Hg Monitoring Options and Requirements

1
Hg Monitoring Options and Requirements

• Region 7
• Kansas City
• August 2007
• Charles Frushour
• USEPA, CAMD

2
Hg Monitoring Options

• CEMS
• Sorbent Trap Appendix K
• Hg LME

3
Hg CEMS

• Initial Certification
• 7-day calibration error test
• Linearity Check (elemental Hg)
• 3-level System Integrity Check (oxidized
  Hg)monitors with Hg converter only
• Cycle Time Test
• RATA
• Bias Test

4
Hg CEMS

• On-going Quality Assurance
• Daily calibration error test (elemental or
  oxidized Hg)
• Weekly 1-level System Integrity Check
• Required when daily calibrations use elemental Hg
  for monitors with Hg converters
• Weekly 168 operating hours
• Quarterly Linearity Check (elemental Hg) or
  3-level System Integrity Check (oxidized Hg)
• Required for each QA operating quarter
• Annual RATA (no semi-annual threshold)
• Bias Test

5
Calibration Tests

• The initial 7-day calibration error test and
  subsequent Daily calibrations may be conducted
  using either elemental Hg or a NIST-traceable
  source of oxidized Hg.
• Specification 5.0 of span value with alt
  spec of 1.0 µg/scm absolute difference
• The form of Hg used for daily calibrations will
  need to be reported in the Monitoring Plan so
  that EPA can track whether Weekly System
  Integrity Checks are required.
• The monitoring plan may be updated as needed

6
Weekly System Integrity Check

• One point version of the 3-level System Integrity
  Check required for initial certification of Hg
  monitors with an Hg converter.
• Required only if daily calibrations are performed
  using elemental Hg and the monitor uses an Hg
  converter to monitor total mercury.
• Required once every 168 unit/stack operating
  hours.
• Specification 10.0 of reference value with
  alt spec of 0.8 µg/scm absolute difference

7
Linearity Check

• Linearity Checks are performed with an elemental
  Hg standard
• Required for initial certification of all Hg
  monitors
• Required for quarterly QA unless a 3-level System
  Integrity Check is substituted
• Specification 10.0 of reference value with
  alt spec of 0.8 µg/scm absolute difference

8
3-level System Integrity Check

• System Integrity Checks are performed with an
  oxidized Hg standard generated from a
  NIST-traceable source
• Required for initial certification of Hg monitors
  with Hg converters
• 3-level System Integrity Check may be used
  instead of a linearity for Quarterly QA.
• Specification 10.0 of reference value with
  alt spec of 0.8 µg/scm absolute difference

9
Relative Accuracy Test Audit (RATA)

• Required for initial certification of all Hg
  monitors
• Required annually (once every 4 QA operating
  quarters) thereafter
• For Reference Methods that require paired
  sampling, the RM value for each run should be the
  average of the values observed from each of the
  paired samples for the run.
• Specification 20.0 RA with alt spec of
  1.0 µg/scm if average RM value is 5.0 µg/scm
• No semi-annual threshold

10
Bias Test

• Use standard bias test logic
• Low Emitter BAF cap is 1.250 for when the average
  RM concentration is lt 5.0 µg/scm

11
Appendix K Sorbent Trap System Basics

• Appendix K Sorbent Trap systems consist of a
  pair of sampling trains that collect an
  integrated sample over a specified collection
  period
• Sorbent traps collect the Hg and have 3 sections.
  
• Section 1 Main Sample Collection Section
• Section 2 Breakthrough Sample Collection Section
• Section 2 Spiked Section for QA
• Each sampling train uses a dry gas meter to
  record the volume of stack gas sampled over the
  course of each sample collection period

Note Agency plans to revise Appendix K in the
upcoming OAQPS direct-final alternative mercury
reference method rule package to allow the use of
any type of gas flow meter that accurately
measures gas volumes within 2 percent. This is
consistent with our performance-based measurement
policy.
12
Appendix K Sorbent Trap Systems

• Initial Certification
• RATA
• Bias Test
• Initial 3-level Dry Gas Meter Calibrations
• Initial DGM Temperature Sensor and Calibration
  Check
• Initial DGM Barometer Calibration Check
• Gas meter calibration procedures are being
  revised as part of the alternative reference
  method rule package.

13
Appendix K Sorbent Trap Systems

• On-going Quality Assurance for Appendix K systems
  and components
• Annual RATA (no semi-annual threshold)
• Annual Bias Test
• Quarterly Gas Meter Calibrations
• Quarterly Gas Meter Temperature Sensor and
  Calibration Check
• Quarterly Gas Meter Barometer Calibration Check
• Quality Assurance for each Appendix K sample
  collection period
• Pre- and Post- sample collection Leak Check
• Monitor Ratio of stack gas flow to sample
  collection flow
• Sorbent Trap Breakthrough check (section 2 of
  trap)
• Percent Trap Agreement
• Spike Recovery (section 3 of trap)

• Gas meter calibration procedures are being
  revised as part of the alternative reference
  method rule package.

14
Appendix K Sample QA

• The following QA is applicable to each sample
  collection period for evaluating the
  acceptability of each sorbent trap sample
  collected
• If any of the following QA checks are failed, the
  sample for that trap is invalidated.
• If data for the other trap is validated, a single
  trap adjustment factor may be used to adjust the
  data from that trap for uncertainty in lieu of
  applying substitute data for the collection
  period.
• If the QA checks are failed for both samples then
  substitute data must be used for the entire
  collection period.

15
Appendix K Sample QA

• Each sampling train must pass a pre- and post
  sample collection leak check. The post
  collection check will be reported to validate
  each sample
• The Ratio of the hourly Stack Gas Flow Rate to
  the hourly Sample Flow Rate must be maintained to
  within 25 of the initial ratio from the first
  hour of each collection period. Samples are
  invalidated if more than either 5 hours or 5 of
  the hourly ratios do not meet this criteria.
• Sorbent Trap Breakthrough check (section 2 of
  trap) The mass of Hg captured on section 2 of
  the trap must be 5 of the total Hg collected
  on section 1.
• Spike Recovery (section 3 of trap) The analysis
  of Hg on section 3 must result in 75 - 125 of
  the spike referenced spike level for the trap.

16
Appendix K Calculating the Collection Period
Concentration

• The Hg concentration for each trap sample is
  calculated by adding the Hg mass from section 1
  and section 2 of the trap.
• There are plans to remove the procedure to
  normalize the Hg mass in the upcoming reference
  method rule package.

• Percent Trap Agreement For each pair of traps,
  the Relative Deviation (RD) must be 10.
  (Alternate Specification 20 if average Hg
  concentration is 1.0µg/scm)
• If the Percent Trap Agreement is within spec.,
  the average Hg concentration is reported for each
  hour of the collection period
• If the Percent Trap Agreement is not within
  spec., the higher of the two Hg concentrations is
  reported for each hour of the collection period

17
Substitute Data for CEMS and Appendix K

• In August 2006, EPA proposed to consolidate the
  Substitute Data requirements for CEMS and
  Appendix K.
• This decision allows for the use of an Appendix K
  system as a back-up to CEMS in a way that does
  not complicate the Substitute Data requirements
  for when neither system is available.
• For hours where there is valid data from the
  CEMS, that data will be used to populate the
  lookback. This is also the value that must be
  used to calculate Hg mass for the hour.
• The Appendix K data will populate the lookback
  period for any hour where CEMS data are
  unavailable.
• For hours where neither system is available,
  substitute data is determined from the
  consolidated lookback of actual concentration
  data used for the previous hourly emissions
  calculations.

18
Substitute Data for CEMS and Appendix K

• The substitute data for Hg has four tiers
• Tier 1 If the PMA is 90 or more
• If the duration of the outage is less than 24 hrs
  substitute the average of the hour before and
  after
• If the duration is greater than 24 hrs,
  substitute the higher of the hour before/hour
  after average or the 90th percentile value from a
  720 hour lookback
• Tier 2 If the PMA is less than 90 and equal to
  80 or more
• If the duration of the outage is less than 8 hrs
  substitute the average of the hour before and
  after
• If the duration is greater than 8 hrs, substitute
  the higher of the hour before/hour after average
  or the 95th percentile value from a 720 hour
  lookback

19
Substitute Data for CEMS and Appendix K

• Tier 3 If the PMA is less than 80 and equal to
  70 or more, substitute the maximum value from the
  720 hour lookback
• Tier 4 If the PMA is less than 70 substitute the
  maximum potential concentration.
• For units with add-on controls, if the controls
  are documented to be operated properly per
  75.34, the unit may substitute the greater of
  the following instead of the maximum potential
  concentration
• The maximum expected concentration as listed in
  the monitoring plan, or
• 1.25 times the maximum controlled value measured
  from the previous 720 hour lookback

20
Hg-LME Monitoring Methodology

• Rather than install Hg analyzers or sorbent
  traps, affected units with lower potential levels
  of Hg may qualify for an excepted monitoring
  methodology
• This method is referred to as the Hg-LME
  methodology
• The Hg-LME excepted methodology is found in 75.81

21
Hg LME

• Low Hg emitting units ( 29 lbs/yr potential
  emissions) may qualify for the Hg-LME methodology
  
• This methodology requires an initial and on-going
  emission testing
• Initial test stack test results are used in Eq.
  1, with the maximum potential stack flow rate to
  demonstrate that the units potential to emit is
  not greater than the 29 lb threshold.
• The Hg concentration from the test is then used
  with actual stack flow data to determine hourly
  emissions
• Ongoing retesting is required annually if the
  calculated potential Hg emissions are 9 lb/yr,
  and semiannually if the potential emissions are
  between 9 and 29 lb/yr (NO GRACE PERIODS!)
• Actual reported Hg mass is used for ongoing
  qualification.

22
Hg-LME Testing

• A minimum of 3 runs at normal load are required
• Minimum run time is 1 hour per run is required
  for the instrumental method
• Run time for other methods to be determined by
  the amount of sampling needed to acquire a
  measurable amount of Hg
• Use the highest Hg concentration from the 3 runs
  or 0.50 µg/scm (whichever is greater) in when
  calculating the potential to emit
• The same default is used for hourly reporting
  until the completion of the next test.

23
Hg LME

• Equation 1 in 75.81 provides the conservative
  estimate of the annual Hg mass emissions from the
  unit (i.e., potential to emit)
• Equation 1 is also used to determine whether
  ongoing test should be performed semi-annually or
  annually
• Equation 1 uses the Hg concentration determined
  based on reference method testing consisting of
  three runs

24
Hg LME

• Equation 1 E 8760 K CHg Qmax
• E Estimated annual Hg mass emissions from
  the affected unit,
  (ounces/year)
• K Units conversion constant, 9.978 x 10-10
  oz-scm/µg-scf
• 8760 Number of hours in a year
• CHg The highest Hg concentration (µg/scm) from
  any of the test runs or 0.50
  µg/scm, whichever is
  greater (paired sampling trains are required for
  each test run)
• Qmax Maximum potential flow rate, determined
  according to section 2.1.4.1 of
  appendix A to this part, (scfh)
• Considering allowing less than 8760 hours if a
  permit condition restricts hours of operation to
  less than 8760

25
Hg LME

• On August 22, 2006 EPA proposed revisions to 40
  CFR 75 and 72.2
• Proposed revisions include the following
  requirements
• Coal combusted during testing must be from the
  same source of supply as the coal combusted at
  the start of the Hg mass emissions reduction
  program
• Test a subset of identical units in accordance
  with 75.19(c)(1)(iv)(B) in lieu of testing each
  unit individually
• The calculated value of E shall be divided by the
  number of units sharing the stack. If the
  result, when rounded to the nearest ounce, does
  not exceed 464 ounces, the units qualify to use
  the low mass emission methodology
• Common Stack testing to be allowed with
  conditions