A procedure to demonstrate that the CEM is suitable for the intended purpose before installation, by meeting required performance standards and the uncertainty allowances specified in EU Directives.

Visit www.MCERTS.net

A procedure to calibrate the CEM once it has been installed, using SRMs and then verify whether it still meets the required uncertainty allowances, once installed.

The preference is for the test laboratory to witness all of the functional tests. From experience, this is not always practical. However, witnessing the tests is the first choice. If this is not possible, then assessment of other evidence that the tests have been performed is required. Whatever is done must be stated in the report. Lastly, it is the operator’s responsibility to ensure that the functional tests are performed, and functional tests usually show whether there are any faults with the analyser, if QAL3 does not reveal such faults.

Some instrument technicians do not seem to have set procedure for performing 14181 FT. If a supplier is performing the functional tests the supplier has to have a management system and procedures to cover the functional tests.

The procedure(s) must be vetted to see if it meets the requirements; if there is not one, then this must be stated this in the report and recommend that the supplier writes such a procedure.

Particulate analysers – quite often only see electronic check via laptop

Functional tests should comply with EN 14181 as far as practicable. Some CEMs are a bit like cars, whereby technicians can connect a laptop and perform a diagnostic check. So look for evidence that the requirements of Annex A of EN 14181 have been performed. If not, then ste in the report what has – and has not – been performed.

Quite often this is only a verbal

Look for the following training on audits (it is usually easy to tell when someone understands what they are meant to do and know), demonstrations, tours of the monitoring systems (ask basic questions, e.g. ‘how does this work?’, ‘what does this do?’, ‘why is this system set up like this?’; documentary evidence to support the above, e.g. training matrices and records.

Make sure the report documents the responses.

If the standards requires linearity checks and response-time checks, then at least the analyser should be checked for all gases. For NH3, HCl and HF, one check through the whole sampling system, followed by checks on just the analyser are acceptable.

Current EA policy is not to allow variation of the emissions, if this results in higher than abnormal emissions, even if below the ELV. If the result is a low-level cluster because of this, then this must be stated the report.

If tests are near zero (e.g. shutdown) oxygen value is near ambient, hence results are very erroneous. (Some results indicated tests during shutdown, but O2 values were still as normal operation this indicates and obvious error).

Permits refer to ‘normal operations’, which excludes start-up and shut-down. High oxygen levels can also occur during these periods, or at the end of a burn cycle. In this case, do not include the levels with very high oxygen. When checking the zero readings of the analyser, do not perform the oxygen correction, or you will get divide by zero/low number values.

Yes – EN 14181 requires you to verify that the CEMs read zero when there are no emissions. In effect, Method B is a rather complicated way of telling you to take an average of the points, and then draw a line from the average point through zero.

You don’t, unless a procedure of a standard states that you must do so. The CEN NOx standard, for example, requires the test lab to check the zero and span setting at the start and the end of the measuring period (at least). If your own procedures require zero and span checks during the middle of the day, you can amend them. If a contract requires this, then please contact the issuer of the contract and cite the applicable standards. Bear in mind, however, under conditions where there are large temperature variations, it may be prudent to perform extra zero and span checks.

Low clusters <30% of ELV – clarification on procedure to follow for calibration to surrogates.

Yes. Theoretically EN 14181 does not allow this, but then EN 14181 does not provide enough eliminating need to repeat visit at later date and hence extra cost for client who has ‘low’ emissions?

If the SRMs show a low level cluster and you cannot apply EN 14181’s calculations sensibly, then state in the report that;

1. the SRMs verify that the CEMs are working,

2. recommend calibrating with surrogates.

Usually the supplier does this, but anyone else who is competent with setting up analysers with surrogates can do this.

In simple terms, you say that you cannot calibrate using a SRM as the data is too low; so the CEMs are calibrated using a surrogate instead. This does not have to be witnessed.

Connect to the DRS and then verify that the DRS reflects the CEM readings.

Random selection of data can often produce different cal factors. This can vary from the difference between a pass and a fail indicating issues with the analyser, to small differences but both passing.

It should not do so; first, eliminate outliers using the attached procedure. Secondly, eliminate any data which is not normal operations. Then use all the remaining data.

If so we could then get down to 15 tests and still have an outlier. Some outliers may be borderline; the procedure estimates the probability that a data point is an outlier or not. You can use your judgment here; Reports have been issued with obvious outliers, which were removed, and then there were less than 15 points (it was HCl – probably rogue chlorides). But report is still accepted the with 13 points.

A procedure to maintain and demonstrate the required quality of the CEM during its normal operation by checking the zero and span readings.

The owner of the CEM(AMS) system e.g. the process operator.

A procedure to evaluate the CEM to show that it continues to function correctly and the calibration function is still valid.