Statutory guidance

Monitoring stack emissions: quality assurance of continuous monitoring

Updated 28 March 2025

Applies to England

1. Guidance on quality assurance

This section provides an overview of the scope of EN 14181 and the roles of those involved.

1.1 Scope and structure of EN 14181

EN 14181 applies to permanently installed continuous emissions monitoring systems (CEMS) used for compliance monitoring.

EN 14181 does not apply to data acquisition and handling systems (DAHS) used with CEMS. EN 14181 consists of quality assurance levels (QAL) 1, 2 and 3 and an annual surveillance test (AST).

CEMS that are certified on the Monitoring certification scheme (MCERTS) for CEMS at an appropriate certification range demonstrates compliance with the QAL1 requirements.

QAL2 specifies procedures to make sure that the CEMS are correctly installed and calibrated using standard reference methods (SRMs).

After installation of the CEMS the operator must carry out functional tests, such as linearity, zero drift, and span drift checks.

QAL3 requires the operator to regularly measure the drift and precision of the CEMS.

The AST consists of the same functional tests as those used in QAL2. But the operator can check the calibration function using a smaller number of repetitions of the SRMs (typically 5). If the calibration function is still valid, no further action related to the AST is required. If the AST shows that the calibration function is no longer valid, the operator must do another QAL2.

Roles and responsibilities

The responsibilities of CEMS manufacturers and equipment suppliers under EN 14181 are:

• achieving and maintaining certification of CEMS to the applicable standards
• supplying, correctly installing and commissioning CEMS
• co-operating with operators and test laboratories to perform the functional tests and calibration of CEMS

The responsibilities of test laboratories under EN 14181 are:

• having accreditation to EN ISO/IEC 17025 General requirements for the competence of testing and calibration laboratories for the MCERTS performance standard for manual stack emissions organisations
• having accreditation for the applicable SRMs and EN 14181
• doing the SRMs for the QAL2 and AST procedures
• reporting the results of the functional tests specified for the QAL2 and AST procedures
• informing the operator that they are responsible for making sure that the functional tests take place before each QAL2 and AST, regardless of who subsequently performs the functional tests

The responsibilities of operators under EN 14181 are:

• using CEMS certified to the appropriate MCERTS performance standards
• performing the QAL3 procedures
• making sure that the functional tests take place before each QAL2 and AST
• submission to us, when required, of information and reports on QAL2, QAL3 and ASTs
• applying a procedure for QAL3
• retaining reports and other records related to EN 14181 for periods specified by us.

Operators have overall responsibility for complying with EN 14181. Test laboratories must carry out the activities specified in QAL2 and the AST. However, CEMS suppliers, operators or test laboratories may perform the functional tests specified in QAL2 and the AST.

Where available, the test laboratory must take part in proficiency testing schemes for EN 14181 and maintain the level of proficiency required to meet the requirements of the scheme.

Spreadsheets used for calculations in QAL2 and ASTs must be controlled documents within the test laboratory’s quality system.

2. Suitability of CEMS

This section provides information on QAL1.

2.1 How to select CEMS

The following applies when selecting CEMS:

• the CEMS must be MCERTS certified for the measurands specified in the permit that require continuous monitoring
• the CEMS must be MCERTS certified for a range that is suitable for the application
• site and stack conditions do not reduce the performance of the CEMS to below the required standard
• the intended CEMS is proven on comparable installations
• all CEMS must have provisions to do zero, span and linearity tests after installation of the CEMS
• extractive CEMS must have the means for leak and response time checks, such as the provisions for applying test gases at the sampling probe

2.2 Measurement ranges and certification ranges

There is a difference between the certification range and the measurement range of the CEMS. The measurement range is the set of values that the CEMS can measure, from the lower detection limit (that is near zero) to a set upper limit. Within this measurement, the certification range is the range over which the CEMS meets the MCERTS performance standards.

The MCERTS certificate states the certified range of the CEMS. The MCERTS performance specifications is expressed as a percentage of the certification range. So, the lower the certified range, the better the performance of the CEMS is likely to be. For example, if the performance requirement for cross-sensitivity is ±4% of the range and a CEMS has a certified range of 0 to 75 mg/m³, then the cross-sensitivity will not be more than ±4% of 75 mg/m³, which is ±3 mg/m³. A CEMS with a certified range of 0 to 200 mg/m³ may have a maximum cross-sensitivity up to ±4% of 200 mg/m³, or 8 mg/m³.

Generally, CEMS with lower certified ranges will perform satisfactorily at higher ranges.

Operators may use a CEMS for higher ranges than those certified, provided CEMS manufacturers have evaluated the linearity over the higher ranges during MCERTS testing.

2.3 Suitability of measuring ranges to capture the expected variations in emissions

CEMS may be auto-ranging with respect to their measurement ranges, or they may have fixed ranges; for example, the range may have settings at 0 to 100 ppm, 0 to 500 parts per million (ppm) and 0 to 1000 ppm. The latter types of CEMS typically have 4 to 20 mA analogue outputs. One drawback with such CEMS is that they have a lower resolution at the higher ranges. Therefore, during performance testing for certification, the range is set at a value that will capture all the typical peaks in the emissions but still maintain the required uncertainty at the emission limit value (ELV). This also applies to installations on site. Incorrect range setting can lead to the capping of peak measurement values, which is something the Environment Agency consider unacceptable. However, they may permit capping of peak values under specific exceptional circumstances that must be agreed with the regulatory officer in advance.

If there is any doubt about a CEMS’ performance for a particular application, refer to the MCERTS test results.

The scale of the CEMS measurement range must be set according to the ELV (for example, 2 times the short-term ELV as defined in the permit, to capture the expected peaks in emissions) and not the range measured during the QAL2. For example, if the ELV is 60 mg/m³ for HCl and the range measured during the QAL2 was varied from zero to 20 mg/m³ expressed as a half-hourly average, then the CEMS should be set to measure of range of zero to 120 mg/m³. Refer to EN 14181 for an example calculation.

2.4 Suitable certification ranges

The certification range can indicate the suitability of a CEMS for a particular application. Legislation for waste incineration and large combustion plants specifies uncertainty budgets for certain measurands, so choose CEMS that will meet (and ideally exceed) these uncertainty specifications.

Our approach for selecting suitable CEMS is to apply range multipliers, whereby the lowest certified range is not more than 1.5 times the daily average (DA) ELV for incineration processes and not more than 2.5 times the DA-ELV for large combustion plant and other types of process (or 48-h ELV for some installations under Chapter III of the Industrial Emissions Directive – IED). As there is a linear relationship between certified ranges and uncertainties, these multipliers provide assurance that CEMS with appropriate ranges will meet the uncertainty requirements specified in the IED. This approach is specified in EN 15267-3 Air quality — Certification of automated measuring systems — Part 3: Performance criteria and test procedures for automated measuring systems for monitoring emissions from stationary sources.

The CEMS must also be able to measure instantaneous values over the expected ranges during all operating conditions. If it is necessary to use more than one range setting of the CEMS to achieve this requirement, the CEMS must be verified for monitoring supplementary higher ranges.

2.5 Selection procedures for CEMS already installed at a site

If CEMS that are already installed at an installation do not meet the requirements for ranges, then the operator may still use them if they fulfil the QAL2, AST and QAL3 requirements of EN 14181. CEMS with ranges higher than those required may still pass the QAL2, AST and QAL3 requirements but the risk of failure increases as the certified range increases. If the CEMS do not meet QAL2, AST and QAL3 requirements, then operators will need to replace them within one year with CEMS that do have suitable ranges based on the ELV multiplier rule.

2.6 Selection procedures for new CEMS

New CEMS must meet the requirements of the ELV multiplier rule. Also, all new CEMS must include the means to allow operators, test laboratories or suppliers to perform zero, span and linearity tests, after the CEMS have been installed.

2.7 Sampling systems for extractive CEMS

The Environment Agency allows some flexibility in the selection of the CEMS’ sampling system. However, the type of sampling system specified on the certificate must be used.

Flexibility could include:

• a different length of sampling line to that tested
• additional manifolds and heated valves to allow more than one analyser to share a sampling system
• a different brand or model of sampling system

The operator can use a different brand or model of sampling system if there is evidence from independent testing, that the alternative components meet the required performance specifications. However, it must use the same general technique. For example, an MCERTS certified permeation dryer may replace an existing permeation dryer. But it is not acceptable to use a chiller-dryer in place of a permeation dryer.

3. Calibration and validation of the CEMS

This section describes the calibration and validation of CEMS. Although this section focusses on the QAL2, it also refers to requirements on the AST because several of the requirements apply to both.

3.1 QAL2 requirements

The requirements of EN 14181 apply. In addition, if the spread and scatter of the SRM data mean that it is unsuitable to derive a calibration function, then the operator, test laboratory or CEMS supplier must calibrate the CEMS using alternative means.

Following CEMS failures, all investigations repairs and re-testing must be completed within 6 months.

3.2 Location of the CEMS

Operators must follow the provisions for location and access described in EN 15259, Stationary source emissions – Requirements for the measurement sections and sites and for the measurement objective, plan, and report.

This enables them to determine the most representative location for the CEMS, according to the homogeneity test described in EN 15259. Operators must locate CEMS at a point where there is safe access and other provisions for the effective and continued operation of the CEMS, and where samples are representative of the emission.

The laboratory must assess the homogeneity of the stack before they carry out the QAL2. On most occasions this requires the operator to carry out a homogeneity test. On some occasions this is not required, for example because of the area of the sample plane. You can find situations where there is no need for a homogeneity test in the Environment Agency method implementation document (MID) for EN 15259.

A homogeneity test involves taking grid measurements of the stack gas at centres of equal area across the sampling plane and comparing the results to a fixed reference point within the sampling plane. Additionally, locate the SRM and the CEMS (or its sampling location) so that they do not interfere with each other.

3.3 Location of the sampling ports

EN 15259 describes the requirements for locating sample ports used for the SRMs.

3.4 Management system requirements

EN 14181 requires operators to have a systematic approach to managing and maintaining the CEMS, documented through procedures within an existing management system, such as those meeting the requirements of EN ISO 9001 Quality management systems – Requirements.

Procedures should include specific provisions for CEMS covering:

• selection
• maintenance and servicing
• responsibilities and training of personnel
• calibration, quality assurance checks and controls
• records and data management
• prevention of unauthorised adjustment of the CEMS and its data recording devices
• maintaining availability – spares, contingencies, and back-up monitoring

3.5 The functional tests

EN 14181 specifies within QAL2 and the AST that CEMS must have a set of annual functional tests.

Operators are responsible for making sure that the functional tests are performed before the parallel reference tests required by the QAL2 and AST.

The test laboratory must inform the operator in advance of the QAL2 or AST that the functional tests are a mandatory requirement within EN 14181 and that the functional tests must be performed before the parallel reference tests with SRMs.

Perform the functional tests no more than one calendar month before the parallel reference tests with SRMs. If the functional tests are performed earlier than this, then the test laboratory must inform the operator that they should provide verifiable evidence to demonstrate that the results of the functional tests are still valid. This evidence should include QAL3 data to demonstrate that the performance of the CEMS has not changed between the functional tests and the parallel reference tests. Such tests would include zero and span data for the entire sampling system and analyser. The test laboratory must document this evidence on the QAL2 or AST report, as applicable.

If the functional tests are not performed, then the test laboratory must include the reasons for this in the QAL2 or AST report.

The test laboratory must report the results of the functional tests in the EN 14181 report. This can be a copy of the report if performed by an organisation other than the test laboratory.

Any organisation with the required competence and training may do the functional tests. This means they must have training plans, evidence of assessments and training records to demonstrate that the people performing the functional tests know, understand, and can apply the procedures required by Annex A of EN 14181.

Functional tests are required for O₂ and moisture monitoring CEMS if these are present. The criteria for the tests must be based upon the requirements of the MCERTS performance standards.

For moisture monitoring CEMS, the linearity and span test will require a moisture generator. The Environment Agency recommend that this test is done but operators may choose not to perform it because it can be difficult to use a moisture generator on site. However, if the CEMS fails the QAL2 variability test because of the failure of the moisture monitoring, then the linearity test for moisture will be required.

It is essential to review historical process data and the CEMS maintenance log before carrying out the functional tests and parallel reference tests. A data review will allow an evaluation of normal instrument operation and details of process peaks. The maintenance log will provide information of any malfunctions and repairs.

3.6 Specific requirements of the functional tests

If the CEMS is equipped with a NOx converter, the operator must make sure that the efficiency of this converter is tested at least once per year. The efficiency must not be less than 95%. The test for converter efficiency must meet the requirements of EN 14792. It may be performed by any competent organisation. The Environment Agency recommend that the test is done using gas-phase titration.

The operator must give the test laboratory access to the documentation specified in Annex A, section A.4 of EN 14181. The functional test must check whether the documentation is controlled, readily accessible, and up to date. The operator is responsible for keeping the documentation controlled, accessible and up to date.

A list of recommended reference materials, tools, and spare parts should be supplied by the CEMS supplier and cross-checked with availability on site or by supply agreement.

The signals from the analyser through to the digital control system or data logger must be checked and confirmed by the test laboratory.

The manual zero and span checks must be performed using the same procedure as for the MCERTS performance tests. Typically, zero and span checks require the use of reference materials. In the case of particulate CEMS, these checks will require the use of appropriate surrogate reference materials. Reference can be made to original QAL1 tests regarding ranges of reference materials.

As a cross reference, test laboratories should conduct a separate span on the operator’s CEMS with their own traceable gases. Treat the data for zero readings in the same way as other data points.

3.6.1 Linearity test

The linearity test may be performed on the analyser alone.

If you can only log CEMS values manually during linearity testing, digital control system trends and logged values can be incorporated into the report. This shows that adequate stability had been achieved and validates each of the points obtained.

The linearity test concludes with the calculation of the residuals of the average concentrations, and a test of those residuals. Initially expressed in concentration units, the residuals must be converted to relative units by dividing by the upper limit of the range used during the linearity test. As the linearity is tested using a range of percentage concentrations of the short-term ELV, then the upper limit of the range is the concentration at 100% of that range.

For the linearity test, the performance of the CEMS is compared with their MCERTS certification performance specifications. The original performance specifications are typically expressed as percentages of a certification range that is based on the daily average ELV.

If a higher range than the daily average ELV is applicable, the linearity check can be done using the short term ELV or 2 times the daily ELV.

The test gases must be labelled with the relevant accreditation logo and number.

3.6.2 Interference

MCERTS certification of CEMS will show which interferents can have a measurable effect on the CEMS response to stack gases other than the target measurand. Therefore, if there is a failure of the QAL2 or AST tests, then the test laboratory must refer to the MCERTS testing data and determine whether any significant interferents are likely to be present in the stack, which could result in a failure of the QAL2. If so, the test laboratory must determine the extent to which such interferents may introduce a bias into the measurements from the CEMS.

3.6.3 Zero and span drift (audit)

The test laboratory must assess whether the operator has a QAL3 procedure in place, and whether the operator has applied this procedure.

3.6.4 Response time

For extractive systems, the CEMS must have an injection point for calibration gases located as close as possible to the sampling point to check the response time. The organisation performing the check needs to note the length of the sample line if the CEMS is an extractive system. When the results of this test are compared to the results of the same span gas fed directly into the analyser, the differences in response time is the lag-time due to the length of the sampling system. This information needs to be considered when the test laboratory performs the parallel reference tests required by the QAL2 and AST of EN 14181, to align the responses of the CEMS and SRMs.

3.6.5 Report

The operator must assess the test results prior to conducting the parallel reference tests required by QAL2 and the AST. Any necessary corrective actions should be addressed before the test laboratory performs the parallel reference tests.

3.7 Verification and calibration

3.7.1 Verification and the spread of data

EN 14181 requires the use of SRMs to verify and calibrate CEMS. It requires that:

• there is a spread of data over the required range of the monitoring system
• the CEMS have a linear response to increasing values of a measurand

Although EN 14181 works best when there is a good spread of data and the CEMS has a linear response to increasing values of the target measurand, it is also common for emissions results to cluster. So, the patterns of emissions that test laboratories encounter are one of the following:

• a linear spread of data across a wide range
• a high-level cluster when steady state readings from the CEMS occur because of stable emissions
• a low-level cluster when the emissions are highly controlled and typically close to zero

A cluster is a set of data whose range of SRM values is less than the maximum permissible uncertainty.

Measurements can range from zero, or near zero, up to the ELV, and the difference between the highest and lowest SRM readings will be greater than the maximum permissible uncertainty.

Figure 1 shows a set of data with a linear spread. Using this data, the test laboratory can derive the calibration function by using linear regression (procedure a in EN 14181, section 6.4.3).

When displaying the results graphically plot the reference material on the Y-axis, and the CEMS data on the X-axis.

Figure 1 – The principle of linear calibration using an SRM

Figure 2 shows a high-level cluster. If the difference between the highest and lowest standardised SRM reading is smaller than the maximum permissible uncertainty and the lowest standardised SRM reading is greater than, or equal to 15% of the daily ELV, then the test laboratory can determine a calibration function using procedure b, EN 14181, section 6.4.3.

EN 14181 requires confirmation that CEMS read zero when the emissions are zero. If the process variations do not provide zero readings, then a surrogate for zero emissions is acceptable. Up to 3 zero readings can be included in the data set when generating the calibration function.

Figure 2 – A high level cluster

Figure 3 shows a low-level cluster. This is a typical pattern of emissions results when the emissions are highly controlled. In such cases, the calibration function is not reliable unless the cluster is highly linear (as indicated by a correlation coefficient of the regression (R2) value of 0.9 or more for gaseous compounds, or 0.5 for dust); this is because EN 14181 was developed for cases where the emissions are towards the ELV, or at least well above zero.

Clusters are classed as low level when the difference between the highest and lowest standardised SRM reading is smaller than the maximum permissible uncertainty and the lowest standardised SRM reading is smaller than 15% of the daily ELV. The calibration function should be derived using procedure c in EN 14181, section 6.4.3.

A valid calibration should ideally have a correlation co-efficient of the regression line (R2 value) of at least 0.9. However, it can be lower than this and still pass the variability test. The variability test should always be considered as the definitive test. It is possible for results containing clusters of dust data to produce an R2 value of 0.5 and still pass the variability test. If the CEMS does pass the variability test but exhibits a low R2 value, then it is advisable to investigate the reason for this. 

If the scatter of data points means that it is not possible to derive a valid calibration function, that is, the difference in the maximum and minimum standardised SRM readings is less than the maximum permissible uncertainty and the minimum standardised SRM reading is less than 15% of the daily average ELV, then:

• the CEMS may be calibrated using additional reference materials
• use the SRM data to verify that the emissions are well below the ELV and that the CEMS is responding with an acceptable degree of accuracy and precision to low levels of emissions
• the readings are acceptable if the average of the CEMS and SRM results do not differ by more than the 95% confidence interval of the daily average ELV.
• when using procedure c, the valid calibration range is based on the highest CEMS calibrated value obtained from the stack gases, not the surrogate gas at the ELV concentration.

Figure 3 – A low level cluster

3.7.2 Monitoring strategy

The expected spread of data, based on process information and prior monitoring data, will dictate the monitoring strategy. Whilst EN 14181 specifies at least 15 parallel reference tests over at least 3 days using SRMs, this depends on there being a linear spread of data over as wide a range as possible. If the results do not fulfil this basic condition, then it is not always possible to carry out all the procedures prescribed in EN 14181. In such cases, especially when there are low levels of clustered data, EN 14181 has more purpose in verifying results, than for calibrating CEMS.

If the emissions are typically below the maximum permissible uncertainty, then an AST is permitted instead of a QAL2. However, operators must first contact their regulatory officer justifying the request for reduced sampling. For low emissions, reduced measurements with longer sampling times may be satisfactory for manual periodic tests, for example, the AST may consist of 5 measurements each with a sampling duration of 1.5 hours, thus providing a total sampling period of 7.5 hours.

When using transportable continuous emissions monitoring systems (T-CEMS) for SRMs, the SRM monitoring-system may be operated continuously during the tests. Zero and span checks on monitoring systems used within SRMs must be taken at the start, end and once per day during the monitoring period. Test laboratories must state in their site-specific protocol the time intervals between the start times of each successive pair of measurements. Furthermore, test laboratories must demonstrate that the interval between each pair of data provides representative samples, considering any process variations.

The Environment Agency recommend that the test laboratory does not perform a QAL2 exercise unless an effective QAL3 procedure is in place at the regulated installation. As a guideline, they recommend that new systems have at least 3 months of QAL3 data before doing a QAL2.

This should demonstrate that the readings from the CEMS are stable and comply with the specifications for drift, as required for the QAL1 tests. If there is not an interval of at least three months, the test laboratory must state the reasons for this in the test report.

Whilst 3 months is the Environment Agency’s recommended minimum interval between commissioning and a QAL2, they recognise that this is not always practical or possible.

3.7.3 Low-level clusters when measuring dust

If the correlation co-efficient (R2) of the regression line for the CEMS and SRM data is 0.9 or more, then the calibration function will most likely be valid. However, it is not possible to apply the above requirements in the following situations:

• low-level clusters of data
• when the R2 value is under 0.9

Whenever there are low-level clusters of data, the uncertainty of the SRM will be proportionally greater with respect to the measured emissions. In such cases, there will be a greater degree of relative scatter of data points, which gives a low R2 value. The test laboratory cannot produce a reliable calibration function using the data. In the case of particulate CEMS, surrogates may be useful for zero, span and linearity tests but this data cannot be meaningfully related to concentrations of dust. There are no surrogates available that can accurately mimic a CEMS’ response to specified concentrations of dust, thus preventing the use of procedure c.

When there is a low-level cluster, it may be possible to use the average dust value to calibrate the CEMS, if the average using the SRM is greater than the uncertainty of the SRM. An uncertainty of 20% of the ELV, or 1 mg/m³ for ELVs below 5mg/m³, can be used for the SRM. If it is not possible to calibrate the CEMS by this method, then the CEMS cannot be used quantitatively, but it can be used qualitatively. So, if the emissions are consistently low, the Environment Agency recommend:

• you use the SRM to verify that the emissions are low
• you use surrogates to check the linearity, zero and span settings of the CEMS
• the CEMS is set on its most sensitive range to alert you when the abatement for dust may need attention if an increase in emissions is observed

3.7.4 Regulation when you cannot calibrate particulate CEMS

If the dust emissions from the CEMS are too low to establish a gravimetric calibration using the SRM, then the operator may configure the CEMS output to the default given on the QAL1 certificate. That is, the CEMS must be configured to output the mg/m³ value specified on the QAL1 certificate rather than, for example, units of scattered light intensity. The operator may propose other site-specific calibration factors with a suitable justification, for example, using historic emissions data from the same plant, from a plant of a similar design or as recommended by the CEMS supplier.

The Environment Agency will regard the CEMS output as indicative but it can be used for compliance assessment and mass emissions reporting in addition to process control (as an indicator of abatement system malfunction). If the indicated dust emission increases during normal operation, such that seven consecutive daily average concentrations are higher than 2.5 mg/m³, then a QAL2 must be done within 6 months. When configured in this way, the CEMS will generally over-estimate mass emissions during normal operation.

Energy from waste plants have an abnormal operation ELV of 150 mg/m³. For CEMS with an output that is indicative, the operator must set a precautionary alarm level of 100 mg/m³ because the Environment Agency deem that dust levels above this may be approaching the abnormal operation ELV.

3.7.5 Calibration functions for NOx

Legislation specifies limits for NOx (NO and NO₂) expressed as NO₂. However, some installations with ELVs for NOx emit mostly NO. In such cases, it is acceptable for the operator to measure NO alone, and then apply a conversion factor to compensate for the small proportion of NO₂ in the stack gases, which is typically around 5% or less.

You can measure NOx with a CEMS in the following ways:

• measure NO alone and then apply a conversion factor to account for NO₂ in the stack gases
• measure NO and NO₂ separately and combine the measurements
• use a NO₂ converter to provide a single combined amount of NOx

The Environment Agency recommend the following approaches when performing QAL2 and AST exercises:

• if you measure NO only, then the test laboratory may either measure NO only or total NOx
• if you measure total NOx, then the test laboratory must also measure total NOx by either measuring NO and NO₂ separately or together as total NOx
• if your plant emits less than 10% NO₂ of the total NOx, across the normal operating range, the QAL2 can be based on NO only, that is, calibrate the CEMS NO using the SRM for NOx

You can use either historic CEMS data or periodic test results to show the normal operating range. You can judge the performance of the NO₂ CEMS at low concentrations, for example, from fluctuating positive and negative values around zero, or excessive drift, or an offset from the SRM reading or high variability.

If you want to discount the measurement of NO₂, you must provide the Environment Agency with information on whether the CEMS and SRM record a percentage NO₂ to NOx that is less than 10%.

The test laboratory must record in the QAL2 and AST reports how NO and NO₂ have been measured by the CEMS and the SRM. All the averaged raw data, before any calculations have taken place, must be provided.

When the QAL2 calibration function for NOx has been generated from separate parallel reference tests of NO and NO₂ in mg/m³, the Environment Agency recommend that the CEMS displays the uncalibrated NO and NO₂ values. You can use these NO and NO₂ values to generate an uncalibrated NOx value by using the following equation:

uncalibrated NO₂ + (uncalibrated NO x 1.53) = uncalibrated NOx

The test laboratory must then apply the calibration function for NOx to the uncalibrated NOx value.

3.7.6 Standard reference methods

Only test laboratories that are accredited to EN ISO/IEC 17025 for the MCERTS performance standards for manual stack emission monitoring for the applicable SRMs may perform the SRM measurements during the QAL2 and AST. Additionally, all test laboratories must have EN 14181 within their scope of accreditation.

Alternative methods to SRMs may be used provided they are approved. SRMs and approved alternative methods are listed in Monitoring stack emissions: techniques and standards for periodic monitoring.

3.7.8 Calibration using an SRM

It is essential that the scatter of the SRM results is less than that of the CEMS being calibrated and validated. So, the test laboratory must characterise the SRM thoroughly and make sure that the uncertainties are understood and are below the uncertainties specified in legislation (ideally, no more than half the uncertainties specified in legislation).

3.7.9 Acquisition of data

When conducting parallel reference tests, EN 14181 requires the test laboratory to take the measured signals directly from the CEMS during the QAL2 and AST. The test laboratory must use a recording system that takes frequent samples in relation to the response time of the CEMS. However, if this is not practicable, then the test laboratory may take the data from a different output, such as a DAHS, if there is evidence to show that the data matches the output from the CEMS. This can be obtained by observing the display of the CEMS at the same time as the display of the DAHS.

3.7.10 Spread of data

The test laboratory must select a set of representative operating conditions that cover as wide a range as possible. Ideally, they should select a time when the emissions are likely to be at their highest and most varied. However, the Environment Agency do not permit you to deliberately modify the process to artificially increase emissions.

A QAL2 requires at least 15 valid data points. At least one of the 15 data points must be at or near zero. If this condition is not fulfilled, test laboratories can use functional test data to attain measurement data at or near zero. EN 14181 states that the test laboratory must verify that the CEMS read zero when the emissions are zero.

3.7.11 Averaging periods for SRMs

The averaging period for each SRM measurement should be 30 minutes for waste incineration plants, and 60 minutes for large combustion plants. If the emissions are low, it is acceptable to use longer averaging periods for manual methods (that is a method that is not based on the use of an on-site analyser) to collect a greater mass of sample. Test laboratories may use averaging periods of less than 60 minutes for large combustion plants if the test laboratory can demonstrate that there is no significant difference between 30 minutes and 60 minutes averaging periods. However, the averaging period must not be less than 30 minutes. If you require shorter periods, then permission must be obtained from us.

If the test laboratory performs the parallel reference tests over a total shorter duration, then evidence must be provided to show that measurements over this duration are representative of the process. If the process is stable and predictable, you could use historical data, supported by statistical analysis for this.

3.7.12 Number of data points and outliers

QAL2 specifies at least 15 sets of valid data when performing the SRMs. It is advisable to obtain at least 18 or 19 sets of data to make sure sufficient valid data sets.

There must also be data at zero or near zero, where near zero is defined as a value that is no more than 5% of the daily average ELV. Ideally, zero values should be measured when the installation is not producing emissions. If this is not possible, then reference materials must be used to determine the CEMS response to zero values of the measurand.

If practicable, the test laboratory can plot the data from the CEMS and SRMs on a chart as the QAL2 tests progress, as this will indicate whether the spread of data is sufficient, whether the data has enough values near zero and whether there are any obvious outliers.

Test laboratories must always consider whether outliers are present. Rejected, or invalid results, known as outliers, will negate the quality of any calibration or verification work.

The Environment Agency recommend the following procedure for identifying outliers:

1. Produce a plot using raw CEMS data, and raw SRM data expressed under the same conditions as the CEMS.

2. As a general guide, if the R2 value for the linear-regression line is equal or more than 0.9, then it is not usually necessary to perform an outlier test. Data points are not likely to be outliers unless they are more than two standard deviations from the regression line.

3. Plotting the data on a graph often reveals whether there are obvious outliers.

4. Calculate the positive and negative differences, Di, between each pair of SRM and CEMS values.

5. Calculate the standard deviation of the differences (sD) of Di.

6. Calculate the average of the differences, Dave.

7. For each paired sample, calculate the difference between Di and Dave and express the results as absolute values.

8. If any absolute difference is more than 2 times the standard deviation (2sD) then that paired sample is most likely an outlier and can be rejected. This procedure does not need to be repeated once any outliers have been removed.

When test laboratories use SRMs based on T-CEMS to carry out a QAL2 exercise, it is a common practice to allow the SRM to run continuously for up to 3 days. In such cases, the test laboratory acquires well over the minimum 15 repetitions and may have at least 2 to 3 times as many pairs of data depending on the time interval between the start of each pair of data. The test laboratories must use all the pairs of data in the calculations for the calibration function and variability test unless the laboratory can justify excluding any pairs of data. The following examples may justify excluding pairs of data:

• one or more of the pairs are outliers, as determined using the outlier test
• measurements for one or more of the pairs were taken when the monitoring was not representative

An alternative approach would be to set a longer time interval between each data point; for example, if the averaging time were 30 minutes, separating the start of each data point by 2 hours in a 24-hour period would result in 12 pairs of data points during a QAL2 exercise.

However, separating the start of each data point by 3 hours would result in 8 data points over 24 hours. When planning a monitoring exercise, it is important that the test laboratory determines how many data points it intends to take and document this plan in their site-specific protocol.

If the test laboratory assesses a data set and considers that some of the data may be invalid then the test laboratory must note in the report the reasons for removing each data-pair (for example, changing process conditions, an error in the SRM or a failure of the SRM instrument).

3.7.13 Use of procedure a, b, or c for calculating a calibration function

EN 14181 defines 3 situations based on the extent of the spread of readings, and then specifies a mathematical method that allows the test laboratory to calculate the calibration function. These mathematical methods are defined in EN 14181 as procedures a, b, and c.

When determining the procedure to use, the test laboratory uses the SRM data and converts this data to values at standard conditions.

Procedure a is a standard linear regression method. Even if the spread of data spans more than the maximum permissible uncertainty, at least one value near zero is required to demonstrate that the CEMS read zero or near zero when the emissions are zero and to estimate the best-fit value of the calibration-line intercept.

Procedure b takes an average of the clustered data points, and then forces the regression line towards zero. To justify this, the test laboratory must verify that the CEMS read zero, or near zero, when the emissions are zero and include the result in the data set.

When performing the calculations to determine whether to use procedure a or b, the test laboratory must not use a surrogate zero-value in these calculations, as this would automatically result in selecting “procedure a” every time.

Use procedure c when a low-level cluster occurs. The test laboratory can examine the historical results, and if the expected emissions are likely to be low-level clusters, then either the test laboratory or operator must contact the Environment Agency to agree to reduce the number of SRM measurements. If the installation is new, then at least 15 repetitions will be required, to thoroughly characterise the plant emissions and effectiveness of the emissions abatement systems. Depending on the concentrations and measurand (that is dust or mercury), you may, if you have approval from us, perform 5 SRM measurements for the QAL2.

For procedure c, you must use the SRM measurements to verify that the CEMS respond to low concentrations of the measurand. The difference between the average of the SRM and CEMS results should not differ by more than half the allowable 95% confidence interval of the daily average ELV.

Surrogate gases used for calibration, when applying procedure c, must be from accredited suppliers. Also, the uncertainty of a calibration gas, including the uncertainty of any gas blender used, must not be more than 2% for all gases.

3.7.14 QAL 2 for dust CEMS

For dust measurements, EN 13284-2, allows fewer but longer measurements using the SRM if you expect the results to be below the maximum permissible uncertainty. Increasing the length of SRM measurements to up to 2 hours improves the limit of quantification of the measurement at low dust concentrations. You need approval from Environment Agency if you wish to do this.

3.7.15 Peripheral CEMS measurements

EN 14181 specifies the following peripheral measurements:

• O₂
• moisture
• temperature
• stack gas pressure

CEMS for O₂ and moisture (if used) must be MCERTS certified. Functional checks should be performed on the CEMS.

Calibration of the peripheral CEMS should be done first. The order should be moisture followed by O₂. The calibrated results are then used for the standardisation of the other measurands.

When performing the variability test for O₂ and moisture measurements, the following virtual ELVs and uncertainty allowances must be applied:

• O₂: ELV = 21%, 95% confidence interval (CI) = 10%
• moisture: ELV = 30%, 95% CI = 30%

EN 14181 allows the operator to use default values for moisture in certain circumstances. For example, in the case of wet abatement techniques with nearly constant water vapour concentration, the moisture concentration may be calculated using other operational data. However, the moisture values must still be validated.

CEMS for temperature and pressure must be cross calibrated using reference instruments that are traceable to national standards.

3.7.16 Sample lines and delays

The test laboratory must have a procedure to determine the lag times and for matching the data from the CEMS and SRMs accordingly.

If the test laboratory uses T-CEMS within SRMs, then differences between the sampling systems of the CEMS and SRMs can result in a difference in integration time. This means that sets of measurements starting and ending at the same time may be uncoordinated. So, the test laboratory must find out if there is a difference in integration time; for example, by injecting a test gas into the CEMS and SRM sampling probes at the same time and determining if there is a significant difference in responses. If this test shows that there is a difference in integration time, then any SRM system may be connected at the same point on the sampling system as the CEMS to align the lag times between sampling and analysis. If this is not practicable, then the test laboratory must measure the differences in lag times and correct the data accordingly.

Alternatively, if both the SRM and CEMS data have been recorded electronically the data can be aligned afterwards by matching the peaks and troughs.

3.7.17 Establishing the calibration function and the test of variability

If a QAL2 and variability test are required for NH₃, and there is no designated ELV, then the test laboratory should use a virtual daily average ELV of 10mg/m³ and a measurement uncertainty of 40%. Similarly, a virtual daily ELV of 10% and a measurement uncertainty of 10% must be used for carbon dioxide (CO₂), and a virtual ELV of 20mg/m³ and a measurement uncertainty of 20% for N₂O. The measurement uncertainties are for variability test calculations only and are not used for subtracting tolerances in the final reported results.

Legislation specifies an uncertainty of 10% for CO. In practice, this is difficult to achieve, whilst the risks of applying a higher uncertainty of 20% are low. So, test laboratories may use 20% for the CO variability test.

For installations that have a low ELV of 5mg/m³ or less for dust, a confidence interval of 50% may be used for the variability and acceptance tests.

The variability test should be based on the results of parallel reference tests only. It must not include data obtained by use of reference materials.

3.7.18 Spare, repaired or replaced CEMS

There will be situations when a CEMS is replaced either temporarily or permanently with a new CEMS.

If the CEMS is a standby system mounted on the same stack as the duty CEMS, then it may either be verified and calibrated at the same time as the primary CEMS during the QAL2 or verified and cross-calibrated over a period of at least 3 days using the calibrated CEMS.

If you replace a CEMS with an identical one, then you may use the same calibration function as the original CEMS and then verify the performance using an AST. If needed, the AST may then be extended to become a full QAL2.

If you replace the CEMS with a different type of CEMS, then this replacement must either undergo a full QAL2 or only the functional tests supported by cross-calibration using a verified and calibrated CEMS. This entails using data from the verified CEMS.

If you use a transportable CEMS for back-up purposes, then this should either have a QAL2 for each stack on which it operates or may be cross calibrated using permanently installed CEMS on the stack, providing that these have been calibrated.

After repairs, or when replacing a CEMS with an identical system, the following procedure is recommended:

1. The same quality of test gases is used when setting up the repaired or replacement CEMS, after which you should apply the same calibration function as before, and then continue with the QAL3 procedure.

2. If there are no visible step changes in either the measured emissions when the process has not changed, and if the QAL3 baselines do not change, then no further action is needed.

3. If there are visible step changes, then the parallel reference tests for an AST are required. If the repaired or replacement CEMS passes both the variability and validity acceptance tests, then no further action is required.

4. If the repaired or replacement CEMS fails either test, then a QAL2 is required.

5. If all or part of the CEMS is repaired, QAL3 data must be collected and inspected in the defined manner to determine whether a new QAL2 is needed.

All parts used for repairs must be covered by the instrument’s QAL1 certificate, so the parts must meet the same performance specifications as those used within the CEMS during its QAL1 testing and approval.

3.7.19 Cross calibrating back-up CEMS

Where plants employ a standby CEMS that is identical to the duty CEMS and both CEMS measure emissions from the same part of the process, then the calibration function from the duty CEMS may be applied to the standby CEMS. However, a check must be made to make sure that the results from both CEMS do not differ by more than half the uncertainty allowance specified by the applicable directive.

Alternatively, if a CEMS has been calibrated and its performance has been verified, then this CEMS may be used to check a back-up CEMS placed on the same stack. Use measurements from the verified CEMS to check (and if necessary, calibrate) the measurements from a back-up CEMS. Although this procedure does not come under MCERTS accreditation the Environment Agency recommend that a company with accreditation to EN14181 performs it. A procedure for cross calibration should involve the following steps:

1. Perform the functional tests as specified in QAL2.

2. Use the same test gases to initially calibrate the CEMS, and to perform the zero and span tests.

3. Make sure that the sampling ports are close enough to sample representatively for both CEMS.

4. Operate the duty and standby CEMS together on the same stack for at least 3 days.

5. Use sets of parallel data as specified in QAL2 to calculate the calibration function and to perform the variability test. This should result in a calibration function for the second CEMS.

6. If the CEMS passes the variability test and the results from both CEMS do not differ by more than half the uncertainty allowance specified by the applicable legislation, then the calibration of the second CEMS is valid, without the need to apply a calibration function.

7. As an additional means of quality assurance, the Environment Agency recommend that the results from the CEMS and back-up CEMS are examined using the procedure for assessing reproducibility, as defined in EN 15267-3.

3.8 Frequency of QAL2 checks

QAL2 tests must be performed:

• when the CEMS is installed
• at least every 5 years, plus or minus 2 months. Similarly, AST checks must be done every 12 months, plus or minus 2 months. If the tests are not done within the 2-month window, then the test laboratory must include the reason for this in the QAL2 or AST report
• if a QAL3 evaluation or AST demonstrates a need for a QAL2
• after a failure of the CEMS that requires significant repairs affecting the calibration
• after a significant upgrade or other significant change to the CEMS affecting calibration
• if there is a change of fuel
• if there is a significant change of process
• if any of the above changes alter the emissions
• when the CEMS exceeds the limits for operating outside its valid calibration range

When a CEMS operates above its valid calibration range, EN14181 specifies conditions that trigger the need to carry out a repeat QAL2. One of the conditions states that a repeat QAL2 will be required if more than 5% of the number of automated measuring systems (AMS) measured values calculated over a weekly period (based on standardised calibrated values) are outside the valid calibration range for more than 5 weeks in the period between ASTs. Experience has shown that this can be problematic because some processes may have relatively low emissions but be subject to occasional spikes causing the CEMS to exceed the limits for operating outside its valid calibration range. This can lead to repeat QAL2s being performed that show no improvement in the calibration function. To alleviate this problem, it is acceptable to increase the trigger level from 5% to 10%. If a CEMS fails the validity acceptance test during an AST, then this implies that a full QAL2 is required. Similarly, if there is a replacement CEMS and following installation and commissioning, there appears to be a step-change in both the measured emissions and QAL3 baselines for span gases, then this also suggests that a QAL2 is required. However, in the first instance, it is acceptable to perform the parallel reference tests required by an AST (using the existing calibration function), before performing a QAL2. If both the variability and validity acceptance tests pass, no further action is needed.

3.9 Performing an AST instead of a QAL2

After the initial QAL2, the Environment Agency may allow ASTs during the following years. This would mean that at year 6 in the QAL2/AST cycle an AST may be performed instead of a QAL2. However, this can only take place if there has been no significant change to plant operation or fuel since the last AST. Also, at least 95% of the CEMS measured values at standard conditions obtained since the last AST and the SRM measured values obtained during the AST, must be less than the maximum permissible uncertainties shown in Table 1.

Table 1 – Measurands and their allowable uncertainties

Measurand	Uncertainties (%)
NOx	20
SO2	20
CO	10
HCl	40
Dust	30
TOC	30
HF	40
NH3	40

3.10 Significant changes to operating conditions and fuels

A QAL2 must be performed within 6 months of a significant change to plant operation or fuel. An operational change is considered significant if it triggers the need for a permit variation and the change alters the emissions profile such that the calibration function is no longer valid.

A change of fuel is considered significant if:

• it results in a change in the emissions profile
• it requires a permit variation and alters emissions
• the change is from any one of the following types to another (gaseous fuel, liquid fuel, or solid fuel) and the alternative fuel is used for more than 10% of the time during a year
• the change is from a single type of fuel to a mixture of more than one type of fuel (or vice versa), and the alternative fuel (or mixture) is used for more than 10% of the time during a year

However, a new QAL2 for changes in the process or fuel will not be needed if:

• the operator can demonstrate that the change in process does not affect the emissions profile and the original calibration factor remains valid
• the thermal input is less than 10% per year for the alternative fuel
• the change in fuel use can be shown to have no significant effects on emissions, when compared to the original fuel

If there are significant changes to plant or fuel, the operator is required to demonstrate that the calibration function is still valid. If the operator is confident that the calibration function has not changed, then an AST can be performed to provide the evidence. If the AST shows that the calibration function is still valid, through a pass of the variability and validity acceptance tests, then a new QAL2 is not required. If it is known or strongly suspected that a change in calibration function has occurred because of changes in plant or fuel, then the operator may choose to perform the QAL2.

High O₂ concentrations can result in apparent breaches of the ELV, as well as apparent failures of the QAL2 and AST exercises. For example, in a batch process, the O₂ concentrations can reach high levels at the end of a process. This problem is known to the Environment Agency and can occur even at very low concentrations, typically when the actual emissions have fallen significantly. In such cases, it is likely that the high-O₂ concentrations are representative of start-up or shut-down modes of operation, rather than normal operations. The ELVs in permits will apply to normal operations and exclude start-up and shutdown, unless otherwise stated.

Therefore, if test laboratories encounter this situation, then the following actions are recommended:

• determine whether results with higher-than-normal O₂ concentrations can be disregarded as outliers, (an O₂ concentration above 18% can be considered as abnormal)
• if the O₂ concentrations are high throughout the QAL2 or AST, then determine whether these concentrations are normal
• if the calculations result in QAL2 or AST failures, then note this in the report together with a caveat explaining the reasons, that is higher than normal O₂ concentrations
• ask the Environment Agency for guidance on a case-by-case basis

3.11 Extending the calibration range

If the data is sufficiently linear to derive a valid calibration function, then based on the calibrated CEMS data, the calibration range for gases may ordinarily be extended by 10%, or to 20% of the ELV whichever is greater; and 100% for dust (if the particulate CEMS is demonstrably linear). This is not applicable when using procedure c.

The calibration range may be extrapolated further for gases using reference materials. The calibration function is extrapolated to the daily ELV. This extrapolation is valid if the difference between a reading from a reference point (for example a span gas) at the ELV, and the extrapolated calibration line at the ELV, is not more than the uncertainty specified by legislation.

However, there may be cases where an operator would need to perform a QAL2 several times because of the unpredictable nature of emissions. If the operator can demonstrate this unpredictability and can use surrogates to extend the valid calibration range to an acceptable degree of accuracy, then it may be extended with the Environment Agency’s consent.

In special situations such as these, the valid calibration range may be extrapolated up to:

• 2 times the daily ELV for gases
• 3 times the daily ELV for dust

The highest reading with a surrogate must not differ from the extrapolated calibration function by more than half the 95% confidence interval of the ELV. This arrangement applies to combined cycle gas turbines and solid fuel power stations with low carbon monoxide (CO) emissions.

3.12 Requirements for test gases

Test laboratories must use traceable gas standards for QAL2 and AST testing. Traceable gas standards are prepared under EN ISO/IEC 17025 accreditation to better than plus or minus 2% uncertainty at a 95% confidence interval. These gases are EN ISO/IEC 17025 accredited gas standards (AGS) and are referred to as certified reference gas mixtures. The analytical result is stated on the calibration certificate with a calculated measurement uncertainty.

The available gases and concentration ranges vary between suppliers and accreditation is not widely available for N₂O. EN ISO/IEC 17025 AGSs do not have a specified stability period; the test laboratory is expected to submit the AGS for re-certification periodically in accordance with the supplier’s recommendations. Under the scope of UK Accreditation Scheme (UKAS) accreditation, the test laboratory is also permitted to use a transferable gas standard on site, having been cross compared with the AGS beforehand.

Operators usually use multi-component working gas mixture standards (WGMS) for QAL3 drift checking, with a concentration tolerance of plus or minus 2%. Certificates of analysis are prepared in accordance with ISO 6141 – Gas analysis – Contents of certificates for gas analysis. These gases may also be used for occasional adjustment of the instrument provided that the CEMS manufacturer accepts the gas mixture as being suitable for adjustment purposes. Adjustment may be required following a QAL3 failure or when the instruments are serviced. WGMS have a declared stability period, ranging from 6 months to 3 years, after which they must be replaced.

If binary gas mixtures are required by the CEMS manufacturer for instrument adjustment, the operator may use the same set of binary gases for QAL3 purposes. EN ISO/IEC 17025 AGSs should be used under these circumstances, if these are available for the required gas components and concentration ranges.

If the site gases are not accredited, it is recommended that the test laboratory checks the site gases using EN ISO/IEC 17025 AGSs. Do this by passing the site gas through the test instruments that have been adjusted using the AGS. The required level of agreement should be within the combined uncertainty of the site and test gases; for example, an uncertainty of 2% for both the site and test gas would require agreement within 2.8%. If there is poor agreement, the test laboratory must report this and decide on the best course of action with the operator. For example, there may be sufficiently close agreement for QAL2 and AST purposes, and the quality of the site gas can be investigated at a later stage. However, if the QAL2 and AST are jeopardised then the CEMS may need to be adjusted using the test laboratory AGS and a new site gas ordered.

An AGS is recommended for the initial calibration of the CEMS following installation or major servicing.

Functional tests may be conducted using the operator’s WGMS site gases.

You must use AGS for procedure c linearity tests and site gas quality checks.

Reactive gases must be traceable analytically to national or international standards.

You can use zero grade N₂ or clean air for zero checks.

You can use an AGS in place of a WGMS.

3.13 Calibrating flow measurement instruments

The methods used for the calibration of flow measurement instruments are based on EN 14181. The standard, EN ISO 16911-2, Stationary source emissions – Manual and automatic determination of velocity and volume flow rate in ducts, Part 2: Automated measuring systems, follows the principles of EN 14181.

An alternative calibration method to procedure a, b, or c (called method d) has been added. It uses linear regression and forces the regression line through the zero point.

Further information on the application of EN 16911-2 is available in the Environment Agency’s method implementation document for EN 16911-2.

3.14 QAL2 report

Both the QAL2 and AST report submitted to the operator and the Environment Agency must comply with the minimum requirements shown in the report template requirements in section 7 of this document. Reports produced by test laboratories must therefore comply with specified headings, structure, and order of items within the template. The reports must be retained by the operator and test laboratory. It is not necessary to include all raw data (as opposed to averaged raw data), although the raw data must be made available to the Environment Agency and the operator upon request. The averaged raw data means the short-term averages, for example half-hourly or hourly averages that are necessary to perform the calculations prescribed in EN 14181.

Furthermore:

• the operator and the test laboratory must retain all raw data for a period of at least 5 years
• operators and test laboratories must have quality assurance provisions to assure the traceability of data
• recommendations and statements must be included in the QAL2 report

Test laboratories should state any actions that the operator needs to perform. For example, some functional tests may have been omitted before the parallel reference tests; or the QAL3 procedure might be incomplete.

Test laboratories must provide the operator (and the Environment Agency if requested) an indication of the results of the QAL2 or AST not later than one calendar month after the completion of the exercise, provided that the test laboratory has all the information required. On receipt of the final report, the operator may be required by the Environment Agency to provide the following information:

• the date of the QAL2 or AST
• the outcome of variability and validity test of an AST
• description of actions taken in the case of an AST failure
• the calibration function
• the date the calibration function was entered in the DAHS

You must implement the results of the QAL2 within 6 weeks of receipt of the test laboratory report and no later than 6 months after the test date.

4. Ongoing quality assurance during operations

This section outlines the ongoing quality assurance procedures for CEMS.

4.1 QAL3 – general

The principle behind QAL3 is that operators should track the stability performance of the CEMS, and not make any adjustments unless necessary. This is because the zero and span readings on a CEMS will typically undergo minor changes because of influence factors such as changes in temperature and pressure. These variations may give a false impression that the CEMS are drifting; whilst there may be no harm in making small changes to the setting of the CEMS in response to this apparent drift behaviour, it is preferable that the operator does not make any adjustments unless the CEMS really has drifted.

During QAL2 and AST exercises, the test laboratory must check whether there is a QAL3 procedure in place, and whether there are data to show that the operator has implemented the QAL3 procedure. The test laboratory must report their findings in the QAL2 and AST report. The main points to note are that:

• the CEMS must be able to indicate negative values
• when using Shewhart charts, no adjustment of the zero and span figures should be made. If the CEMS has been calibrated over an expanded range, then the zero and span checks will be over this set range

The standard deviation of the CEMS (referred to in EN 14181 as the SAMS) can either be calculated using MCERTS test data, approximated based on the ELV multiplied by a factor, or determined by using span test gases. Multiples of SAMS are then used to set warning levels and alarm levels.

When calculating SAMS, the method described in the examples in EN 14181 for determining SAMS can be used. The Environment Agency recommend using the following influence factors:

• effect of ambient temperature
• effect of stack gas pressure for on-site CEMS
• effect of voltage
• cross-sensitivity to other measurands
• detection limit

When using test gases, several readings are taken, and the standard deviation is calculated from these readings. If the operating conditions at the time result in a high precision, this can result in artificially low warning and alarm limits.

Two limits are set on the control charts, which are a warning limit to show that the CEMS is starting to drift out of control and an action or alarm limit to show that the CEMS has drifted beyond specifications and corrective actions are needed.

Whilst auto-corrections before the CEMS drifts out of the control range are not recommended, such auto-corrections may take place if the CEMS still meet the MCERTS specification for zero and span drift. The QAL3 is a test that the system is remaining under control in the operating environment, particularly if the sampling system is tested by the QAL3 and the auto-corrections are based on the analyser performance only.

Instead of combining uncertainties, the French national standards body (AFNOR) have determined SAMS values for waste incinerators based on a fraction of the uncertainty specified in applicable directives (see table 2). This approach produces similar results to combining uncertainties.

Table 2 – Suggested method for calculating SAMS for waste incinerators

Measurand	Uncertainty allowance (%)	DA ELV (mg/m3 unless stated)	SAMS (%)	SAMS (mg/m3 unless stated otherwise)
CO	10	50	2	1
NOx	20	100	2	2
SO2	20	40	2	0.8
TOC	30	10	3	0.3
HCl	40	10	10	1
HF	40	1	20	0.2
Dust	30	5	10	0.5
N2O	20	20	2	0.5
NH3	40	10	10	1
CO2	10	10%	2	0.5%

Operators may also develop their own control charts and limits based on experience.

The procedure for determining the upper and lower limits on exponentially weighted moving average charts differs from those used for CUSUM and Shewhart charts.

4.2 Zero and span checks

Zero and span checks must be performed using several means and reference materials, such as test gases for gas monitoring CEMS and filters for particulate CEMS. If this is not practicable or possible, then the CEMS supplier may provide surrogates such as filters which should be traceable to national standards in the case of gases. In the case of dust, because of their complex sizes, shapes and behaviour, there is no true surrogate for this. However, any surrogate which remains constant for a given, known time (ideally at least a year) may be suitable for span measurements under QAL3.

4.3 Test gases and reference materials

If test gases are used, then they should be traceable to primary national standards and should have certificates which meet the requirements of ISO 6142 Gas analysis – Preparation of calibration gas mixtures. Test gases are required for all gases with ELVs, unless the operator or CEMS supplier can demonstrate a linear relationship between the drift effects of different gases. Such a relationship would have to have a correlation coefficient of the regression line of at least R2 = 0.99.

Gas-mixing systems can be used, as these are particularly useful for multi-point span checks. Such systems should meet the performance standards specified in United States Environmental Protection Agency (USEPA) Method 205 – Gas dilution calibration.

Surrogate reference materials are required for performing zero and span checks on particulate CEMS and these should be assessed, as part of the MCERTS testing, for their validity in providing an appropriate QAL3 check. However, it is permissible to use surrogates and alternative devices, such as filters or electronic simulations, for dust if these have been validated during the MCERTS testing for QAL1.

In terms of span gas concentrations, there are no firm rules on the actual concentrations. However, a good starting point is to use a span gas which has a concentration of 80% to 100% of the short-term ELV.

4.4 Requirements on the CEMS and data recording systems

To carry out zero and span checks, the CEMS and the data recording systems must be able to record:

• both positive and negative values.
• any changes in readings from the previous zero and span checks.
• zero and span data results for greater than one year.

4.5 Frequency of checks

If operators are using cumulative sum (CUSUM) charts, then weekly zero and span checks will be required. If operators are using Shewhart charts, then the frequency may be based on the maintenance interval determined during testing for MCERTS certification, although the Environment Agency recommend using shorter intervals until sufficient data is available to lengthen the time between checks.

Users have the option to use instruments with either automatic or manual QAL3 checks. Most instruments use automatic self-checks since these tests can be conducted without additional work from personnel.

4.6 Adjustments to span readings

Once the CEMS has been calibrated and passed the QAL2 tests, the initial span readings are used to set the baseline for the control charts. Minor adjustments to zero and span settings of the CEMS must not be made unless several span readings are outside the action limits set within the control charts. If the span readings are within the action and alarm limits on the control charts, then no action is required and the CEMS should not be adjusted.

Some CEMS have automated zero and span functions, together with small automatic adjustments. This approach is acceptable if the CEMS was tested and certified using this approach, and if the applied auto-corrections are recorded.

If the QAL3 results are approaching the action limits at the time of a CEMS service-visit, and there is a clear trend of drift, then the CEMS supplier conducting the service visit may be requested to adjust the CEMS during the visit, rather than arrange for a separate visit soon afterwards.

4.7 Replacing gas cylinders or other surrogates

When replacing gas cylinders, differences in the concentrations of cylinders may mislead an operative into believing that the CEMS has drifted. This is because 2 gas cylinders with seemingly identical contents can produce different readings in a CEMS because of the uncertainty of the concentrations. This results in a step change in the CEMS readings when a cylinder is changed for another. Hence it is important to differentiate and account for such step changes instead of mistaking such changes for drift. Therefore, when changing gas cylinders, the following steps are recommended:

1. Take at least 5 span readings with the current gas cylinder and then take an average of the readings.

2. If the span readings using the current gas cylinder show that the CEMS has not drifted beyond the action limits since the last span readings, then go to step 4.

3. If the CEMS has drifted, then carry out any necessary actions to remedy this; then proceed to step 4.

4. Take at least 5 span measurements using the replacement span-gas cylinder and then set a new baseline for the control-chart span-level using an average of the 5 measurements.

Sets of readings with an existing cylinder, followed by an equal number of readings with a second cylinder, will establish the magnitude of any step-changes.

4.8 Control charts

Under QAL3 the operator regularly checks the response of the CEMS to zero and span reference materials. If these readings are repeated over a sufficiently short period of time, and the CEMS has not had a chance to drift, then the actual readings will be caused by variations in precision and allowable effects of influence quantities. Over time, as you collect more data, there is only a small chance that the readings will change by more than 3 standard deviations, unless the CEMS has drifted. The purpose of control charts is to plot these trends and give an indication of actual or forthcoming drift.

You must calculate or determine the standard deviation for the operation of the instrument under anticipated stack conditions and then use multiples of this standard deviation to set warning levels and alarm (or intervention) levels.

Annex C of EN 14181 provides examples of 3 types of charts.

4.9 Reporting

QAL3 records should include:

• CEMS monitoring approach and technique, operating range, make and model
• any change in make, model, and serial number of the CEMS through the year
• manufacturer’s service visit records and routine maintenance
• manufacturer’s call out records and corrective actions taken
• operator’s routine maintenance and corrective actions
• summary of QAL3 baseline re-sets
• zero and span drift plots
• zero and span drift tabulation

5. Annual surveillance test

This section describes an annual test to check the CEMS calibration function remains valid.

5.1 Purpose of the annual surveillance test

The AST is a mini-QAL2 whose purpose is to verify the continuing validity of the calibration function.

5.2 Functional tests

The requirements and responsibilities for carrying out the AST are the same as for QAL2.

5.3 AST failures

If a CEMS fails the validity or acceptance test during an AST, then an investigation should be conducted to determine the cause of the failure. Once you establish the cause of failure it might be necessary to do corrective maintenance. On completion, the CEMS must undergo a QAL2. A report must be produced for the AST. When a QAL2 is subsequently done, a separate report must be generated for this.

During an AST, an operator may request that the test laboratory collects enough data to cover the requirements of a QAL2. If an AST fails and it is not possible to provide a satisfactory explanation for an AST failure (recognising that this might be caused by poor historic QAL2 implementation), then the additional data collected during the AST test may be used to generate a new QAL2 calibration function. Separate reports must be generated for the AST and QAL2. A summary of the AST failure and investigation must be included in the CEMS maintenance log.

The following aspects should be considered in an AST failure investigation, which may be conducted whilst the test laboratory is on-site or retrospectively:

• historic AST data since the last QAL2
• functional test results from the previous and the current AST
• historic QAL3 data over the previous year
• AMS service records over the previous year
• the performance of the SRM
• when 2 AMS are installed, the agreement between the duty and standby AMS can indicate that the performance of one AMS has shifted compared with the other. Or that the performance of the SRM is lacking and in the case of low concentrations (that is less than 20 mg/m³) of SO₂, HCl and NH₃, the SRM data should be inspected to check that:
  • i) the stabilisation period of the T-CEMS is sufficient and the concentration is stable during the test period
  • ii) there are no losses in the T-CEMS sampling lines. An extended stabilisation period of up to 2 days may be required prior to testing with new or recently cleaned sampling lines

If the AST passes but there are indications of a calibration shift, the test laboratory may provide a new QAL2 calibration if the following conditions are fulfilled:

• the performance of the AMS and SRM are proved satisfactory
• the deviation is checked to make sure it is greater than so/2
• the operator is consulted and agrees

5.4 Parallel reference tests with an SRM

Five tests are the minimum requirement. However, the test laboratory is advised to carry out a greater number in case any tests are deemed invalid. Additionally, the length of the sampling day should not compromise safety. Therefore, the length of the sampling day should balance the requirements for representative data against the requirements for safety.

The test should be based on the results of parallel reference tests. It must not include data obtained by use of reference materials.

5.5 Extending the valid calibration range after an AST

Data from the AST may be used to extend the valid calibration range if the AST data are within the 95% confidence interval of the calibration range. The valid calibration range can be extended up to the maximum measured calibrated CEMS value at standard conditions determined during the AST plus 10%, provided this does not exceed 50% of the ELV.

The QAL2 or AST tests must state the scope of application of the tests, including the type of process.

6. Assessing and reporting the results of functional tests

The following is a list of checks for assessing and reporting the results of functional tests.

6.1 Alignment and cleanliness

A visual inspection, with reference to the CEMS’ manuals must be done on the following when applicable:

• internal check of the CEMS
• cleanliness of the optical components
• flushing of air supply
• obstructions in the optical path

After re-assembly at the measurement location at least the following must be checked:

• alignment of the measuring system
• contamination control (internal check of optical surfaces

6.2 Sampling systems

A visual inspection of the sampling system must be done, noting the condition of the following components, when fitted:

• sampling probe
• gas conditioning systems
• pumps
• all connections
• sample lines
• power supplies
• filters
• NOx converters – if the sampling system contains a NOx converter, then the test laboratory must record when the last efficiency test was performed, and the result of this test.
• the sampling system must be in good condition and free of any visible faults which may decrease the quality of data

6.3 Leak testing

Leak testing must be done according to the CEMS’ manual. The test must cover the entire sampling system.

6.4 Zero and span check

Reference zero and span materials must be used to verify the corresponding readings of the CEMS.

For non-extractive CEMS, zero and span checks must be performed using a reference- path free of flue gas before and after readjustment, and after re-assembly of the CEMS at the measurement location.

6.5 Linearity

During the calibration and linearity tests the applied concentrations must be logged onto the DAHS to prove the complete system (that is concentration applied to the instrument is represented by the instrument output and identical to the value logged on the DAHS). DAHS logged values must be included in the instrument service report.

The linearity of the CEMS’ response must be checked using 5 different reference materials, including a zero concentration.

The reference material with zero concentration, as well as the reference materials with 4 different concentrations, must have a verifiable quantity and quality.

In case of gaseous reference materials, these 4 reference materials can be obtained from different gas cylinders or can be prepared by means of a calibrated dilution system from a single gas concentration.

The reference material concentrations must be selected such that the measured values are at approximately 20%, 40%, 60% and 80% of a range that is at least the short-term ELV. It is necessary to know the values of the ratios of their concentrations precisely enough so that an incorrect failure of the linearity test does not occur. The dry test reference material must be applied to the inlet of the CEMS.

The individual CEMS must be tested using the following concentrations of reference materials applied in a randomised sequence at:

• zero concentration
• approximately 20% of the range
• approximately 40% of the range
• approximately 60% of the range
• approximately 80% of the range
• zero concentration

After each change in concentration, the first instrument reading should be taken at a time equal to at least 3 times the response time of the CEMS. At each reference material concentration, at least 3 readings must be made, 6 readings must be taken at zero. The time between the start of each of the 3 readings must be separated by at least 4 times the response time.

A risk-based approach to linearity testing may be applied to reduce the time for the tests. For example, the readings may be taken after less than 3x the response time, however, if the CEMS fails the linearity test, then the test must be repeated after a period of at least 3x the response time as stated above. Alternatively, the number of repetitions of the test may be reduced if the CEMS passes the required performance criteria by a factor of at least 2 (that is, half the allowable residual). Increasing the waiting time to 5x the response time, for example, may be a means of meeting this requirement.

Where no other method is possible, the linearity can also be performed with the aid of reference materials such as grating filters or gas filters.

The linearity must be calculated and tested using the procedure given in EN 14181, Annex B. If the CEMS does not pass this test, then the problem must be identified and rectified.

6.6 Interference

A test must be done if the process gases to be monitored contain components that are known interferences, as identified during QAL1 and there is a failure of the QAL2 or AST which could be due to interferents.

6.7 Zero and span drift (audit)

The test laboratory must assess whether the operator has a QAL3 procedure in place, and whether the operator has applied this procedure. The evidence for this is a documented procedure, zero and span data, and control charts.

6.8 Response time

The response time of the CEMS must be checked. This can be performed, if appropriate, by feeding of the reference material at the end of the sampling probe. The response time must not exceed the performance requirement applied during the QAL1 tests.

6.9 Service report

As a minimum requirement the service report should include the:

• document reference for work being done
• instrument manufacturer
• instrument type
• instrument model
• instrument serial number
• operating principle
• operating range
• certification details
• compliance with MCERTS (including certificate number)
• date and time work was done
• equipment used – type, serial numbers, calibration dates
• gases used – certificate numbers, expiry dates, binary or mixed gases
• NOx converter efficiency test, if applicable
• calibration and linearity data (as required by EN 14181) where you do linearity testing during the service
• logged data for period of calibration and linearity where linearity testing is done during the service. There may be gaps in the data, for example, if the CEMS are removed from the stack for the linearity test. In such cases, the test laboratory must state why there are gaps in the data.
• name and signature of service engineer.

7. Template for a report

This template specifies the minimum requirements for reports for QAL2 and ASTs, as required by EN 14181. It is a specification for both the contents of a report, and the order of the contents. This means that each item included in this template must be included in the test reports for QAL2 and ASTs and use the same section numbers and headings. The Environment Agency may reject any test reports that do not comply with these requirements. Test laboratories may include additional information at the end of the main report. They may present the information in tables. The full data and supplementary information should be included at the end of the report.

Additional reporting requirements for the calibration of CEMS that measure flow are specified in the Environment Agency’s method implementation document for EN 16911-2.

All pages must have a header and footer, which includes the:

• report reference number
• permit number
• operator and installation name
• year of monitoring
• sequential number of the visit in the year (if applicable)
• version number
• page number (in the format page x of y)

Cover sheet must include the:

• MCERTS and UKAS logos, and registration number of monitoring organisation
• title including type of report (QAL2 or AST)
• permit number
• operator name
• installation name
• dates of monitoring visits
• contract number of reference (if applicable)
• name and address of client organisation
• name and address of monitoring organisation
• date of report
• report approver’s name, MCERTS registration number, function and signature
• signature of person approving the report

Summary of results (report section number 1.1) 

Summary of the results must include:

• whether the test is an AST or a QAL2
• the stack designation
• the measurands
• the value for a in the calibration function
• the value for b in the calibration function
• the valid calibration range based on calibrated CEMS data from the QAL2
• the valid calibration range based on calibrated CEMS data from the AST
• the extrapolated range based on reference materials
• a statement of a pass or fail for the variability test (QAL2 and AST)
• a statement of a pass or fail for the calibration test (AST)
• recommendations where applicable
• the previous calibration function
• has the new calibration function varied by more than 10% from the previous one
• a statement that the calibration function is only valid if the QAL3 data remains within control limits
• a statement that the calibration function is only valid if there are no manual adjustments made to the CEMS other than those allowed to bring the settings back within the QAL3 control limits

Deviations (report section number 1.2) 

This section must state:

• deviations from the SRMs, and reasons for this
• deviations from EN 14181, and reasons for this
• any impacts on the results
• actions required

Information about the regulated installation (report section 2)

Regulatory information (report section 2.1)

This section must include the:

• name of the installation
• address of the installation
• sector for the installation
• date of the last QAL2 or AST
• measurands and ELVs

Operational information and site monitoring provisions (report section 2.2)

Details of operational information and site monitoring provisions must be provided in this section.

Process type and variations in emissions (report section 2.2.1)

This section must include:

• a description of the operating phases of the process
• a statement of the percentage of the load of normal runs and expected variations of emissions
• an explanation of how the expected emissions and variations in the emissions influence the sampling times and duration, to capture a representative set of samples
• any factors that would affect the monitoring results (for example, automatic zero and span operations or low-emissions values)
• whether the emissions are at or near zero
• if the emissions are at or near zero, then state how these are dealt with
• if the CEMS is reading zero, then investigate to make sure that the CEMS is working. An agreement with the client that the implications are understood and that these discussions and findings are documented

Type of fuel (report section 2.2.2)

This section must include:

• a description of the types of fuels and their proportions used during the QAL2/AST, and during a normal operating year
• whether multiple calibration-functions are required
• if the process is co-incineration, then what types and proportions of fuels were used

Abatement (report section 2.2.3)

Type of abatement plant and how this affects emissions.

Periodic monitoring measurement location (report section 2.3)

Information about the measurement location for periodic monitoring must be included in this section.

Stack and sampling ports (report section 2.3.1)

This section must include:

• if the measurement location is in a rectangular or circular stack
• the diameter or dimensions of the stack
• the location of sampling ports
• the number of sampling ports
• Include diagrams or photographs of the emission point, platform, and location.

Monitoring platform and site-provisions (report section 2.3.2)

This section must include:

• confirmation that there is a safe working environment with sufficient space and weather protections.
• confirmation of safe access to the CEMS
• whether there are adequate supplies of reference materials, tools, and spare parts
• whether there are facilities to introduce the reference materials for gaseous- monitoring systems, both at the inlet of the sampling line (where present), and at the inlet of the CEMS.
• the degree of compliance with the requirements of EN 15259

Representativeness of the sample location (report section 2.3.3)

This section must include:

• grid measurements – compliance with EN 15259. State whether the site has had a homogeneity assessment to EN 15259 and when and where this is reported
• the ratio of highest to lowest flow-rates

CEMS (report section 2.4)

Information about CEMS must be included in the following section.

Types of CEMS for each main measurand (report section 2.4.1)

This section must include:

• type, for example cross-duct, on-site, or extractive
• brand
• model
• certification range
• measurement principle
• location of sampling and measurement
• statement of QAL1 compliance
• statement whether moisture is by measurement or calculation – if there are CEMS for moisture, then state the measurement technique.

Types of monitoring for peripheral measurands (reports section 2.4.2)

This section must include monitoring for temperature and pressure, and a statement whether temperature and pressure are recorded.

Information about the monitoring campaign (section 3)

Stack emission monitoring personnel (report section 3.1)

This section must list the names and MCERTS certification status of the people who did the stack emissions monitoring. Certification status include their MCERTS certification level and their technical endorsements, along with the expiry dates for these.

Standard reference methods (report section 3.2)

This section must include:

• measurand
• SRM used
• measurement principle of the SRM
• operational range of the SRM
• certification range of transportable continuous emissions monitoring system used
• uncertainty associated with the SRM
• UKAS accreditation status

Data and calculations – QAL2 (report section 4A)

This section specifies the minimum number of tables and charts, and the minimum requirements for each table. Test laboratories may combine tables where data is repeated.

Raw monitoring data (report table 4.1)

This table must contain:

• start and end times of each pair of data
• raw CEMS results
• stack or CEMS peripheral measurands for temperature, pressure, O₂, and moisture (if measured)
• raw SRM results
• SRM peripheral measurands for temperature, pressure, O₂, and moisture (if measured)
• SRM results expressed under the same conditions as the CEMS results

Standardised monitoring data (report table 4.2)

This table must contain:

• standardised CEMS results (that is standard temperature and pressure, dry and to the reference O₂ concentration)
• standardised SRM results (that is standard temperature and pressure, dry and to the reference O₂ concentration)

Time series plot (report plot 1)

This is a time series of standardised CEMS versus standardised SRM data.

Elimination of outliers

Outliers must be clearly indicated in the averaged raw data set.

Determination of procedure a, b, or c (or method d for flow CEMS)

Data used to determine the calibration function (report table 4.3)

This table must include:

• SRM results expressed under the same conditions as the CEMS results
• raw CEMS results

Calculation of the calibration function

This section must include the calculation of the calibration function.

Calculation of calibrated CEMS values (report table 4.4)

This section must include:

• raw CEMS values
• calibrated CEMS values at CEMS conditions
• peripheral measurands for CEMS
• calibrated standardised CEMS values

Plot of CEMS versus SRM data (report plot 2)

This section requires:

• an x-y plot of CEMS versus SRM data, both at conditions measured by the CEMS, and not standardised
• a plot of calibration function, including R2 value

Data used for the variability test (report table 4.5)

Table 4.5 must include the following:

• calibrated standardised CEMS values
• standardised SRM values
• difference between each pair of values
• difference minus the average of the differences
• difference minus the average of the differences, squared

The variability test

This section must include the calculation of the variability test and a statement of the results.

Plot of CEMS data versus SRM data (report plot 3)

This section requires a plot of the following:

• x-y plot of calibrated, standardised CEMS data versus standardised SRM data,
• Indication of the valid calibration range
• Extrapolation of the valid calibration range, using surrogates
• Parallel lines above and below the regression line through the standardised, calibrated CEMS values and standardised SRM values. The parallel lines should indicate the derived uncertainty (σo) of the allowable 95% confidence interval of the daily average ELV (sometimes called tramlines)

Data and calculations – AST (report section 4B)

This section specifies the minimum number of tables and charts, and the minimum requirements for each table. Test laboratories may combine tables where data would be repeated.

Raw monitoring data (report table 4.1)

This section must include a table of raw monitoring data.

Monitoring data and calculations

This section must include:

• start and end times of each pair of data
• raw CEMS results
• CEMS peripheral measurands for temperature, pressure, O₂, and moisture (if measured)
• raw SRM results
• SRM peripheral measurands for temperature, pressure, O₂, and moisture (if measured)
• SRM results expressed under the same conditions as the CEMS results

Standardised monitoring data (report table 4.2)

This section must include:

• standardised CEMS results (that is standard temperature and pressure, dry and to the reference O₂ concentration)
• standardised SRM results (that is standard temperature and pressure, dry and to the reference O₂ concentration)

Time series plot (report plot 1)

This is a time series of standardised CEMS versus standardised SRM data.

Elimination of outliers

Outliers must be clearly indicated in the averaged raw-data set.

Data used to calculate calibrated values (report table 4.3)

This table must include:

• raw CEMS values
• the original calibration function from the previous QAL2
• calibrated CEMS values at CEMS conditions
• peripheral measurands for CEMS
• standardised calibrated CEMS values

Data used for the variability test (report table 4.4)

This table must include:

• calibrated standardised CEMS values
• standardised SRM values
• difference between each pair of values
• difference minus the average of the differences
• difference minus the average of the differences, squared

Variability and validity acceptance tests

This section must include:

• the calculations set out in EN 14181
• the variability test
• the validity acceptance test
• statement of the results

Plot of CEMS verses SRM data (report plot 2)

This section requires:

• an x-y plot of calibrated, standardised CEMS data versus standardised SRM data
• an indication of the valid calibration range
• a plot with parallel lines above and below the regression line through the calibrated, standardised CEMS values and standardised SRM values. The parallel lines should indicate the derived uncertainty (σo) of the allowable 95% confidence interval of the daily average ELV
• extrapolation of the valid calibration range, using surrogates, if applied

Results of the functional tests (report section 5)

This section must include the results of the functional tests including linearity data, information on test personnel, the names of staff performing the functional tests and the name of person witnessing the tests, where applicable.

8. Quality assurance and quality control procedures for automated dust arrestment plant monitors

The quality assurance requirements for monitors that measure dust but are not required to meet the requirements of EN 14181 are specified in EN 17389 – Stationary source emissions – Quality assurance and quality control procedures for automated dust arrestment plant monitors.

EN 17389 applies to filter dust monitors that can be calibrated but have a larger measurement uncertainty than specified by EN 14181. These monitors can demonstrate that dust emissions are below an ELV and that dust arrestment plants are working properly.

EN 17389 also applies to filter leakage monitors installed at dust arrestment plants. These indicatively monitor changes in dust pulses created by the dust arrestment plant cleaning process. These monitors can demonstrate that a dust arrestment plant is working properly.

Filter dust and filter leakage monitors are performance tested and certified according to EN 15859 – Air Quality – Certification of automated dust arrestment plant monitors for use on stationary sources – Performance criteria and test procedures.