On this page
When is a performance test done?
- A new boiler or a new plant: the supplier has guaranteed emissions, output and efficiency at a certain load and with a certain fuel. The acceptance test shows whether the guarantees are met (EN 12952-15 and EN 12953-11).
- New or rebuilt flue gas cleaning: a bag filter, electrostatic precipitator, scrubber, SCR or SNCR. The guarantee covers the concentration after the cleaning or the collection efficiency, and for NOx reduction often the ammonia slip too.
- A new fuel or a rebuild: a new burner, a new fuel mix or a modified furnace changes the emissions. The measurement shows whether the plant meets the guarantee and the limit values with the new fuel. If the boiler is a medium combustion plant, the supervisory authority must be informed (section 19).
- The permit's first measurement: a medium combustion plant (1–50 MW) must be measured within four months of registration or commissioning (section 45 of the Swedish Ordinance on Medium Combustion Plants, SFS 2018:471). For plants that require a permit, the time is set in the permit or the monitoring programme.
- Guarantee claims: if the plant does not perform as promised, an accredited measurement is the evidence you need against the supplier, with traceable measured values and a stated measurement uncertainty.
What is checked, and against what?
The guarantee values are contract values. They are given in mg/Nm3 dry gas at a reference oxygen content, at a set load and with a set fuel, and they differ between suppliers. Output, steam data and efficiency are tested according to EN 12952-15 or EN 12953-11; METLAB measures the flue gas side of that test.
| Plant | What is usually guaranteed | What METLAB measures |
|---|---|---|
| New steam or hot-water boiler | Output, steam pressure and steam temperature, efficiency or flue gas temperature, unburnt matter in the ash and the emissions in the flue gas (EN 12952-15) | The emissions: dust, NOx, CO and SO2 as agreed, plus O2, CO2, moisture, flue gas temperature and gas flow. The flue gas loss in the efficiency comes from the flue gas temperature, O2 and CO. Output and steam data are read from the plant's own instruments |
| Dust collector: bag filter, electrostatic precipitator, cyclone | Dust concentration after the collector, or collection efficiency in per cent | Dust after the collector according to EN 13284-1. For the collection efficiency, dust is measured upstream of the collector at the same time (ISO 9096 at high concentrations), with gas flow on both sides |
| Acid gas cleaning: scrubber, dry sorbent injection | HCl, SO2 and HF after the cleaning | HCl according to EN 1911, SO2 according to EN 14791 and HF by wet chemical sampling, with O2 and moisture for the conversion |
| NOx reduction: SCR, SNCR, low-NOx burners | NOx after the reduction and ammonia slip (NH3) | NOx according to EN 14792 and NH3 according to EN ISO 21877, with O2 |
| Change of fuel or modified furnace | The emissions with the new fuel at the agreed load | The same parameters as for a new boiler, with the fuel and the load that the contract specifies. Fuel samples are taken during the measurement |
The guarantee applies to steady operation unless otherwise agreed (EN 12952-15). In EN 12953-11 operation counts as steady when the flue gas temperature varies by at most 10 °C and the oxygen content by at most 0.5 % from the mean during the test.
How a performance test is done
Go through the contract
The guarantee states which parameters apply, at which load and with which fuel, the reference oxygen content, the averaging time and the number of sub-samples. If any of that is missing, or a rule for how the measurement uncertainty is to be counted, we agree it with you and the supplier before the measurement. The acceptance test standards want it settled when the boiler is ordered.
Measurement programme and measurement site
We write a measurement programme according to EN 15259: measurement objective, operating conditions, parameters, methods, measurement site and timetable. The measurement site needs sampling ports on a straight run of duct and a safe working platform, and for the collection efficiency measurement sites are needed both upstream and downstream of the cleaning. A preliminary meeting with the supplier settles most things before the measurement day.
Steady operation at the guaranteed load
The supplier tunes the plant in. During the test it runs at the load and with the fuel that the guarantee specifies, and operation must be steady before the measurement starts. Load, steam flow and O2 are logged from the plant's own instruments, and fuel samples are taken.
Measurement with standard reference methods
Dust according to EN 13284-1, NOx according to EN 14792, CO according to EN 15058, SO2 according to EN 14791, O2 according to EN 14789, moisture according to EN 14790 and gas flow according to EN ISO 16911-1. Normally three sub-samples per operating condition, with the sampling time adapted to the guarantee value so that the uncertainty is small.
Evaluation against the guarantee value
The measured values are converted to mg/Nm3 dry gas at the guarantee's reference oxygen content and reported with their measurement uncertainty. The comparison with the guarantee value follows the decision rule that has been agreed, see below.
The report as evidence
You receive an accredited report with operating data, measured values, measurement uncertainty and, if you have asked for it, a statement on the guarantee with the decision rule stated. If the measurement was also the permit's first measurement, you send the same report to the supervisory authority.
Guarantee value, measurement uncertainty and decision rule
Every measured value has a measurement uncertainty, and we report it as the expanded uncertainty at a 95 % confidence level (k = 2). For dust and gases it is often around 10 % of the measured value under good measurement conditions. If the measured value lies far below the guarantee value, the uncertainty makes no difference. If it lies close, the uncertainty decides whether the guarantee is to be considered met, and then a rule is needed. ISO/IEC 17025 calls it a decision rule: a rule for how the measurement uncertainty is taken into account when the laboratory makes a statement of conformity. It must be agreed with the customer before the measurement and stated in the report.
ILAC G8 describes the common rules. Simple acceptance: the measured value is compared directly with the guarantee value, and the risk is shared between the parties; a value right at the limit can lie above it with a probability of up to 50 %. Guard band: the guarantee is considered met only if the measured value plus the uncertainty lies below the guarantee value, which gives at most a 2.5 % risk of a false acceptance, and correspondingly for a rejection. Four-level assessment: pass, conditional pass, conditional fail or fail, depending on where the measured value and its uncertainty lie in relation to the limit.
Which rule applies is set out in the contract. If it is missing, we propose one before the measurement, so that both parties know what the result means. The acceptance test standard for boilers says the same: the acceptable level of measurement uncertainty and the accuracy of the instruments should be agreed, preferably before the contract is signed (EN 12952-15). If the contract requires a certain maximum measurement uncertainty, we assess the suitability of the method according to EN ISO 14956. In practice we lower the uncertainty with longer sampling times, more sub-samples and a steady load.
The performance test and the permit
If the plant is a medium combustion plant (1–50 MW), the first measurement under the Swedish Ordinance on Medium Combustion Plants (SFS 2018:471) must be made within four months of registration, or of commissioning if that comes later (section 45). The measurement is made at steady operation with a representative and even load (section 46) and, if you burn several fuels, with the fuel that gives the highest emissions (section 49). Methods according to the EN standards meet the ordinance's requirements (section 50), and CO is always measured (section 43). The guarantee test and the first measurement are often done on the same visit. The operating conditions can differ, though: the guarantee may apply to full load with a certain fuel, the ordinance to a representative load. Then we measure both.
For plants that require a permit, such as large combustion plants, the permit sets out how the emissions are to be checked: measurement method, measurement frequency and evaluation method (chapter 22, section 25 of the Swedish Environmental Code). The monitoring programme describes the self-monitoring (chapter 26, section 19). The time for a first measurement after commissioning is set in the permit or the monitoring programme. At waste incineration, the Swedish Ordinance on Waste Incineration (SFS 2013:253) also requires the residence time, minimum temperature and oxygen content to be checked when the plant is put into operation, and metals, mercury and dioxins to be measured every three months during the first twelve months of operation (sections 42 and 43).
The difference between the tests is who the result is for. The guarantee test is a matter of contract between you and the supplier, and the values are compared with the guarantee according to the decision rule. The measurement for the authority is your self-monitoring, and the values are compared with the limit value as the rules say: at periodic measurement under SFS 2018:471 without any deduction for measurement uncertainty (section 51). If the measurement shows that a limit value is not complied with, you must inform the supervisory authority as soon as possible (section 55). The same measured values can therefore be judged differently against the guarantee and against the ordinance.
What does a performance test cost?
The price depends on the number of operating conditions and parameters, how many measurement sites are needed (upstream and downstream of a cleaning stage makes two), the sampling time the guarantee value requires and the travel distance. Waiting time is added if the plant does not reach steady operation while we are on site. If the guarantee test and the permit's first measurement are done on the same visit, they share the cost of travel and set-up. Request a quote – we reply within two working days. Approximate prices for the most common measurements are in the price guide.
Frequently asked questions about performance tests
Are a performance test, a guarantee test and an acceptance test the same thing?
Yes. Performance test, guarantee test, guarantee measurement, guarantee check and acceptance test are different names for the same check: a measurement that shows whether the plant meets the values the supplier has guaranteed. In the boiler standards it is called an acceptance test (EN 12952-15 and EN 12953-11).
Who should order the test, we or the supplier?
The contract decides. The acceptance test standards say that the parties should decide at the time of ordering whether the supplier or an independent body is to carry out the test, and who is to attend. An independent laboratory gives a result both parties can use, and the supplier is welcome to be present during the measurement.
Does the laboratory have to be accredited?
That is for the contract to decide, but EN 15259 presupposes a competent testing laboratory according to ISO/IEC 17025. An accredited laboratory reports the measurement uncertainty and applies a documented decision rule when it makes a statement on the guarantee, which makes the report usable if the parties disagree.
Which load and which fuel apply during the test?
Those the guarantee specifies, at steady operation. Usually that is the maximum continuous load, where the emissions of most parameters are highest. CO and NOx often move in opposite directions when the combustion changes, though, so a guarantee test may need more than one operating condition. The operating conditions are written into the measurement programme.
Can the performance test replace the permit's first measurement?
Often, if the same visit covers both requirements. For a medium combustion plant the measurement must be made within four months of registration, at a representative load, with the fuel that gives the highest emissions and always with CO. If that matches the guarantee test's operating condition, one visit is enough. Otherwise we measure both operating conditions on the same visit.
What happens if the measured value lies close to the guarantee value?
Then the agreed decision rule decides. With simple acceptance the measured value is compared directly with the guarantee. With a guard band equal to the measurement uncertainty, the guarantee is met only if the measured value plus the uncertainty lies below the guarantee value. If the contract has no rule, we propose one before the measurement and state it in the report.
The new boiler has a permanently installed measuring system (AMS). How is it checked?
A new automated measuring system is calibrated with a QAL2 according to EN 14181 once it is installed, with parallel measurements against the reference methods. The QAL2 can be done on the same visit as the guarantee test, so that the plant's own measured values are calibrated from the start. A major change in operation, such as new flue gas cleaning or a new fuel, also requires a new QAL2. Read more on the page about QAL2 and AST.
What should the contract say about the measurement?
Which parameters are guaranteed and the values in mg/Nm3, at which reference oxygen content, load and fuel; the averaging time and the number of sub-samples; which measurement methods are to be used; how the measurement uncertainty is to be counted (the decision rule); who measures and who attends; and where the sampling ports are to be placed.