Regressionsmodellering av dynamiska råemissioner från statiska mätningar
By using steady-state measurements for predicting emissions under a dynamic drive cycle wouldsave a lot of time and money for the exhaust aftertreatment specialists at Volvo cars. The idea forthis thesis has been to investigate if statistical regression models can be used with good accuracy.Questions included are for example if common operating variables such as engine speed, air-fuelratio etc. is sufficient to predict engine-out emissions over the engine operating range with goodaccuracy. Focus was set on the modelling of warm engine, but also the more complex engineheat-up phase was investigated since it is a great contributor to total emissions. While NOxcouldnot be measured because of malfunctioning measurement equipment, only HC, CO andtemperature at inlet of first catalytic converter has been modelled. For the experiments a SI6naturally aspirated petrol engine was used, and drive cycle tests were run with a S80 withautomatic gearbox.MATLABs? tool for statistical modelling,Model-Based Calibration Toolbox, were used since itincludes everything needed both for building and evaluation of the models. To broaden theanalysis, two separate test plans were made with different approaches regarding spark advancewhich experienced difficulties. Much time was spent evaluating both polynomial and RBFmodels. Tools as PRESS RMSE, R2and graphical residual analysis were used.Many interesting discoveries were made, including a very good prediction of CO with a R2valueof 0.95. Temperature at the inlet of first catalyst was slightly worse with a R2value of 0.9 whereHC reached only 0.6. Further findings include that different correction factors were needed to geta good drive cycle prediction. Except for model errors this discrepancy can come from responsetimes of the measurement equipment and environmental differences between the two rigs used.Tests with cold start showed good agreement with findings from other researchers; that it ispossible to scale emissions from fully warm engine proportional to coolant temperature for agood prediction during the engine heat up.Model wise were RBF Gaussian and 4thorder polynomial models the best, but RBF Modelsshould not be used for predicting the CO response. The correlations between the rigs are not fullyknown, and without more investigation of this and the response time phenomena it is altogetherhard to get any further than this.