A New Auxiliary Decision Method For Power Equipment Maintenance Based On On-line Monitoring Data Analysis

With the development of society,the requirements of the whole society for power reliability have also increased.As the largest power supply company in China,State Grid Corporation is committed to improving the reliability of power supply by updating management methods and strengthening technical means.Talking about the technical means,the infrared thermal image detection and partial discharge detection of switch cabinets has been widely used as a daily monitoring method for equipment in power system of China.However,the above-mentioned charging detection means still plays a supporting role in daily maintenance,and is not fully integrated into the work flow of power equipment detection and maintenance.Currently,the operation and maintenance of substation equipment still heavily relies on the operational experience of the operating personnel.The advanced scientific methods are rarely used in the maintenance or overhaul of the substation equipment,and the decision-making process is still based on the experience.On the one hand,the level of equipment operation and maintenance has not been significantly improved.On the other hand,it has caused a waste of many advanced means or resources,and contradict with the management methods such as “3+N” duty and “intensive management” implemented by State Grid Corporation.In order to solve the above problems,this paper has carried out the implementation specification of the charging detection,the analysis of the detection data and the prediction method of the equipment defect development,and the power equipment maintenance decision-making method based on the analysis result of the charged detection data.Content of each section are as follows:1)Based on relevant literatures at home and abroad,the current status of live detection methods such as infrared thermal image detection and transient ground voltage detection in practical power production applications is briefly introduced.The related research contents and engineering application examples of power equipment maintenance decision-making model are summarized.It is found that there are few research or application cases that take the charging test result as the consideration factor of power equipment maintenance decision-making,but with the development of technology,the decision-making of power equipment maintenance based on the mining of big data results of electrification detection is an inevitable development direction.2)In terms of the accumulation of valid data for live detection,this paper proposes a method that select the shortest walking path of the tester as the objective function,including safety distance constraint,detection distance constraint,detection angle constraint and detection space constraint to find the optimal detection location.After the substation actual application detection,the method can effectively overcome the adverse effects of personal behavior preference on the charging detection result,and can obtain the tracking of massive charging detection data with strong continuity,high accuracy and good consistency.3)In the field of live detection big data mining,the analysis methods and defect development prediction methods for different components of power equipment are studied.A method for calculating the correlation coefficient between the test result and the abnormal potential influence factor of the power equipment is used to reflect the correlation between the potential influence factor and the abnormality of the power equipment.Based on wavelet decomposition method,least squares vector machine and autoregressive modeling,the prediction of the development trend of power equipment defects is realized.4)In the application of the big data mining,the sustainable operation time of the power equipment is predicted as the reference element of the maintenance time window in the decision of the maintenance plan,a multi-objective optimization model considering maintenance economy,grid reliability and balanced workload for power equipment maintenance is built,including power flow constraint,overhaul time window constraint,overhaul sequence coordination constraints and grid reliability constraint.At the same time,the model solving algorithm based on Pareto optimal concept and differential evolution algorithm is given.Based on the RBTS-BUS6 system example,the planned maintenance program and the maintenance plan based on the correction of the charging detection result are optimized.The results of the example confirm the effectiveness of the proposed model and the superiority of the power equipment maintenance decision-making based on the analysis of the charged detection data compared with the general cycle planning and maintenance.