefing. Consider changes to test plan As needed. Utility Infrastructure Support Continuously. Operate Flare Test Equipment As needed. Instrumentation Readiness Beginning of the day. Direct Measurements of Flue Gas Continuously during each test. Revision No. 0 Page 12 of 13 8/10

38 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A4 Table A4.1 (Continued) Responsibilities of Project Participants During Flare Tests Company/Entity TRC Direct Flare Stack & Flue Gases Measurements Leak Surveys, Inc. (LSI) IR & Visible Wavelength Video Camera Recordings of Flare Flue Gas Industrial Monitor and Control Corporation (IMACC) Flare Flue Gas Remote Sensing Measurements Telops Flare Flue Gas Remote Sensing Measurements Primary Responsibility/Measurement to be Performed During Flare Tests Ensure that all instrumentation is ready to conduct direct measurement of stack and flue gas constituents before each flare test begins. During each test, make, record, and send to control room display direct measurements of flare emissions per Appendix G. Ensure that all instrumentation is ready to make IR, visible frequency spectrum video camera images and videography of flue gas constituents before each flare test begins. During each test, make, record, and send to control room display IR, visible frequency spectrum digital and camera images of flare emissions per Appendix G. Ensure that all instrumentation is ready to conduct remote sensing measurement of flue gas constituents using active and passive FTIR spectroscopy before each flare test begins. During each test, make and record active and passive FTIR spectroscopy measurements of flare emissions per Appendix G. Ensure that all instrumentation is ready to conduct direct measurement of stack and flue gas constituents before each flare test begins. During each test, make and record measurements of flare emissions per Appendix G. What and Frequency Instrumentation Readiness Beginning of the day. Direct Measurements of Stack and Flue Gases Continuously during each test. Instrumentation Readiness Beginning of the day. IR, visible frequency spectrum and digital camera images of Flare Emissions Continuously during each test. Instrumentation Readiness Beginning of the day. Active and Passive FTIR Measurements of Flare Emissions Continuously during each test. Instrumentation Readiness Beginning of the day. Infrared Hyper- Spectral Imaging Measurements of Flare Emissions Continuously during each test. Revision No. 0 Page 13 of 13 8/10

39 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A5 A5 PROBLEM DEFINITION/BACKGROUND A5.1 Background The TCEQ contracted National Physical Laboratory (NPL), based in the United Kingdom, to perform differential absorption lidar (DIAL) measurements on industrial emissions sources located in a refinery and a storage terminal near Houston during Measurements focused on those industrial sources that are difficult to measure using conventional sampling techniques. Specifically, the study involved: 1. Identifying potentially under-reported industrial emissions sources, 2. Conducting remote sensing measurements of these sources, 3. Collecting process and operational data from these sources, and 4. Comparing emissions determined using conventional EPA-approved determination methods to the remote sensing measurements. TCEQ 2007 Remote Sensing Study Results NPL submitted a final report to EPA in An independent third party is currently comparing remote sensing measurements to conventionally determined emissions. Although these results are still being analyzed, based upon the preliminary total volatile organic compounds (VOC) measurements, flare emissions may potentially be under-reported when emissions are determined using EPA or TCEQ material balance calculation methods. Additionally, preliminary results indicate flare destruction and removal efficiency (DRE) may be reduced during certain operating conditions, such as combusting small volumes of waste gas, and during flare air- or steam-assist operations. These preliminary results indicate the need to conduct a study that determines the relationship between flare design, operation, and DRE. Purpose The purpose of this study is to measure flare flue gas and collect required process and operational data in a semi-controlled environment to determine the relationship between flare design, operation, and DRE. The ambient air conditions, i.e., temperature, humidity, wind speed and wind direction will not be controlled. Direct measurement techniques of flare emissions as well as remote sensing measurement techniques, will be employed in the semi-controlled environment. Analysis of collected process and operational data will permit comparisons between traditional flare material balance emissions determinations, process stream and air measurements, and the emissions rates and concentrations measured by the direct and remote sensing technologies. The TCEQ anticipates that the results of the controlled tests will be broadly applicable and provide insight to operational conditions that may impact flare VOC, DRE and flare combustion efficiency (CE), such as steam- and air-assist rates or waste gas volumetric flow rates. For this project, the following definitions will be used. (1) Destruction Removal Efficiency (DRE) is the percent of the waste gas molecules that are removed or destroyed, relative to the number of the waste gas Revision No. 0 Page 1 of 3 8/10

40 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A5 molecules that enter the flare. (EPA web site) (2) Combustion Efficiency (CE) is the percent of the waste gas molecules converted to carbon dioxide and water.(john Zink Handbook) (3) Visible emissions are the populations of smoke particles that can be seen with the naked eye. (4) Waste gas flow rate turndown is the ratio of the maximum flow rating over the minimum flow occurring during the measurement collection time. (5) A flame is the visible part of a fire. (6) Blow off means to come off due to a strong force. (7) Flash back is when the flame goes into the tip. (8) Ignition is the process of setting the vent gas on fire. (9) A stable flame is when the flame velocity (rate of burning) is matched by the velocity of the material fed. If it is not possible to match the rate of burning and the imposed velocity at some position, the flame will either blow off or flash back. Therefore, a stable flame is one that has been heated to the temperature where the heat generation is balanced by the heat loss (ignition), and the rate of consumption is balanced by the imposed velocity. All calculations for this project are contained in Appendix B. Study Objectives Primary study objectives in order of importance include: 1. Assessing the potential impact of waste gas flow rate turndown on flare DRE and CE. 2. Assessing the potential impact of steam- and air-assist on flare DRE and CE at various operating conditions, focusing exclusively on low flow rate conditions. 3. Assessing whether flares operating over the range of requirements stated in 40 Code of Federal Regulations (CFR) achieve the assumed hydrocarbon DRE of 98 percent at varying flow rate turndown and assist ratios as well as variable waste gas heat content. 4. Identifying and quantifying the hydrocarbon species in flare flue gas currently visualized with passive infrared technology. A5.2 Conclusions to be made 1. The impact on flare CE and VOC DRE of low waste gas volumetric flow rates. 2. The impact on flare CE and VOC DRE of excess use of steam-assist and air-assist at low waste gas volumetric flow rates. 3. The applicability of remote sensing technologies for measurement of flare flue gases. A5.3 Uses of Data The potential uses of the data are listed below: To compare flare flue gas rates and concentrations determined using conventional EPA-approved determination methods and remote sensing methods to direct measurements using material balance methods for low waste gas flow rates. To better understand the use of steam- and air-assist at low waste gas flow rates in controlling flare emission rates and VOC DRE. To determine if flare emissions are underreported using current conventional reporting practices. To determine if additional air pollution control strategies are required. To assess the use of the remote sensing technologies included in this study for measurement of flare flue gases. Revision No. 0 Page 2 of 3 8/10

41 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A5 A5.4 Decision Makers Ms. Danielle Nesvacil, Project Manager, Air Quality Division, Texas Commission on Environmental Quality. Mr. Russ Nettles, Project Manager, Air Quality Division, Texas Commission on Environmental Quality. Dr. David Allen, Principal Investigator, the University of Texas at Austin. A5.5 Principal Customers for the Results Texas Commission on Environmental Quality. Environmental Protection Agency. Chemical and Petrochemical Industry. The University of Texas at Austin. Local city and county health departments. Texas citizens. Revision No. 0 Page 3 of 3 8/10

42 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A6 A6 PROJECT/TASK DESCRIPTION This section provides a description of the work to be performed, an overall view of the project objectives, activities, assessments, and outputs of the project, identification of potentially applicable ambient air quality regulations and standards, and an implementation schedule for the project. The measurements to be made during the project are identified in Appendix G. The data for this project will be produced during controlled flare tests over approximately a 2 to 3 week period in the late-summer of The project will be completed by March A6.1 Project Overview The University of Texas at Austin has developed and will oversee a prescribed series of flare tests (Appendix D), all of which will be conducted at the John Zink Company flare test facility in Tulsa, Oklahoma. Appendix D has been developed to provide data that will fulfill as many of the Study Objectives as possible within the budgetary constraints of the Project. The study objectives are contained in Section A5.1. During each series, there will be direct measurement of the flare flue gas for many parameters. Remote sensing measurements of flare flue gas will also be made using remote sensing technologies specified by the TCEQ. All direct and remote sensing technology measurements that will be made during the flare test series are described in Appendix G and tabulated. The detailed schedule for conducting the flare tests, including pretest activities, is found in Appendix K. The test plan includes flare operating conditions that attain the incipient smoke point. The definition of the incipient smoke point is included in Appendix N. In addition, two flare flame conditions, stable flame and wind-influenced flame, will potentially be observed during the operation of the flare. For this project, a stable flame will be defined as a visible, orange to white in color, area in the flue gas that does not go out of view for the observer while operating the flare at the test point. For this project, also defined, a wind-influenced flame will be defined as intermittently visible, orange to white in color area in the flue gas that goes out of view for the observer while operating the flare at the test point. In the event that it is determined that a modification to the flare operation plan should be considered, a Test Series Modification Process will be followed. The Test Series Modification Process is detailed in Appendix E. After the test series is completed, the resulting data will be used to improve the understanding of flare operations and the impact of steam- and air-assist and waste stream turn down rate on DRE and CE. After completion of the flare operation plan, The University of Texas at Austin will compile all data from the flare series, analyze the data and produce a final project report by March, A6.2 Sampling and Measurement Activities Sampling activities will include remote and direct measurements of the flare flue gas. These activities will be performed to measure concentrations and visible emissions of hydrocarbons, carbon monoxide, particulates, flared gas heat content and flared gas exit velocity to help understand DRE and CE on a typical steam-and air-assisted flare tip rated at 937,000 and 144,000 pounds per hour respectively, but operated nominally at 0.25 and 0.1 % of rated design. Revision No. 0 Page 1 of 3 8/10

43 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A6 In addition, a combustion zone heating value (CZHV) will be calculated for each test. For the study it will be referred to as the Flare Combustion Zone Net Heating Value and will be defined as follows: NHVFCZG = [(VG)(NHV VG )( /MWVG)]+ [(PG)(NHV PG )( [(VG)( / MWVG) + (PG)( / MWPG) + (S)( / 18)] Parameter Description (Unit) Source NHV FCZG = Flare Combustion Zone net heating value (BTU/sft 3 ) VG = Vent Gas mass flow rate (lb/hr) NHV VG = Vent Gas Net Heating Value (BTU/sft 3 ) MW VG = Vent Gas molecular weight (lb/lb-mole) PG = Pilot Gas mass flow rate (lb/hr) NHV PG = Pilot gas net heating value (BTU/sft 3 ) MW PG = Pilot Gas molecular weight (lb/lb-mole) S = Actual total steam mass flow rate (lb/hr) = Constant (sft 3 68 F and 1 atm) Result From ultrasonic flare gas flow meters Calculated from GC analysis Calculated from GC analysis As measured by flow meters Calculated from GC analysis Calculated from GC analysis From ultrasonic steam flow meter Ideal Gas Law This calculation is being used at the direction of this projects sponsor. This is the same definition that was used by the Marathon Petroleum Company, LCC during the flare testing program in Texas City, Texas that was conducted September 15 24, A6.3 Standards and Screening Levels This section references some federal statutes for which data generated by this project may be compared. 40 CFR, Part Determination of tip velocity, heating value, pilot flame requirements for flares 40 CFR, Part 63.11(b) Determination of Flare exit velocity, visible emissions, pilot flame presence, minimum heating value A6.4 Assessment Tools Revision No. 0 Page 2 of 3 8/10

44 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A6 Assessment tools that will be used are described in this section. A6.4.1 Performance Evaluations Performance evaluations will not be able to be performed by the respective project participants on the instruments and sampling systems provided by each project participant due to the fact that this is a first of its kind research project, therefore not lending itself to commercially available performance standards. A6.5 Project Reports The following reports will be produced. See Section C2.4 for more detailed information. Daily Field Activity Reports will be provided by each project participant. Comprehensive Flare Study Draft and Final Report. Daily Quality Assurance Reports will be provided by each project participant. Final Project Report. Final Quality Assurance Report will be provided by each project participant. Revision No. 0 Page 3 of 3 8/10

45 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A7 A7 DATA QUALITY OBJECTIVES (DQO) FOR MEASUREMENT DATA This section presents the data quality objectives for the project. The results of the DQO process include: 1. specifying the intended use of the data; 2. defining the type of data needed to support the decision; 3. identifying the conditions under which the data should be collected; and 4. Specifying tolerable limits on the probability of making a decision error due to uncertainty in the data. The quality control indicators of each measurement technique for this project are presented in Appendix G. A7.1 General Project Objectives Provide direct and remote sensing measurements of flare flue gas to be used in calculating CE and DRE for low waste gas flow rates and typical steam- and air-assist operating conditions. Provide data and information on flare flue gas to guide in the minimization of flare flue gas during periods of low waste gas flow rates. Guide in the use of remote sensing technologies and design of future flare test work. A7.2 Measurement Quality Objectives The approaches used to assess data uncertainty and the measurement quality objectives for each type of measurement are addressed in this section. Total Measurement Error is the combination of the published instrument error and the potential error from installation and application of the equipment to this project. Appendix G presents the measurement quality objectives for each measurement that will be employed. A7.2.1 Detection Limits Detection limits are expressed in units of concentration and reflect the smallest concentration of a compound that can be measured with a defined degree of certainty. The detection limits for each parameter measured during the project are provided in Appendix G. A7.2.2 System Contribution to the Measurement A blank or zero air level is part of each calibration and span check for each measurement that is reported in units of ppm, ppb or ppt. The units for each project measurement are contained in Appendix G. As part of the calibration, the zero level is used along with the upscale (span) concentrations to establish the monitor s calibration curve. As Revision No. 0 Page 1 of 2 8/10

46 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A7 part of the span check, this upscale level is used as a quality control check for a monitor s zero drift. A7.2.3 Precision Precision is a measure of the repeatability of the results. Estimates of precision are assessed in different ways for different measurement technologies. Each project participant s precision criteria are presented in Appendix G. A7.2.4 Accuracy Accuracy is the closeness of a measurement to a reference value, and reflects elements of both bias and precision. Each project participant s accuracy criteria are presented in Appendix G. A7.2.5 Completeness Data completeness for all measurements is calculated on the basis of the number of valid measurements made out of the total possible number of measurements. Data completeness is calculated as follows: % Completeness = Number of valid measurements x 100 Total possible measurements Each contractor will provide an assesment of data completeness in the draft and final reports provided to the UT Austin Project Manager. A7.2.6 Representativeness Representativeness is the extent to which a set of measurements reflects actual conditions for a specific application. The representativeness objective for the data is not stated numerically as a quality assurance objective because quantitation is generally not possible. The measurement results from each test will represent the actual flare flue gas under low flow conditions. A7.2.7 Comparability Comparability is achieved when the results are reported in standard units to facilitate comparisons between the data. In order to accomplish this objective, the reporting units for all measurements of this project are contained in Appendix G. Comparison of many measurements will be reported by UT Austin in the draft report documents, which will then be incorporated into the final report. Revision No. 0 Page 2 of 2 8/10

47 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A8 A8 SPECIAL TRAINING REQUIREMENTS/CERTIFICATION Specialized training/certification required to operate any instruments will be the responsibility of each subcontractor. Revision No. 0 Page 1 of 1 8/10

48 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A9 A9 DOCUMENTATION AND RECORDS All subcontractor personnel working on this project are expected to maintain records for three years from date of termination of the subcontractors contract, that include sufficient information to reconstruct each final reported measurement from the test data originally gathered during the flare tests conducted at the John Zink Company test facility for a period of no less than three (3) years. These records include but are not limited to information (raw data, electronic files, and/or hard copy printouts) related to media preparation, sampler calibration, sample collection, sample handling (Chain-of-Custody and processing activities), measurement instrument calibration, quality control checks of sampling or measurement equipment, "as collected" measurement values, an audit trail for any modifications made to the "as collected" measurement values, and traceability documentation for reference standards. Difficulties encountered during data collection, sampling or analysis need to be documented and must clearly indicate the affected measurements. All electronic versions of data sets should reflect the limitations associated with individual measurement values. A9.1 Mechanisms for Documentation of Procedures and Objectives Comprehensive Flare Study Project Quality Assurance Project Plan. Published guidance (Code of Federal Regulations, U.S. Environmental Protection Agency [EPA] documents, and EPA Quality Assurance Handbooks). Method-specific Standard Operating Procedures (SOP s) A9.2 Mechanisms for Record Keeping The following electronic or hard copy documents are maintained by the analysts (e.g., Chainof-Custody forms in the laboratory with final data), field operators (e.g., activity logs), or data managers (e.g., electronic logs). All hard copy documentation is recorded in non-erasable ink, with any changes denoted by a single line through the entry, the initials of the person making the change, and the date. Sampling information and Chain-of-Custody forms; Instrument calibration data forms; Electronic run logs; Electronic and manual daily activity logs; Electronic and manual data processing and validation logs; Electronic and manual data management activity logs; Records of assessment, such as performance evaluation records; and Exception reports. Revision No. 0 Page 1 of 2 8/10

49 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section A9 A9.3 Data Reporting Turnaround Time After the end of the flare test, all measurements made as specified in the contract shall be provided in a preliminary form 15 days after the flare tests are completed. The final report of measurements will be due 30 days after the flare tests are completed. A9.4 Data Storage All data shall be stored by the contractor for no less than three years from termination of the contractor s contract with UT. Electronic copies of all measurements shall be provided with the project final report in MS Excel and MS Word 2003 or newer. The format for all video images shall be MS Movie (.wmv), Apple QuickTime (.mp4 or newer) or other format approved in advance of collection of any data by the UT Project Manager. Revision No. 0 Page 2 of 2 8/10

50 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B1 B1 SAMPLING PROCESS DESIGN (EXPERIMENTAL DESIGN) B1.1 Study Site Design The laboratory that has been selected to conduct the field tests is the John Zink Company, LCC flare test facility in Tulsa, Oklahoma. The John Zink Company, LCC has a long history of testing combustion equipment. Built in 1991 and the subject of several major expansions, the current John Zink Company, LCC test facility provides rare capabilities. The facility includes multiple furnaces for the testing of process burners, boiler burners, and duct burners. It also includes an area for thermal oxidizer technology testing as well as a flare testing area. The flare testing area is a comprehensive, industrial-scale research and development test center with stateof-the-art equipment that can accommodate steam-assisted flares, air-assisted flares, enclosed flares, and high pressure flare arrays. The flare testing capabilities include large and small flow rates, a variety of fuels and fuel mixes, and a data acquisition system which records everything from flows, pressures, and temperatures, to radiation and noise. This facility is capable of characterizing the performance and operation of flares over a wide range of operating conditions and also has the flexibility and its personnel have the expertise to design and conduct flare tests safely at conditions never used before, i.e., low waste gas flow rates with extensive instrumentation of the waste gas flow and the steam flow rates, while at the same time accommodating the remote sensing technologies concurrently during the flare tests. This flexibility and expertise will be critical to the success of this project. The following URL has more details about the Zink facility. The flare test equipment utilized in the Comprehensive Flare Study Project consists of a fuel supply system, fuel metering system, steam supply system, steam metering system, steam assisted flare, air supply and measuring system, air assisted flare, and a data acquisition system. The study site is designed so that an unobstructed view of the project s test flare s flue gas may be seen from all directions within an arc of at least 180 about the centerline of the flare burner. The challenge is to allow for in situ sampling of the flare flue gas during each test that will not block the view required by the remote sensing technologies. See Appendix A for the study site diagram and satellite aerial photo. All measurements taken are classified as critical to meet project objectives. All steam-assisted and air-assisted flare tests proposed comply with 40CFR, Part requirements for BTU and exit velocity for the flare flue gas emissions and presence of a pilot flame on each flare tip. (refer to Appendix D). At the sponsor s direction, a 36 diameter flare burner with tip exit 13 feet above the ground will be used for the steam-assisted flare and a 24 diameter flare burner with tip exit 33 feet above the ground will be used for the air-assisted flare tip. These flare burners are designed for maximum capacities of 937,000 and 144,000 pounds per hour, respectively. At these heights, both direct and remote sensing technology measurements are possible. The steam-assisted flare tip is a John Zink model QSC. This tip design has an upper ring for injecting steam around the perimeter of the tip. It also has a center steam injection nozzle for injecting steam inside the body of the tip. This tip is equipped with three natural gas pilots. The air-assisted flare tip is a John Zink model LHTS. This tip design receives the fuel gas through a central riser. The assist air is delivered in an annulus around the fuel gas riser. This tip is equipped with three natural gas Revision No. 0 Page 1 of 4 8/10

51 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B1 pilots. Drawings of the steam- and air-assisted flare tips are included in Appendix L and L1. In Appendix A are the diagrams and aerial photos of the John Zink facility showing the site layout and locations of the project participants making direct and remote sensing measurements of the flare flue gas during each test. The control room that is marked in the photo of Appendix A will be where the flare system operator will monitor steam, air and waste gas flow to each flare. Only one flare burner will be operated during any test. Operation of the equipment consists of manually adjusting the fuel control valves until the desired flow rate is achieved. Steam flow is also manually controlled by adjusting the steam control valves. Manual control is used on both the fuel and the steam in order to achieve fine control of flows. Air flow to the air assisted flare is manually controlled by the operator. The operator will command between 10 to 100% rotation via the computer. Waste gas flow control will be monitored through the use of calibrated orifice plates and transmitters. Mixing of the waste gas components (Tulsa Natural Gas, propylene and nitrogen) is accomplished by injecting the components into a mixing manifold which contains a mixing device. The flow of each component is carefully monitored to achieve the desired lower heating value (LHV) for each test. Temperature of the steam will be monitored through the use of calibrated thermocouples and steam pressure will be monitored using calibrated pressure transmitters. The flare flue gas will be characterized by direct, real-time measurements and by remote sensing technologies. The flare flue gas for this study is defined as the exhaust gas that exits from the flare tip. Direct measurements will be conducted by continuously extracting samples of the flue gas using the plume sampling device shown in Appendix s J and J1. The device consists of, an inlet cone, a sample preparation section, a sample extraction section, and eductor. The inlet face of the cone is 20 diameter tapered to a 12: outlet. The 12 outlet is connected to a 90 degree elbow which in turn is connected to the inlet of the sample preparation section. There are three exposed junction temperature elements located equidistant around the perimeter of the inlet to the 20 diameter cone to measure the temperature of the flue gas as it enters the cone. Attached to the elbow is the sample preparation section, which consists of a 12 diameter Vortab insertion type flow mixer to homogenize the sample. This mixer is 3 feet into the inlet of the 9.5 long sample preparation section which consists of a 12 diameter stainless steel straight pipe. The sample preparation section will mix the gases to obtain a uniform composition. At the exit of the sample preparation section is the extractive sampling section. The sampling section is 1.5 feet long and consists of a pitot tube for measurement of the flue gas velocity in the apparatus, an exposed junction temperature element, and two flue gas sampling probes. The flare flue gas sample that will be analyzed for emissions composition will be obtained from this location. Downstream of the sampling section 7.5 feet is the end of the flare flue gas sampling device where the eductor is attached. The eductor utilizes compressed air to induce a flow through the apparatus. By varying the pressure of the compressed air at the eductor, the flue gas eduction rate can be varied. Pipe clamps are used to lift the apparatus with a crane. Rotation of the pipe in the clamps allows orientation of the cone inlet so the inlet plane can be positioned either horizontal to the ground to collect flue gas exiting vertically from the flare, perpendicular to the ground to collect flue gas exiting and traveling horizontally from the flare due to strong cross winds, or at any angle in-between. During the flare operation, the sampling device cone inlet will be positioned outside the visible flame and downwind of the end of the visible flame approximately one flame length. The Revision No. 0 Page 2 of 4 8/10

52 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B1 position of the inlet to the cone will be adjusted so that the average temperature readings of the three inlet temperature thermocouples is no more than 500 F or the temperature determined to be the point where all the combustion has ceased by O 2 concentration measured in the sampling device using a O 2 monitor and CO 2 concentration measured in the sampling device using a CO 2 monitor. It is expected that the oxygen content of the homogenized gases will be between 18.0% and 20.5%. Lower oxygen levels could indicate that the sampling is being conducted in an area where combustion reactions may still be occurring. The carbon dioxide concentration will also be monitored to help position the entrance to the cone of the sampling device. A low O 2 concentration and a high CO 2 concentration will be an additional indication that the sampling device is located at a point in the flare flue gas where no further combustion or oxidation will occur. Analysis of the flue gas samples collected at this location by Aerodyne and TRC will be used to establish the true DRE and CE of each test condition. The correct positioning of the sampling apparatus will be determined by UT Austin and Aerodyne after careful review of the sample probe inlet temperature readings, CO 2 and O 2 concentrations, and the concentration of the other products of combustion determined by the carbon fraction analysis method as described in Appendix B8. If it is determined through this analysis that samples from multiple points in the flare flue gas are needed, then multiple samples will be collected and the analysis from the multiple samples integrated to calculate a DRE for the test condition. The remote sensing technologies will be located with clear line-of-sight to the flare tip at various (as needed by each vendor) distances away from the flare test apparatus. The remote sensing technologies will be allowed maximum flexibility to relocate instruments to account for wind direction or other physical attributes of the flare test setup so as to maximize performance of their instruments. LSI will be collecting digital IR, and visible frequency ranges of images of the flare flue gas constantly during each air and steam flare test. These images will be posted in the control room and observers conference room live during each test. One set of cameras consisting of digital IR, and visible frequency range camera will be aimed perpendicular to the flare flue gas exit from the tip. A second set of digital IR and visible frequency range cameras will be mobile around the flare testing pad and used primarily to view the flare flue gas from and angle perpendicular to the first set of cameras. The lenses for the digital IR cameras will be determined on a case by case basis. Selection of the lens depends on many factors that can influence the quality of the image desired. B1.2 Study Site Design Rationale The study site design was developed to conduct tests on an air-assisted and steam-assisted flare tips at low waste gas flow rates that will provide data to answer as many of the study objectives as possible. This site design will also allow up to 6 types of remote air quality sensing (monitoring) devices the ability to collect data at the same time without interfering with each other as well as the continuous direct measurements of the flare flue gas. B1.3 Measurement Validation Revision No. 0 Page 3 of 4 8/10

53 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B1 Appendix B references the measurement methods used to obtain data in this project. Standard methodology has been followed whenever possible. Sampling and validation efforts are described in Sections B2 and Appendix B. All data will be reviewed by the respective subcontractor s quality assurance officer for acceptable data quality and compliance with project objectives before inclusion in the final project report. The meteorological (met) data used for this project will be that measured using Aerodyne Research instrumentation. Aerodyne will use the adjacent John Zink Company, LCC and Tulsa airport met data as a data validation tool. If there is a difference in the Aerodyne, Zink, Tulsa airport meteorological data, the Aerodyne validated met data will be used as the valid data set for this project. UT Austin will conduct an on-site audit of all the Aerodyne meteorological equipment prior to beginning testing and after the testing ends. Project data may be invalidated due meteorological data on a case by case basis. Meteorological data will not be used to disqualify flare emissions data. When the acceptance criteria for a measurement are not achieved, the corrective action criteria will be followed by each project participant as noted in Appendix H. Revision No. 0 Page 4 of 4 8/10

54 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B2 B2 SAMPLING METHODS REQUIREMENTS This section addresses the sampling methods for the equipment and their operating procedures. It should be recognized that some of the procedures might change over the course of the project if logistical, technical or quality related difficulties are encountered. A tabulation of all measurements that will be made during this project, the instrument that will be used to make the measurement, manufacturer of the instrument, model number and measurement location, and subcontractor who will be making the measurement, is contained in Appendices G and H. A schematic showing the location where these measurements will be made is included in Appendix C1, C2, C3. B2.1 Continuous Methods B2.1.1 John Zink Company, Inc. The following flare operational parameters will be monitored and reported to the project members by the John Zink Company, Inc., to verify that the flare system is producing the predetermined flare conditions during every operational plan series. Waste gas stream total flow, total steam flow rate (upper plus center steam), flue gas temperature (three measurements at entrance of sampling device), air flow rate for air-assisted flare, and pilot gas flow rate will be measured by the flare operation facility. The flare waste gas average tip exit velocity (FV) will be calculated by the flare operation facility. The quality control assessment methods of the instruments necessary for proper operation of the flare facility are contained in Appendix B5. These data will be recorded by the data acquisition system (DAS) located in the flare operation facility control room. Reporting of these data will be through and compact disk (CD)/flash drive media to the project participants and UT Austin project manager at the end of each test day and as part of the John Zink Company, LCC final report. The John Zink Company will make video recordings of the flare flue gas during each test series. These videos will be used as visual examples of flare flue gas conditions during each operational plan series. Additionally, the direct and remote sensing technology companies will use these videos during analysis of their data. Refer to Appendix L for the Steam Tip Diagram, Appendix L1 for the Air Tip Diagram, and Appendix M for the ZINK Facility Waste Gas and Steam Flow Diagram. B2.1.2 Aerodyne Research, Inc. Aerodyne Research Inc. (ARI) will use one-second detection methods (continuous) in conjunction with GC methods to determine the flare flue gas constituents in real time through direct extractive sampling of the flue gas. The continuous measurements will allow immediate review of the data to determine if the sample is representative of the flue gas or a portion of the flue gas. The indicators that determine if the location of the sample collector is representative of either the whole flue gas or a portion of the flue gas is described in Section B1. The methods to make these measurements are contained in Appendix B4. Revision No. 0 Page 1 of 6 7/10

55 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B2 The direct measurement of VOC will help close the carbon mass balance and produce VOC/CO, VOC/CO 2 ratios in real time through the use of the in-the-flue gas extraction device. The device was designed by UT, UT s flare consultant, Aerodyne, and Zink. A diagram of the completed sample extraction device is included in Appendixes J and J1. A summary of the chemical compounds, if present, in the flare flue gas that will be measured by Aerodyne and the instrument to be used for each is summarized below in Table B2.A. Table B2.A Instrument Species Timeresponse Detection limit (Based on manufacturers literature) QC-TILDAS CO, NO 2 1 sec ~400 ppt HCHO, HCOOH, CH 3 CHO, HC 2 H CH 4, C 2 H 4, C 3 H 6, HC 2 H TNMHC ThermoElectron 42i NOx analyzer NO 1 sec 500 ppt LI-COR CO 2 analyzer CO 2 1 sec 2 ppm Thermoelectron SO s < 1 ppbv 2B Tech 205 O 3 2 sec 4 ppbv PTR-MS Acrolein,Benzene,Toluene,C2 -Benzenes, C3- Benzenes,Acetaldehyde,Acet one,1,3 butadiene 1-15 s ~0.1-1 ppb Auto-GC EPA TO-14 analysis list 30 min 1 ppb Aerosol Mass size-resolved chemical 5 sec 10 Spectrometer composition and mass sec loadings of PM1 MAAP black carbon 2 sec SMPS PM size distribution 2 min CPC Particle number (D p >7 nm) 2 sec Dustrak Particle mass (80 nm < D p < 2.5 µm) 2 sec QC-TILDAS: pulsed quantum cascade tunable infrared laser differential absorption spectrometer MAAP: Multi-angle absorption photometer SMPS: Scanning mobility particle sizer CPC: Condensation Particle Counter Dp: Particle Diameter Revision No. 0 Page 2 of 6 7/10

56 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B2 B2.1.3 Telops Telops will be making remote measurements of the flare flue gas with their FIRST, Fieldportable Imaging Radiometric Spectrometer Technology, instrument. This instrument uses a Fourier transform, infrared electronics, onboard data processing system and a visible bore sight camera for detection of hydrocarbon concentrations in flare flue gas. The methods and QC protocols that will be employed by Telops are included in Appendix B12. A summary of the chemical compounds, if present, in the flare plume that will be measured by Telops is shown below in Table B2B. Table B2.B Chemical species Chemical Formula Butane C 4 H 10 Formaldehyde HCHO Formic Acid HCOOH Carbon Dioxide CO 2 Water H 2 0 Ozone O 3 Sulfur Dioxide SO 2 Ethylene (Ethene) C 2 H 4 Propylene (1-Propene) C 3 H 6 B2.1.4 Industrial Monitor And Control Corporation Industrial Monitor And Control Corporation (IMACC) will be using Passive Fourier Transform Infrared (PFTIR) and Active Fourier Transform Infrared (AFTIR) spectroscopy to remotely detect hydrocarbon concentration in the flare flue gas. The PFTIR operates by receiving the infrared radiation emitted by hot gases and producing an infrared spectrum from this radiation. The AFTIR operates by retro reflecting a source infrared radiation and receiving the infrared radiation emitted by hot gases and producing an infrared spectrum from this radiation. These spectrums then allow for evaluation of species that are present and their concentration in part per million volume (ppmv). The methods for this technology are contained in Appendix B9, B10 and B11. Prior to deploying the PFTIR and AFTIR instruments, IMACC will calibrate the instrument to known concentrations of the analytes of interest in the ranges of concentrations expected to be encountered during the flare operation series. In addition, IMACC will evaluate, in a laboratory setting, the maximum Combustion Efficiency (CE) that can be calculated using Revision No. 0 Page 3 of 6 7/10

57 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B2 the PFTIR and AFTIR technologies at various gas temperatures to be prepared for the exact flare operation conditions during the field campaign. B2.1.5 Leak Surveys, Inc. Leak Surveys, Inc, (LSI). will provide real-time visualization of the flare flue gas hydrocarbons recorded in standard digital video formats. Using two sets of Forward Looking Infrared (FLIR) technology along with digital videography a video record of the flare test series will be made for future analysis. One set will be aimed perpendicular to the flare flue gas emission point for each flare tip and the second set will be mobile and perpendicular to the first set of cameras but aimed at the flue gas and understand background sky conditions on all images being collected. Gas imaging will be performed from approximately 100 feet from the target at 12 ms time resolution per frame, lens size (25mm, 50mm, 100mm) will be determined during each test series. A 60 frames/sec digital camera will be collecting images of the flare flue gas for each steam and air assisted test. Figure B2.A shows typical black and white images that are produced using the IR technology. Figure B2.A Revision No. 0 Page 4 of 6 7/10

58 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B2 B2.2 Non-Continuous Methods B2.2.1 TRC Concentrations of methane, ethane and propylene in the waste gas fed to each flare and in the flare flue gas will be reported every 10 minutes through the use of a Gas Chromatograph (GC) equipped with a flame ionization detector (FID) and an electronic integrator. Appendix B1 contains this analysis method. A National Institute of Standards and Technology (NIST) traceable standard containing a known concentration of methane, ethane and propylene will be used to calibrate the GC. The samples will be collected in the flare flue gas and inside the stack, prior to combustion, using a heated transfer line to transport the sample to the GC inside the TRC trailer. Reporting of these data will be through and compact disk (CD)/flash drive media to the project participants and UT Austin project representative. Mass emission rates will be calculated using a modified EPA Method 19 that is contained in Appendix B2. This reference is written to calculate sulfur compound concentrations emitted from flares and applied to methane, ethane and propylene for this project. Preliminary calculations will be available immediately after each flare operation series. The method for determining the oxygen and carbon dioxide concentrations of the flare flue gas are contained in Appendix B. There will be three sample lines used in this test. The plume sampler will draw and mix a large volume of air and two of the sample points will draw sample from this mixed flow. As the test matrix proceeds, a third probe, which does not premix portions of the plume may be sampled in order to address the scale of combustion dynamics. Sample Locations Sample probe after mixing, non dilution tip Sample probe after mixing, dilution tip Sample probe before mixing, dilution tip The non-dilution probe tip refers to whole sample being drawn through a temperaturecontrolled manifold to the instrument packages. The temperature will be kept high to prevent combustion water or added steam from condensing. The dilution tip probes add nitrogen within 1 mm of the sample entrainment in the probe. The dilution ratio will be adjusted to keep the sum of the measured CO and CO 2 less than 2000 ppm by volume. Based on the anticipated C:H ratio in the fuel stocks for this test, this will keep the level of H 2 O below the condensation point for the anticipated ambient temperatures. The level of dilution by nitrogen will not affect the calculated DRE based on carbon mass balance. The dilution ratio will be measured by the difference between the total flow (in-line venturi flow meter) and the added nitrogen. The dilution level estimates will be corroborated by flow calculated dilution factor to the effect of dilution on stable atmospheric species such as CO 2 and CH 4. Past experience with the dilution system suggests it is an effective way to arrest trace combustion and preserve the hydrocarbon speciation (particularly oxygenated compounds such as formaldehyde) through the sample lines. Revision No. 0 Page 5 of 6 7/10

59 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B2 B2.3 Corrective Actions During the field testing, each subcontractor is responsible for operating their respective equipment and initiating minor corrective actions on equipment when required. Equipment problems are generally detected through a failed sample run or through performing routine quality control (QC) checks. The QC checks that are performed on the sampling equipment are identified in Section B5. When a major equipment problem is encountered, the company operating the equipment has the responsibility to follow up on restoring the equipment to its proper operating status. The University of Texas at Austin Project Manager shall also be informed of major problems and the corrective action employed to solve the problem and documented in the daily field inspection report. Appendix I has an example of a daily field inspection report. Any equipment problems that can result in the loss of data are addressed as high priority. All situations requiring corrective action will be documented in the site activity logs. Section B4.2 contains additional information on documentation of corrective action. B2.4 Equipment Failure In the event that one of Aerodyne s, LSI s, Zink s, or TRC s equipment fail, the UT project manager will consult with the project sponsor to determine whether to continue the flare test series, postpone, or cancel the testing. If a remote sensing technology instrument fails, testing will continue. Flare testing will not be terminated if a remote sensing technology is unable to continue making measurements. Revision No. 0 Page 6 of 6 7/10

60 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B3 B3 SAMPLE HANDLING AND CUSTODY There are no discrete samples handled by individuals for the methods in this project with one exception. There may be a need for Aerodyne to collect Summa polished evacuated canister samples. If there is a need to collect Summa polished evacuated canister samples for this project, Aerodyne will T off the main flare flue gas sampling line to collect a 6 liter sample. The canister will be shipped over night for analysis at the UT Austin, NELAC accredited, laboratory in Austin, Texas. In addition, if there is a need for an ambient air sample collected in a Summa polished evacuated canister, Aerodyne will collect a grab sample and ship it over night to the UT Austin, NELAC accredited laboratory in Austin, Texas. All analysis results will be reported to the project manager for inclusion into the final project report. Revision No. 0 Page 1 of 1 8/10

61 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B4 B4 ANALYTICAL METHODS B4.1 Analytical Procedures B4.1.1 Aerodyne Research, Inc. The analysis of any Summa polished evacuated canister samples will be accomplished through the use of the EPA TO-14 method. B4.2 Corrective Actions Documentation of the problem will be through using the site activity logs. Generally the subcontractor is responsible for or arranging for the repair of all equipment. The subcontractor shall notify the UT Austin Project Manager of all corrective action. Revision No. 0 Page 1 of 1 8/10

62 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B5 B5 QUALITY CONTROL (QC) The QC protocol for the project is discussed in this section. An attempt has been made to provide adequate information from which to estimate the uncertainty and potential limitations of measurements generated by the instruments. In an ambient air regulatory network the minimum expectation is that the QC protocol should address: Matrix effects on the measurements; Sampling system contribution to the measurements; Measurement system contribution to the measurements; and Qualitative performance of the method. Since this project is a research program, many of the methods that will be utilized are being developed for this project. In some cases, limitations of project resources restrict the ability of UT Austin to make certain quality assessment measurements. The specific QC procedures and acceptance criteria for each company referenced are contained in Appendix H. B5.1 John Zink Company, LCC The flow rate of each waste gas component (propylene, Tulsa Natural Gas or nitrogen) will be measured by in-line metering systems. The flare flue gas sample temperature will be measured by a Type K thermocouple located at the entrance to the flare flue gas sampling device. The exact quality control parameters for the measurement techniques contained in the flare flue gas sampling device are included in Appendix H. A Fluke Model 743 will be used to calibrate the pressure sensor/transmitter (PT) or differential pressure sensor/ transmitter (DPT) used in flare operation system. An ASME dimensional flow nozzle will be used to calibrate the air flow nozzle used to measure the amount of air assist provided to the air assist flare burner. The orifice diameter will be measured and compared to the markings on the orifice plate. The proper mounting will be checked visually for the flare air blower inlet pressure sensing array. Data quality control parameters for this technology are contained in Appendix G. A proposed plan to assess the Zink facility is contained in Appendix K. John Zink will use the instantaneous data feed from all the flow control devices to determine that the specified test condition has been achieved to within ± 5% of the requested test conditions. A test condition is not considered stable until the flow control devices are within ± 5%. B5.2 TRC The gas chromatorgraph s (GC) response will be checked and adjusted in the field prior to the collection of data using a multi-point calibration error test. The linearity of the GC instrument will be checked by first adjusting the zero and span responses to zero (nitrogen) and an upscale calibration gas of propane in the range of the expected concentrations. The GC s response can then be challenged with other calibration gases of known concentration and accepted as being linear if the response of the calibration gases agrees to within ± 2% of range of the predicted value if funding allows for this type of check. Before and after each test run, the analyzers will be checked for zero and span drift Revision No. 0 Page 1 of 2 8/10

63 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B5 (calibration drift checks). This check allows each test run to be bracketed by calibrations and documents the precision of the data just collected. Anytime an adjustment is made to an analyzer, a calibration error test will be performed. Measurement data quality objectives for this technology are contained in Appendix G. B5.3 Aerodyne Research, Inc. Please refer to Appendix F for the Aerodyne quality assurance step for their analytical methods. Measurement data quality objective for this technology are contained in Appendix G. B5.4 Telops The QC checks will be performed using a black body on the flare stack. Measurement data quality control objectives for this technology are contained in Appendix H. B5.5 Industrial Monitor and Control Corporation A log will be kept with the beginning and ending times of activities. Sample, background and calibration measurement times will be recorded. Start time, time of stable IR signal and test end time will be recorded for each test condition for time averaging of the data. Before the unit is shipped to John Zink, a laboratory calibration of the unit will occur to allow comparison to the field calibration so that any instrument adjustment that may be necessary will take place. On a daily basis temperature calibration of the IR will conducted using a black body source at known temperatures. Before and during each test series, a sky background spectrum will be collected to enable correction of the IR signal to the background signal. During all tests measurements, the data will be manually checked for completeness and accuracy to insure that the IR system is operating in a quality manner. Measurement data quality control objectives for this technology are contained in Appendix H. B5.6 Leak Surveys, Inc. Daily, the camera systems will be allowed to condition for 30 minutes before actual video footage will be recorded. If during this conditioning time it is discovered that the images are not clear and concise by operator review, adjustment to the systems will be accomplished to achieve a quality image determined by the operator. Before each test, a propane cylinder will be opened and the camera will be aimed at the cylinder so that the camera can prove it is seeing hydrocarbons. Revision No. 0 Page 2 of 2 8/10

64 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B6 B6 INSTRUMENT/EQUIPMENT TESTING, INSPECTION, AND MAINTENANCE REQUIREMENTS This section describes the procedures to ensure and maintain the readiness of the field equipment throughout all phases of the project. B6.1 Instrument Testing/Inspection Each contractor that is providing an instrument to assess the flare flue gas will test and inspect their monitors in accordance with company standards. B6.2 Preventive Maintenance Procedures This section describes the routine preventive maintenance procedures performed on equipment used to assess the flare flue gas. This project will be conducted over a two week period, due to the short operation of the equipment, preventive maintenance may not be necessary for all instruments during the testing period. B6.2.1 John Zink Company, Inc. Flow metering system maintenance will be conducted, as necessary, in accordance with the manufactures requirements. B6.2.2 TRC Routine preventative maintenance procedures and schedules for the Gas Chromatograph (GC) are described in the instrument service manuals. The TRC Standard Operating Procedures for monitoring using a GC are contained in Appendix B. B6.2.3 Aerodyne Research, Inc. If preventative maintenance is required, it will be in accordance with recommendations from the manufacturer of the analysis and meteorological equipment. B6.2.4 Telops Routine preventive maintenance procedures are not available since this is an experimental technology. B6.2.5 Industrial Monitor And Control Corporation Routine preventive maintenance procedures are not available since this is an experimental technology. Revision No. 0 Page 1 of 2 8/10

65 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B6 B6.2.6 Leak Surveys, Inc. If routine preventive maintenance procedures are required, they will be in accordance with the instrument manual direction. B6.3 Corrective Maintenance Procedures This section describes the corrective maintenance procedures performed on the technologies participating in this flare project. B6.3.1 John Zink Company, Inc. Corrective maintenance procedures for the flow monitors follow the Manufacturer s recommendations in the device service manuals. B6.3.2 TRC Corrective maintenance procedures for the gas chromatograph monitor follow the manufacturer's recommendations in the instrument service manuals. B6.3.3 Aerodyne Research, Inc. Corrective maintenance procedures for the mobile lab follow the manufacturer's recommendations in the service manuals. B6.3.4 Telops Corrective maintenance procedures for the spectral imaging equipment follow the manufacturer's recommendations in the instrument service manuals. B6.3.5 Industrial Monitor And Control Corporation Corrective maintenance procedures for the infra red camera follow the manufacturer's recommendations in the service manual. B6.3.6 Leak Surveys, Inc. Corrective maintenance procedures for the infra red camera will follow the manufacturer's recommendations in the service manual. B6.4 Availability of Spare Parts A minimum stock level shall be maintained by the subcontractors and stored in the area designated by each subcontractor. Revision No. 0 Page 2 of 2 8/10

66 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B7 B7 INSTRUMENT/EQUIPMENT CALIBRATION AND FREQUENCY This section identifies the instruments, tools, and standards whose quality must be controlled, the methods and frequency of calibration, the calibration and performance standards, and the traceability of the standards. It is the responsibility of each participant to maintain documentation regarding the traceability of the standard materials used as references for calibration purposes via logbooks or electronic logs. B7.1 Calibration B7.1.1 John Zink Company, Inc. Second source standards are used by Zink for this project. Calibration frequency is contained in Appendix H. B7.1.2 TRC The methane, ethane and propylene calibration gases will be derived from secondary standard span gas bottles that have been certified by the vendor. The primary cylinders are traceable to a National Institute of Standards and Technology (NIST) Standard Reference Materials (SRMs). Calibration frequency is contained in Appendix H. B7.1.3 Aerodyne Research, Inc. The calibration of the formaldehyde measurement is performed using a calibrated permeation source diluted into zero air. NIST traceable sources of Carbon Monoxide, Carbon Dioxide, Nitrogen Oxide and Sulfur Dioxide are used as standards. An effusive source of naphthalene, mounted inside the ionization region of the mass spectrometer provides a low concentration constant background signal which is used to provide a mass calibration marker at 128 amu. The PTR-MS (Ionicon Analytic GMBH) is a chemical ionization based mass spectrometry method that utilizes H 3 O + as a reagent ion. These H 3 O + reagent ions are pulled through the air sample by an electric field where they can react via proton transfer reactions. The reagent ions and the resulting proton transfer reaction products are mass selected and detected using the mass spectrometer. Calibration frequency is contained in Appendix H. B7.1.4 Telops The gas signatures are from a library of spectral data. Calibration frequency is contained in Appendix H. B7.1.5 Industrial Monitor And Control Corporation Before the unit is shipped to John Zink Company, Inc., a laboratory calibration of the unit will occur to allow comparison to the field calibration so that any instrument adjustment that may be necessary will take place. On a daily basis, temperature calibration of the IR will conducted using a black body source at known temperatures. Revision No. 0 Page 1 of 2 8/10

67 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B7 Before and during each test series, a sky background spectrum will be collected to enable correction of the IR signal to the background signal. Calibration frequency is contained in Appendix H. B7.1.6 Leak Surveys, Inc. There are no standards used by the Leak Surveys company for this project. B7.2 Traceability Each contractor will provide documentation as requested of the traceability of their standards. B7.2.1 John Zink Company There are no standards used by the J. Zink company for this project. B7.2.2 TRC The primary cylinders are traceable to National Institute of Standards and Technology (NIST) Standard Reference Materials (SRMs). B7.2.3 Aerodyne Research, Inc. NIST-traceable certified sources of Carbon Monoxide, Carbon Dioxide, Nitrogen Oxide and Sulfur Dioxide are used as standards. An effusive source of naphthalene, mounted inside the ionization region of the mass spectrometer, provides a low concentration constant background signal which is used to provide a mass calibration marker at 128 amu. The PTR-MS (Ionicon Analytic GMBH) is a chemical ionization based mass spectrometry method that utilizes H 3 O + as a reagent ion. B7.2.4 B7.2.5 B7.2.6 Telops The gas signatures are from a library of spectral data. Industrial Monitor And Control Corporation The gas signatures are from a library of spectral data. Leak Surveys, Inc. There are no standards used by the Leak Surveys, Inc. for this project. B7.3 Documentation It is the responsibility of each subcontractor to maintain documentation regarding the traceability of the standard materials used as references for calibration purposes. Site logbooks, electronic logs, and data are maintained by each company s operators. Revision No. 0 Page 2 of 2 8/10

68 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B8 B8 INSPECTION/ACCEPTANCE REQUIREMENTS FOR SUPPLIES AND CONSUMABLES This section identifies the quality objectives for supplies and consumables to ensure high, valid data return. B8.1 Sampling Supplies Each subcontractor is required to order/obtain, prepare, and store all required sampling materials and supplies for operating analyzers and samplers. B8.2 Standards Standards for the analytical calibrations shall be ordered by the subcontractor on an as needed basis either from EPA or from commercial suppliers who provide standards meeting applicable EPA requirements. Standards are either traceable to National Institute of Standards and Technology (NIST) or are certified by the vendor and certification of traceability is kept on file by each laboratory. B8.3 Spare Parts Each subcontractor shall procure, store and maintain an inventory of spare parts for all equipment based on equipment manufacturer s recommendations, experience, and project history. Revision No. 0 Page 1 of 1 8/10

69 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B9 B9 DATA ACQUISITION REQUIREMENTS (NON-DIRECT MEASUREMENTS) All data for this project are expected to be direct measurements. Revision No. 0 Page 1 of 1 8/10

70 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B10 B10 DATA MANAGEMENT Continuously sampled data are managed by each subcontractor within their own data management system. Subcontractor data will be handled in flat file format in Excel 2003 or newer spreadsheets. The UT Quality Assurance officer will compile all data collected under this QAPP so that the data may provide for analysis of DRE and CE. B10.1 Zink Data collected by ZINK will be stored at the ZINK facility for up to three years. The data will be reported to UT in Excel 2003 or newer spreadsheets after the subcontractor validates the data. In addition, the subcontractor s data validators keep individual notebooks of corrections to data files. B10.2 TRC Data will be collected on the TRC data acquisition system and reported to UT in Excel 2003 or newer spreadsheets after the subcontractor validates the data. In addition, the subcontractor s data validators keep individual notebooks of corrections to data files. After the validated data have been archived, project personnel continue to review the data for higher levels of data validation. If there is clear evidence that a problem exists that was not detected by earlier stages of data validation, then the Project Quality Assurance Officer may choose to invalidate the data. B10.3 Aerodyne Data will be collected on the Aerodyne mobile lab data acquisition system and reported to UT in Excel 2003 or newer spreadsheets, after the subcontractor validates the data. In addition, the subcontractor s data validators keep individual notebooks of corrections to data files. After the validated data have been archived, project personnel continue to review the data for higher levels of data validation. If there is clear evidence that a problem exists that was not detected by earlier stages of data validation, then the Project Quality Assurance Officer may choose to invalidate the data. B10.4 Telops Data will be collected on the Telops data acquisition system and reported to UT in Excel 2003 or newer spreadsheets after the subcontractor validates the data. In addition, the subcontractor s data validators keep individual notebooks of corrections to data files. After the validated data have been archived, project personnel continue to review the data for higher levels of data validation. If there is clear evidence that a problem exists that was not detected by earlier stages of data validation, then the Project Quality Assurance Officer may choose to invalidate the data. B10.5 IMACC Data will be collected on the IMACC mobile lab data acquisition system and reported to UT in Excel 2003 or newer spreadsheets after the subcontractor validates the data. In addition, the Revision No. 0 Page 1 of 3 8/10

71 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B10 subcontractor s data validators keep individual notebooks of corrections to data files. After the validated data have been archived, project personnel continue to review the data for higher levels of data validation. If there is clear evidence that a problem exists that was not detected by earlier stages of data validation, then the Project Quality Assurance Officer may choose to invalidate the data. PLAN to Ensure Data Processing Without Multiple Algorithms Purpose: To insure the raw spectral data collected during each test are post-processed with 1) the normal industrial process data usually provided, i.e., gas flow and gas composition, as if these tests were conducted at an industrial facility 2) that a copy of the algorithm used by each remote sensing technology is provided to the UT project manager prior to the field campaign so that an independent post processing of test data may occur. Method: The UT flares test coordinator will follow up to be sure each remote sensing technology provides the post processing algorithm prior to the first day of testing. The flares test coordinator will monitor the communication between the test facility personnel, direct measurement personnel and remote sensing measurement personnel before, during and after each test sequence. Status Checks: During the field inspection data acquisition each morning, a visual review of the remote sensing measurement stations will be made to verify that there are no cell phone or walkie talkie devices to allow for communication with anyone who could relay test condition information to the remote measurements teams. If these types of devices are discovered, the flares test coordinator will have these devices removed by the measurements company. Periodically during the day, the field tests coordinator will have the remote measurements companies download recently generated data to a removable drive for later comparison to the data presented in the company s preliminary report. If the comparison proves that the preliminary post processed data does not match the independent post processing of the data, then this company s data will be deleted from the project final report with a footnote replacing the data stating post processed data were not acceptable. B10.6 Leak Surveys Data will be collected on the Leak Surveys data acquisition system and reported to UT in a video format after the subcontractor validates the data. In addition, the subcontractor s data validators keep individual notebooks of corrections to data files. After the validated data have been archived, project personnel continue to review the data for higher levels of data validation. Revision No. 0 Page 2 of 3 8/10

72 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section B10 B10.7 Acceptability of the Hardware/Software Configuration Each project participant will determine the process and equipment they require to produce quality data. B10.8 Data to Users Data will be provided to users through the issuance of the final project report within 30 days after the last test. Revision No. 0 Page 3 of 3 8/10

73 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section C1 C1 ASSESSMENTS AND RESPONSE ACTIONS Review of process performance is done on a continuous basis. This section addresses the assessment and response actions for this project. C1.1 Technical Systems Audit Due to limitations in the project funds, technical systems audits are not scheduled to be performed. If funds become available, technical systems audits may be performed by UT Austin or a subcontractor to UT Austin during the project field campaign. C1.1.1 Field Technical Systems Audit Due to limitations in the project funds, field technical systems audits are not scheduled to be performed. If funds become available, field technical systems audits may be performed by UT Austin or a subcontractor to UT Austin during the project field campaign. C1.1.2 Field Inspections UT Austin will work with each contractor to develop a company specific field inspection checklist. See Appendix I for an example of a field inspection check list. C1.2 Performance Evaluations Due to limitations in the project funds, performance evaluations are not scheduled to be performed. If funds become available, performance evaluations may be performed by UT Austin or a subcontractor to UT Austin during the project field campaign. C1.2.1 Field Assessment Due to limitations in the project funds, field assessments are not scheduled to be performed. If funds become available, field assessments may be performed by UT Austin or a subcontractor to UT Austin during the project field campaign. C1.3 Assessment of Data Quality Indicators Assessment of data quality indicators consists of (1) performance evaluations to establish data accuracy; (2) repeatability checks and collocated samplers to establish data precision; and (3) valid data return calculations to determine data completeness. Revision No. 0 Page 1 of 3 8/10

74 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section C1 C1.3.1 Specific Procedures for Assessment of Data Quality Indicators C Data Precision Assessment Precision is a measure of agreement among two or more determinations of the same parameter under similar conditions. The precision acceptance criteria for each parameter measured are presented in Appendix G. C Data Accuracy Assessment Accuracy is the closeness of a measurement to a reference value, and reflects elements of both bias and precision. The accuracy acceptance criteria for each parameter measured are presented in Appendix G. C Data Completeness Assessment For this project per cent completeness is calculated on the basis of the number of valid data runs divided by the total number of expected data runs for a given period of time. Data completeness is calculated as follows: % Completeness = Number of valid measurements x 100 Total number of scheduled measurements C1.4 Audits of Data Quality The audit of data quality (ADQ) is an examination of data after they have been collected and verified by project personnel. The ADQ documents and evaluates the methods by which decisions were made during the treatment of the data. The project Quality Assurance Manager will perform post field campaign data quality assessments. Zephyr Environmental will perform data analysis that will include audits of data quality. Revision No. 0 Page 2 of 3 8/10

75 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section C1 C1.5 Corrective Actions Corrective action is an essential part of any quality system and involves those procedures and actions taken to correct situations causing data quality to fall below established expectations. The need for corrective actions will be minimized by the implementation of the Comprehensive Flare Project Quality Assurance Project Plan, project standard operating procedures (SOPs), and the application of statistical quality control to establish appropriate data quality limits for measurement activities. Refer to Appendix H for the list of corrective action documentation for each parameter measured during this project. Corrective action can be initiated by any project participant by contacting the UT project manager. The UT project manager will make a determination on how to proceed to implement the corrective action on a case by case basis. Revision No. 0 Page 3 of 3 8/10

76 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section C2 C2 REPORTS TO MANAGEMENT C2.1 Quality Assurance (QA) Audit Reports Due to funding limitations, performance evaluations and technical systems audits are not planned. C2.2 Annual Project QA Report This type of report will not be required by this project. C2.3 Data Reports Each of the subcontractors is responsible for preparing a preliminary and final data report to the Project Manager daily and 15 days after completion of the field campaign. C2.3.1 Field Activity Reports Daily field reports for technologies that generate 10 minute averages will consist of all 10 minute averages of each parameter collected during the field campaign. Daily field reports for technologies that generate continuous data will consist of all data points collected for each parameter measured during the field campaign. C2.3.2 Quality Assurance Reports Daily instrument quality assurance reports will be provided by the project participants. Final instrument quality assurance reports will be provided by the project participants. C2.4 Reporting Schedule Each subcontractor will be required to report their data on an as needed basis during each test, at the end of each test before the daily debriefing and within 15 days after completion of the field campaign to the UT Austin Project Manager. Revision No. 0 Page 1 of 1 8/10

77 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D1 D1 DATA REVIEW, VALIDATION, AND VERIFICATION D1.1 Data Validation Data validation is an integral part of quality management. All data and the conditions under which they were recorded shall be reviewed closely by the subcontractor s data validators to determine the validity of the data and whether individual measurements can be included for statistical analysis. The criteria for validity of each measurement made during the project will be a combination of the Measurement Quality Control objectives contained in Appendix G and the Data Quality Activities contained in Appendix H. When a subcontractor determines that data are invalid, these data will not be reported to the project. If one subcontractor s data set is integral to another subcontractors set of data, neither set of data will be reported to the project. D1.1.1 John Zink Company, Inc. Initial data review and validation is performed by the test system control operator. If a problem is noticed, according to the measurement quality control objectives contained in Appendix G or the data quality objectives contained in Appendix H, the operator contacts the UT Austin project manager to discuss the next step. If determined, corrective action will be implemented by John Zink before any more tests begin. D1.1.2 TRC Initial data review and validation are performed by the test measurement equipment operator. If a problem is noticed, according to the measurement quality control objectives contained in Appendix G or the data quality objectives contained in Appendix H, the operator contacts the UT Austin project manager to discuss the next step. If determined, corrective action will be implemented by TRC before any more tests begin. D1.1.3 Aerodyne Research, Inc. Initial data review and validation are performed by the test measurement equipment operator. If a problem is noticed, according to the measurement quality control objectives contained in Appendix G or the data quality objectives contained in Appendix H, the operator contacts the UT Austin project manager lead to discuss the next step. If determined, corrective action will be implemented by Aerodyne before any more test begin. Revision No. 0 Page 1 of 3 8/10

78 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D1 D1.1.4 Telops Initial data review and validation are performed by the test measurement equipment operator. If a problem is noticed, according to the Measurement Quality Control Objectives contained in Appendix G or the Data Quality Objectives contained in Appendix H, the operator contacts the UT Austin project manager to discuss the next step. If determined, corrective action will be implemented by Telops before any more tests begin. D1.1.5 Industrial Monitor And Control Corporation Initial data review and validation are performed by the test measurement equipment operator. If a problem is noticed, according to the Measurement Quality Control Objectives contained in Appendix G or the Data Quality Objectives contained in Appendix H, the operator contacts the UT Austin project manager to discuss the next step. If determined, corrective action will be implemented by IMACC before any more tests begin. D1.1.6 Leak Surveys, Inc. Initial data review and validation are performed by the test measurement equipment operator. If a problem is noticed, according to the Measurement Quality Control Objectives contained in Appendix G or the data quality objectives contained in Appendix H, the operator contacts the UT Austin project manager to discuss the next step. If determined, corrective action will be implemented by LSI before any more tests begin. D1.2 Data Custody Custody of data is maintained by each subcontractor. All data submitted to the project manager will be archived at UT for at least 3 years after the project terminates. D1.2.1 John Zink Company, Inc. Data custody of all measurement data generated by the John Zink Company, Inc. is maintained and managed by the John Zink Company, Inc for at least 3 years after completion of testing. After all the John Zink Company, Inc. data have been reported to UT, UT will archive the data for at least 3 years after termination of the project. D1.2.2 TRC Data custody of all measurement data generated by TRC is maintained and managed by TRC. After all the TRC data have been reported to UT, UT will archive the data for at least 3 years after termination of the project. D1.2.3 Aerodyne Research, Inc. Data custody of all measurement data generated by Aerodyne Research, Inc. is maintained and managed by Aerodyne Research, Inc.. Once all the Aerodyne Research, Inc. data have been reported to UT, UT will archive the data for at least 3 years after termination of the project. Revision No. 0 Page 2 of 3 8/10

79 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D1 D1.2.4 Telops Data custody of all measurement data generated by Telops is maintained and managed by Telops. After all the Telops data have been reported to UT, UT will archive the data for at least 3 years after termination of the project. D1.2.5 Industrial Monitor And Control Corporation Data custody of all measurement data generated by Industrial Monitor And Control Corporation is maintained and managed by Industrial Monitor And Control Corporation. After all the Industrial Monitor and Control Corporation. data have been reported to UT, UT will archive the data for at least 3 years after termination of the project. D1.2.6 Leak Surveys, Inc. Data custody of all measurement data generated by Leak Surveys, Inc. is maintained and managed by Leak Surveys, Inc. After all the Leak Surveys, Inc. data have been reported to UT, UT will archive the data for at least 3 years after termination of the project. Revision No. 0 Page 3 of 3 8/10

80 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D2 D2 VALIDATION AND VERIFICATION METHODS The objective of the data processing and validation effort is to obtain quality assured databases containing the monitoring data in a consistent format. The procedures that will be implemented for data processing and validation will ensure that reported data are valid and comparable to those collected by other programs contributing data to this effort. After the validated data have been archived, data analysts continue to review the data for higher levels of data validation. If there is clear evidence that a problem exists that was not detected by earlier stages of data validation, then the Quality Assurance Officer may choose to invalidate the data. D2.1 John Zink Company, Inc. The acceptance criteria for the measurement methods conducted by the John Zink Company, Inc. are presented in Appendix H. The John Zink Company, Inc. will provide validated data to UT in an Excel spread sheet in their final report 15 days after the end of the field campaign. D2.2 TRC The acceptance criteria for the measurement methods conducted by TRC are presented in Appendix H. TRC will provide validated data to UT in an Excel spread sheet in their final report 15 days after the end of the field campaign. D2.3 Aerodyne Research, Inc. The acceptance criteria for the measurement methods conducted by Aerodyne Research, Inc. are presented in Appendix H. Aerodyne Research, Inc. will provide validated data to UT in an Excel spread sheet in their final report 15 days after the end of the field campaign. D2.4 Telops The acceptance criteria for the measurement methods conducted by Telops are presented in Appendix H. Telops will provide validated data to UT in an Excel spread sheet in their final report 15 days after the end of the field campaign.. D2.5 Industrial Monitor And Control Corporation The acceptance criteria for the measurement methods conducted by Industrial Monitor And Control Corporation are presented in Appendix H. Industrial Monitor And Control Corporation will provide validated data to UT in an Excel spread sheet in their final report 15 days after the end of the field campaign.. Revision No. 0 Page 1 of 2 8/10

81 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D2 D2.6 Leak Surveys, Inc. The acceptance criteria for the measurement methods conducted by Leak Surveys, Inc. are presented in Appendix H. Leak Surveys, Inc. will be provided validated data to UT in a PDF format in their final report 15 days after the end of the field campaign. D2.7 Data Review If any contractor s equipment operator notes any unusual or nonstandard conditions during the collection of the data, the operator enters the information in the contractor site activity log, which may be reviewed by the contractor s data validator during the data validation process. If during the data validation process the contractor s data validator determines that these conditions impact the data in a negative manner, and then the contractor s data validator may reject the data based on entries in the contractor site activity log on a case-by-case basis. If, during a review of the test data, the data validator discovers abnormal concentrations as compared to expected values based on knowledge of past data, meteorology, and other conditions, the data validator checks logs, and quality control records to determine if there is a reason to invalidate the data in question. The Project Quality Assurance Officer (or their designee) reviews all data for anomalies and asks contractor s staff to investigate the value for an assignable cause. Any findings are documented as to the reason for invalidation and limitations on use of the data, if appropriate. Revision No. 0 Page 2 of 2 8/10

82 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D3 D3 RECONCILIATION WITH USER REQUIREMENTS Problems with potential limitations of the data are handled at two different levels: (1) by the contractors data validators communicating with the monitoring technician by telephone, ; and (2) by users of data, who may question or want to verify the data quality objectives with a contractor data validator at a later date after data are processed. Issues are reconciled at the lowest level and earliest time possible. The mechanisms for communication between the producers and users of data are the telephone, the contractor site activity log, and electronic systems. The QA officers, validators, analysts, and data managers are empowered to review and question any part of the measurement process and may initiate data reviews and corrective actions to bring the process back into compliance. D3.1 Detection Limits Instrument method detection limits for each measurement are provided by each contractor. They are presented in Appendix G. The methods contained in this project are experimental; method detection limits have not been developed. D3.2 Precision Precision for each instrument method will be determined using the procedures developed by each study participant. If precision is reported, it is presented in Appendix G. Precision between in situ instruments measurements will be presented in the project final report. Precision between remote sensing technologies will not be possible since these experimental measurements do not have a precision metric reported in any literature. D3.2.1 John Zink Company, Inc. Precision for all device measurements made by the John Zink Company, Inc. is presented in Appendix G. D3.2.2 TRC Precision for all instrument measurements made by TRC are presented in Appendix G. D3.2.3 Aerodyne Research, Inc. The Precision for all instrument measurements made by Aerodyne Research, Inc. is presented in Appendix G. D3.2.4 Telops Precision for all instrument measurements made by Telops is presented in Appendix G. Revision No. 0 Page 1 of 3 8/10

83 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D3 D3.2.5 Industrial Monitor And Control Corporation Precision is not a metric reported by Industrial Monitor And Control Corporation since their measurement is made with an experimental instrument. D3.2.6 Leak Surveys, Inc. Precision is not a metric reported by Leak Surveys, Inc. since their measurement is made with a video instrument.. D3.3 Accuracy Accuracy is the closeness of a measurement to a reference value, and reflects elements of both bias and precision. D3.3.1 John Zink Company, Inc. Accuracy for all device measurements made by the John Zink Company, Inc. is presented in Appendix G. D3.3.2 TRC Accuracy for all instrument measurements for all measurements made by TRC is presented in Appendix G. D3.3.3 Aerodyne Research, Inc. Accuracy for all instrument measurements made by Aerodyne Research, Inc., is presented in Appendix G. D3.3.4 Telops Accuracy for all instrument measurements made by Telops is presented in Appendix G. D3.3.5 Industrial Monitor And Control Corporation Accuracy is not a metric reported by Industrial Monitor And Control Corporation since their measurement is made with an experimental instrument. Revision No. 0 Page 2 of 3 8/10

84 The University of Texas at Austin Comprehensive Flare Study Project Quality Assurance Project Plan Section D3 D3.3.6 Leak Surveys, Inc. Accuracy is not a metric reported by Leak Surveys, Inc. since their measurement is made with a video instrument. D3.4 Completeness For all data completeness calculations, see Sections A7.2.5 and C Revision No. 0 Page 3 of 3 8/10

85 APPENDIX A Overall Plan View of Flare Operation Facility Area The drawing in this appendix is provided by the John Zink Company as a tool to help project reviewers understand the locations of several items of interest to this project. Some items in the drawings are not part of the Comprehensive Flare Study Project but are not excluded from this standard John Zink drawing. Of interest are the two small circles in the lower middle that depict the steam and air flare locations (labeled on drawing). In addition, the facility control room (labeled on drawing) location is depicted as a square box in the middle left side of this drawing. Each of the in situ and remote sensing measurement technologies will have a clear view of the steam and air flare tips to allow for quality measurements. Revision No. 0 Page 1 of 1 8/10

86

87 APPENDIX A1 Satellite Photo of Flare Operation Facility Area, Subcontractor Location The photo in this appendix is provided by Google Earth, it is presented as a tool to help project reviewers understand the approximate locations of the major items of the Comprehensive Flare Study Project. The facility control room, identified on the center left of the photo, is where the steam and air flare (lower center of photo) operations are controlled. In the center of the photo is a depiction of the location of the crane that will move and support the flare flue gas sampling device. The other icons that are shown are representative of the approximate location of where the subcontractors will be making measurements during the flare tests. As can be seen in the photo, the Aerodyne and TRC companies are very close to the flare stacks since they will be making the in situ measurements for this project. The LSI, Telops and IMACC companies are at a greater distance from the flare stacks, since these companies will be making the remote sensing measurements for the project. Revision No. 0 Page 1 of 1 8/10

88

89 APPENDIX A2 Satellite Photo of Flare Operation Facility Area, Observer Room Propylene Storage Location The photo in this appendix is provided by Google Earth, it is presented as a tool to help project reviewers understand the approximate locations of the major items of the Comprehensive Flare Study Project. The facility control room, identified on the center left of the photo, is where the steam and air flare (lower center of photo) operations are controlled. In the top center is the location of the observers when outside. The far left center shows the location of the observers room. The location center top left shows the propylene storage area. As can be seen in the photo, the Aerodyne and TRC companies are very close to the flare stacks since they will be making the in situ measurements for this project. The LSI, Telops and IMACC companies are at a greater distance from the flare stacks, since these companies will be making the remote sensing measurements for the project. Revision No. 0 Page 1 of 1 8/10

90

91 APPENDIX B TRC Modified EPA Method 3A O 2 and CO 2 The method presented in this appendix is the documentation provided by the TRC Company, which will be making the flare flue gas in situ post-combustion oxygen and carbon dioxide concentration measurements, for the Comprehensive Flare Study Project. The difference between the EPA Method 3A and the TRC Modified Method 3A is that the TRC Modified Method 3A is applicable to all TRC measurement programs using Method 3A. Revision No. 0 Page 1 of 1 8/10

92 APPENDIX B1 TRC Modified EPA Method 18 GC Analysis The method presented in this appendix is the documentation provided by the TRC Company, which will be making the in situ flare flue gas pre-combustion and postcombustion gas chromatography (GC) analysis of the methane, ethane and propylene concentrations, for the Comprehensive Flare Study Project. The difference between the EPA Method 18 and the TRC Modified Method 18 is that the TRC Modified Method 18 is applicable to all TRC measurement programs using Method 18. Revision No.0 Page 1 of 1 8/10

93 APPENDIX B2 TRC Modified EPA Method 19 Mass Emissions Calculations The method presented in this appendix is the documentation provided by the TRC Company, which will be making the in situ flare flue gas pre-combustion and postcombustion methane, ethane and propane mass emissions calculations for the Comprehensive Flare Study Project. The difference between the EPA Method 19 and the TRC Modified Method 19 is that the TRC Modified Method 19 is applicable to all TRC measurement programs using Method 19. Revision No.0 Page 1 of 1 8/10

94 APPENDIX B3 Heating Value Calculations The spreadsheet presented in this appendix is the documentation provided by the TRC Company, which will be making the in situ flare flue gas pre-combustion and postcombustion methane, ethane and propylene lower heating value calculations for the Comprehensive Flare Study Project. Revision No.0 Page 1 of 1 8/10

95 Company: Sample ID: Time: Date: Cylinder # Gas Fuel "F Factor" & Heating Value Calculation CALCULATION OF DENSITY AND HEATING 60 F and 30 in Hg % volume Component Gross Volume % Molecular Density x Gross Weight Heating Value Fract. Component Volume Wt. (lb/ft3) Density weight % Btu/lb Fract. Btu (Btu/SCF) Btu Hydrogen #DIV/0! #DIV/0! Oxygen #DIV/0! 0 #DIV/0! Nitrogen #DIV/0! 0 #DIV/0! CO #DIV/0! 0 #DIV/0! CO #DIV/0! 4347 #DIV/0! Methane #DIV/0! #DIV/0! Ethane #DIV/0! #DIV/0! Ethylene #DIV/0! #DIV/0! Propane #DIV/0! #DIV/0! propylene #DIV/0! #DIV/0! Isobutane #DIV/0! #DIV/0! n-butane #DIV/0! #DIV/0! Isobutene #DIV/0! #DIV/0! Isopentane #DIV/0! #DIV/0! n-pentane #DIV/0! #DIV/0! n-hexane #DIV/0! #DIV/0! H2S #DIV/0! 7100 #DIV/0! Total 0.00 Average Density = #DIV/0! Gross Heating Value Gross Heating Value Specific Gravity = Btu/lb = #DIV/0! Btu/SCF = 0 CALCULATION OF F FACTORS Weight Percents Component Mol. Wt. C Factor H Factor % volume Fract. Wt. Carbon Hydrogen Nitrogen Oxygen Hydrogen #DIV/0! Oxygen #DIV/0! Nitrogen #DIV/0! CO #DIV/0! #DIV/0! CO #DIV/0! #DIV/0! Methane #DIV/0! #DIV/0! Ethane #DIV/0! #DIV/0! Ethylene #DIV/0! #DIV/0! Testing By Cubix Corporation

96 APPENDIX B4 Aerodyne QC The descriptions presented in this appendix are the analytical method quality control documentation provided by Aerodyne Research, Inc., which will be making the in situ flare flue gas measurements for the Comprehensive Flare Study Project. Revision No.0 Page 1 of 1 8/10

97 Aerodyne Specific Instrumentation Methods: 1) Quantum Cascade Tunable IR Laser Differential Absorption Spectrometer (QC- TILDAS) Quantum cascade, tunable infrared laser differential absorption spectroscopy (QC- TILDAS) has been used in numerous research contexts for the absolute, unambiguous detection of small molecules with rotationally resolved infrared vibration spectra. The wavelength region is known as the fingerprint region for its ability to be highly specific. QC-TILDAS will be used to measure methane (CH 4 ), NO 2 formaldehyde (HCHO), and CO with 1 s time resolution to fingerprint combustion sources. The Limit of Detection (LOD) at 1 s is typically ppb, except for methane, for which it is 3 ppb. The LOD is much smaller at 1 min time resolution. Although the QC-TILDAS method is fundamentally an absolute Beers-law absorption technique that does not require calibration, Aerodyne has found that the use of calibration procedures, either in the field or in the laboratory prior to deployment is a mandatory practice. The analysis software records all direct absorption spectra, which can facilitate archival re-analysis that retains substantial value if there are any future discoveries about the spectroscopy. There are some fundamental sources of systematic error though that have to be checked and monitored that can foil this method. Typically the calibration is used to fix the source of the problem. The analytical method is a number-density detection scheme, that is to say it counts the molecules along the beam path in the absorption cell. In order to convert this measurement to a mixing ratio, the cell pressure and temperature need to be accurately measured. The temperature measurement is known to be good to +/- 1K which is less than 1%. The pre-measurement checks for pressure are to compare the three manometers in the truck during stopped or very shallow flow to ensure they agree. Periodically the pressure heads are calibrated by the factory. They do require as much warm up time as the instrument itself for true readings. In previous measurement campaigns, the calibration of the HCHO measurement is preformed using a calibrated permeation source diluted into zero air. The instrument has been seen to hold its calibration within the current uncertainty in the absorption line strengths (7%). Because the calibration source in this case, the permeation tube, requires substantial time to achieve its steady state delivery, it will not be practical as an on-site calibrant. This will be done prior to the campaign. We will calibrate the carbon monoxide and nitrogen dioxide compounds during the premission set up period on-site and prior to packing up the mission at departure. We use a certified source of CO, NO and SO 2 together with a dynamic dilution system. The NO in this certified source is converted to NO 2 using an ozonizer in the Chemiluminescence analyzer described later. Plots such as that depicted in the inset for the AAFEX campaign are generated during the calibration procedures. This analysis will be in the preliminary report to the UT/TCEQ on the results of this campaign. Page 1 of 8

98 We acquired a new laser device that operates at a more favorable wavelength for the detection of C 2 H 4. The old laser was calibrated against gas chromatography during an eight week campaign and found to be within 3% of those measurements once the laser linewidth was properly measured. The ethene measurement that will be done under this work has not been calibrated yet, but similar to the formaldehyde measurement, we will perform a laboratory calibration prior to the mission deployment. The preliminary results from the new laser at AAFEX (January 2009) were quite favorable, however the calibration will be needed to verify the retrieved concentrations. 2) LiCor CO 2 Analyzer A fast (1 s time resolution) CO 2 analyzer will enable the use of above background concentrations of CO 2 + CO to identify combustion sources and calculate fuel-based emission indices for NOx, CO, HCHO, other trace gases and particulate matter (PM) from these sources. We use several instruments for monitoring CO 2. When the concentration range is in excess of 2500 ppmv, we use the Licor model 820 instrument equipped with the short absorption cell. When the concentration range is less than 2500 ppmv, we rely on two Licor model 6262 instruments. Occasionally, Licor model 840 are also used when the Licor 6262 is not available on a particular sampling line. The on-site calibration procedures are similar in principal for all of the units. The Licor 6262 have additional sorbent chemicals that are changed in the laboratory prior to deployment. In the field, the CO 2 instruments are calibrated semi-daily with the use of a CO 2 free zero air (also used to zero other instruments in the suite) and a certified calibration span gas. Typically we calibrate all of the instruments with a span gas less than 2000 ppmv. The reference tank is checked several times during the campaign, while awaiting engine-on or during the end of day data archival. When available, the 8x0 units are checked at a higher concentration tank. Experience has taught us that the instruments typically hold their calibration very well when they are not subjected to caustic pressure changes and when particulate matter is filtered from the sample gas. The sample gas is moved through the instrument in two Page 2 of 8

99 ways depending on the experiment. In the dedicated engine testing on the 1 meter probe, there is little hope of having the manifold pressure be constant. To prevent the instrument from experiencing the pressure fluctuations in the manifold, we use a sealed diaphragm pump to compress sample and push it through the instrument. The flow through the instrument is monitored with a venturi flow meter on the outlet to ensure that the instrument is always sampling continuously. In the mode when sampling from atmosphere, such as is the case in a 30 meter probe or when sampling from the front of the truck, we use the vacuum pump and a critical orifice to draw sample through the instrument. Generally, in this mode the QC-TILDAS instrument, described earlier, control valving which periodically floods the inlet with zero air. This serves several functions. The licors are zeroed by this process, the time response is evaluated and any sample line temporal offset (1-2 seconds) is measured. 3) Proton Transfer Mass Spectrometer (PTR-MS) A PTR-MS will measure various VOCs, including the HAPs: benzene, toluene, the sum of ethylbenzene and the xylenes (BTEX), as well as additional HAPs such as acetaldehyde, acetone, acetonitrile and methanol. All measurements have 2 s time resolution. The LOD for benzene and other HAPS is typically ~2 ppb at 1s and ~0.3 ppb at 1 min. The PTR-MS (Ionicon Analytic GMBH) is a chemical ionization based mass spectrometry method that utilizes H 3 O + as a reagent ion. A schematic of the PTR-MS is provided in inset. The instrument consists of an ion source, a drift tube reaction region and a quadrupole mass spectrometer. H 3 O + reagent ions formed in the hollow cathode discharge ion source are electrostatically injected into the drift tube through which the sampled air stream is continuously passed at reduced pressure, 2.0 mbar. These H 3 O + reagent ions are pulled through the air sample by an electric field where they can react via proton transfer reactions, depicted in the equation below, with those components (R) in the sample having proton affinities greater than that of water. The reagent ions and the resulting proton transfer reaction products are mass selected and detected using the mass spectrometer. Page 3 of 8

100 H 3 O + + R RH + + H 2 O Aircraft turbine engine exhaust contains a complex mixture of hydrocarbons.. In a 1994 study there were identified 57 different organic compounds in the exhaust of a CFM-56 engine of which 45 are expected to be ionized and detected by the PTR-MS. Many of these components do not provide unique ion signatures in the PTR-MS and thus are not distinguishable or identifiable on the basis of ion mass alone. Interpretation of the PTR-MS analysis of aircraft turbine engine exhaust is aided considerably and shaped heavily by detailed chemical composition information taken from former studies. By knowing what compounds are present along with the ionization dynamics for those compounds, it is possible to interpret the mass spectrum and reasonable decisions can be made regarding which exhaust compounds can be accurately quantified using the PTR- MS. A compilation of the compounds and their emission data that were examined by the PTR-MS technique is summarized below. The data in the comparison table has been normalized to formaldehyde (measured independently by tunable infrared laser differential spectroscopy) because it has been found to remove the effects of ambient temperature and variations in power (4% - 7% rated thrust) and allows for the examination of the relative distribution of the different VOC emission products present in the exhaust. In the section that follows, each of the compounds is critically evaluated in terms of which neutral components contribute to the intensity of that ion signal. Where possible, an estimate of the fraction of the signal attributable to the different components is also provided. Proton transfer reaction product branching fractions are taken primarily from independent laboratory studies performed under experimental conditions. Similar results have been observed in all of the subsequent aircraft exhaust emission measurement campaigns. In some cases the branching fractions and ionization efficiencies are taken from other studies determined under different experimental conditions and are not necessarily totally reflective of the measurements reported within. Comparison of PTR-MS measurements with that of a 1994 campaign for CFM-56 engine at idle. Neutrals ER x /ER HCHO ER x /ER HCHO % EI x (a) g/kg LD (b) PTR-MS Spicer et al. Deviation g/kg Slope (r 2 ) Methanol 0.17 (0.97) Na (c) Na Propene (estimated see text) 0.31 (0.97) Acetaldehyde 0.18 (0.97) Page 4 of 8

101 Butenes + acrolein 0.18 (0.98) Acetone + propanal + glyoxal (0.94) Acetic acid (0.64) Na Na Pentenes 0.28 (0.98) NR (d) - Butanal + methylglyoxal (0.95) Benzene (0.94) Hexenes 0.13 (0.98) NR - Toluene (0.93) Phenol (0.88) Heptenes 0.62 (0.89) NR - Styrene (0.95) C2-benzenes + benzaldehyde (0.95) C3-benzenes (0.92) Na Na Naphthalene (0.92) C4-benzenes (0.95) Methylnaphthalenes (0.88) C5-benzenes (0.86) Dimethylnaphthalenes (0.81) Na Na (a) based on ER of formaldehyde of 5.72 x 10-4 (mole HCHO/mole CO 2 ) from Spicer et al. (b) calculated based on the PTR-MS detection limit and a DCO 2 = 2000 ppmv. (c) Na no data available (d) NR not reported 4)Nitric Oxide (NO) Chemiluminescence A commercial, high performance instrument will measure NO. The LOD is 0.5 ppb at 1 s, and 0.06 ppb at 1 min. This instrument is calibrated using the instrument manufacturer s protocol. An onboard dynamic dilution system is used with zero and span gasses that are certified prior to use. The calibration of NO (CL) is coupled with the calibration of NO 2 and CO in the QCL instruments. (5) Ozone (O 3 ) Ultraviolet Absorbance Photometer A commercial dual-beam absorbance instrument (2B Tech model 205) operating at 254 nm will measure ambient ozone with an accuracy of 2% and a time response of 2 Page 5 of 8

102 seconds. This instrument will be used to monitor ONLY the ambient O 3. It is not suited to measuring O 3 in the concentrated exhaust. If samples are collected in the highly diluted plume (dilutions of 10,000 fold) this instrument will be used to assess the conservation of odd oxygen (O 3 + NO 2 ) in order to evaluate the effective chain length in the near-field flare. Direct NO 2 can be considered as an emission of O 3, but this can be determined directly from the concentrated exhaust. The measurement of effective O 3 downwind allows for an investigation into other radical cycles that may produce O 3 very quickly. 6) Automatic Gas Chromatograph (Auto-GC) An auto-gc coupled with a flame ionization detector (FID) will measure about 60 hydrocarbons with a 5 minute sampling cycle and 25 minute analysis time (LOD of typically a few ppb) to characterize both emission plume and ambient background VOC concentrations. Although slower than other instruments, the auto-gc will be used for more detailed fingerprinting of sources. Emission plume fluxes for the VOCs measured in GC whole air samples can be estimated from the VOCs fluxes deduced from real-time TILDAS (CH 4 ) and PTR-MS measurements and the plume excess VOC concentration ratios provided by the GC whole air concentration measurements. No real-time instrument will be configured for organic sulfur species detection, but some light mercaptans and dimethylsulfide may be quantified in the GC-FID samples. Lastly, two types of Aerosol Mass Spectrometer (AMS) will be used to measure organic, sulfate, nitrate, and chloride components of particulate matter (PM) at 1 min time resolution (LOD down to a few ng m -3 ), while a Multi-angle Aerosol Absorption Photometer (MAAP) will be used to measure black carbon at 2 s time resolution (LOD of 5 µg m -3 ). Besides the meteorological instrumentation on-board the mobile laboratory, ARI also has two free standing monitors that record atmospheric pressure, temperature, humidity, GPS position, unit orientation and wind speed/direction. These units output data to a USB memory stick, are powered by automotive batteries, and are easily portable. They will be used to augment meteorological parameters upwind and downwind from the flare. During the John Zink flare tests Aerodyne Research, Inc (ARI) will be using the Davis Anemometer #7911 for wind speed and direction measurements. Wind speed is derived using the following formula: V = P(2.25/T), where V = Wind Speed in mph T = Sample speed in seconds Page 6 of 8

103 P = Number of pulses in the sample period ARI will be using a sample speed of 1 second; therefore the simplified formula is, V = P(2.25) Every rotation of the anemometer generates one pulse, and the number of pulses is recorded digitally, ARI will then take that number and multiply it by 2.25 to obtain the wind speed. In conjunction with that, Davis provides a wind speed look up table, which ARI will refer to once ARI calculates a wind speed. This table provides the true wind speed and accounts for any offset error which may be present. Wind direction is calibrated by manually. An operator will adjust the anemometer so that the dead-band of the sweep is pointing directly at the flare. ARI then will find true north and record that digital output. The #7911 anemometer has a 10 bit digital output. A fullscale output would equal 1024 bits. Therefore, (360 o )/(1024 bits) = degrees/bit ARI will then take the given output from the anemometer and multiply it by to obtain the direction in degrees. Furthermore, ARI has designed the electronics associated with the anemometer to record Relative Humidity (RH) as well as ambient Temperature (T). This is done using the Sensirion SHT1x RH and T sensor. ARI will use the high resolution mode of this device; therefore, ARI will use the following RH formula: RH Linear (%) = ( * SO RH ) + ( e-6 * SO RH ) 2 Page 7 of 8

104 Where, SO RH is the RH output from the Sensirion SHT1x. For temperatures much different than 25 o C the RH sensor requires temperature compensation. This is done with the following formula: RH True = (T C 25) * ( * SO RH ) + RH Linear The output for temperature from the SHT1x is much more linear. To calculate the true temperature from this device we will use the following formula: T = * SO T Where SO T is the digital output from the SHT1x. Page 8 of 8

105 APPENDIX B5 John Zink QC The quality control methods presented in this appendix are the documentation provided by the John Zink Company, which will be controlling all the pressure, temperature, and mass flow conditions for the Comprehensive Flare Study Project. Revision No.0 Page 1 of 1 8/10

106 John Zink Process Instrumentation Assessment Methods A. The instruments to be used by the John Zink Company to gather data for the Comprehensive Flare Study fall into 5 general groupings: 1. Pressure transmitter or multivariable mass flow transmitter for an orifice 2. Temperature Elements with or without transmitter 3. Flow Element (orifice plate, pitot tube) 4. Ambient Conditions 5. Other instruments (Thermal mass flow meter, Ultrasonic flow meter) The instruments used by John Zink to gather data for the project are to be calibrated or verified prior to the start of testing. The procedure to be used for each group is set forth below. The procedures reflect the fact that these are industrial instruments in an industrial setting. A calibration log will be kept for each instrument. B. Calibration of instruments in Group The primary instrument used in calibrating a pressure sensor / transmitter (PT) or a multivariable mass flow transmitter (MVT) is a Wika model , digital electro-pneumatic calibrator. This calibrator is certified by a third party lab on an annual basis. The accuracy of the calibrator is ±0.02% of measured value. 2. Calibration procedure for a PT a. Determine the span of the transmitter. b. Disconnect the PT from service. c. Connect PT to Wika calibrator and verify transmitter is still connected to data acquisition system. d. Connect instrument air to Wika calibrator. Utilized precision pressure controllers internal to the calibrator to dial in various pressures. e. Pressure the transmitter to 0%, 25%, 50%, 75%, and 100% of transmitter span. Allow each reading to stabilize. f. After each point has stabilized, record the value from the calibrator and the value displayed on the data acquisition system. g. If the data acquisition system value is ±0.5% of transmitter span of the calibrator value, for all five points, the transmitter passes. h. Reconnect the PT to service and attach calibration / date sticker to PT. 3. Calibration procedure for a MVT a. Determine the spans of the transmitter, both gage pressure and differential pressure. b. Disconnect the MVT from service. c. Connect MVT to Wika calibrator and verify transmitter is still connected to data acquisition system. Page 1 of 4

107 d. Connect portable computer to transmitter via HART connection and run Rosemount MVT software. e. Connect instrument air to Wika calibrator. Utilized precision pressure controllers internal to the calibrator to dial in various pressures. f. Pressure the transmitter to 0%, 25%, 50%, 75%, and 100% of transmitter spans, both gage and differential. Allow each reading to stabilize. g. After each point has stabilized, record the value from the calibrator and the value indicated for that parameter on the computer. Also record the mass flow value from the computer and the mass flow value displayed on the data acquisition system. h. If the parameter value is ±0.5% of transmitter span of the calibrator value, for all five points, for both gage and differential pressure, and the mass flow values displayed on the data acquisition system are ±0.5% of the mass flow value from the computer, then the multivariable transmitter passes the pressure tests. i. Reconnect the MVT to service. j. Temperature testing is required since an RTD input is part of the multivariable transmitter. See Calibration of instruments in Group 2 section for details. C. Calibration of instruments in Group The primary instrument used in calibrating a temperature elements (TE) and/or transmitters (TT) is a Fluke model 9102S dry well. This device is certified by a third party lab on an annual basis. The accuracy of the calibrator is ±0.45 F. 2. Calibration procedure for a TE. a. Insert element into Fluke dry well and verify element is connected to data acquisition system. b. Adjust drywell temperature to 100 F and allow sufficient time for temperature reading to stabilize. c. Record the value from the calibrator and the value indicated on the data acquisition system. d. Adjust drywell temperature to 150 F and allow sufficient time for temperature reading to stabilize. e. Record the value from the calibrator and the value indicated on the data acquisition system. f. Place element into an ice water bath and allow sufficient time for the temperature reading to stabilize. g. Record the value from the data acquisition system. h. If all three values from the data acquisition system are ±5 F of the dry well reading and 32 F, the element passes. Page 2 of 4

108 3. Calibration procedure for a TT. a. Note that the only temperature transmitter utilized for the project is connected to the GE ultrasonic flow meter. b. Verify the element is connected to the transmitter and the transmitter is connected to the flow meter. c. Insert element into Fluke dry well. d. Adjust drywell temperature to 100 F and allow sufficient time for temperature reading to stabilize. e. Record the value from the calibrator and the value indicated on the GE flow meter display. f. Adjust drywell temperature to 150 F and allow sufficient time for temperature reading to stabilize. g. Record the value from the calibrator and the value indicated on the GE flow meter display. h. Place element into an ice water bath and allow sufficient time for the temperature reading to stabilize. i. Record the value from the GE flow meter display. j. If all three values from the GE flow meter display are ±5 F of the dry well reading and 32 F, the element / transmitter passes. 4. Calibration procedure for MVT. a. Connect portable computer to MVT via HART connection and run Rosemount MVT software. b. Insert element into Fluke dry well c. Adjust drywell temperature to 100 F and allow sufficient time for temperature reading to stabilize. d. Record the value from the calibrator and the temperature value indicated on the portable computer. e. Adjust drywell temperature to 150 F and allow sufficient time for temperature reading to stabilize. f. Record the value from the calibrator and the temperature value indicated on the portable computer. g. Place element into an ice water bath and allow sufficient time for the temperature reading to stabilize. h. Record the temperature value from the portable computer. i. If all three values from the portable computer are ±5 F of the dry well reading and 32 F, the element / transmitter passes. j. Reconnect the TE to service. k. If the transmitter has passed the pressure tests, attach calibration / date sticker to MVT. D. Calibration of instruments in Group Orifice meters a. For orifice meters, the orifice diameter is measured and verified to match the markings on the orifice plate. Page 3 of 4

109 b. The orifice is inspected for damage to the inlet or outlet edges. If damage is found, the orifice is replaced. c. The diameter of the orifice run is measured and recorded. d. Utilizing Rosemount MVT software, verify the orifice size, line size, tap type, and gas are correctly input into the transmitter. 2. Pitot tubes a. The pitot tube is inspected for damage or blockage. b. Orientation of the pitot in the sample collector is verified. E. Comparison of ambient meteorological readings to those at the Tulsa International Airport. 1. Ambient temperature, humidity, barometric pressure, wind velocity, and wind direction at the test site will be recorded using existing site instruments. 2. The site data will be compared to the data recorded by the National Weather Service at the Tulsa International Airport. (Note: The Tulsa airport is approximately 3 miles west of the John Zink site. F. Calibration of ultrasonic flow meter and thermal mass flow meter. 1. Ultrasonic flow meter a. The ultrasonic flow meter is assembled by the manufacturer and sent to a 3 rd party laboratory for flow testing. b. With ultrasonic flow meters, the only parameter that affects accuracy that can be adjusted in the field is the distance between transducers. Consequently, once the transducers are set and the accuracy is determined by the flow lab, there is no drift. c. The flow testing for the ultrasonic meter will occur in July 2010 and a calibration certificate will then be available. 2. Thermal mass flow meter a. The thermal mass flow meter is calibrated in the manufacturer s flow lab. N.I.S.T. traceable calibration certificate is supplied. b. Adjustment of calibration in the field is not possible. c. The flow testing for the thermal mass flow meter occurred in May / June d. The internal diameter of the air flow duct into which the flow meter will be installed will be verified to match with the 59.5 the transmitter was calibrated to. Page 4 of 4

110 APPENDIX B6 Zephyr Calculations The calculations presented in this appendix are the documentation provided by Zephyr Environmental Company, which will be providing the emission estimates using conventional methods for the Comprehensive Flare Study Project. Revision No. 0 Page 1 of 1 8/10

111 Zephyr Environmental Description of Calculations Hourly Emission Rate: VOC VOC (lb/hr) = mass flared (lb/hr) * (100 - %DRE) / 100 Other Compounds (THC, NOx, CO) THC (lb/hr) = flow rate (scf/hr) * Heat Content (BTU/scf) * Emission Factor (lb/mmbtu) NOx (lb/hr) = flow rate (scf/hr) * Heat Content (BTU/scf) * Emission Factor (lb/mmbtu) CO (lb/hr) = flow rate (scf/hr) * Heat Content (BTU/scf) * Emission Factor (lb/mmbtu) Page 1 of 1

112 APPENDIX B7 John Zink Calculations The calculations presented in this appendix are the documentation provided by the John Zink Company, which will be providing the average exit velocity and lower heating value of the fuel being combusted for the Comprehensive Flare Study Project. Revision No. 0 Page 1 of 1 8/10

113 Description of Calculations John Zink Company LLC 20 Aug Nomenclature SCF = standard cubic foot SCFS = standard cubic foot per second SCFH = standard cubic foot per hour TNG = Tulsa natural gas Calculation of Tip Exit Velocity Assume that no steam condenses when mixed with the fuel. Assume the fuel constituents and steam all behave as an ideal gas. Measured Variables: C3H6 = Mass flow of propylene lb/hr N2 = Mass flow of nitrogen lb/hr STM = Mass flow of center steam lb/hr TNG = Mass flow of TNG lb/hr Constants and Conversion Factors Exit area of steam flare tip = ft 2 Fuel exit area of air flare tip = 1.68 ft 2 Molecular weight of propylene = Molecular weight of steam = Molecular weight of nitrogen = Molecular weight of TNG = SCF = 1 lb-mole 3600 seconds = 1 hr Standard pressure = psia Standard temperature = 68 F Calculated Values F_SCFS = Fuel mixture flow in standard cubic feet per second ft 3 /sec Flow = Total mass flow rate of steam / fuel mixture lb/hr M_SCFS = Steam / fuel mixture flow in standard cubic feet per second ft 3 /sec Page 1 of 3

114 Steam Flare Step 1: Convert steam / fuel mass flows to standard cubic foot per second units. C3H6 N2 TNG STM x = M _ SCFS Step 2: Divide the volumetric flow of steam / fuel at standard conditions by the area of the steam flare tip. Air Flare Exit velocity of the steam assisted tip (ft/sec) = M_SCFS / ft 2 Step 1: Convert fuel mass flows to standard cubic foot per second units. C3H6 N2 TNG x = F _ SCFS Step 2: Divide the total fuel volumetric flow by the area of the air assisted flare tip Exit velocity of the air assisted tip (ft/sec) = F_SCFS / 1.68 ft 2 Page 2 of 3

115 Calculation of LHV of Fuel Mixture Assume the material is mixed homogeneously. Known Parameters Mass flow of propylene lbs/hr, (C3H6) Mass flow of nitrogen lbs/hr, (N2) Mass flow of TNG lbs/hr, (TNG) Constants and Conversion Factors Molecular weight of propylene = Molecular weight of nitrogen = Molecular weight of TNG = SCF = 1 lb-mole LHV of propylene = 2152 BTU/SCF LHV of nitrogen = 0 BTU/SCF LHV of TNG = 899 BTU/SCF Standard pressure = psia Standard temperature = 68 F Step 1: Convert the mass flow of each component to SCFH C3H 6 x = C3H 6( SCFH ) N 2 x = N 2( SCFH ) TNG x = TNG( SCFH ) Step 2: Calculate the total energy content of the mixture. Note that nitrogen has zero contribution. C 3 H 6( SCFH ) x TNG( SCFH ) x899 = TotalBTU / Hr Step 3: Divide the total BTU content by total volume of mixture TotalBTU / Hr C3H 6( SCFH ) + N 2( SCFH ) + TNG( SCFH ) = BTU / SCF( mixture) Page 3 of 3

116 APPENDIX B8 Aerodyne DRE The description presented in this appendix are the destruction removal efficiency calculation documentation provided by Aerodyne Research, Inc., which will be making the in situ flare flue gas measurements for the Comprehensive Flare Study Project. Revision No.0 Page 1 of 1 8/10

117 Aerodyne Approach using Carbon Fraction to compute DRE Using the carbon fraction metric to characterize emissions allows for varying dilution of the emissions. The essential tenant in this methodology is to measure all possible routes for carbon. In a highly efficient combustion process, carbon dioxide and carbon monoxide tend to account for the majority of the fuel or process gas burned. As the combustion becomes increasingly inefficient the un-burned and partially burned compounds become important to measure. The carbon fraction CF for a specific compound Q is defined as: q moles(q) CF(Q) = (1). all carbon atoms Where q is the number of carbon atoms present in species Q. For example propane and propylene (propene) each have three carbon atoms which implies q = 3. The normalizing quantity depicted as the sum of all carbon atoms need only carry the same units as the numerator. Alone, the carbon fraction isn t useful, but for mixtures it is essentially the carbon weighted fractional content associated with species Q. Consider a fuel stock mixture that consists of 25% methane and 75% propylene by volume. If this mixture is being prepared by measured flow rates the carbon fractions are represented as: CF(methane) = = 0.1 and CF( propene) = = 0.9. This example shows that 90% of the carbon atoms in the mixture are present in propene and 10% are from methane. The destruction and removal efficiency (DRE) can also be defined for species Q as: DRE =1 CF flare _ exhaust() Q (2). CF initial _ fuel () Q The definition of CF is unitless and the CF initial exhaust can be computed using metered flow rates and verified by measurements of the pre-combustion sample stream. The CF flare_exhaust must be measured with a robust characterization of all fuel species, partial and complete products of combustion. Assuming the same fuel mixture of 25% methane and 75% propene as above, the CF flare_exhaust could be represented by the following expression. Page 1 of 4

118 3 propene CF post combustion (propene) = 1 methane + 3 propene +1 CO 2 +1 CO + q trace intermediates The expression for the post combustion CF for propene depicted above does not account for any possible entrainment of ambient air into the pre and post combustion air mass. A flare is designed to use atmospheric oxygen to mix into the fuel and provide the needed oxidant for combustion. The ideal combustion equation in a general form is, ~CH 2 (fuel or process gas) + 3/2 O 2 (ambient) H 2 O + CO 2. The background mixing ratio of CO 2 is typically 370 to 390 parts per million by volume (ppmv). In the absence of any combustion, sampling the unburned fuel+process gas following emission from the stack will contain a fraction of ambient CO 2. Once combustion is initiated, however, the mixing ratios of CO 2 (and likely CO) will be completely dominated by carbon from the fuel and process gas combustion. The mixing ratio of the fuel+process gas is initially 100%. As this is mixed with ambient air in order to be combusted the fuel+process gas mixing ratios necessarily drop. They do, however continue to exceed background CO 2 by a large margin. For example if in an extreme case, the fuel+process gas is diluted by 90 parts of ambient air as combustion is initiated. Assume that combustion efficiency is very high and the 1% fuel+process gas is converted to CO 2. In this highly diluted combustion example, the CO 2 mixing ratio would be 0.9 * 380 ppmv * 10% = ,000 = 10,342 ppmv CO 2 If a calculation merely approximated the CO 2 from combustion by subtracting the entire measured background CO 2 from the measurement, in this case it would only result in a minor error. (10, )/10,000 = 99.6%, as opposed to 100%. A very straightforward method can be used, however that does not approximate the CO 2 due to combustion but requires a simple rearrangement of the CF definition. To simplify the expression, let the prime notation denote the measurement of a particular species multiplied by the carbon number, eg. 3 * propylene (ppbv) = propylene (ppbc). Page 2 of 4

119 CF post combustion (propene) = Dividing the numerator and denominator by measured CO 2 rephrases the post combustion CF in emission ratios. This is extremely useful when variable dilution is taking place. In the schematic, the figure depicts the concomitant rise in compound Q that is associated with the increase in CO 2. This schematic depicts how the ratio in the formulation of CF above can be determined exactly using the slope of this line. propen e /CO 2 methan e /CO 2 + propen e /CO CO/CO 2 + intermediate s /CO 2 The choice of CO 2 in this illustration is appropriate for efficient combustion (high DRE). The post combustion CF expression can be rewritten as emission ratios with respect the fuel for very inefficient combustion (low DRE). The schematic emission ratio figure assumes that the DRE was constant at the various sampled points (different dilutions). The strength of this assumption for flare combustion is not known. One method of smoothing the variance in the combustion process will be to sub-sample a large volume of the flare and let it mix for ~1.5 seconds prior to measurement. In order to ensure that the sampling device does not induce a perturbation to the combustion process, only a small portion of the flare combustion exhaust will be sampled. In order to desensitize the time response in the sampling apparatus to the conditions on the flare, a large over-pull is depicted in the sample hood schematic depicted below. Page 3 of 4

120 The schematic figure depicts a cone and chimney sampler. Compressed air induces an over pull which effectively extends the sample footprint of the cone. The beam will be positioned using feedback from the monitors and chemical sensors throughout the plume depending on ambient wind conditions. Page 4 of 4

121 APPENDIX B9 IMACC CE The mathematical equation presented in this appendix are the combustion efficiency calculation documentation provided by IMACC, which will be making a remote measurement of the flare flue gas to determine the efficiency of combustion for the Comprehensive Flare Study Project. Revision No.0 Page 1 of 1 8/10

122 Appendix B.9 IMACC Combustion Efficiency Calculation Combustion efficiency is defined as: Combustion Efficiency [ CO2] = [ CO] + [ CO2] + [ THC] (1) Where [X] is the concentration of compound X. The fundamental output of a PFTIR is gas concentration times the path length of the gas or ppm*m. If the actual path length of the PFTIR beam through the plume is known, the ppm*m value can be divided by the path to provide gas concentration in ppm. Fortunately, the path length of all gases in the plume is the same. This means the path length cancels in the ratio given by equation (1). Consequently, for combustion efficiency, the actual path length does not need to be known. A more subtle issue is that of gas temperature. To first order, the temperature affects the gas density and does so linearly with absolute temperature. However, the densities of all gases in the plume are influenced in the same way; so this factor also cancels in equation (1). A less significant effect of temperature is the variation it causes in spectral band shapes. This does not cancel and it must be corrected for independently. The PFTIR was developed to provide Combustion Efficiencies not actual gas concentrations. To get actual gas concentration, accurate knowledge of both path length and temperature is required. Consequently, evaluation of gas concentrations is much more difficult than evaluation of combustion efficiency.

123 APPENDIX B10 IMACC Procedures The descriptions presented in this appendix are the measurement procedures for the active and passive Fourier Transform Infrared (FTIR) technologies. Also included is the policy for safeguarding that the algorithm for post processing all spectral data is not altered in any way once the data results are provided to UT Austin. IMACC will be making remote measurements of the flare flue gas to determine the concentration of the compounds emitted and their respective combustion efficiencies for the Comprehensive Flare Study Project. Revision No.1 Page 1 of 1 8/10

124 IMACC Active and Passive FTIR Measurement Procedures FTIR monitors can be operated in active or passive mode. In active mode, as shown in Figure 1, the FTIR has an internal infrared light source and its energy is propagated through the region to be analyzed. This energy is then captured and analyzed to determine what wavelengths of the IR have been absorbed. Absorption patterns are unique to each molecule, so the presence of a particular pattern is proof of the presence of a particular compound. In addition, the strength of the absorption is proportional to the concentration of the gas producing it. Consequently, in active monitoring, the absorption patterns are used to both identify the compounds present and to quantify their concentrations. In passive mode, as in Figure 2, there is no light source in the instrument. In this case hot gases ( 150C) external to the instrument emit their own infrared light and this light is simply captured by the instrument. It turns out; the patterns with which gases emit when they are heated are identical to those they absorb. Consequently, the same patterns observed in active or passive monitoring provides identification of the gases present. The difference in passive monitoring is that the strength of the pattern depends on gas concentration, as in active monitoring, but also on the gases temperature. This makes passive monitoring a little more difficult because there are two unknowns: concentration and temperature. However, where active absorption measurements are impractical, passive monitoring frequently allows for measurements to be made. A prime example is an elevated flare. In this case, it is all but impossible to get an IR beam through the plume and receive it afterwards. Passive monitoring only has to have a view of the plume to measure it. Figure 1 Active FTIR monitoring: propagating a beam through a gas cloud for detection. Page 1 of 8

125 Figure 2 Passive FTIR monitoring: receiving infrared light from an elevated temperature source for gas detection. Radiance Calculations Radiance measurements are difficult because the signal received is a compound one. Figure 3 shows the sources of radiation which comprise the complete FTIR signal. The first arises from the sky behind the flare. Because most flares are elevated, radiation from the sky is transmitted through the flare plume, through the intervening atmospheric path and then detected by the FTIR. The hot gases in the flare itself emit radiation. This is the signal of most interest and it is propagated through the intervening atmosphere before being detected by the FTIR. The atmospheric path between the flare and the FTIR also emits as does the FTIR instrument itself. Although these emissions are small they must be accounted for. The sum of all these individual signals comprise the spectrum seen by the FTIR. The flare radiance is the component needed because it contains information about the gases present in the flare exhaust and their concentrations. Extracting this signal from the total compound spectrum of the FTIR is the analytical problem. Several individual measurements are made to deduce the flare radiance. These are shown in Table 1. The first two are measured by aiming the FTIR at the flare and then at the sky to the side of the flare. The remaining measurements use a collimating telescope with various radiation sources. This telescope is shown in figure 4, without its enclosure which protects it from the elements. The platform on the back of the collimator accepts various types of sources. For measuring M bb a black body emission source is used; for M ir a standard infrared source is used; and for M n a liquid nitrogen cooled source is used. Page 2 of 8

126 Background Radiance Flare Radiance Atmospheric Transmission & Radiance Figure 3 Sources of the passive FTIR signal when monitoring an elevated flare. Table 1 Individual measurements Required to Deduce plume transmittance Measurement Description The observed flare radiance measured as a single M obs beam spectrum with the FTIR The measured sky background radiance M b collected by looking to the upwind side of the flare M The measured radiation from a black body bb calibration source in the collimator. M The measured radiation from an infrared source ir in the collimator. This term allows for computation of atmospheric transmission. The measured collimator and air radiance M n obtained using a liquid nitrogen source in the collimator Page 3 of 8

127 Figure 4 Calibration telescope on portable cart with hot cell. Environmental enclosure is removed. Mathematically, the total Radiance seen by the PFTIR, denoted M obs, is given by: M R τ τ R τ R R = (1) obs bkg air flr flr air air FTIR Here R is the radiance from the various sources. The term,, is given by the air measured infrared source spectrum M ir divided by a smoothed version of this spectrum tracing the baseline but skipping over all molecular absorption features. This is called a synthetic I o and software to generate it is included in the standard Imacc FTIR software. is given by: τ air M SYN τ ir τ = (2) air Io The measured sky background M b is measured adjacent to the flare plume so it is given by: M b = + + (3) air FTIR R bkg The radiance produced by an object is its absorbance or (1 transmittance) times a function called the Planck function. This function describes the radiance emitted by a totally absorbing object as a function of temperature and wavenumber. This function is given by: τ air R R Page 4 of 8