Grey Modeling Method of Wear Prediction for Reciprocating Machinery Based on Spectrum Analysis

Abstract

Oil spectrum analysis technology is an important method to research wear of mechanism without unpacking, based on the characteristics of limited amount of oil analysis samples and randomness of sample data changes to apply gray theory to prediction of wearing capacity of reciprocating engine. When gray modeling, if sample data does not meet smoothness conditions, firstly parametric process is carried out for the data in order to weaken randomness of the original random sequence, then equal interval process for initial data sequence is conducted and sequence of gray values is determined according to variation rules of sample data, finally, base on four-dimensional data to carry out unequal interval process by timing fitting method. Test confirms that gray modeling by timing fitting methods based on wearing law to determine gray values point can significantly increase prediction accuracy of wearing capacity.