The Effects of Aircraft Engine Pollutant Emission Measurement Variability on Engine Certification Policy

Abstract

Emission measurements on a sample of nominally similar aero-gas-turbines—a single engine type—exhibit scatter as a result of assembly and manufacturing tolerances and of instrumentation and sampling errors. For some pollutant species, the standard deviation can be a quarter of the mean. The paper shows how to obtain an estimate of the value below ehich, say, 95 percent of the population will lie, by fitting an appropriate statistical distribution to the data. When the amount of data is small so that the precision of such estimates is not good, it is not meaningful to predict a limit which contains more than about 95 percent of the population. On entering service, each new engine type will require a certificate, similar to an Airworthiness Certificate, recognizing compliance with the emissions regulations. The authors show that, despite the large variability in the measurements, it is possible to devise a certification procedure which requires a test on only one engine. It would have to demonstrate emissions levels rather better than the maximum permitted and the results would have to be supported by additional information on the variability of related standards of engine, obtained perhaps during the development program. Some engine results available to the authors have been analyzed for variability and the paper recommends values of the parameter σ/mean to be used in setting certification standards for the different species of pollutant.