A support vector machine model to predict a spark ignition engine performance and emissions
2022-01-0454
03/29/2022
- Event
- Content
- The internal combustion engines are still the mainstream of the transportation industry in the foreseeable future. Spark ignition engines are widely used in cars, medium and small trucks and military off-road vehicles due to its low noise during operation conditions, low manufacturing and maintenance costs and so on. However, due to the complexity of its structure and internal combustion mechanism, numerical simulation optimization and in-depth experiments of the engines consume a large amount of time or computing resources. In this study, a faster method based on support vector machine (SVM) model are proposed to predict the engine performance and emissions of a gasoline engine. The indicated mean effective pressure (IMEP), indicated thermal efficiency (ITE), indicated specific fuel consumption (ISFC), carbon monoxide (CO), nitrogen oxides (NOx), unburned hydrocarbons (UHC) and exhaust gas temperature (EGT) are estimated by the machine learning model at engine operating points that cover almost the engine operating regime. A validated one-dimensional (1D) computational fluid dynamics (CFD) model was operated at different spark timings, engine speeds and loads to provide dataset for training (2767 steady state points) and testing (285 steady state points) the established support vector machine model. The calibrated SVM model showed good agreement with the validated performance and emissions results, with the correlation coefficients greater than 95% and small root mean square error. As a result, the established SVM regression model can be used to accurately predict the various performance parameters and emissions indicators of the spark ignition engine, which can replace more complicated, time-consuming engine models and engine bench tests with acceptable prediction accuracy, thus accelerating the process of future engine optimization and improvement.
- Citation
- Yang, R., Yan, Y., Sijia, R., Liu, Z. et al., "A support vector machine model to predict a spark ignition engine performance and emissions ," SAE Technical Paper 2022-01-0454, 2022, .