The Design And Implementation Of Intelligent Video Monitoring System On High-voltage Electrial Partial Discharge

The smooth operation of high-voltage electrical equipment is the precondition of regular power supply of electric power system. Since the country’s increasing requirement for the intelligent power grid, the implementation of real time monitoring on high-voltage electric equipment becomes urgent. The main failures of electric equipment are insulation problems, which are primarily caused by partial discharges of electric equipment. By monitoring the partial discharges, insulation problems can be timely analyzed and predicted, and further accidents can be prevented. Hence, it is of both significant engineering and practical value to study the partial discharge monitoring system.This paper mainly designs an intelligent video monitoring system on high-voltage electrical equipment partial discharge, which integrated technologies relating to computer vision and image processing, to implement discharge target automatic detection and video segments storage in surveillance video, and accomplishes system implementation and test. For none of the domestic researchers have conducted in-depth studies or fulfilled the system implementation of the method of monitoring partial discharge exploited by this paper, this study is of great exploratory value. The main work is outlined as follows.The causes and types as partial discharge are discussed. The features of discharge are analyzed. In order to attain massive discharge video experimental data, a laboratory simulation device that can generate corona discharge, arc discharge and pont discharge is designed and made.For implementing real-time detection and identification of discharge video, the improved Gaussian mixture algorithm and two types of improved three-frame difference methods are discussed and analyzed. By testing a variety of types of discharge through the three algorithms, and balancing between real-time processing speed and accuracy in the discharge target test, an improved three-frame difference algorithm that combines the Sobel edge detection methods is eventually decided to be used. This algorithm helps determine the actual resolution parameters of the camera equipment.In the former monitor systems, videos are directed stored without filtering out theirrelevant information, which leads to a huge waste of storage space. This paper proposes corresponding solutions to this problem. The target detection algorithm identifies the discharge phenomena in the video, and then the monitoring system cuts and stores the video segments those contain only discharge phenomenon. This selective storage method effectively lowers the request for the capacity of hard drives of the system, which saves the cost and brings convenience to the staffs.At last, this paper completes the hardware and software implementation and practical testing of the intelligent video monitoring system on partial discharge. The experimental result shows that the partial discharge monitoring system designed in this paper can accomplish the real-time target detection and identification for electrical equipment partial discharge video, moreover it can cut and save the needed video segment. In short, the system meets the design task requirements.