Knowledge data hybrid-driven multi-channel lubricating oil abrasive particle abnormity monitoring method

The invention relates to a knowledge data hybrid-driven multi-channel lubricating oil abrasive particle abnormity monitoring method. The method comprises the following steps: constructing an abrasive particle confidence rule base based on expert knowledge; different rules of the abrasive particle belief rule base are activated based on the triangular membership degree and a rule activation formula through the operation data, and activation weights are obtained; fusing different rules by using an evidence reasoning mode on the basis of the activation weight so as to obtain a preliminary abrasive particle abnormity monitoring result; calculating the accuracy rate of the preliminary abrasive particle abnormity monitoring result, ensuring that the accuracy rate reaches a set value or above, inputting historical data when the monitoring accuracy rate is smaller than the set value, and optimizing expert knowledge by adopting a sequential least square method; and inputting the optimization expert knowledge into the