Boosting CEMS performance

• By Steven A. Jaasund, PE
• Oct 01, 1999

Electronic water condensers are an effective means of removing water from combustion waste gas streams and process samples, in preparation for performing gas analysis. In recent years, electronic water condensers have also found prevalent use in continuous emissions monitoring systems (CEMS). Due to recent air quality regulations requiring lower emission levels for criteria pollutants (nitrogen oxides (NO_x), sulfur oxides (SO_x), carbon monoxide (CO) and particulate matter), environmental agency personnel have focused more of their attention on component absorption in the sample conditioning system. Electronic condensers provide high-speed water separation from stack gases, reducing the component absorption as compared to other dewatering methods.

The U.S. Environmental Protection Agency (EPA) Region 6 and the Texas Natural Resources Conservation Commission (TNRCC) began investigating criteria gas absorption in CEMS in 1994. TNRCC found CEMS Relative Accuracy Test Audit (RATA) reports were just making the +_15 percent EPA required in 40 Code of Federal Regulations (CFR), Part 60, Appendix F, Quality Assurance Procedures. The test audit verifies with field testing if CEMS are in or out of control. Out of control is defined as excessive audit inaccuracy as determined by failing the RATA testing. The out of control period is defined as beginning upon failure of the RATA, and ending upon the completion of the sampling of the subsequent successful audit. In between these two periods, the emitter is subject to whatever punitive action the local air quality district's rules and regulations prescribe. Because the RATA testing as defined in Appendix F verifies all of the components of a CEMS, including the data acquisition system (DAS), every source of error comes under scrutiny. All errors add up, requiring a close watch on the quality control (QC) procedures at the source and the equipment make-up of the CEMS.

For these reasons, EPA Region 6 was interested in establishing what the performance levels were for each of the CEMS components. A comparison of the gas absorptions caused by separating the water vapor from the stack gas was requested. This test compares the absorption characteristics of a thermoelectric cooler and the standard EPA, Method 5, impinger sampling train, as described in 40 CFR, Part 60, Appendix A, Method 6. A similar water removal device, refrigeration condenser described in Method 6C is also commonly used by stack testers, who are considered the reference method the source CEMS performance is measured against for determining in or out of control status.

The thermoelectric cooler produces consistent outlet dew points, making the test easy to perform. The thermoelectric heat exchanger exhibits lower gas absorption than the Method 6 impinger sampling train for sulfur dioxide (SO₂) and nitrogen dioxide (NO₂) concentrations of 200 to 400 parts per million (ppm) levels, based on initial test results started for the Method 6 impinger train absorptions. The Method 6 test results are rather narrow in scope, so testing engineers plan to conduct additional testing and anticipate that further test data will demonstrate much higher SO₂ and NO₂ absorption levels for both the Method 6 and 6C water removal systems in comparison with the standard, thermoelectric heat exchanger. Field stack reference method testing problems have been documented for SO₂ and NO₂ levels of 20 to 40 ppm at wate r concentration levels of 15 to 30 percent. We intend to add more field comparison testing on actual stack tests to augment the initial findings found in our laboratory test results.

Due to very low NO_x emissions required in many air quality districts in the United States, particularly in the Los Angeles basin, and taking into consideration the lower absorption exhibited by the thermoelectric cooler, the South Coast Air Quality Management District (SCAQMD) issued their Rule 100.1, which requires the use of a thermoelectric cooler in combination with a single impinger, for all 40 CFR, Part 60, Appendix F, RATA and compliance testing within the district.

By integrating the low gas absorption feature of a thermoelectric cooler with its rapid load response, M-Class modular design, and efficient power consumption, into a gas sample conditioning system, a flexible, reliable pollutant measurement can be made on a continuous basis with greater than 98 percent uptime, which will pass the most stringent of performance testing.

Sample conditioner description
A dry, extractive, sampling system consists of a heated probe to extract the sample from the stack or source; a heated sample line to transport the sample to the water condenser; and a sample conditioning panel where the sample is filtered, dewatered, and pumped to the gas analyzers. A typical flow diagram for a stack sampling system is depicted in Figure 1.

The sample conditioner consists of an electronic water condenser, a diaphragm gas sample pump, a peristaltic drain pump, a low micron in-line sample filter, water slip sensor, and total sample flowmeter. The components are mounted on a plate or bracket, for easy integration into a continuous gas measurement system.

M-class sample cooler design
To sample combustion product stack gas or exhaust from any internal combustion process, a method to remove moisture from the sample, without removing the components to be analyzed, is a must. This new, compact water condenser design is an ideal way to decrease the dew point of the combustion gases to a repeatable, stable, constant low dew point. Most gas analyzers have some interference due to water vapor, and in the presence of acid gases, are subject to corrosion. By reducing the moisture to a low level, and maintaining stability of the dew point, the gas analyzers can make a more accurate analysis on a long term, repeatable basis.

The M-class electronic water condenser lowers the incoming gas sample dew point to 5(C (41(F). Particulate matter is partially removed by the heat exchangers. Particulates that pass through the condenser are removed by a low micron in-line filter downstream of the condenser.

Standard thermoelectric coolers are normally supplied with the following integrated features: LCD display of operating temperature; voltage output proportional to operating temperature; LED display of key operating characteristics; dry contact alarms; sample pump on/off control; heated line temperature interlock; and current output for exit sample gas dew point temperature, cooler operating temperature and sample gas inlet temperature.

The M-class cooler separates these standard features into optional features. Separate control boards are supplied for the LCD display, the alarm card and the current output card.

By making these features modular, the sample system design engineer can pick and choose the options required for the application, as well as the process loading in a single thermoelectric cooler package. These features provide a broad range of flexibility in a single enclosure. Application of an electronic water condenser to stack sampling requires thorough knowledge of the sample's characteristics. The M-Class water condenser is available in a number of capacities and configurations and can cover a wide range of applications and loads within the same physical package. The mounting footprint stays the same no matter what model number and load capacity is chosen. This feature makes the cooler applicable for system integrators and large company purchases. A single cooler model series covers a broad range, with common spare parts and utility connections.

Gas absorption in the laminar heat exchanger
Many users are interested in quantifying the criteria pollutant (absorptive gases) attenuation (losses) in thermoelectric coolers. A test apparatus was designed and constructed in 1997 to investigate the losses in high speed heat exchangers (impingers) for various materials of construction.

A Baldwin Environmental Model M325 electronic sample cooler was tested for gas absorption within its rated flow and load ranges, using stainless steel, Kynar and glass heat exchangers.

The test apparatus was operated using Protocol 1 gases, delivered to the boiler through a Horiba S-tec capillary gas divider. Due to boiler attenuation, the calibration gas was injected into the flowing ambient air stream directly above the boiler. Thus several dilutions took place. The ambient air/water enriched stream diluted the calibration gas approximately 50 percent. Since this test is a relative one; i.e., knowing the beginning and ending gas values without water, the testing engineers were able to measure gas absorption as a function of water concentration injected, and materials of construction used.

Test results
Figure 3 summarizes the test results for SO₂, using stainless steel, Kynar and glass heat exchangers. Glass appears to attenuate SO₂ the least for water concentrations below 30 percent. It is interesting to note that stainless steel performs quite well at water concentrations of 15 percent and below, where the majority of applications fall. Due to time constraints, SO₂ concentrations tested were at the 300 to 400 ppm level. Future tests will be at lower SO₂ concentrations and high water volume percentage such as found in waste incinerator applications. Testing engineers do not yet know the relationship of lower concentrations, residence time and water concentrations. These absorption tests will assist the application engineer in choosing the correct heat exchanger materials of construction to guarantee a successful installa tion considering the stack gas source, water concentration, S O₂ expected concentration, and system flow rate requirements.

Figure 4 summarizes the test results for NO₂. NO₂ was found to have negligible absorption under all conditions and materials of construction, so the test results summarize the NO₂ absorption data. NO₂ calibration gas was Protocol 1 grade, 2,000 ppm, supplied by Scott Company. The NO₂ gas was gas divided with nitrogen to achieve the lower level test concentrations. Again, as in the SO₂ test data, time constraints prevented further investigations at low NO₂ levels. This data used 200 to 400 ppm NO₂ levels for the test, concentration levels rarely found in actual practice. The testing engineers' intent was to develop a worse case scenario. Again, NO₂ absorption follows the same curve shapes as SO₂. Glass performs quite well at water concentrations to 50 percent.

Conclusion
The test data shows a thermoelectric cooler absorbs criteria pollutants to a much lesser degree than the EPA impinger train as shown in Table 1.

Table 1

EPA Method 6 glass impinger train gas absorption test

Test condition	%H20	Indicated ppm SO2	Indicated ppm NO2	Indicated ppm NO	Corrected ppm SO2	Corrected ppm NO2	Corrected ppm NO	% SO2 Absorbed	% N02 Abs.	% NO Abs.
Sample	0.0000	173.0000	475.0000	101.0000	173.0000	475.0000	101.0000	0.0000	0.0000	0.0000
flowrate	10.0000	166.0000	476.0000	100.0000	167.6800	475.0000	100.0000	3.0751	0.0000	0.9901
4 liters per	20.0000	156.0000	461.0000	98.1600	157.9900	466.6000	99.3000	5.7789	1.7684	0.7000
minute	30.0000	148.0000	444.6900	98.0000	149.9000	444.6900	99.2700	5.1206	4.6957	0.0302
(L/m)	50.0000	136.1600	399.6000	97.1600	138.0700	404.7500	98.3200	7.8919	8.9815	0.9570
Sample	0.0000	173.0000	475.0000	101.0000	173.0000	475.0000	101.0000	0.0000	0.0000	0.0000
flowrate	10.0000	169.0000	475.0000	101.0000	170.5700	475.0000	101.0000	1.4046	0.0000	0.0000
8 L/m	20.0000	160.0000	464.0000	99.8000	162.0700	470.0000	100.5000	4.9833	1.0526	0.4950
	30.0000	153.0000	451.0000	99.2000	154.9900	456.8900	100.4000	4.3685	2.7894	0.0995
	50.0000	139.0000	403.0000	98.0000	140.9500	408.6800	99.3800	9.0586	10.5518	1.0159

The thermoelectric heat exchanger also exhibits lower absorption rates for the stack tester, performing the Appendix F CEMS Relative Accuracy Test Audit. A successful CEMS monitoring and reporting system requires all components in the system to perform to high standards, continuously, to ensure a successful RATA, in control, and to provide greater than 98 percent operating uptime. Air quality districts are under increasing pressure by EPA and environmentalists to improve the air quality. State implementation plans (SIPs) rely upon the data produced by sources' CEMS to assess the state's air pollution inventory. EPA uses the SIPs to gradually restrict air pollutant emissions to achieve the air quality goals established by Congress in the Clean Air Act Amendments of 1990. The source's responsibilities are to generate accurate summaries of its emissions, as an equal partner towards reducing air emissions to mandated standards. Use of a thermoelectric condenser in both the CEMS and the RATA test apparatus provides a much greater degree of pollutant measurement accuracy than previous stack gas water separation techniques.

Impinger (heat exchanger) design

The heat exchanger is the work horse of the sample cooler. It removes water from hot, wet, dirty stack gas, with minimal gas component absorption (see Figure 2.)

The hot, wet, stack gas enters the exchanger from the top, and travels down through the thermally isolated center tube. The thermal isolation is achieved by a vacuum jacket around the inlet tube.

This jacket serves two purposes: to keep the sample gas temperature above the dew point while the gas transits the inlet tube, and to minimize the heat loss to the cold outer wall of the heat exchanger.

Once the gas reaches the bottom of the inlet tube, water condensation takes place immediately upon reaching ambient temperatures. The water nucleates into droplets, where gravity takes hold, quickly removing the droplet, with minimum gas contact. The gas, not affected by gravity, makes a 180° turn at the bottom of the tube and travels up the cold wall. This rapid separation greatly reduces the gas absorption in the water, verified by empirical test results. Further condensation of water occurs on the cold heat exchanger wall, until an exit dew point of 5(C is achieved. Sample gas "superheating" occurs when it flows over the non-insulated inlet tubing. This superheating results in an exit gas temperature of 55°F to 60°F. Further condensation of the remaining moisture in the sample gas is minimized by the higher sample temperature.

An optimal design for the heat exchanger calls for minimal surface area, low volume, and flexible materials of construction. The minimal surface area is required to minimize component "attenuation," loss in the heat exchanger, while allowing enough surface area for complete water condensation to take place. Minimal volume is required to minimize residence time, and to provide minimal pressure drops across the exchanger. Flexible materials of construction allow the application engineer a choice of materials in contact with the sample depending upon the acidic components likely to be present, the water volume to be removed, and the desired flow rate to be achieved. By constructing the heat exchanger in a variety of materials, many application design requirements can be achieved.

We have settled upon two heat exchanger options. These heat exchangers are identical to Figure 2, differing only in length. Empirically, we have found a relationship of residence time, water concentration, and gas component removal is optimal at a rule of thumb of 0.5 L/m per inch of heat exchanger length. Thus we use a 5-inch and a 10-inch heat exchanger. The rule of thumb of 0.5 L/m per inch holds true in this figure for water concentrations below 50 percent. TE cooler performance curves have great flexibility with the heat exchanger design. Various increasing flow rates can be achieved depending upon the inlet process load (inlet water concentration).

Thermoelectric coolers use high speed heat exchangers to make the water separation. The design of the impinger, limits the condensing, liquid water contact with the highly absorptive criteria pollutants in the exhaust gas. By coupling this high speed heat exchanger design with rapid electronic temperature control, the thermoelectric cooler reacts rapidly and directly to changing process loads, and can be sized to the application, reducing energy consumption in the sampling system. The refrigeration type condenser, on the other hand, must be oversized to the application, because it operates on a steady state control principle. By definition, it cannot rapidly react to process input loads.

This article originally appeared in the 10/01/1999 issue of Environmental Protection.

About the Author

Steven A. Jaasund, PE, is the president at Geoenergy International Corp. in Seattle, Wash. The author can be reached via their Web site at www.geoenergy.com.