A1 Methodology
A1.1 Introduction
In addition to lowering the air pollutant emission levels, the Emission Control Project will also result in lowering the temperature and exit velocity of the gases leaving the CPB stacks. This creates the potential that at some near-field receivers, the ground level concentrations of SO2 and NOx may be higher due to the lower plume rise, despite the reductions in emissions. The Study Brief stipulates that “The Applicant may carry out a comparative study to demonstrate if the stack emission impacts of the “B” Units before and after the Project will lead to lower air quality impacts at the Air Sensitive Receivers (ASRs) by using either a simple screening tool such as ISCST3 Gaussian model or a more sophisticated tool, such as wind tunnel test, if necessary”. The following sub-sections describe the methodology and results of such comparative assessment, using a Wind Tunnel testing methodology.
A1.2 Wind Tunnel Methodology (General)
The spatially and temporally variable meteorological and atmospheric dispersion conditions associated with complex terrain pose several challenges to assessing the dispersion of airborne pollutants in a coastal, mountainous region such as the Study Area. Physical scale wind tunnel modelling accounts for building wake and complex terrain effects, and is one of the most accurate methods for the simulation of these near-field influences for neutrally stable atmospheric conditions.
In general, wind tunnel air quality studies involve placing a physical model of the emission sources and surrounding terrain in a wind tunnel, emitting a passive tracer from the sources and measuring its concentrations at a number of receivers inside the wind tunnel for different wind speeds and directions. The raw results come in the form of Concentration Ratios expressing the dilution of the pollutant from a source to the identified receptor for a given wind speed and direction. The concentration ratios depend on the source and receptor locations and the source characteristics, such as release height and exit temperature and velocity, and on the meteorological conditions tested, but do not depend on the emission levels of particular pollutants. However, the concentration ratios can be easily converted into the predicted concentrations of different pollutants by a simple scaling procedure based on the tracer concentration at the source and the corresponding pollutant concentrations.
A1.3 The Present Study
Wind tunnel tests for this
study were conducted by Rowan Williams Davies & Irwin Inc. (RWDI) of
It should be noted that, as discussed by Snyder, it is generally impossible to match all dimensionless parameters (eg Reynolds number, Froude number, Rossby number and Peclet number) at the same time for model scales greater than 10:1. The guidance provided by Snyder and briefly summarized below gives a detailed evaluation of the appropriate criteria for non-dimensional parameters and discusses their limitations and conditions for their relaxing.
The Rossby number is an indicator of how important the Coriolis accelerations are when compared with advective accelerations. A large Rossby number means small Coriolis accelerations, and therefore the enhanced dispersion due to wind shear may be ignored. Published literature and data provide information regarding the length scales considered borderline for consideration of Rossby number. These distances are in the range of 5 to 12 km from the source under neutral or stable atmospheric conditions in flat terrain. The Coriolis forces may also be ignored when modelling rugged terrain, in which case the flow is dominated by the advective forces. As the modelling area for this Study is in line with these criteria, the Coriolis forces may be ignored.
The square of the Froude number represents the ratio of inertial forces to buoyancy forces. A large value of the Froude number means that buoyancy forces are small compared to inertial forces and as such, simulation of the nondimensional parameter is not required and this is the case for this Study. In addition, the dispersion in the study area is dominated by mechanical mixing as a result of the wind interaction with the rugged terrain.
The Peclet and Reynolds-Schmidt numbers are products of Reynolds number and a ratio of molecular transport coefficients. Turbulent diffusion in the atmosphere dominates molecular diffusion in turbulent flows, therefore the effect of not matching the Peclet and Reynolds-Schmidt numbers of the prototype in the model is not significant.
The Reynolds Number is a measure of the turbulence of the flow. The scale reductions typically result in model Reynolds numbers that are three to four orders of magnitude lower than in the atmosphere. Flow modelling is made possible by the Reynolds number independence theory that stipulates that for a specified flow system, in the absence of Coriolis and thermal effects, the turbulent flow structure is similar at all sufficiently high Reynolds numbers. Essentially, geometrically similar flows are similar at all sufficiently high Reynolds numbers. Therefore, rather than reproducing the Reynolds number exactly, it is sufficient to ensure that it remains above a critical threshold, which can be lowered by placing trip wires and flow restrictions inside the model stack. This method was used in this project: the stacks were tripped, and the experimental data confirm that adequate Reynolds numbers were maintained during all wind tunnel tests.
A 1:2000 scale model of the site including the existing plant and all surrounding terrain was constructed and put on a disk (see Figure A.1). In order to quantify impacts on regions that cannot be accommodated on the disk, extensions were constructed. Another disk was constructed for some areas where the separation distance between the source and receptor location was greater than 12 km.
The completed model and extensions were placed in RWDI’s Boundary layer Wind Tunnel (BLWT) and rotated to simulate any desired wind direction. The diameter of the turning table was 4.87 m corresponding to a full scale value of approximately 9.7 km. Thus, the actual constructed (physical) model was in fact much larger and sections of the model had to be moved during the testing procedures. To facilitate measurements at the receptors located farther away, the exhaust stacks were placed upwind on the disk and ASRs were installed on downwind model extensions.
Similarly, the fan speed was adjusted to simulate a variety of wind speeds. Turbulence generators were placed in the tunnel upwind of the model to simulate the mean and turbulent characteristics of the approaching wind in an open setting.
The actual measurements using carbon monoxide were preceded by flow visualization tests, performed as a qualitative assessment to understand how the exhaust plume was affected by different meteorological conditions and terrain effects.
The wind tunnel simulated the winds approaching the Study Area, the exhaust discharged from the source being tested and the dispersion of the exhaust in the atmosphere. The wind tunnel tests were conducted by emitting a tracer gas at a scaled concentration and flow rate for each of the scenarios tested. Carbon monoxide (CO) was used as the tracer gas in this study and it was mixed with helium to simulate the buoyant characteristics of the exhaust plume. The CO was introduced into each exhaust source at an initial concentration, Co, of approximately 6% (60,000 ppm). Mean concentrations of CO were measured at identified sensitive receivers by drawing samples through tubes (receptors) leading to a bank of infrared analyzers stationed outside the tunnel. The measured CO concentration at each receptor location was compared to the CO concentration at the stack and the results are presented as concentration ratios. The concentration ratios presented in Annex A represent the percentage of the initial source concentration, measured at each receptor, for a wide range of conditions.
Quantitative tracer gas testing was performed for five wind speeds (2.8, 5.1, 7.4, 9.7 and 12 m s-1 referenced to 10 m full scale above ground in open terrain) and all wind directions necessary to assess the plume impacts at all receptor locations. Wind directions were adjusted in 5 degree increments. Receptor locations were selected according to the Study Brief.
In order to accurately simulate the exhaust plume for each scenario, detailed scaling of the full scale exhaust parameters and approaching wind was performed. A combination of both momentum and buoyant scaling, following the classic approach of Snyder [1]was implemented for the modelled exhaust.
A1.4 Exhaust parameters Studied
The exhaust parameters for the modelled exhaust source before and after the retrofit are summarized in Table A.1. Scenario 1 is representative of operating conditions before the implementation of the retrofit programme. All four flues have the same exhaust parameters for this scenario. Scenario 2 is representative of operating conditions after the implementation of the retrofit programme. For this scenario, there was a decrease in the exhaust flow rate, temperature and exit velocity, as shown in Table A.1. All four flues have the same exhaust parameters for Scenario 2.
Table A.1 Source Exhaust Parameters
|
Scenario |
Number of Stacks |
Chimney Height (m) |
Inside Flue Diameter (m) |
Flow Rate (m3s-1) |
Efflux Temp. (oC) |
Efflux Velocity (m s-1) |
|
Scenario 1 - Before Retrofit |
4 flues in one chimney |
250 above ground |
6.6 each flue |
821 each flue |
110 each flue |
24.0 each flue |
|
Scenario 2 - After Retrofit |
4 flues in one chimney |
250 above ground |
6.6 each flue |
756 each flue |
80 each flue |
22.1 each flue |
The stack exit concentrations of SO2, NOx and particulates before and after the retrofit were assumed according to the current licence limits[2] and the anticipated emission reduction efficiencies of 90%, 80% and 60% respectively.
A1.5 Air Sensitive Receivers in the Wind tunnel
A total of 36 ground level and elevated receptors, were installed on the scale model to represent 23 locations of Air Sensitive Receivers (ASRs) at which the comparison between Scenario 1 and 2 was measured. Thirteen of these locations are represented by two receptors installed at different levels above the ground. The ASRs are listed in Table 3.3 of the main report; their locations shown in Figure A.2. Note that the location of the exhaust source as well as the two modelling disks configuration is also shown in Figure A.2.
Figure A.1 Physical Model inside the Wind Tunnel
Figure A.2 Exhaust Source and Locations of Wind Tunnel Air Sensitive Receivers (ASRs)
A2 Results
A2.1 Measured Concentration Ratios
Measured concentration ratios (%) for each of the tracer gas tests performed before and after the retrofit are presented in Tables A2 and A3 below. These tables summarize concentration ratios (presented as a percent) for each receptor location and for each scenario tested. The meteorological conditions of wind speed and wind direction at which each measurement was made are also identified. The concentration ratios, together with the assumed source concentrations of each pollutant form the basis for predicting of pollutant concentrations at each receptor before and after the retrofit.
Table A.2: Concentration Ratios(%) – Before Retrofit
Table
A.2 (continued): Concentration Ratios(%) – Before Retrofit
Table
A.2 (continued): Concentration Ratios(%) – Before Retrofit
Table A.2 (continued): Concentration Ratios(%) – Before Retrofit
Table
A.2 (continued): Concentration Ratios(%) – Before Retrofit
Table A.2 (continued): Concentration Ratios(%)
– Before Retrofit
Table
A.3: Concentration Ratios(%) – After Retrofit
Table
A.3 (continued): Concentration Ratios(%) – After Retrofit
Table
A.3 (continued): Concentration Ratios(%) – After Retrofit
Table
A.3 (continued): Concentration Ratios(%) – After Retrofit
Table
A.3 (continued): Concentration Ratios(%) – After Retrofit
Table
A.3 (continued): Concentration Ratios(%) – After Retrofit
A2.2 Air Quality Improvements under Worst-Case Measured Concentration Ratios
The estimated relative reductions in concentrations of SO2, NOx and particulates after the retrofit for the worst-case meteorological conditions (ie those for which the highest concentration ratios were measured) are presented in Table A.4. The wind speeds shown in the table are referenced to 10 m above grade, in open terrain. The concentration estimates are based on pollutant concentrations obtained for each receptor by multiplying the assumed stack exit concentrations corrected for the actual efflux temperature, by the worst-case concentration ratios before and after the retrofit obtained from Tables A.2 and A.3. The calculation steps are as follows:
Step 1
The stack exit concentration before the retrofit expressed in mg Nm-3 shown in Table A.1 is first converted to concentration expressed in mg m-3:
e1’ = e1 x 273/383
where e1 = stack exit concentration in mg Nm-3 before the retrofit
e1’ = stack exit concentration in mg m-3 before the retrofit
Step 2
The worst-case concentration before the retrofit at a given sensitive receiver is obtained by multiplying the stack exit concentration before the retrofit by the highest concentration ratio at the same sensitive receiver (from Table A.2):
c1 = e1’ x r1
where c1 = worst-case concentration before the retrofit at receiver in mg m-3
r1 = highest concentration ratio for the same receiver before the retrofit
Step 3
The stack exit concentrations after the retrofit expressed in mg Nm-3 shown in Table A.1 are first converted to concentrations expressed in mg m-3
e2’ = e2 x 273/353
where e2 = stack exit concentration in mg Nm-3 after the retrofit
e2’ = stack exit concentration in mg m-3 after the retrofit
Step 4
The worst-case concentration after the retrofit at the same sensitive receiver considered in Step 2 is obtained by multiplying the stack exit concentration after the retrofit by the highest concentration ratio at the same sensitive receiver (from Table A.3):
c2 = e2’ x r2
where c2 = worst-case concentration after the retrofit at receiver in mg m-3
r2 = highest concentration ratio for the same receiver after the retrofit
Step 5
The relative reduction in concentration of an air pollutant at the sensitive receiver given above is obtained as follows:
R = (c1 - c2) / c1
where R = relative reduction in concentration at the given receiver
Table A.4 Modelled Air Quality Improvement at ASR under Worst-Case Concentration Ratios
|
ASR |
Wind Speed (m s-1) |
Wind Direction (deg) |
Modelled Change in SO2 Concentration (%) |
Modelled Change in NOx Concentration (%) |
Modelled Change in Particulate Concentration (%) |
|
A1 |
NA(a) |
NA(a) |
NA(b) |
NA(b) |
NA(b) |
|
A2 |
12 |
180 |
-85.9 |
-71.8 |
-43.7 |
|
A3 |
12 |
175 |
-87.4 |
-74.8 |
-49.7 |
|
A4 |
12 |
200 |
-89.0 |
-78.0 |
-56.0 |
|
A5 |
9.7 |
205 |
-88.3 |
-76.7 |
-53.3 |
|
A6a |
7.4 |
235 |
-89.5 |
-78.9 |
-57.9 |
|
A6b |
7.4 |
235 |
-89.5 |
-79.0 |
-57.9 |
|
A7a |
7.4 |
235 |
-90.0 |
-80.0 |
-60.0 |
|
A7b |
7.4 |
235 |
-90.3 |
-80.7 |
-61.4 |
|
A8a |
12 |
230 |
-89.6 |
-79.2 |
-58.3 |
|
A8b |
12 |
235 |
-89.6 |
-79.2 |
-58.3 |
|
A9a |
7.4 |
250 |
-89.3 |
-78.6 |
-57.2 |
|
A9b |
5.1 |
250 |
-89.6 |
-79.2 |
-58.4 |
|
A10a |
12 |
270 |
-89.4 |
-78.8 |
-57.5 |
|
A10b |
12 |
265 |
-89.5 |
-79.0 |
-58.0 |
|
A11a |
9.7 |
270 |
-89.8 |
-79.6 |
-59.3 |
|
A11b |
7.4 |
270 |
-90.3 |
-80.6 |
-61.2 |
|
A12a |
7.4 |
270 |
-91.4 |
-82.6 |
-65.7 |
|
A12b |
7.4 |
275 |
-89.6 |
-79.2 |
-58.4 |
|
A13 |
12 |
270 |
-88.7 |
-77.4 |
-54.8 |
|
A14a |
12 |
280 |
-86.9 |
-73.9 |
-47.8 |
|
A14b |
12 |
280 |
-86.6 |
-73.2 |
-46.5 |
|
A15a |
NA(a) |
NA(a) |
NA(b) |
NA(b) |
NA(b) |
|
A15b |
NA(a) |
NA(a) |
NA(b) |
NA(b) |
NA(b) |
|
A16 |
7.4 |
310 |
-89.6 |
-79.1 |
-58.2 |
|
A17 |
9.7 |
330 |
-89.7 |
-79.3 |
-58.7 |
|
A18a |
5.1 |
345 |
-90.0 |
-80.0 |
-60.0 |
|
A18b |
7.4 |
345 |
-89.7 |
-78.2 |
-58.6 |
|
A19a |
9.7 |
345 |
-89.7 |
-79.5 |
-59.0 |
|
A19b |
7.4 |
345 |
-89.8 |
-79.5 |
-59.1 |
|
A20 |
5.1 |
10 |
-88.8 |
-77.6 |
-55.2 |
|
A21a |
7.4 |
220 |
-88.9 |
-77.8 |
-55.7 |
|
A21b |
7.4 |
220 |
-88.5 |
-77.0 |
-53.9 |
|
A22 |
7.4 |
265 |
-90.0 |
-80.0 |
-60.0 |
|
A23a |
12 |
260 |
-89.9 |
-79.7 |
-59.4 |
|
A23b |
9.7 |
260 |
-90.2 |
-80.5 |
-61.0 |
|
Notes: (a) 0.0010 is the lowest detection limit for concentration ratios. When all the readings in the wind tunnel test measurements are below the detection limit, establishing which wind speed and direction results in the highest concentration is not possible. (b) With concentration ratios below the detection limits, changes in concentrations cannot be estimated. |
|||||
The results shown in Table A.4 represent the resultant reduction of SO2, NOx and particulate concentrations, hence air quality improvements at the ASRs, under the worst-case concentration ratios measured in the wind tunnel for each of the two scenarios tested. The results demonstrate that the percentage reductions in the measured SO2, NOx and particulate concentrations at the ASRs are very similar to the corresponding reductions of SO2 (90%), NOx (80%) and particulate (60%) emissions from the CPB units after retrofit. The slight changes in the percentage reduction of the measured concentrations can be explained by the changes in the plume characteristics (i.e. a lower exit velocity and a lower efflux temperature after implementation of the retrofit programme) and the effects of the complex terrain on the exhaust dispersion.
In conclusion, the effects of changes in dispersion characteristics (due to changes in flue gas physical properties) on the pollutant concentrations at ASRs after the retrofit are much lower in magnitude when compared with those of the expected emission reductions of SO2, NOx and particulates. As can be seen in Table A.4, the combined effects of the emission reductions and the changes in physical characteristics of the plume result in air quality improvements at all the identified ASRs ranging from 86% to 91% for SO2, 72% to 83% for NOx. Similarly, assuming a 60% reduction of particulate emissions, corresponding reductions in particulate concentrations ranging from 44% to 66% at ASRs are expected.
It should be noted that the above reductions in predicted worst-case concentrations are related to the Castle Peak Station “B” emissions only and do not include cumulative effects. While the assessment of cumulative impacts is not the focus of this EIA Study, it is anticipated that (with all other emissions assumed constant) the Project will result in an improvement in SO2, NOx and particulates levels in the vicinity of the Castle Peak Power Station. The magnitude of such improvements will of course be lower for sensitive receivers located further away from the CPPS.