Engine Performance Degradation Trend Prediction Method

Abstract

The invention relates to a prediction method for the performance degradation trend of an engine. The prediction method comprises the following steps that 1, a monitoring data sequence is reconstructed; 2, a gray relevant vector machine prediction model is constructed; 3, an improved relevant vector machine model is constructed; 4, a correction model of oil monitoring data obtained after oil change is constructed. Improved RVM model construction is suitable for a prediction method of OLVF monitoring data, implicit information in the OLVF monitoring data is deeply mined, overall trend prediction on the engine can be achieved, and the technical problem that the performance degradation trend of the engine cannot be accurately predicted through a previous method and means is solved. Meanwhile, the improved RVM model is used for reconstructing noise data in the engine oil monitoring process, noise points are eliminated, overall trend and sudden change information is reserved, and under the condition of progressive increase of the volume of prediction data, especially the continuous data volume of online monitoring, the accurate prediction result is still kept.

Description

technical field [0001] The invention belongs to the cross-technical field of tribology and fault diagnosis, and in particular relates to a method for predicting engine performance degradation trend. Background technique [0002] With the development of economy, my country has become a big country of automobile production and consumption. Automobile engine is a complex mechanical system consisting of thousands of parts. The main failure mode is wear or wear-related failure, accounting for 47.2% of engine assembly failures. [0003] But at present, the main means of engine wear state monitoring is shutting down and dismantling for inspection. Shutdown inspection is more direct to find the cause of the fault, but it is inefficient and takes a long time. Looking at it from another angle, wear and tear will produce wear debris, and the wear debris will carry machine wear information into the lubricating oil. On-line monitoring of lubricating oil is expected to make real-time j...

AI Technical Summary

Problems solved by technology

The invention adopts the data processing method to correct the noise data in the engine oil monitoring process, reconstructs the monitoring data; builds a prediction model suitable for OLVF to realize the evaluation and prediction of the performance degradation trend of the main equipment of the automobile, thereby solving the problem that the previous methods cannot be accurate Technical Problems in Realizing Engine Performance Degradation Trend Prediction

Images