ETV Test Design. The data of primary interest in this ETV are the reduction in emissions of NOx, HC, PM, and CO. (Numerous other measurements, including CO2 emissions, will also be made as described in later sections. However, these are not considered critical, and the methods and data quality objectives will be stated in the test/QA plan for each technology.) NOx emissions will be used to compute the number of tests required unless the applicant specifically requests that the reduction in another of the pollutants be used to control test design. The number of and type of FTP tests required for ETV will be as follows: A minimum of three tests is required to provide a mean emission reduction and the 95% confidence interval on that mean based on measured variability for each of the measured emissions and test parameters. For highway engines this minimum is satisfied with one cold start test and three hot start tests. For nonroad engines the minimum is satisfied with three replicates of the test sequence appropriate for the engine classification for which the technology is intended (i.e., three 8-mode tests for applicable engines, three 5-mode test for constant speed engines, three 6-mode test for variable speed engines under 19 kW.) Additional tests may be required to meet the ETV requirement that the test/QA plan provide a 90% probability of detecting the expected emissions reductions when computed using the expected experimental errors for the various measurements. These criteria become controlling for low emissions reductions and/or high test variability for the controlling pollutant, which will be NOx unless another pollutant is selected. This is a planning requirement for the test/QA plan. Section 5.1.1 describes the calculation procedure. Relative to the minimum number, additional tests may be desired by the applicant to reduce the width of the 95% confidence interval on the mean emission reduction. Section
