Research on Wearing Prediction of the Cylinder Surface Based on Time Series Autoregressive (AR) Model

Abstract

Samples gotten from lubricating oil used in the marine diesel engine cylinder piston ring were analyzed with the spectrum analyzer to get the value of wearing elements. The time series autoregressive (AR) prediction model was established with monitoring data to predict the wearing data. Actual results showed that, AR model can handle non-linear dynamic data, reflect the nonlinear relationship among the various elements, and the prediction accuracy was high and with a strong practicability.