Gear residual life predicting method based on long short-term memory

Abstract

The invention provides a gear residual life predicting method based on long short-term memory, and belongs to the field of big data and intelligent manufacturing. The problems that the existing RNN algorithm cannot solve long-term dependence and recurrent neural network is prone to gradient exploding or gradient vanishing are solved. The specific steps of using the algorithm are as follows: 1, monitoring the degradation of a power gear in real time with a sensor; 2, extracting features of the gear fatigue state, and evaluating the gear wear degradation performance; 3, establishing a real-timemonitoring data prediction model of the gear bending fatigue based on LSTM; 4,optimizing parameters based on the LSTM prediction model; 5, predicting the gear residual life based on gear state estimation and a known gear fault threshold. The gear residual life predicting method based on long short-term memory has the advantages that the gear degradation state and the real-time residual life can beaccurately predicted, and the basis for predictive maintenance of the gear is provided.

Description

technical field [0001] The invention belongs to the field of big data and intelligent manufacturing, in particular to a method for predicting the remaining life of a gear based on a long-short-term memory network. Background technique [0002] The gearbox is the main mechanical component in the mechanical field such as the engine block and the wind power system, and the gear transmission is accompanied by the entire steel industry. With the expansion of the integration scale of modern mechanical equipment and the continuous improvement of precision, the gears are prone to wear and tear during long-term operation, resulting in thinning of the teeth or stress concentration, and the decrease in the fatigue strength of the tooth root reduces the life of the gear, and even occurs under heavy load. The tooth breaks off. Monitoring the health status of gears is an important means to ensure the reliable operation of mechanical equipment. [0003] Gears are generally enclosed in a ...

AI Technical Summary

Problems solved by technology

[0005] The present invention overcomes the deficiencies in the prior art, solves the problem that the existing BPTT algorithm cannot solve the problem of long-term dependence and the recurrent neural network is prone to gradient explosion or gradient dissipation, and aims to provide a method based on improved long-term short-term memory The gear residual service life prediction method of network, the present invention includes the detailed design of three-layer (input layer, hidden layer and output layer) network structure, and the realization algorithm of network training and network prediction etc.

Images