So You Have To Do A Stack Test…

This article will briefly describe some best practices to determine particulate matter (PM) emissions using United States Environmental Protection Agency (U.S. EPA) Methods 5 and 202 for filterable particulate matter (FPM) and condensable particulate matter (CPM), respectively. The best practices provided below address sensitivity, contamination, representativeness, and consistency of implementation. While this article focuses on particulate matter, the concepts presented are applicable to source testing in general, regardless of the analyte or methodology.

Background

Particulate matter in stationary source emissions is determined gravimetrically. That is, gas is withdrawn from the duct and passed through a filter. The weight gain on the filter is related to the volume of gas withdrawn to determine FPM loading. This approach (i.e., Method 5) is completed by rinsing the sampling system upstream of the filter and combining the weight gain from the rinse with the weight gain from the filter. Determination of CPM is completed by measuring any matter that condenses after the filter. This is done in a (nominally) dry impinger. CPM is recovered by a rinse of the appropriate contact surfaces with water and with solvent.

Best Practices

Sensitivity

The result of the measurement should be in a robust portion of the analytical system. Any analytical system has upper and lower thresholds (think about measuring length with a yard-stick, anything below ~¼ inch and above ~35 ¾ inches is outside the meaningful measurement range). So, we want the results of our measurement to be in a well-characterized portion of the analytical range. And if the results are below that level, we want to feel confident that the conclusion around our decision point is minimally impacted by the uncertainty associated with measurements near zero. This is accomplished by collecting enough sample gas during testing to put the decision point (i.e., the emissions limit) an order of magnitude above the detection limit. Here’s an example calculation to illustrate this concept:

• Target limit: 0.02 lb/hr
• Gas flow rate: 12,000 scfm
• Method detection limit (MDL): 2 mg/sample (this is for Methods 5/202 – the detection limit or DL is 0.5 mg/fraction and there are 4 fractions in a sample train).
• Targeted mass: 20 mg/sample (one order of magnitude increase from MDL)
• Calculations:
  • 2 lb/hr à 1.5 g/min
  • 5 g/min and 15,000 scfm à 0.00013 g/cf or 0.13 mg/cf
  • To get 20 mg in a train, 20 mg and 0.13 mg/cf à 180 scf
  • Because the sampling rate is 0.75 cfm (typical for M5 sampling), we want to sample for 240 minutes, or 4 hours.

The result from this calculation must guide the sampling specifications, to ensure the data are usable for the intended purpose.

Contamination

As described above, emissions may be very near the lower threshold of the measurement system. As such, it is imperative that background and contamination be minimized:

• Reagent selection: Use high quality reagents. For a stand-alone test program, consider using unopened reagent bottles.
• Media selection: Use high-quality quartz filters only.
• Document glassware system cleaning activities.
• Collect and analyze blank samples. These can include proof blanks, field blanks, reagent blanks, media blanks.

Representativeness

It is crucial that any testing be conducted under the appropriate operating conditions:

• Operating conditions should be specified, numerically, in advance, in more detail than “within 10% of full load”. For example, “4500 watts” or “22 gpm of fuel flow”.
• Operating conditions are often reflected in stack gas volumetric flow rate; verify that the flow rate is as expected.
• Communication between all parties is necessary; the test planner needs to communicate with the facility operator and the stack tester.
• Define and understand steady state before testing commences; wait the appropriate and necessary times.
• If a test occurs after any kind of delay (overnight, for example), return facility operations as closely as possible to the earlier operating scenario.
• Any field-expedient changes (e.g., not achieving targeted load) need to be made carefully and documented appropriately.

Consistency of Implementation

At first glance, every test effort is a standalone effort. Resources are mobilized and testing happens in a fairly short timeframe (i.e., hours or days). Retesting can be another standalone effort. It can be problematic if testing devolves from a demonstration of compliance to a diagnosis of unexpected results. This type of situation can be minimized with the following considerations:

• Address all the issues above (sensitivity, contamination, representativeness) before the initial test effort. Don’t change the approach without a valid reason; document differences between subsequent test efforts.
• Do not change reagents or media. Ask the testing firm to acquire media and reagents specific to your program and to use them for all testing efforts.
• Do not change human resources. As best possible, use the same facility operations personnel and the same test personnel.
• Ambient conditions change. Review testing and operational details to minimize any differences due to weather.

ALL4 has decades of experience with simple and complex stack testing efforts, from concept through planning, execution, and interpretation of data. If you have any questions, please contact Gene Youngerman at 512.649.2571 or eyoungerman@all4inc.com.