Knock-Prediction System for Kerosene Engines Using In-Cylinder Pressure Signal

Piston engines fueled by kerosene have a strong application prospect in special vehicles and small aircrafts, but the low amount of octane fuel causes its knock combustion phenomenon to be particularly serious. A will deteriorate power economy engine damage serious cases. Therefore, knocking is key problem with engines. We propose knock-prediction system for based on in-cylinder pressure signals. Firstly, intrinsic mode function (IMF) caused resonance extracted from signal via empirical decomposition (EMD) time–frequency domain analysis. time-domain statistical analysis (TDSA) combined principal component (PCA) used extract features IMF. Finally, data collected test bench are trained support vector machine obtain model. Compared other technical combinations training, proposed scheme achieved more accurate results prediction. Considering working characteristics engines, slight can increase engine. some incorrectly predicted cycles (slight-knock cycles) do not affect normal operation