A real-time early warning method for electrical equipment temperature based on abnormal factor extraction

Abstract

The invention discloses an electrical equipment temperature real-time early-warning method based on abnormal factor extraction. The method comprises the following steps: (1) collecting temperature data of the electrical equipment in running in uniformly spaced manner by using an infrared sensor, acquiring a temperature data flow ti (i=0,1,2,...n); (2) preprocessing the collected temperature data flow to obtain an input data set (3) in a unified way; (3) computing a local abnormal factor value LOF of each Tk in the data set; and (4) comparing the computed local abnormal factor value LOF with apreset value, analyzing and correspondingly performing real-time early-warning. The electrical equipment running state can be real-time monitored and early-warned in non-supervision manner based on the temperature data flow collected by the infrared sensor, and an operation of adjusting according to the specific model specification performance parameter of the monitored equipment is unnecessary ingeneral; in the face of multiple monitoring data flows, a certain abnormal signal can be well represented and early-warned.

Description

technical field [0001] The invention belongs to the field of infrared fault diagnosis of electrical equipment, in particular to a real-time early warning method for electrical equipment temperature based on abnormal factor extraction. Background technique [0002] With the development of society and economy and the proposal of "Industry 4.0", more and more electrical equipment are widely used in HVAC systems of rail transit and large buildings, key parts of aerospace systems and various power systems, as well as industrial production sites and its equipment room, etc. There are a large number of these electrical devices and their types are complex. Once the key devices operate abnormally, it is very likely to cause serious safety accidents, threatening the benefits of social and economic development and the safety of public life and property. Therefore, through the real-time early warning and fault diagnosis of the key equipment of the system, it is very important to realiz...

AI Technical Summary

Problems solved by technology

[0004] However, the above-mentioned real-time early warning and status monitoring methods for equipment operation inevitably have the following similar problems: they all belong to the category of supervised learning, and the successful construction of their models depends on a large number of labeled data sets, which limits the further promotion and use of the models ; and, when in use, it is necessary to make corresponding adjustments according to the specific model specifications and performance parameters of the tested equipment, which also restricts the generalization ability of the model to a certain extent; The fusion of data streams (such as principal component analysis) is used as the input of the model to realize the online detection and real-time early warning of the important electrical equipment of the key system, which will inevitably lose part of the information, so the abnormal signal collected by the sensor may not be correct. will be detected and characterized, affecting the reliability and robustness of these real-time early warning methods

Images