The use of computational models for internal combustion engine development is ubiquitous. Numerical simulations using simpler to complex physical models can predict engine’s performance and emissions, but they require large computational capabilities. By comparison, statistical methodologies are more economical tools in terms of time and resources. This paper investigated the use of an artificial neural network algorithm to simulate the nonlinear combustion process inside the cylinder. Three engine control variables (i.e. spark timing, mixture equivalence ratio, and engine speed) were set as the model inputs. Outputs included peak cylinder pressure and its location, maximum pressure rise rate, indicated mean effective pressure, ignition lag, combustion phasing, burn duration, exhaust temperature, and engine-out emissions (i.e. nitrogen oxides, carbon monoxide, and unburned hydrocarbons). Eighty percent of the experimental data from a heavy-duty natural gas spark ignition engine were utilized to train the model. The perceptions accurately learned the combustion characteristics and predicted engine responses with acceptable errors, evidenced by close-to-unity coefficient of determination and close-to-zero root-mean-square error. Moreover, the regressors captured the effect of key operating variables on the engine response, suggesting the well-trained models successfully identified the complex relationships and can help assist engine analysis. Overall, the neural network algorithm was appropriate for the application investigated in this study.
Machine learning (ML) models can accelerate the development of efficient internal combustion engines. This study assessed the feasibility of data-driven methods towards predicting the performance of a diesel engine modified to natural gas spark ignition, based on a limited number of experiments. As the best ML technique cannot be chosen a priori, the applicability of different ML algorithms for such an engine application was evaluated. Specifically, the performance of two widely used ML algorithms, the random forest (RF) and the artificial neural network (ANN), in forecasting engine responses related to in-cylinder combustion phenomena was compared. The results indicated that both algorithms with spark timing, mixture equivalence ratio, and engine speed as model inputs produced acceptable results with respect to predicting engine performance, combustion phasing, and engine-out emissions. Despite requiring more effort in hyperparameter optimization, the ANN model performed better than the RF model, especially for engine emissions, as evidenced by the larger R-squared, smaller root-mean-square errors, and more realistic predictions of the effects of key engine control variables on the engine performance. However, in applications where the combustion behavior knowledge is limited, it is recommended to use a RF model to quickly determine the appropriate number of model inputs. Consequently, using the RF model to define the model structure and then employing the ANN model to improve the model's predictive capability can help to rapidly build data-driven engine combustion models.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.