Heterogeneous data cooperative modelling industrial fault detection method based on neighborhood projection preservation

Abstract

The invention provides a heterogeneous data cooperative modelling industrial fault detection method based on neighborhood projection preservation, and belongs to the technical field of the fault monitoring and diagnosis. The method comprises the specific steps as follows: collecting N groups of current and image data; solving a structure weight coefficient of each image data; establishing a current neighborhood projection map dovetailing matrix; rejecting the data greater than the current neighborhood weight threshold, and performing normalization; constructing a target function, and solving aprojection vector optimal solution by using a manifold dimension reduction method; performing dimension reduction on the image data produced in the electrolytic magnesium process after graying; and performing fault detection on the dimension-reduced image data by using a SVM classification method. A heterogeneous data modelling problem is mainly solved, the fault detection is performed when the image information and the current information are combined, a process monitoring result for the method shows that the false alarm is greatly reduced, and the accuracy of the fault detection is furtherimproved.

Description

technical field [0001] The invention belongs to the technical field of fault monitoring and diagnosis, and in particular relates to an industrial fault detection method based on collaborative modeling of heterogeneous data maintained by neighborhood projection. Background technique [0002] Large crystal fused magnesium is made of pure natural light-burned magnesium powder through ultra-high temperature electric melting and firing. Due to the characteristics of stable physical and chemical properties, large crystal fused magnesium has become an excellent high-temperature electrical insulation material, which is mainly used in metallurgy, chemical industry, electrical appliances, aerospace and national defense and other industrial fields. Ultra-high temperature metallurgy products are a major demand for China's industrial development. Compared with the low-temperature industrial process, the ultra-high temperature smelting process is more complicated, and the production cond...

AI Technical Summary

Problems solved by technology

However, the existing research results are limited to separate modeling and diagnosis of physical variable data and multimedia heterogeneous data, which ignores the inevitable connection between big data and loses the inherent advantages of big data

Moreover, the straightening process of the traditional machine learning method when processing the image is easy to lose the structural information of the image.

Images