Precision & Accuracy of Odor Control System Performance Tests

Odor control system performance is measured by testing odor samples collected from system outlets and sometimes from system inlets as well.Odor testing is a precise testing method, however, precision depends on the quality assurance/quality control (QA/QC) statistics associated with the odor laboratory. The European odor testing standard, EN13725:2003, describes a standard method using n-butanol for monitoring the performance of the odor panel members (assessors), the odor laboratory, and the odor test results (CEN, 2003).EN13725 defines precision, bias, and accuracy as they relate to odor testing. Precision measures the scatter of data. Bias is a measure of systemic errors. Accuracy is the closeness of a value to the true or correct result. Accuracy is a function of both precision and bias. The assumption is made in EN13725 that the laboratory's precision and accuracy for the standard odorant, n-butanol, is transferable to environmental odor samples tested by the laboratory (CEN, 2003).A 95% confidence interval can be applied to odor control system performance testing. EN13725:2003 Annex H “Example of the calculation used to determine the number of odour concentration measurements required to detect a difference between two means” presents the statistics and the 95% confidence interval for remove efficiency of an odor control system.Performance testing, when conducted with a set of samples (inlets and/or outlets) evaluated by the same odor panel in one odor laboratory, uses the laboratory's within panel precision data to determine the confidence interval of odor performance and/or reduction efficiency. Performance testing, when conducted with sets of samples (inlets and/or outlets) evaluated by more than one odor panel in the same laboratory, uses the laboratory's inter-panel precision data to determine the confidence intervals of odor performance and/or reduction efficiency.Examples in this paper show how the statistics of precision and accuracy are used to understand odor control system performance testing results. Depending on the sampling and laboratory protocols, different levels of precision and accuracy may be used for computing confidence intervals and performing statistical comparisons. If an odor control system performance test requires samples sent to a laboratory over several days, the precision and accuracy of the odor results would be defined by the laboratory's inter-panel standard deviation. On the other hand, if performance test samples are collected on the same day and evaluated within 30-hours by one odor panel, the within panel standard deviation of the laboratory would define the precision and accuracy of the results.