Audio-video-heat integrated anomaly detection and alarm method based on artificial intelligence

Abstract

The invention provides an audio-video-heat integrated anomaly detection and alarm method based on artificial intelligence. The method comprises the following steps: 1, collecting the temperature, sound and image data of a capacitor and a reactor, and carrying out the 24-hour uninterrupted data collection of a reactive electrical device of a power transformation and distribution station through a medium-wave infrared detector, a directional sound pickup and a high-definition video camera; step 2, by analyzing the thermal imaging picture, denoising the sound by using a clustering method, and comparing the deformation of the video image, the preprocessing of the audio, video and thermal information is achieved; 3, according to the three types of data anomalies, abnormal state recognition is conducted through a fuzzy neural network method, and the collected data are subjected to classified macroscopic judgment and output; and step 4, comprehensively considering three types of data anomalydetection results to form a comprehensive criterion and an early warning threshold, and further formulating a corresponding intelligent early warning mechanism. The method has an efficient and accurate anomaly detection alarm effect, can reduce the labor cost, and facilitates the development of unmanned safe operation of the power transformation and distribution station.

Description

technical field [0001] The invention relates to the field of distribution network operation and maintenance, in particular to an artificial intelligence-based audio-visual-heat integrated anomaly detection and alarm method. Background technique [0002] When reactive devices such as capacitors and reactors in substations and distribution stations are abnormal or faulty, various characteristics will appear, such as abnormal heating, sudden noise, oil leakage and deformation, etc., but there is currently no effective means to simultaneously take into account temperature, sound and Image detection. Based on the characteristics of fault heating and abnormal sound of specific frequency, it is proposed to simultaneously use thermal imaging and pickups to capture the characteristics of abnormal situations, and use intelligent analysis to identify, warn and alarm faults or abnormalities. Based on sound, temperature, and image multi-directional intelligent detection, early warning a...

AI Technical Summary

Problems solved by technology

[0002] When reactive devices such as capacitors and reactors in substations and distribution stations are abnormal or faulty, various characteristics will appear, such as abnormal heating, sudden noise, oil leakage and deformation, etc., but there is currently no effective means to simultaneously take into account temperature, sound and Image detection

Images