Multivariable-grey-neural-network-based prediction system for residual life of industrial equipment

Abstract

The invention, which belongs to the technical field of industrial control, especially relates to a multivariable-grey-neural-network-based prediction system for a residual life of industrial equipment. The system comprises a residual life precision platform for predicting the residual life of industrial equipment in real time; and when the residual life of the industrial equipment is predicted each time, industrial equipment residual life processing is executed. The processing for the industrial equipment residual life is executed by six steps. The system has the following beneficial effects: problems that the prediction precision level and the prediction capability of the modeling method can not be guaranteed and the influence on the equipment or system operation by field working condition factors are not taken into consideration can be solved; the original sequence trend becomes obvious; key performance parameters can be predicted; the precision level, the prediction capability, and the applicability of the model are enhanced; and an RBF neural network error correction model is obtained.

Description

Technical field [0001] The invention belongs to the technical field of industrial control, and in particular relates to a system for predicting the remaining life of industrial equipment based on a multivariable gray neural network. Background technique [0002] The remaining life prediction of the equipment or system is an important basis for formulating equipment replacement and maintenance strategies, which has a very important impact on the long-term stable operation of the equipment or system. The remaining life of a device or system is defined as the time difference between the degraded performance parameter reflecting the operating state of the device or system from the current moment to the moment when it first reaches its threshold. The definition of a single performance degradation parameter: [0003] URL j =(t d -1)-t i , J = 1, 2,..., N [0004] URL=min(URL j ) (t i Is the current moment, t d Is the moment when the performance degradation parameter reaches the threshold)...

AI Technical Summary

Problems solved by technology

[0005] In many existing studies, the remaining life prediction of equipment or systems generally only uses the performance degradation data of the equipment itself. In this context, in order to obtain higher prediction accuracy, more performance degradation parameters are needed. To establish a remaining life prediction model, the modeling method of this type of model has the following two shortcomings. First, when considering the individual differences of equipment or systems, the amount of data on performance degradation parameters obtained at the initial stage of operation is small. At the same time, it is difficult to obtain the performance parameters of the model equipment or system, so it is difficult to guarantee the prediction accuracy level and prediction ability of this modeling method. Second, the influence of on-site working conditions on the operation of the equipment or system is not considered.

Images