EN14181 was first published in 2004 and the power industry subsequently highlighted areas in which the standard needed further development. A CEN Technical Report, EN/TR 15983, was issued in order to provide further guidance on implementation.
A new version of EN14181 was published in November 2014 in order to consolidate and extend the additional interim guidance which has been withdrawn. Note that Continuous Emissions Monitoring (CEM) systems are referred to as Automated Measuring Systems (AMS) within the standard and throughout this article, which also makes reference to recent legislative developments for large combustion plant and waste incinerators.
Regulatory framework for combustion and incineration plant
The regulatory framework for large combustion plant defines Emission Limit Values (ELVs) which are of key importance when implementing EN14181. The Industrial Emissions Directive (IED), published in December 2010, replaces numerous European Directives that govern the environmental regulation of process plant and this requires the application of Best Available Techniques (BAT). The IED applied, in entirety, to new large combustion plant (≥50 MW net thermal input) permitted from 7 Jan 2013 and to existing plant from 1 Jan 2016, subject to a number of temporary derogations not described here. Incineration and co-incineration plant have been fully regulated under the IED from 7 January 2014.
All plant have to comply with the general permitting and BAT provisions in Chapter II of the IED. Large combustion plant must comply with Chapter III and Annex V of the IED. ELVs are defined for SO2, NOx and dust for large combustion plant fired by solid and liquid fuels and also CO for gas fired plant (Table 1). ELVs are also defined for gas turbines (Table 2). ELVs apply during normal operation only for all plant types (excluding start-up and shut-down) and only above 70% of ISO base load for gas turbines. The tabulated ELVs apply as monthly averages and a daily ELV is also specified as 110% of these monthly ELVs. Peripheral measurements, which are required to correct the emission concentration to reference conditions, must also be measured continuously where applicable: oxygen, water vapour (not required if the pollutant is measured on a dry basis), temperature and pressure.
Continuous monitoring is required for large combustion plant with a rated thermal input ≥100 MW, although there are specific derogations for SO2 and dust that instead require six monthly periodic monitoring, noting that mercury shall also be measured at least once per year for coal and lignite fired plant. However, alternative procedures can be proposed in place of periodic monitoring of the main pollutants, e.g. calculation of SO2 from fuel sulphur content.
The IED strengthens the importance of Best Available Techniques (BAT) and permit conditions must comply with BAT Conclusions, drawn from the BAT Reference note (BREF) for each industrial sector, within four years of publication. Effectively, tighter ELVs and a wider range of pollutants and additional monitoring requirements may be defined within the sectoral BREF. The final draft of the Large Combustion Plant BREF was published in mid-2016 and this is unlikely to change substantially before final publication of the BAT Conclusions, anticipated in late-2017 which would trigger implementation no later than late-2021.
The BREF introduces substantially lower ELVs and a wider range of continuously monitored species than the IED. As an example, new combined cycle gas turbines will be subject to an annual ELV in the range 10 – 30 mg/m3 and a daily ELV in the range 15 – 40 mg/m3 (15% O2 dry). New large pulverised coal fired plant will be subject to annual ELVs for NOx (65 – 85), SO2 (20 – 75), Dust (2 – 5), HCl (1 – 3) and HF (<1 – 2) mg/m3 at 6% O2 dry, respectively. Also, an annual mercury ELV in the range <1 – 2 mg/m3.
Incineration plant must comply with Chapter IV and Annex VI of the IED. Additional species are regulated and monitored continuously for incineration plant: CO, Total Organic Carbon (TOC), HCl and HF, with various derogations from continuous monitoring for HCl and HF which must instead then be monitored periodically along with trace metals, dioxins and furans. The Waste Incineration BREF is currently under revision and this, again, may result in lower ELVs than specified in the IED.
The IED defines a confidence interval (CI) to account for measurement uncertainty. The confidence interval is subtracted from the reportable hourly average concentration for combustion plant and the half-hourly average for incineration plant. The CI is also used as the basis for many of the subsequent QA assessments described below. The defined confidence intervals are given in Table 3.
Both combustion plant and incineration plant will be regulated with a daily ELV that is used when applying the QA requirements described below. The reduction in the absolute value of the daily ELV is the main regulatory impact on the QA requirements, as discussed below.
Overview of EN14181
This section provides an overview of EN 14181 which defines three Quality Assurance Levels (QALs): QAL1, QAL2 and QAL3, and an Annual Surveillance Test (AST). The basic structure of the QA process is shown in the flow diagram.
The operator has the following general responsibilities:
- Installation of compliant equipment (QAL1)
- In-situ calibration of AMS using an accredited test laboratory (QAL2)
- Annual check of the in-situ calibration using an accredited test laboratory (AST)
- Performing ongoing Quality Assurance based on regular zero and span checks (QAL3)
- Submission of QAL2, QAL3 and AST reports and ongoing maintenance of records
- Checking of hourly averages against the valid calibration range (weekly)
QAL1 requires an assessment of the suitability of the AMS equipment. New analysers should be certified and the certification field trial should be carried out on a similar process to that under consideration. The monitoring equipment must also be placed so that a representative reading can be obtained.
QAL2 requires calibration of the monitors against analytical methods – standard reference methods (SRM) – applied by a test laboratory accredited to ISO 17025. The straight line calibration relationship between the AMS and the SRM test data is established by taking at least 15 pairs of measurements obtained across at least three days of operation. Any scatter in the data comparison is assumed to be caused by the plant monitor and this scatter (variability) must be below a threshold in order to pass.
Prior to a QAL2 calibration, various functional tests must be performed. As a minimum, the test laboratory must audit the functional tests. The operator must also perform a weekly check of the reported emissions data to determine if this lies outside the Valid Calibration Range (VCR) established during the test campaign.
QAL2 is intended to take account of any bias related to the specific monitoring equipment or the sampling location and must be conducted every five years (combustion) or three years (incineration) or following a significant change to the process, the fuel mix or the AMS.
ASTs are intended to validate the calibration established under QAL2 by, again, employing an accredited test laboratory to take a reduced number of parallel measurements. The tolerance applicable to the
data scatter is widened and an additional test compares the mean deviation from the calibration line with the 95% confidence interval mentioned above.
QAL3 is intended to provide an audited check of ongoing performance by conducting regular zero and span checks of the monitors and comparing the measured drift against pre-defined warning and action limits using a control chart approach.
Changes to EN 14181 General provisions
Appendix K of EN 14181:2014 describes the main technical changes between the first and second editions of the standard.
The main purpose of the amended standard remains the same, that is, the AMS meets the regulatory uncertainty requirements, now specified in the Industrial Emissions Directive for combustion and incineration plant. The IED confidence interval, evaluated at the Daily ELV, is now referred to as the Maximum Permissible Uncertainty (MPU) within EN 14181:2014. All of the uncertainty requirements within the standard are replaced by the MPU. As an example, for a plant with a daily ELV of 100 mg/m3 NOx, MPU is 20% of 100 mg/m3, i.e. 20 mg/m3.
All QAL2/AST testing shall be conducted by a test laboratory that is accredited, under ISO 17025, to make the measurements or one that is approved directly by the competent authority. The measuring range of the AMS can now be selected to match the anticipated maximum short term (hourly or half-hourly) ELVs.
Quality Assurance Level 1
A normative reference to the EN 15267 series of standards for the certification of CEMs has been added. This means that new AMS installations must be certified to EN 15267 as demonstration of QAL1 compliance. For large combustion plant, the certification range should be below 2.5*ELV and for incinerators below 1.5*ELV. The QAL1 uncertainty should be less than 75% of the IED confidence interval. In the above example, less than 75% of 20 mg/m3, i.e. 15 mg/m3.
Existing uncertified AMS installations may be approved by the competent authority, for example, on the basis that the remaining QA requirements are satisfied (QAL2, AST and QAL3), noting that the pass criteria become more stringent as the daily ELV is reduced.
Existing certified instruments that cannot meet the formal uncertainty requirements due to a reduction in the ELV, e.g. due to new IED requirements, may be approved by the competent authority, e.g. on the basis that the remaining QA requirements are satisfied (QAL2, QAL3, AST).
For existing instruments that do not formally meet the certification requirements, the operator should consider measures to reduce the measurement uncertainty, e.g. air conditioning of the AMS shelter.
Normative reference to EN 15259 has been added. Representative sampling points for AMS and SRM must be defined and there must be no interference or disturbance between the two.
In some cases, the existing duct work may be horizontal and readily accessible (Figure 2). The standard does not address the problems at existing plant where it can be impractical to meet the full requirements of EN 15259 with regards to sample ports and sampling platforms, and the operator will then need to agree alternative provisions with the competent authority, noting that the IED (Article 38) requires the competent authority to ‘determine the location of the sampling or measurement points’.
Quality Assurance Level 2 and the annual surveillance test
Functional tests are now identical for both QAL2 and AST testing and are also recommended for peripherals (O2, H2O). It is recommended that functional checks are carried out not more than one month prior to the QAL2 or AST. Functional testing must be performed by an experienced testing laboratory that has been recognised by the Competent Authority.
The linearity test range is ‘at least the short term ELV’ rather than being ‘two times the emission limit’, thus providing greater flexibility in the concentration range than can be considered in relation to concentration excursions.
The response time of the AMS must continue to meet the EN 15267-3 performance criterion, rather than the historic actual QAL1 result, that is, degradation of the instrument in-service is now recognised, along with the need to remove in-situ analysers from the stack to conduct a zero and span check. Documentation must now include details of the AMS certification.
An AST may replace a repeat QAL2, provided that the AST measured values and at least 95% of the AMS short term averages, reported since the previous AST, are both less than MPU. There is a recommendation that QAL2 can be applied to peripherals (O2, H2O). However, the standard does not point out that there may be genuine differences between peripheral measurements at two locations, so it may not be appropriate or necessary to apply QAL2 factors under these circumstances.
Water vapour values from a calibrated AMS or a calculated water vapour (for wet abatement systems) can be used to correct SRM measurements to a wet or dry basis as required.
Excluded outliers must be identified, reported and retained in data tables and calibration graphs, noting that the assessment method, and the justification for excluding outliers, are to be documented in the QAL2/AST report. It is recommended that additional test points, beyond the minimum requirements, are obtained to allow for outliers.
Exclusion of invalid data from the initial data set, prior to the outlier analysis, must also be justified in the QAL2/AST report, e.g. it is easy to justify the exclusion of start-up or transient data, but all stable operating data during normal operation should usually be included.
QAL2 calibration function
There is a requirement that the QAL2 calibration shall be implemented within six months, rather than it simply being reported within six months. However, it is noted that the member states may allow the continued use of the previous calibration function if it can be proven by use of a specified statistical procedure that the new calibration function does not significantly differ from the previous one.
A new analysis procedure is defined for the treatment of low level data clusters (Method C). A calibration function is derived from a combination of data from the parallel reference tests and reference materials. Two reference points (at zero and close to ELV) are added to the QAL2 data set prior to deriving the calibration function. However, the reference points are not included in the variability test. If appropriate reference materials, e.g. test gases, are not available, an alternative procedure can be approved by the competent authority and this must be fully documented in the QAL2 report.
The criterion for selecting the calibration procedure (method A, B or C) has changed. This was previously based on the SRM test data range (max – min) being less than 15% of ELV. It is now based on the data range being less than MPU.
Valid Calibration Range
The Valid Calibration Range (VCR) established under QAL2 continues to be defined as the highest, calibrated, AMS QAL2 concentration, at standard reporting conditions, plus 10%. If the emissions data, reported subsequently, lie outside of this range for more than a specified proportion of the time point, a repeat QAL2 is required. Industry has long argued that this range extension is insufficient to cover normal process variability. The VCR can now be extended using the highest AST data point (plus 10%) provided that the extension is less than 50% of ELV. The VCR can always now be extended to 20% of ELV (minimum VCR).
As before, reference materials, e.g. test gases, can be used to improve confidence in data that lie beyond the VCR, although the emphasis is placed on concentrations close to the ELV rather than a wider extrapolation. The agreement between the high test gas concentration and the calibration line must be within the MPU. The agreement between the zero point and the calibration line must be within 10% of ELV. The competent authority must therefore determine if reference materials can be used to actually extend the VCR in addition to improving confidence. An extension to the short term ELV would be reasonable since this is the range across which the plant is expected to operate during normal operation.
Weekly data inspection (Monday to Sunday), to check that reported data are within the VCR, can now be based on the most recent operational hours (168 hours), rather than a calendar week, when the plant is not run continuously. Also, exceedances of VCR caused by plant failures do not now trigger a QAL2.
Quality Assurance Level 3
QAL3 requires operators to have a procedure that describes the requirements for:
- Measuring zero and span values
- Plotting these values in control charts
- Using the control charts to determine if there is a systematic deviation (drift)
There is an enhanced description of QAL3 requirements including the selection and use of control charts and the execution of zero and span measurements. Three examples of control chart are described in detail in Annex C of the standard: Shewhart and CUSUM (included in EN 14181:2004) and an additional Exponentially Weighted Moving Average (EWMA) approach. In all cases, including dust monitors, alternative reference materials can be used, such as optical filters, provided that these are certified as being QAL3 compliant under EN 15267-3.
There is a new general requirement on the operator to ensure that AMS internal checks or compensation systems are active and operational and that AMS alarms are received by plant personnel so that corrective action can be taken at all times. Any type of manual or automated control chart can be used. Built-in procedures are allowed as an alternative to an external QAL3, provided that the data is available to the operator to allow for annual auditing of the QAL3 data.
The required frequency is at least once per maintenance interval and at least once per month for uncertified AMS. For maintenance intervals greater than one month, justify with internal checks and, for multi-component analysers, a monthly span check of at least one gas component is required.
Hot spares (fully independent back-up AMS) can be used to extend QAL3 to an annual check provided that an alarm is raised when the difference between the two AMS is > 5% of the short-term ELV for more than five consecutive measurements.
Control chart limits can be based on the MPU rather than a detailed uncertainty assessment. The alarm limit cannot be higher than 50% of the MPU. The warning limit may be set at 25% MPU.
For audit purposes, QAL3 documentation must include a history of the checks and the actions taken when exceeding control chart limits. Control charts should be stored for five years.
The amendment of EN 14181 clarifies and improves the applicability of this Quality Assurance standard and this is a positive development that is welcomed by industry. There are a few remaining issues, which have been identified previously, that have not been addressed and that will require approval by the relevant Competent Authority for specific industrial sectors. These include: extending the Valid Calibration Range beyond the daily Emission Limit Value using reference materials; and simplification of linearity test requirements for multi-component analysers. It should be noted that the progressive reduction of Emission Limit Values increases the difficulty of passing each of the QA requirements. Further details can be found in a related conference paper.
Recently, the Article 13 Forum, established under the IED to review and approve the BREFs, has concluded that the measurement uncertainty is an implementation issue that must be addressed by Member States. Measurement uncertainty has not been taken into account when determining the emission levels associated with BAT, but it must be taken into account when assessing compliance against ELVs. Further work is therefore required to properly assess the various contributions to measurement uncertainty that arise at each point in the measurement and calibration chain.