Estimation of engine friction using vibration analysis and artificial neural network

This paper deals with the estimation of engine friction by vibration analysis. Vibration signals and friction of an IC engine were measured under motored condition. To analyze the vibration signals, time-domain and frequency- domain methods were used. The results showed that the second harmonic of the fundamental frequency was dominant at some engine speeds. The engine friction demonstrated an increasing trend with the speed rise. In order to find a relation between the engine vibration and friction, three different characteristics were extracted from the vibration signals. A procedure based on ANN was also proposed to estimate the engine friction using the vibration signals. The results showed that the proposed procedure can estimate and predict the engine friction with the correlation coefficient of 0.9735, RMSE of 0.0523 bar, MRE of 8.2204% and accuracy of 91.7796%.