A Reliability Growth Prediction Method Based on Ga-elman Neural Network

Abstract

The invention discloses a reliability growth prediction method based on a GA-Elman neural network. The method comprises the following steps: 1, collecting fault data; 2, arranging the fault data intoa training data group; 3, setting GA-Elman neural network parameters; 4, establishing a reliability growth model; 5, performing reliability growth prediction on the product; 6, carrying out reliability growth tracking prediction on the product; 7, analyzing and discussing a result. Through the above steps, the reliability growth prediction method based on a GA-Elman neural network is established.On the basis of historical fault data, fault diagnosis is carried out; every time new fault data is generated. The new state of reliability increase is realized. A growth prediction model is constructed by using neural network nonlinear fitting, model updating is realized by using a self-learning capability. The problems of limited application range of a traditional model, complex parameter solution, incapability of updating the model in time and the like are solved. The prediction accuracy and tracking effectiveness in the growth process are improved, and guidance is provided for reliabilitygrowth management.

Description

technical field [0001] The invention provides a method for predicting reliability growth based on GA-Elman neural network, which belongs to the field of reliability growth Background technique [0002] In recent years, with the rapid growth of my country's economy, my country's manufacturing industry has continued to develop rapidly. At present, my country has established an independent and complete industrial system with complete categories in the world, which has effectively promoted the development of my country's industrialization and modernization. However, it is comparable to the world's advanced level. Compared with other countries, my country's manufacturing industry is still large but not strong, and the gap in quality and efficiency is still very obvious. For this reason, our country puts forward the concept of "Made in China 2025", and is committed to improving the quality level of products in an all-round way. In "Made in China 2025", it is pointed out that we mu...

AI Technical Summary

Problems solved by technology

However, this approach also has certain problems: that is, too much emphasis on prior information, or the prior distribution of assumptions, which will have a great impact on the evaluation and prediction of reliability growth

However, the BP neural network is actually a static feedforward neural network. When it is applied to reliability growth, it will transform the dynamic time series in the growth process into static modeling, which will not be able to objectively describe the reliability growth process. dynamic characteristics

Images