Summary of draft updates to PG7 2005 (October 2021)

Summary of key changes
No. Draft section Summary of draft updates to PG/7 2005 (October 2021) Pros Cons
1 1.0 Introduction Applicable to monitoring emissions of SO2, NOx and CO2 monitoring from large combustion sources, which may be extended to other contaminants and point sources. The PG/7 2005 version was heavily focused on SO2 and NOx emissions from thermal power generation. Yet some provincial regulations, such as Ontario Reg. 127/01, applied PG/7 to other processes. The updated version included CO2 and CO emissions and the potential extension to other combustion processes. n/a
2 2.0  Summary of specifications Specifications and protocols are presented in narrative detail, as well as in summary tables that may serve as a quick reference to those familiar with the subject (fox example, Tables 3, 6 and 7). The document structure of PG/7 2005 was maintained. A few sections deemed confusing were reworded. Tables were reformatted. For practitioners tables 3, 6, and 7 present a reasonable summary of the specifications, with further explanations in the internal references. Choice of wording identifies mandatory from discretional clauses (that is, "must" and "shall" vs "may" and "could"). n/a
3 3.1.1 Location of the calibration gas injection port All CEMS installed after 2023 must be able to conduct daily calibration drift and quarterly linearity tests using as reference flowing calibration gases. The eventual replacement of some Canadian CEMS for others than can be calibrated with reference gases will place CEMS and reference methods in the same calibration footing. EPA instrumental reference methods for SO2, NOx, CO2 require pre-test EPA Protocol gas calibration and post-test drift check. This is how the test results are validated. CEMS fitted only with internal calibration may find that following a failed CGA or RATA, large data sets are questionable. Alternate CEMS calibrations for these CEMS have been maintained in the revised PG/7, but any system installed from 2023 onwards must be calibrated with flowing reference gases. This is in generation agreement with sections 4.2, 3.1A, and 3.1B of the AB CEMS code 2021. The timeframe of this clause, however, tolerates existing systems that might not meet quality assured data targets, and therefore facilitate the planned extension to their operating life.
4 3.2.1 Operating range The measurement range (MR) of each analyzer must be slightly higher than the maximum probable exhaust concentration level. Full scale (FS) is defined as a subset of MR, such that most of normal operation measurements fall within 20% and 80% of it. This FS must be adjusted annually before RATA, as it is used to define some specifications. PG/7 2005 requires that the average (1 value) of the CEMS measurements to be within 40% and 75% of the set FS. Instead the revised PG/7-8 requires the majority (> 50%) of the CEMS measurements to fall within 20% and 80% of the FS, which is a more comprehensive criterion. The modified requirement is in line with the US EPA full scale policies for CEMS and instrumental reference methods for SO2, NOx and CO2. Note, Alberta estimates that approximately 120 CEMS units currently in service in the province, have an oversized MR that cannot be altered and that cannot meet this criterion (the measurement range is much higher than the majority of current measurements. Therefore, the recently updated AB CEMS Code will not require "to make adjustments to current operating range or full scale of analyzers".
5 3.2.2 Interference and 3.2.3 Temperature-response drift The analyzer manufacturer must certify that the sum of combustion exhaust interferences is ≤  4.0% of the expected FS, as well as meeting certain temperature response drifts. This modification reduces the number of tests that the user must comply with at pre-certification of a new unit. The requirement is transferred from the user to the manufacturer, which may be better equipped to test one or more identical units of a production batch and provide a certificate of conformance to the user. This change does not apply to existing units.
6 3.2.4. NOx converters If NOx analyzer rely on NO2 converter then the efficiency must be tested quarterly, as per defined protocols This requirement is similar to that of EPA RM 7E, except that in CEMS must be done quarterly as opposed to pre- and post- each field test, as per Sections 8.2.4 and 8.2.4.1 of EPA reference method 7E. n/a
7 3.2.5 FTIR extractive CEMS An FTIR may be used as a CEMS component to monitor NO2, SO2 and CO2 concentrations of the exhaust gases of a combustion source, providing that the FTIR met the applicable analyzer specifications of this document, including the required daily, quarterly, annual or semi-annual QA/QC tests. The special characteristics of this technology are acknowledged. The driving force for the installation of FTIR is, for example, the required monitoring of hazardous contaminants such as HCl in cement kilns, for which in U.S. the CEMS must meet Performance Specification 18. The same CEMS may be used in Canada to monitor SO2, NOx and CO2 emissions but is not granted privileged exemption to daily, quarterly, annual or semi-annual QA/QC tests. This level field decision regarding QA/QC requirements is consistent with current US. Part 75 acid rain program policies but may disappoint the expectations of FTIR vendors, their clients and some practitioners.
8 3.4 DAHS specifications Enhanced capabilities to compile monthly CEMS reports, and annual third-party audit. DAHS must archive monthly and annual hourly data, including availability, certification tests, RATA, CGA, BAF, and backfilling. DAHS requirement specifications are similar to AEP, except in the formats and codes for electronic data reporting which are spelled out in a separate AEP 55 page document (aep-cems-information-user-manual-version-2-4.pdf). We are not aware of CEMS users in provinces other than AB that are doing the monthly CEMS reports, or the prescribed annual third party audit (PG/7 2005, Table 5, Quality control procedures, subsection 16) so they may object upgrading their data handling system for doing something that they and the appropriate regulatory agency currently ignores.
9 3.4.3 Backfilling of missing data Examples of backfilling options are presented. Specifics must be described in the Quality Assurance Plan (QAP) and be accepted by the appropriate regulatory authority. The objective of CEMS is to produce a complete record of hourly emissions. However, updated PG/7 specifies CEMS availability (≥ 90% in 1st year, ≥ 95% thereafter) which implies that data from 5% to 10% of the hours can be expected to be missing because equipment malfunctions or routine maintenance. Programs that use CEMS data determine what to do with the missing hours. The U.S. cap-and-trade missing hours backfilling policies, for example, are designed to provide conservatively high substitute data values, so as to motivate a quick response to the cause of the missing data. The backfilling during the 1st to 3rd month following certification are likely to be close to the expected values if the CEMS had operated. Later on, the backfilling becomes more punitive, particularly for longer and more frequent episodes. In Canada, CEMS data likely are used for purposes other than cap-and-trade, so the revised PG/7 only provides examples of short interval backfilling using the most probable values according the historical database of the source. Source specific backfilling schemes should be approved by authority that request and collects the data. This applies to the CO2 emission data collected by CEMS and subject to Carbon Tax. Backfilling when the CEMS availability targets are not met is a related issue.
10 3.5.1 Time-shared systems After 2023, new time-shared CEMS will be limited to two (2) adjacent sources. PG/7 2005 allows an extractive CEMS to be shared by various adjacent sources, providing that all of them were measured every 15 minutes. As the response time of the CEMS may be up to 200 seconds (3.33 minutes), a cycle time of 15 minutes for two sources would barely leave enough time for 1 minute average readings and spare time for daily calibration. Therefore the revised PG/7 limits time-sharing to two adjacent sources and provides alternate schemes to do RATA as per US. EPA Part 75 Emission Monitoring Policy Manual (2013), page 8-20 and 8-21, questions 8.33 to 8.35. We are not aware of any CEMS time-shared between more than two sources; nevertheless, this change should not apply retroactively.
11 3.6 Test procedures for verification of design specifications Analyzer interference rejection, temperature response and system cycle time (for time-shared CEMS) to be tested by the CEMS manufacturer, with certificate of conformance archived in the QAP manual. This modification reduces the number of tests that the user must comply with for pre-certification of a new CEMS. The manufacturer should perform them on one or more units of a production batch of identical units. The certificate of conformance should be supplied for the user's records. These are pre-certification test requirements that should not apply retroactively.
12 4.2 Representativeness EPA Methods 2G (two dimensional probes) or 2F (three dimensional probes) may be used as flow reference methods for certification and RATA when cyclonic flow cannot be otherwise corrected. U.S. EPA Methods 2G and 2F were developed to address complex flow patterns where conventional Method 2 is known to overestimate stack gas flow. These patterns are encountered more frequently in non-ideal sampling locations than in the ideal locations (straight section of 10 diameters length). A 1% reduction in the flow measured with Method 2 or ECCC Method B may be applied automatically (Method 2H, for velocity decay near walls), but then these methods or adjustments must be used for subsequent RATAs. Methods 2G and 2F are more complicated than ECCC Method B or Method 2 and require different probes and differential pressure measurement system.
13 Table 3, page 19 Reduced the alternate limit to ≤ 5 ppm avg. absolute difference for SO2 & NOx. Limited bias to ≤ 4.0% FS for contaminants, diluents, and flow. Limited daily calibration drift to ≤ 2.5% FS for all levels. The removal of the < 250 ppm clause eliminates a sizable "loophole" for passing SO2 RATA. The 10% RA and the alternate low level accuracy limits are similar to those in 40 CFR 75 and the AB CEMS Code 2021. The latter, however, defines Full Scale (FS) as the upper value of the analyzer operating range (MR in PG/7), whereas for 40 CFR 75 and PG/7 FS is a subset of MR that encompass the majority of measured values between 20% and 80% of it. MR is ≥ FS, therefore the PG/7 allowable drift (2.5% FS), and bias (4% FS) may be lower than those of the AB CEMS Code 2021, depending on how oversized is the measurement range of the analyzer or monitor. PG-7 Bias and calibration drift limits may be lower than those of the AB CEMS Code 2021, as in that document FS means the analyzer operating range that encompass all anticipated concentrations or values (including emission limit exceedances).
14 5.3.2.2 Calibration drift and linearity check protocols Narrative for calibration drift protocol (7 days, 2 level test per day) was separated from the linearity CGA protocol (3 runs, 3 levels per run). Both these tests may be performed during the operation test period (OTP) The procedures of section 5.3.2.2 interweaved the execution of these two flowing gas tests, which was confusing to some practitioners. In the revised PG/7 text, they are explained separately in section 5.3.2. n/a
15 5.3.6 Bias test calculations Bias ≤ 4.0 %FS are acceptable, and subsequent measurements must be corrected by a Bias Adjustment Factor (BAF) that compensate for the bias. Alternatively, Bias are acceptable if the average absolute difference (|d|) between CEMS and RM average is less than certain minimum value. In these cases, however, no BAF must apply to subsequent data. RATA produces 9 -12 CEMS vs RM comparison, and the main objective is to determine with 95% probability that the CEMS measurements are within ± 10% of the RM data. Statistical bias between the two datasets is calculated and a ≤ 4.0 FS limit and an alternate average absolute difference (|d|) are applied. These limits coincide with the U.S. EPA limits, which only allows BAF ≥ 1.0.  This is consistent with other EPA measures that link performance with punishment or concessions (that is, punitive backfilling of missing data or reduced RATA frequency if RA ≤7.5%). PG-7 allows BAF >1 as well as BAF <1. Our numerical modeling indicate that BAF should be applied only when the ≤ 4.0 FS limit is met. The AB CEMS 2021 Code does not allow BAF.
16 6.3.1.6 Alternate quarterly analyzer audit Deleted former method EPS 1/RM/15 reference to perform alternate quarterly audits to CEMS installed before 2023 that cannot be calibrated with flowing gases. Replaced by abbreviated RATA using portable analyzer that meets RM specifications. Alternate acceptable RA ≤ 15% for pollutants and diluents, or 1.0% absolute difference for O2 and CO2. EPS 1/RM/15 is an old EC electrochemical method no longer available in the ECCC web sites. The proposed alternate quarterly audit (CGA alternate) is based on an abbreviated RATA (6 tests of 21 minutes each) using portable analyzers (for example, Horiba PG350) that meet RM specifications. The expanded acceptable alternate (≤ 15% RA) is partially due to the higher t0.025 for 6 test series. The abbreviated RATA using RM-capable portable analyzers is more expensive that the former alternate quarterly audit.
17 6.3.2.4  Flow-to-load flow monitor audit Maintained the quarterly database to 168 hours and, in the case of peaking units expanded the potential data base to the preceding 12 months of operation. Peaking was defined as 1,500 of operation per calendar year. The quarterly Flow-to-Output compares the last RATA average stack gas flow (a) with the average stack gas flow of ≥ 168 quarterly hours (b) in which the average electric or steam output was within ± 10% of the average electric or steam output of the last RATA. If flow (b) is within ±10% of (a) the audit is passed, for output levels ≥ 60 MJ/s (± 15% for ≤ 60 MJ/s). This paper exercise is an acceptable quarterly audit of the stack gas flow monitor, and require sorting ~2,160 to 8,640 hours of DAHS data. This audit avoids the costly reference method quarterly measurement of stack gas flow, but may be too cumbersome for CEMS not equipped with a suitable DAHS.
18 Table 6, page 37, Daily and quarterly performance evaluations summary Removed the redundant daily drift limit  for dry O2 - wet O2 moisture monitors (but maintained drift limit for O2 and CO2 analyzers). Linearity CGA limits relative to the reference gas value, as opposed to FS. Alternate audit for CEMS that cannot perform CGA. These are minor changes or corrections to the reformatted Table of Daily and Quarterly Performance Evaluations, which were previously discussed. n/a
19 6.5.2 Annual inspection Defined the scope of the annual 3rd party inspection of the CEMS QA/QC program, based on the review of the QAP and the DAHS data: hours of quality assured data, out-of-control hours, backfilled hours, in addition to the results of quarterly, semi-annual or annual performance tests (CGAs, flow audits, and RATAs). Narrative discussion of non-compliance issues, corrective actions to out-of-control occurrences and recommendations to improve CEMS performance. The annual inspection by a 3rd party is a recourse that regulators may use to gain insight to the quality of data produced by CEMS (that is, NOx, CO2, or other regulated contaminant), in the absence of a montlhy data reporting program (such as Alberta's). If the regulatee does not have a customized DAHS, the preparation of such a report (from 8760 data hours) by an off-site independent party may be quite onerous. However, if the site installed a DAHS as per Sections 3.4 to 3.4.3, then an off-site annual review of the CEMS QA/QC program is estimated to cost approximately $ 3,500 (from budgetary quotes received).
20 Table 7, page 40, Semi-annual or annual performance evaluations summary Reduced the alternate limit for SO2 & NOx (5 ppm RATA avg. absolute difference), but included an alternate limit for stack gas H2O (1.5% RATA avg. absolute difference). Bias limited to ≤ 4.0% FS for contaminants, diluents, and flow. Removed the mass emission RA specifications. The alternate limit for SO2 and NOx RATA acceptance was lowered from 8 ppm to 5 ppm average absolute difference to reduce the possible use of this alternate limit as a "loophole" for acceptance of RATA at low emitting sources (that is, stationary gas turbines rated at 10 to 25 ppm NOx). Relative accuracy limits provide a consistent accuracy benchmark but become tougher to meet at low level, as both the CEMS and Reference Methods increase their uncertainty. Adequate alternate limits provide relief to this situation. n/a
21 7.0 Determination of carbon dioxide emissions (new) Four acceptable CO2 CEMS fitted with stack gas flow monitors are described, with calculating formulas and limitations, including the use of O2 analyzers in lieu of CO2 analyzers. This chapter completes the "PG/8" merging. The formulas provided were derived from US 40 CFR Part 98, c) Tier 4 Calculation methodology, and  40 CFR Part 75 Appendix F, 4. Procedures for CO2 Mass Emissions. F factors were adjusted to the ECCC reference temperature. The CEMS fitted with O2 analyzers in lieu of CO2 analyzers shall perform daily and quarterly QA/QC tests as O2 analyzers, but RATA performance is determined from the calculated CO2 value. CO2 emission data collected by CEMS may be subject to Carbon Tax, so it is paramount that the annual data set is complete. Source specific backfilling schemes should be approved by the pertinent authority, including cases when the 80 - 95% availability was not met.
22 Glossary, page 44 New peaking unit definition: combustion unit operated ≤ 1500 hours within a calendar year. Various peaking concessions included in sections 5.1.1, 5.1.2, 6.3.1, 6.3.2.4, 6.3.3.1, and 6.5.1 The updated definition is simpler and of faster application than the previous, and is compatible with other ECCC documents. The concessions strike a balance between the performance of QA/QC audits and the intermittent and sporadic operation of these units. Peaking units may be exempted from the test performed during the 7-day operational test period; only annual audits (as opposed to quarterly) of stack gas flow, CGA, F-factor (Appendix A systems); 12 month database for flow-to output test; 80% minimum availability; may not calibrate the CEMS during extended downtime periods, providing that it is done upon restart; valid completion of 24-hour calibration drift test even if process is down by then. These concessions are similar to those of the U.S. Acid Rain Program. n/a
23 Appendix A, Emission calculation by combustion F-factors Applicable to combustion sources CEMS where the desired result is emission per heat input (without flow monitor). The updated Appendix A expands the previous F-Factors Table; corrects a formula error (Kx values); and recommends remedies for transition periods of extreme high O2 and low CO2 levels. Appendix A was edited to correct a formula error (Kx values), Table A-1 was expanded to include all combustion F factors in Table 1 of U.S. 40 CFR 75 Appendix F, which were converted to ECCC units and reference temperature. Appendix A formulas such as Eqn. A-1, A-3 and others with (20.9 - %O2) denominator may calculate abnormally high emission SO2 or NOx values when the stack gas oxygen level approaches the level of ambient air (fox example, start ups or shut downs) which would affect hourly averages. This is avoided by setting minimum CO2 and maximum O2 levels, which were adopted to coincide with U.S. 40 CFR 75, Appendix F, Section 3.3.4.1. n/a
24 Appendix B, Determination of mass emission rate Minor editing to this Appendix, except for the addition of linear equations to calculate combustion stack gas moisture from measured O2 wet (for NG, Oil and Coal), and average monthly air moisture on Canadian provincial capitals, which combined may result in acceptable stack gas moisture levels. In Canada, the use of dry SO2 and NOx analyzers in mass emission CEMS is not uncommon, in contrast to the U.S. power sector where dilution CEMS predominate. The stack gas moisture value necessary to combine the dry concentration with stack gas flow, is often determined during certification or RATA and maintained constant throughout the year, regardless the excess combustion air level or ambient air moisture. This approach may become a significant source of inaccuracy if the source varies the excess combustion air levels. Stack gas moisture can be estimated or measured in different ways. A simple one, based on the measurement of wet O2 (ZrO2 analyzer) is presented in Appendix B which includes linear formulas applicable to the combustion of NG, Oil and Coal. Ambient air moisture levels may be added to the calculated combustion moisture to produce a reasonable stack gas moisture estimate. Some may oppose the stack moisture monitoring and adjustment, as they pass annual RATA the same month of the year and at the same capacity and excess air level, regardless of other operating conditions.

Page details

Date modified: