Equipment residual life prediction method and system

Abstract

The invention belongs to the technical field of mechanical equipment state monitoring and life prediction, and discloses an equipment residual life prediction method and system. The method comprises the steps: firstly, screening historical monitoring signals of multiple physical quantities of equipment need to be acquired, and signals with high relevancy from the historical monitoring signals; setting a normalization label for the screened signal and converting the normalization label into a matrix form; constructing a nested long-short-term memory neural network, and training by using a knownsignal matrix to obtain an equipment residual life prediction model; finally, further optimizing model output through a Gaussian filtering method, and guaranteeing that an output result is stable andreliable. According to the method and the system adopting the method, the residual service life and the degradation state of the equipment, especially the mechanical equipment which works for a longperiod and has a coupling fault mode, can be accurately predicted in real time, the faults of the mechanical equipment can be perceived in advance, and safe, stable and long-term operation of the equipment is guaranteed.

Description

technical field [0001] The invention belongs to the field of state monitoring and life prediction of mechanical equipment, and relates to a method and system for predicting the remaining life of equipment, and more specifically, to a fusion nested long-short-term memory neural network (Nested Long Short-Term Memory, NLSTM) algorithm and Gaussian smoothing filter (Gauss filter) equipment remaining life prediction method. Background technique [0002] With the development of the equipment manufacturing industry in the direction of automation, networking, greening, and intelligence, the structure of mechanical equipment is becoming more and more complex, and the relationship between the components is getting higher and higher, and the functions are compounded. Once it fails, it will be difficult to diagnose the cause of the failure in a timely and accurate manner. In addition, the operating environment of mechanical equipment is complex and the working conditions are changeabl...

AI Technical Summary

Problems solved by technology

[0003] Whether it is a recurrent neural network (Recurrent Neural Network, RNN) model, or a long-term short-term memory neural network (Long Short-Term Memory, LSTM) and its related derivative network structures, data processing is performed in the order of time; this means Long-term information will gradually disappear over time, causing the problem of vanishing gradients

Although the information that LSTM can remember has been improved by an order of magnitude compared with RNN, in the field of actual state monitoring and life prediction technology, the length of the time series sequence that needs to be processed is much greater than the magnitude that LSTM can remember

In addition, if the time span is large and the network is deep, then training RNN or LSTM will consume a lot of resources

These have brought great challenges to the accurate prediction of the remaining life of mechanical equipment

Images