Method and system for fault feature information extraction and early fault early warning

Abstract

The invention discloses a method and a system for fault feature information extraction and early fault early warning. The method comprises the steps of: acquiring system state data, preprocessing the system state data, extracting feature information, acquiring standardized variables, and establishing a covariance matrix of the standardized variables; reducing dimensionality of the covariance matrix by combining a PCA principal component analysis method, and selecting a principal component with the largest contribution rate; monitoring and acquiring data of the principal component with the largest contribution rate in real time as time sequence data, and acquiring a time sequence model and predicted error data through fitting and modeling; and judging the predicted error data, and sending out an early warning signal. According to the method, the data is subjected to dimensionality reduction by adopting the PCA principal component analysis method, the principal component with high contribution rate is extracted, the predicted error data is obtained on the basis of the time sequence data of the principal component, and the predicted error data is judged to perform timely prediction and diagnosis on a fault, thereby reducing the difficulty of diagnosis and manual inspection cost to a great extent.

Description

technical field [0001] The invention relates to the technical field of fault analysis and early warning, and more specifically relates to a method and system for extracting fault feature information and early fault early warning. Background technique [0002] With the rapid development of injection molding industry technology, how to ensure the safe and stable operation of industrial machines has become an important issue, and the real-time monitoring and fault warning of injection molding industrial machines is the key to ensure safe and reliable operation of injection molding production and reduce defective products The essential. [0003] Most of the traditional fault diagnosis methods in the injection molding industry are manually diagnosed, which is time-consuming and laborious, and the accuracy and reliability are very low. Timely detection of faults, the accuracy of fault diagnosis is not high, let alone pre-discovery, reducing the generation of defective products fr...

AI Technical Summary

Problems solved by technology

[0003] Most of the traditional fault diagnosis methods in the injection molding industry are manually diagnosed, which is time-consuming and laborious, and the accuracy and reliability are very low. Timely detection of faults, the accuracy of fault diagnosis is not high, let alone pre-discovery, reducing the generation of defective products from the root

At the same time, there are many detection factors for judging the faults of injection molding machines, and the most relevant detection factors cannot be determined, and the accuracy of fault targeting is not high.

Images