A fault classification and diagnosis method based on weighted nearest neighbor decision

Abstract

The invention discloses a fault classification and diagnosis method based on weighted nearest neighbor decision, aiming at weighting variables of various faults by using nearest neighbor component analysis algorithm, and diagnosing fault types by weighted nearest neighbor distance on the basis thereof. Specifically, the method of the invention firstly optimizes corresponding weighting vectors foreach reference fault type one by one by using the nearest neighbor component analysis algorithm. Then, the weighted nearest neighbor distance between the weighted samples is calculated, and the faulttype of the on-line fault data is diagnosed in real time. Compared with the traditional method, the method of the invention is based on the nearest neighbor relationship, and does not need a sufficient number of available training samples, no matter whether the weighted vector of each fault is found or the fault type is diagnosed on line. In addition, the main idea of the method of the invention is to optimize the weighting coefficients of each variable corresponding to each fault type, so as to highlight the variation of the characteristic variables of each fault. As such, that method of thepresent invention is a more prefer fault classification and diagnosis method.

Description

technical field [0001] The invention relates to a data-driven fault diagnosis method, in particular to a fault classification diagnosis method based on weighted neighbor decision-making. Background technique [0002] In order to ensure safe production and maintain stable product quality, it is necessary to accurately diagnose the faults that occur during the operation of the production object. It can be said that as long as there is production, process monitoring will always be a research topic of widespread concern in the industry and academia. In the existing scientific research literature and patent materials, research on fault detection emerges in an endless stream. In contrast, there are only a handful of research results on fault diagnosis. From the requirements of process monitoring tasks, both fault detection and fault diagnosis are indispensable. Generally speaking, the task of fault detection is to tell us that there are abnormal conditions in the production pro...

AI Technical Summary

Problems solved by technology

This is mainly because these algorithms need sufficient training data to ensure the accuracy of the model when establishing a classification model, and they are usually not suitable for fault classification and diagnosis

Images