Use of a Repeatable Car to Improve Intra-Lab Variability of Emissions Tests 930077

A Repeatable Car (REPCA) program has been developed at the Environmental Protection Agency's National Vehicle and Fuel Emissions Laboratory (NVFEL) as part of an ongoing effort to improve the precision of fuel economy and emissions measurements. This concept of using a repeatable car as an integrated system diagnostic tool is not a new idea in the emissions testing field; however, our statistical analyses and organizational approach may be different from what other laboratories are using. Furthermore, given the NVEFL's role in automotive emissions testing, we felt it appropriate to provide related industries a detailed account of our standard laboratory practices, both for informational and comparative purposes.

In order to separate vehicle and measurement variability in a relatively simple manner, a process was developed to track REPCA data based on Statistical Process Control principles using the calculation of individual site offset values from two week moving averages. Eleven individual parameters are monitored for each test site, and the moving averages of these eleven parameters are based on the previous two week's REPCA data. Therefore, a minimum of 44 REPCA control charts are generated and analyzed each week. In conjunction with this, routine equipment diagnostics are also recorded, thus producing an additional 34 control charts. Based on the maintenance history of a specific site and a review of the weekly control charts, recommendations regarding corrective action or preventive maintenance are formulated. Since implementing this program in August of 1990, we have seen a 65% reduction in our site to site testing variability as measured by a comparative inter-laboratory round-robin test program conducted in 1991.