Modelling a quality assurance standard for emission monitoring in order to assess overall uncertainty

摘要

Emissions of many pollutants to the atmosphere are controlled to protect human health and the environment. Operators of large combustion plants are required under the industrial emissions directive to continuously measure various pollutants and report their measurements. The directive and associated standards set uncertainty requirements for this process. The main governing standard is EN 14181:2014, covering the required quality assurance regime for automated measuring systems (AMS). Uncertainty assessment on single measurements are fairly straightforward, but understanding overall uncertainty for annual mass emissions from an instrument only calibrated every 3-5 years requires significant effort. NPL has modelled a gas analyser running according to the quality assurance scheme described in EN 14181:2014, with each measurement including a suite of uncertainties. These measurements include recorded values and those from the quality assurance tests, QAL2 (calibration), QAL3 (regular drift correction test) and AST (annual surveillance test to check the calibration function), both from the AMS and a standard reference method (SRM) where appropriate. The model uses a Monte-Carlo approach to provide an overall uncertainty for the annual mass emission. The model has been tested with data from hypothetical good, ok and barely meeting requirements AMS, illustrating the effects of improvements in uncertainty from monitoring equipment on overall reported emission totals. While meeting measurement uncertainty limits has to be demonstrated, operators are not required to report emission totals with uncertainties. The result is little understanding of the impact this uncertainty has. The NPL model demonstrates this impact, encouraging industry to improve their measurement quality.