With EN ISO 16911-1 in the spotlight thanks to a recent Technical Report, Dave Curtis revisits the 2013 standard on flue gas flow rate measurement.
With a Technical Report published in March 2017 issuing guidance on the application of EN ISO 16911-1, this article looks back at the 2013 standard on flue gas flow rate measurement: EN ISO 16911 ‘Stationary Source Emissions – Manual and automatic determination of velocity and volume flow rate in ducts.
The Technical Report, PD CEN/TR 17078:2017, does not follow the numbering of EN ISO 16911-1:2013; however, for easier handling it uses the same headings and sub-headings as EN ISO 16911-1:2013. It does not repeat text, tables or diagrams from EN ISO 16911-1:2013, instead it refers to the relevant sections of the Standard.
It is essential, therefore, that the user has a copy of the Standard to refer to. For sections of the Standard where this Technical Report does not provide any text or guidance it is deemed that the relevant section does not require any additional clarification.
Combustion plant operators need to know the flue gas flow rate in order to calculate the mass release of pollutant emissions. To give the mass release rate in mg/s the flue gas flow rate (m3/s) is multiplied by the concentration (mg/m3) of pollutant, e.g., NOx.
Information on the mass release rate is required for numerous purposes, including:
- Emissions trading
- Compliance or inventory reporting
- Air quality modelling purposes
The current standard on flue gas flow rate measurement was published in 2013: EN ISO 16911 ‘Stationary Source Emissions – Manual and automatic determination of velocity and volume flow rate in ducts. As based on the original European Union mandate, the scope of the standard is linked to the requirements of European Directives, such as the EU Emissions Trading System (EU ETS) and Industrial Emissions Directive (IED). These allow for the alternative ‘measurement’ approach for CO2 and requires it for emissions of N2O and CH4 from other sectors, all subject to defined uncertainty requirements. European Directives require the use of CEN standards when available.
Divided into two parts, the standard first defines in Part 1 the manual Standard Reference Methods (SRM) to be used for the calibration of continuous stack flow monitors as well as for other compliance purposes, such as periodic testing. Following this, Part 2 of the standard specifies the requirements for the certification, calibration and ongoing control of continuous flow monitors, as applies to continuous monitoring.
“3D Pitot measures all three velocity components, including the axial velocity that is required for an unbiased flow rate determination”
Manual reference method
Part 1 of the standard is performance based, that is, a number of different techniques may be used as the manual reference method provided that the specified performance requirements are satisfied.
Alternative techniques include: velocity traverses with Pitot probes (various designs) or vane anemometers; tracer (dilution) and tracer (time-of-flight) methods. Under certain circumstances, flow calculation from fuel consumption can be used to perform compliance checks and a mandatory calculation approach is also provided in Part 1 (Annex E).
When measuring the velocity profile point velocity measurements are required in order to determine, for example, if a given measurement plane is suitable for the installation of a flow monitor. For this purpose any type of Pitot tube or vane anemometer with a traceable calibration can be used, with the proviso that the level of swirl is low (nominally less than 15° swirl angle at all traverse points). If the level of swirl is significant, however, then the traverse must be conducted using a 3D or 2D Pitot. It should be noted that a conventional S-type Pitot can be operated as a 2D Pitot with measurement of the swirl angle. As given away by its name, the 3D approach measures all three velocity components, including the axial velocity that is required for an unbiased flow rate determination.
One example of a 3D device is a spherical (5-hole) Pitot. This would be inserted into the flow and turned until one of the ΔP measurements is nulled. Wind tunnel calibration relationships are then used to calculate all three velocity components from the various measured ΔPs. For further information on this, the operation of 3D Pitots is described in detail in US EPA Method 2F.
The S-type Pitot, on the other hand, is commonly used to establish iso-kinetic sampling conditions when measuring dust concentrations. This is normally inserted into the flow so that the ‘impact’ orifice faces into the flow and the ‘wake’ orifice is then positioned at 180° to this. Operation as a 2D Pitot is described in detail in US EPA Method 2G. It is worth noting that if a Pitot tube is used in a configuration with a closely coupled gas-sampling probe, then the device must be calibrated in this configuration.
“EN15259 specifies the procedures for determining the required number and location of points. It notes that when it comes to EN ISO 16911, the ‘tangential method’ is required”
If it’s average velocity that you’re seeking to determine, the traverse points are located at centres of equal area so that a simple average of the point readings gives an area weighted average in a duct of circular cross-section. EN15259 specifies the procedures for determining the required number and location of points. It notes that when it comes to EN ISO 16911, the ‘tangential method’ is required, i.e., the centre-line of the duct cannot be included. In large ducts, 20 measurement points are normally sufficient.
The field trial validation indicated that lack of uniformity of the flow profile (Figure 3) caused by a poor measurement location did not significantly affect the average velocity determination. That is, a 20 point average from a poor flow profile gave the same result as a 20 point average from a uniform flow profile.
For each technique performance and quality assurance requirements are specified. For Pitot tubes, a pre-test leak check is required and, when using an electronic pressure reading device, a daily calibration check is required using a liquid manometer device (temperature corrected) or a calibrated pressure sensor with an uncertainty better than the test device. The repeatability also needs to be determined at a single measurement point (the standard deviation of five consecutive one minute velocity readings). Each point velocity measurement must be obtained from a one minute average ΔP based on a continuous measurement or at least three separate readings.
A velocity traverse to EN 15259 does not have sufficient resolution to capture the very low velocity boundary layer at the duct wall. For a large duct, this can optionally be measured according to US EPA Method 2H. However, the correction is usually very small and it is normally sufficient to multiply the measured average velocity by a Wall Adjustment Factor of 0.995 for a smooth duct or 0.99 for a rough (brick-lined) duct of circular cross-section. This is a requirement when calibrating a flow monitor.
Tracer transit time methods determine the bulk (average) velocity directly by recording the time taken for a tracer material to travel between two measurements stations (Δt). The distance between these two stations, situated in duct work of constant cross section, is divided by the measured time-of-flight to obtain the average velocity. The example in the standard is based on the injection of a radioactive tracer, upstream of the flue. Two sets of clamp-on detectors are then used to detect the arrival of the tracer at two different heights within the flue. The medians of the recorded tracer concentration peaks are extracted so that the shape of the detector response is taken into account to obtain an accurate Δt.
In order to obtain the volumetric flow rate the average velocity must be multiplied by the duct’s cross-sectional area. EN ISO 16911 requires the Test Laboratory to measure the duct dimensions, across at least two axes, rather than simply relying on plant drawings.
The tracer dilution method directly determines the flue gas flow rate and does not, therefore, require the cross-sectional area to be known. A tracer is injected into the flue gas, for a short period of time, well upstream of the flue, so that the tracer is intimately mixed with the flue gas. The concentration of tracer in the flue gas is then measured. A oneoff EN 15259 concentration traverse must be performed to demonstrate that the tracer is well mixed for the given injection configuration. Simple dilution relationships are then used to calculate the flue gas flow rate from the tracer injection flow rate and concentration.
If all the above techniques are regarded as different implementations of the same method, the ensemble average uncertainty, based on validation field trials, is estimated to be ± 5% at 95% confidence, assuming that the flow is non-swirling. However, it is anticipated that a lower uncertainty can be obtained using a specific technique in a given application. The Test Laboratory must calculate the uncertainty of the method, using the approaches described in the standard, and ensure that this complies with the requirements of the Test Objective.
Automated measuring systems
As with Part 1 of the standard, Part 2 is also performance based. That’s to say, any continuous monitoring technique can be employed – provided the specified performance requirements are satisfied; for example, single point or averaging Pitot tubes, hot wire or hot film sensors, point or cross-duct ultrasonic devices (Figure 4) or correlation (pattern matching) devices. It is recognised, however, that the uniformity of the velocity profile at the monitoring location, combined with the stability of this profile with regard to plant operations, may affect the choice of flow monitor and how this is configured.
A pre-investigation of the velocity profiles at the proposed monitoring location is encouraged by the standard, based on point velocity measurements (see Part 1). For a new plant, this can be conducted using Computational Fluid Dynamics. The survey needs to be performed at the normal base load operating condition and the minimum stable operating condition.
To assist in the selection of a flow monitor, informative guidance is presented in Table 2. The profile is assessed by means of three parameters:
- Reproducibility – the deviation in the normalised velocity profile shape between the minimum and maximum plant flow rates
- Crest factor – the ratio of the maximum to average velocity
- Skewness – the ratio of the average velocities either side of the duct centre-line
If a pre-investigation (velocity survey) is performed, the plant does not then have to operate at minimum load when the monitor is calibrated.
The Quality Assurance (QA) approach defined in the standard is based on EN 14181 which defines three Quality Assurance Levels (QALs). QAL1 requires that the instrument is fit for purpose and this is satisfied by an appropriate instrument certification. QAL2 requires in-situ calibration of the CEM using parallel test data obtained by an accredited Test Laboratory using Standard Reference Methods (SRMs) defined in Part 1. The calibration must also be checked annually by the Test Laboratory by means of an Annual Surveillance Test (AST). QAL3 requires the ongoing monitoring of instrument zero and span drift.
QAL1 defines additional certification requirements and emphasises the need to have an appropriate reference material, or surrogate approach, for checking the zero (or low level) and span (high level) instrument capability. For example, a Pitot tube would require the capability to check the ΔP measurement combined with procedures to ensure that the pressure tappings remain blockage free. The instrument configuration, and sensitivity to changes in flue gas conditions and velocity profile shape, must also be audited by the Test Laboratory during the certification field trial.
The approach to be taken for in-situ calibration of the flow monitor is defined by QAL2. EN14181 employs Emission Limit Value (ELV) and an uncertainty level specified in the relevant European Directive when assessing the quality of the calibration. Since these parameters are not defined for flue gas flow rate, surrogate values are defined in the standard for the ELV (120% of the maximum measured value) and the uncertainty (σo = 4%). Testing does not have to meet any particular time constraints, e.g., a QAL2 can potentially be conducted in one day, and the number and range of the measurement points can be reduced if a pre-investigation of the flow profile is conducted, as noted above. In addition to the usual variability (QAL2) and bias (AST) assessments, the quality of the linear regression between the test results and continuous monitoring results must be good (R2 > 0.9).
Calculation of the flue gas flow rate from fuel consumption can be also employed for continuous monitoring purposes (according to Part 1 Annex E) subject to QAL2/AST verification. QAL3 requires the usual control chart approach for the assessment of instrument drift using the internal reference points established under the QAL1 certification.
Applying this standard to existing combustion plant poses a number of challenges relating to a) sample port provision and access, b) choice of manual test method and c) implementation of the QA requirements in a consistent and meaningful way. However, the standard provides a framework for improving the quality of flue gas flow rate monitoring for emissions reporting and other purposes.