Annual Guide 2023 I SOURCE TESTING ASSOCIATION
The importance of representative sampling is highlighted in the Industrial Emissions Directive which specifi es that the Competent Authority is required to approve the sampling location.
Standards and Acceptance Criteria
Prior to the adoption of EN 15259, relatively simple criteria were used for the assessment of sampling representativeness and these are considered fi rst since they are relevant when considering the positional uncertainty, in the pollutant concentration, that is associated with single point sampling.
ISO 10396 [7] notes that, whilst it is important that the sampling point is not located near any obstructions that could seriously disturb the gas fl ow in the stack, the pollutant of interest may still have ‘cross-sectional variation’. ISO 10396 therefore requires that a stratifi cation survey is conducted using an automated analytical method to measure the concentration of the pollutant (restricted to NOx
STANDARDS AND ACCEPTANCE CRITERIA
multiple points of equal area across the measurement plane (two minutes sampling duration at each point). The pollutant or diluent gas is considered to be ‘non-stratifi ed’ if the concentration at all of the individual traverse points is within ±10% of the arithmetic average concentration. Temporal fl uctuations are monitored by simultaneously sampling from the centre of the stack. If the gas concentration is non-stratifi ed, any sampling point located between one-third and half of the stack diameter can be selected. If the gas concentration is ‘stratifi ed’ an alternative monitoring location or a permanent multi-point grid sampling system is required.
, SO2 or CO), and the diluent concentration (O2 or CO2
US EPA Method 7E [8] recommends that measurement planes are located at least eight stack diameters downstream and at least two diameters upstream of the nearest fl ow disturbances. Method 7E also requires a stratifi cation survey is conducted, for a single pollutant or diluent, in order to determine the appropriate number of representative sampling points. The pollutant is considered to be ‘non-stratifi ed’ if the concentration at all of the individual traverse points is within ±5% of the average concentration and is ‘minimally stratifi ed’ if all of the traverse point concentrations are within ±10% of the average.
Prior to the adoption of EN 15259, relatively simple criteria were used for the assessment of sampling representativeness and these are considered first since they are relevant when considering the positional uncertainty, in the pollutant concentration, that is associated with single point sampling.
), at
ISO 10396 [7] notes that, whilst it is important that the sampling point is not located near any obstructions that could seriously disturb the gas flow in the stack, the pollutant of interest may still have ‘cross-sectional variation’. ISO 10396 therefore requires that a stratification survey is conducted using an automated analytical method to measure the concentration of the pollutant (restricted to NOx, SO2 or CO), and the diluent concentration (O2 or CO2), at multiple points of equal area across the measurement plane (two minutes sampling duration at each point). The pollutant or diluent gas is considered to be ‘non-stratified’ if the concentration at all of the individual traverse points is within ±10% of the arithmetic average concentration. Temporal fluctuations are monitored by simultaneously sampling from the centre of the stack. If the gas concentration is non-stratified, any sampling point located between one-third and half of the stack diameter can be selected. If the gas concentration is ‘stratified’ an alternative monitoring location or a permanent multi-point grid sampling system is required.
US EPA Method 7E [8] recommends that measurement planes are located at least eight stack diameters downstream and at least two diameters upstream of the nearest flow disturbances. Method 7E also requires a stratification survey is conducted, for a single pollutant or diluent, in order to determine the appropriate number of representative sampling points. The pollutant is considered to be ‘non-stratified’ if the concentration at all of the individual traverse points is within ±5% of the average concentration and is ‘minimally stratified’ if all of the traverse point concentrations are within ±10% of the average.
EN 15259:2007 [4] recommends that measurement planes are located at least five stack diameters (5D) upstream and downstream from the nearest flow disturbances or at least two diameters (2D) upstream of the stack exit. In common with the methods described earlier, for the purposes of conducting a duct survey, a stack of greater than 2m diameter (D), with a circular cross-section, is sub-divided into equal areas giving the sample points shown in Figure 1, located at xi. The larger the duct cross-sectional area, the larger the number of sampling points, noting that 20 points are considered to be sufficient for large stacks. A similar approach is adopted for rectangular exhaust ducts, as shown in Figure 2.
a) l1/l2 ≤ 2 b) l1/l2 > 2
Figure 2. Sample point positions at centres of equal area in rectangular ducts.
Figure 2 Sample point positions at centres of equal area in rectangular ducts
It should be noted that the recommended minimum lengths of straight ductwork, either side of the measurement plane, are focussed on fl ow uniformity, i.e., the avoidance of fl ow recirculation, swirling fl ow and extremes in maximum to minimum velocity ratios. These guidelines do not address inhomogeneity in gas concentration. If an abatement unit, such as a Flue Gas Desulphurisation (FGD) unit at a coal fi red power plant, is located only 5 diameters upstream of the measurement plane then signifi cant stratifi cation/inhomogeneity in concentration may still be present.
EN 15259 specifi es a different approach for assessing homogeneity of the gases when compared with previous standards. The standard deviation of the spatially varying
measuring instrument. If the concentration distribution is deemed to be homogeneous, an SRM sample can be extracted from any point within the measurement plane.
describes the spatial variation across the measurement plane but also incorporates the temporal variations that are described by sref
. The quantity sgrid . Both sgrid and sref
If the spatial variation is less than the temporal variation (sgrid
homogenous. Otherwise, an F-factor is calculated as the ratio of the two variances (F-factor = sgrid2
≤ sref
be more than about 50% higher than the temporal variation in order for the spatial distribution to be considered homogenous.
The distribution is then homogenous when sgrid2 equivalent to sgrid
< 1.47 * sref a) General method (including centre-point) circular ducts.
sectional area, the larger the number of sampling points, noting that 20 points are considered to be suffi cient for large stacks. A similar approach is adopted for rectangular exhaust ducts, as shown in Figure 2.
EN 15259:2007 [4] recommends that measurement planes are located at least fi ve stack diameters (5D) upstream and downstream from the nearest fl ow disturbances or at least two diameters (2D) upstream of the stack exit. In common with the methods described earlier, for the purposes of conducting a duct survey, a stack of greater than 2m diameter (D), with a circular cross-section, is sub-divided into equal areas giving the sample points shown in Figure 1, located at xi
. The larger the duct cross- b) Tangential method (excluding centre-point)
Figure 1: Sample point positions at centres of equal area in large circular ducts Figure 1. Sample point positions at centres of equal area in large
Because the EN 15259 acceptance criterion is based on the relationship between spatial and temporal deviations, rather than the absolute thresholds discussed previously, it is possible for this test to deliver a false FAIL when both the spatial and temporal (process) variations are very low. This is often the case at Combined Cycle Gas Turbine (CGGT) plant, requiring additional national guidance to address this issue. For example, in England, the location is considered to be homogeneous when sgrid the Daily ELV or sgrid
< 5% of
has also highlighted the additional possibility of a false PASS of the homogeneity test when the spatial and temporal variations are both high [10].
< 0.3 %O2
dependent variance from the spatial variance: spos This positional uncertainty is considered later.
Following a failure of the homogeneity test, there is a further test to determine the severity of the inhomogeneity. The standard uncertainty relating to the spatial deviation alone, i.e., the positional uncertainty, spos
, is calculated by subtracting the time- = √(sgrid2
- sref2 for oxygen [9]. The power industry
homogeneous. The critical F-factor is defi ned as a function of the number of grid points (N). For the most common power plant case of N = 20, that is, the maximum number of points within a large duct, the value of FN-1;N-1;0,95
than a critical F-factor (FN-1;N-1;0,95 is 2.17. . That is, the spatial variation cannot /sref2 < 2.17,
) then the distribution is automatically assumed to be /sref2
It should be noted that the recommended minimum lengths of straight ductwork, either side of the measurement plane, are focussed on flow uniformity, i.e., the avoidance of flow recirculation, swirling flow and extremes in maximum to minimum velocity ratios. These guidelines do not address inhomogeneity in gas concentration. If an abatement unit, such as a Flue Gas Desulphurisation (FGD) unit at a coal fired power plant, is located only 5 diameters upstream of the measurement plane then significant stratification/inhomogeneity in concentration may still be present.
EN 15259 specifies a different approach for assessing homogeneity of the gases when compared with previous standards. The standard deviation of the spatially varying concentration measured across the above traverse points, sgrid, is first calculated. This is compared with the standard deviation of the time-dependent concentration measured during the traverse at a fixed reference point, sref. The quantity sgrid nominally describes the spatial variation across the measurement plane but also incorporates the temporal variations that are described by sref. Both sgrid and sref include the analytical repeatability of the measuring instrument. If the concentration distribution is deemed to be homogeneous, an SRM sample can be extracted from any point within the measurement plane.
concentration measured across the above traverse points, sgrid, is fi rst calculated. This is compared with the standard deviation of the time-dependent concentration measured during the traverse at a fi xed reference point, sref
If the spatial variation is less than the temporal variation (sgrid ≤ sref) then the distribution is automatically assumed to be homogenous. Otherwise, an F-factor is calculated as the ratio of the two variances (F-factor = sgrid2/sref2) and this must be less than a critical F-Factor (FN-1;N-1;0,95) for the distribution to be deemed homogeneous. The critical F-factor is defined as a function of the number of grid points (N). For the most common power plant case of N = 20, that is, the maximum number of points within a large duct, the value of FN-1;N-1;0,95 is 2.17. The distribution is then homogenous when sgrid2/sref2 < 2.17, equivalent to sgrid < 1.47 * sref. That is, the spatial variation cannot be more than about 50% higher than the temporal variation in order for the spatial distribution to be considered homogenous.
nominally include the analytical repeatability of the
Because the EN 15259 acceptance criterion is based on the relationship between spatial and temporal deviations, rather than the absolute thresholds discussed previously, it is possible for this test to deliver a false FAIL when both the spatial and temporal (process) variations are very low. This is often the case at Combined Cycle Gas Turbine (CGGT) plant, requiring additional national guidance to address this issue. For example, in England, the location is considered to be homogeneous when sgrid < 5% of the Daily ELV or sgrid < 0.3 %O2 for oxygen [9]. The power
) and this must be less ) for the distribution to be deemed
).
3
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48
