Prediction of combustion and performance parameters in a homogeneous compression ignition (HCCI) engine using genetic programming method

As a result of the increase in commercial and personal needs around the world, the use of road vehicles is increasing day by day. Due to limited fuel resources and increasing environmental pollution, scientists are constantly working to improve vehicle engines. With the advancement of technology, various researches are carried out on internal combustion engines in order to provide higher thermal efficiency and lower emission conditions. In line with the findings obtained, progress has been made in this field by focusing on low temperature combustion engines. HCCI engines, known as one of the combustion modes with low temperature combustion engines, provide higher thermal efficiency, lower NOx and soot emission values compared to traditional internal combustion engines. However, due to some adverse operating conditions of HCCI engines, such as knocking at high load values and misfiring at low loads, engine operating ranges are more limited compared to traditional internal combustion engines. HCCI engines have not fully taken their place in the industry due to problems such as knocking at high loads and misfiring at low loads. In this study, the numerical analysis study was verified by comparing the numerical verification results with the Genetic Programming method using the experimental study data found in the literature and the open source GPTIPS tool set written in MATLAB 2018a. Using the multi-gene genetic programming method, which is a branch of Genetic Programming, the HCCI engine combustion performance parameters, namely maximum in-cylinder pressure, pressure increase rate (MPRR), average indicated pressure (IMEP), CA10, CA50 and CA90, were estimated and compared with the values in the experimental data set. The relationships between them were examined by comparison. The coefficient of determination, RMSE, MSE, MAE, SSE performance values of the results were compared. Sensitivity analysis was performed according to the maximum and minimum variability levels of the 4 input parameters in the Pmax, MPRR, IMEP, CA10, CA50 and CA90 values, which were estimated with the help of experimental data, and their behavior as a result of the analysis was examined.