Advanced Air Quality Control

MEE 6501, Advanced Air Quality Control 1

Course Learning Outcomes for Unit VII Upon completion of this unit, students should be able to:

1. Describe methods for monitoring air pollution.

2. Critique air pollutant modeling equations and software. 2.1 Discuss statistical methods for modeling air pollutants. 2.2 Discuss software options for modeling air pollutants. 2.3 Calculate operational air emission rates for a selected scenario.

3. Assess health effects of air pollution.

4. Examine causes of indoor and outdoor air pollution.

5. Evaluate health risks of air pollution exposure.

6. Estimate the impact of air pollution on the environment.

7. Evaluate air pollution control technologies.

Course/Unit Learning Outcomes

Learning Activity

1 Unit VII Course Project

2.1

Unit Lesson Chapter 3, pp. 77-98 Chapter 7, pp. 269-277 Unit VII Mini Project

2.2

Unit Lesson Chapter 3, pp. 77-98 Chapter 7, pp. 269-277 Unit VII Course Project

2.3 Unit Lesson Unit VII Course Project

4 Unit VII Course Project

5 Unit VII Course Project

6 Unit VII Course Project

7 Unit VII Course Project

Reading Assignment Chapter 3: Atmospheric Dispersion, Transport, and Deposition, pp. 77–98 Chapter 7: Air Quality and Emissions Assessment, pp. 269–277

Unit Lesson In Unit VI, we touched briefly on the need to understand the fundamentals of statistical data analysis, given that it impacts our ability to read and understand our laboratory analysis reports. However, in this unit, we are

UNIT VII STUDY GUIDE

Engineering Air Quality Monitoring Systems, cont.

MEE 6501, Advanced Air Quality Control 2

UNIT x STUDY GUIDE

Title

about to realize an even more pronounced need to understand these basic concepts. We are now going to learn to understand how our engineered air quality interacts, and subsequently impacts, our own atmosphere. We are now ready to investigate how these statistical tools are inherent in our mathematical and software models, to which we have access when working with air quality studies. Models Godish, Davis, and Fu (2014) focus our attention on Gaussian models (specifically Gaussian plume models) for dispersion modeling, given regulatory authorities’ propensity for requesting analysis by this model type. However, there are many more dispersion models that we need to consider in order to understand our options, with or without software assistance. We need to understand that there are models better suited for different atmospheric and structural considerations. It may be easier to tabulate these first, then encourage you to investigate further with a cursory Internet search for each software solution identified from the literature (Godish et al., 2014; Phalen & Phalen, 2013; Gurjar, Molina, & Ojha, 2010):

Dispersion Model Structural Basis

Box model Mass-balance equation of pollutants

Gaussian plume model Steady-state atmospheric conditions

Gaussian puff model Temporal/spatial/variable atmospheric conditions

Lagrangian model Specific fluid pollutant trajectory

Eulerian model Specific fluid properties in a control volume at a fixed point

CFD model Fluid pollutant flows in a complex geometric space

Further, each of these dispersion models is supported with commercial software for air quality engineers like us. These may be tabulated by dispersion model type. However, the engineer must be very careful to select software that either supports or does not support aerosol dynamics, depending on the specific monitoring situation (Gurjar et al., 2010):

Dispersion Model Commercial Software Solution

Box model AURORA, CPB, PBM

Gaussian plume model CALINE4, HIWAY2, CAR-FMI, OSPM

Gaussian puff model CALPUFF, AEROPOL, AERMOD, UK-ADMS, SCREEN3

Lagrangian model GRAL, TAPM, FARM

Eulerian model GATOR, MONO320SPM, UHMA, CIT, RUM-1ATM, RPM, AEROFOR2, UNI-AERO, CALGRID, STEM, WRF-Chem, CMAQ, MADRID, CAMx, PMCAMx

CFD model ARIA, MISKAM, MICRO-CALGRID, ATMoS

Approaches Functionally, there are several different ways to approach the air data that differ by mathematical theory. From our reading of Godish et al. (2014), we understand that the goal of modeling is to calculate concentrations of a given contaminant or pollutant for a known set of variables. As such, we can analyze data using the following approaches (Gurjar et al., 2010): (a) artificial neural network, (b) fuzzy logic, (c) ranking, or (d) time series. Let’s look at each one of these conceptually and briefly. The artificial neural network (ANN) approach (available in a software solution) is helpful when known variables are available to calibrate regression analysis tools. For example, we might consider using this approach if we are attempting to quantify air quality downstream of a known polluting operation just a few miles from the source (Gurjar et al., 2010).

MEE 6501, Advanced Air Quality Control 3

UNIT x STUDY GUIDE

Title

The fuzzy logic approach is commonly used for ranking multiple air quality models in order to normalize and subsequently formalize limits (Gurjar et al., 2010). In other words, we might consider using fuzzy logic in order to establish upper limits of model forecasts (not unlike probability charts in quality engineering for control charts with upper control limits and lower control limits) when evaluating different data outcomes and projected emission forecasts from two or more air models, as described above in our first table. The idea is to make the decision-making easier with rather imprecise data represented with relatively higher levels of uncertainty (such as forecasting air quality with no known background data for a new site). Other iterations of fuzzy logic include fuzzy inference that approaches the data similarly, but with the use of weighted variables (Gurjar et al., 2010). The ranking approach affords the engineer the ability to rank the air models by known variables of concern or focus (Gurjar et al., 2010). For example, if we were to use all four of the listed software solutions in the second table for calculating a plume concentration (CALINE4, HIWAY2, CAR-FMI, OSPM), we could use the ranking approach to statistically evaluate which models were more appropriate to our situation (by level of appropriateness). With this, we could still utilize all of the resulting data from all four models instead of attempting to use only one model and throwing out the data from the other three models as being irrelevant. This is accomplished by mathematically evaluating the statistical index scores and subsequently comparing indices among the ranked software solutions.

Finally, we could consider using the time series approach of analysis when we are attempting to accurately forecast pollutant concentrations in areas of high pollutant susceptibility as a function of time (Gurjar et al., 2010). For example, we might be studying the impacts of a refinery near a neighborhood in an attempt to discourage people from exercising or working outdoors during air quality alerts. Analyzing air quality against time provides a high level of statistical reliability and is consequently effective in working within behavioral engineering modification controls, such as limiting or discouraging exposures during time-specific points. With this new information regarding air quality modeling, let’s focus again for a few minutes on our spray booth for our course project. We need to finish just a few more calculations in order to be able to adequately model our anticipated air quality emissions from our engineered design. Given that our interior lining cure equipment (natural gas-fired heater) matches the state specifications, we are going to reference the state’s natural gas unit emission factors tabulated as “Firing Rate Between 0.3 MMBtu/hr (lb106 scf) and 100 MMBtu/hr (lab/106 scf)” that would have also been provided to us for our calculations.

Atmospheric conditions cause plume variations not factored in the Gaussian model. (Santa Maria, 2010; Chrishowey, n.d.)

MEE 6501, Advanced Air Quality Control 4

UNIT x STUDY GUIDE

Title

We will use the following referenced steps to calculate our air contaminant analyses generated from our natural gas-fired cure heater for our air permit application. We will be using the units in our calculations for million British thermal units/hr (MMBtu/hr), thousand British thermal units/hr (MBtu/hr), and standard cubic feet/hr (SCF/hr). For example, 1.0 lbs/MMscf would be 1.0 lbs of contaminant (pollutant) per one million standard cubic feet. For our first set of calculations, we will be calculating the short-term (hourly) emissions generated from the heater. The following formula would apply:

Lbs of air contaminants/hour = lbs air contaminant x 1.0 scf x 2.1 MMBtu MMscf 1,020Btu 1.0 hr

First, we reference our scenario for the technical information referenced for the Interior Liner Cure heater and see that our natural-gas fired cure heater has a firing rate of 2.1 MMBtu/hr, and that we anticipate firing liners in the curing process for a maximum of 2,500 hours/year. We then reference each contaminant limit as tabulated here for our course project (use these tabulated values in your course project calculations):

Contaminant Categories Mass Equivalent (lb/MMscf)

NOx 100.0

CO 84.0

PM 7.6

VOC 5.5

SO2 0.6

Note that NOx content is tabulated as100 lb/MMscf (lbs/million scf). Next, we multiply 100 lb NOx/MMscf by 1 scf/1,020 Btu by 2.1 MMBtu/hr to derive a value for lbs of NOx/hr. For example, for a tabulated 200.0 lb/ MMscf, [Note: The actual scenario needs to be calculated with actual tabulated value of NOx at 100.0 lb MMscf]:

Lbs of air contaminants/hour = lb air contaminant x 1.0 scf x 2.1 MMBtu MMscf 1,020Btu 1.0 hr

= 200.0 lb/MMscf x 1.0/1,20 Btu x 2.1 MMBtu/1.0 hr

= 0.412 lb of NOx/hr

Now, we simply do the same calculation for each of the remaining four individual contaminant categories tabulated for our scenario (lb of CO/hr, lb of PM/hr, lb of VOC/hr, and lb of SO2/hr). For our second set of calculations, we will be calculating the long-term (annual) emissions generated from the heater. In order to accomplish this, we simply convert the hourly emissions (performed above) to annual emissions (2,500 hours/yr), and then to tons (2,000 lbs/ton) to derive an annual tons of air contaminant per year. The following formula would apply:

Tons of air contaminant/hour = lbs air contaminant x 2,500 hr x 1.0 ton 1.0 hr 1.0 yr 2,000 lb

For example, our first unit conversion would be for NOx. First, we note that our calculated hourly NOx is 0.412 lb/hr. Second, we multiply the calculated lb of NOx/hr by 2,500 hrs/yr by 1 ton/2,000 lb to derive a value for ton NOx/yr. For example, for a calculated 0.412 lb of NOx/hr, [Note: The actual scenario needs to be calculated with actual calculated value of NOx at 100.0 lb/MMscf or 0.206 lb of NOx/hr]:

MEE 6501, Advanced Air Quality Control 5

UNIT x STUDY GUIDE

Title

Tons of air contaminant/hour = lbs air contaminant x 2,500 hr x 1.0 ton 1.0 hr 1.0 yr 2,000 lb

= 0.412 lb of NOx/1.0 hr x 2,500 hr/1.0 yr x 1.0 ton/2,000 lb

= 0.515 ton of NOx/yr

Now, we simply do the same calculation for each of the remaining four individual contaminant categories tabulated for our scenario (ton of CO/yr, ton of PM/yr, ton of VOC/yr, and ton of SO2/yr). After we complete all of the short-term air pollutant generation rate conversions to long-term generation rates, we are finished with our air permit evaluation math. I think you will agree that none of the math was overly complicated, but rather all of the steps were understandable after we broke up the process over several units. What you have effectively done is to mathematically model the air emissions from the interior spray booth operations, but without the booth even being in operation for a single second. This is the power of mathematical forecasting with models! In your air quality engineering work in industry, you will often find that your ability to break up these complicated mathematical models into smaller subsets of calculations is one of the most powerful engineering skills that you can employ. Remember that our role as engineers is to effectively forecast emissions before the operation begins and be able to model the environmental impacts before humans or ecological life is injected into the system and subsequently put at risk. We statistically model air quality in order to mitigate air pollution risks to all life in our precious environment. You have done a wonderful job in protecting lives with your work on this air permit evaluation project!

References

Chrishowey. (n.d.). Polluting smokestack, (ID 3629517) [Photograph]. Retrieved from https://www.dreamstime.com/royalty-free-stock-photography-polluting-smokestack-image3629517

Godish, T., Davis, W. T., & Fu, J. S. (2014). Air quality (5th ed.). Boca Raton, FL: CRC Press. Gurjar, B., Molina, L., & Ojha, C. (2010). Air pollution: Health and environmental impacts. Boca Raton, FL:

CRC Press. Phalen, R. F., & Phalen, R. N. (2013). Introduction to air pollution science: A public health perspective.

Burlington, MA: Jones & Bartlett Learning. Santa Maria, R. G. (2010) Smokestack billowing smoke, ID 16425832 [Photograph]. Retrieved from

https://www.dreamstime.com/stock-photography-smokestack-billowing-smoke-image16425832