Recommended unit: State Grid Sichuan Electric Power Company Yibin Power Supply Company.
Authors: Yang Xin, Tang Long, Zhong Rui, Li Xiaohang, Sun Xuedong.
Abstract: In order to promote the digital transformation of the production business of power enterprises, improve the level of digital operation and maintenance of grassroots teams. In this paper, through the analysis of the needs of the front-line team for the application scenarios of substation inspection and equipment fault judgment, combined with the advantages of RPA+AI technology, the substation ** surveillance camera fault inspection and equipment defect judgment architecture model design is carried out. Painful and difficult problems such as manual judgment have greatly improved the efficiency of operation and maintenance inspections, and helped grassroots teams to digitally transform and reduce burdens and increase efficiency.
Keywords: RPA technology; AI technology; intelligent patrols; Repeat the operation.
Introduction. With the advancement of the digital transformation of the power industry, technologies such as big data, artificial intelligence (AI), and robotic process automation (RPA) are playing an important role in the digital transformation of the power industry.
First, the repetitive and tedious transactional work has been liberated, which has promoted the cost reduction and efficiency increase of the power industry.
The integration of RPA and AI technologies can provide a perfect solution to the problems of repetitive, cross-system operation, and human judgment in the power industry. RPA realizes the reengineering of business processes by simulating the operation of manual labor in digital systems or equipment such as computers, greatly reducing highly repetitive, logically determined, and large-scale work tasks, and achieving process automation of production business. Through the construction of algorithm models, AI uses robot learning and deep learning to continuously iteratively optimize algorithm models to achieve the ability to replace manual analysis, decision-making and judgment. Therefore, the comprehensive application of RPA and AI technologies can empower the business of the power industry and accelerate the digitalization process of the power industry.
01 The development status and trend of RPA+AI in power enterprises.
For example, the application of RPA in marketing business helps grassroots employees solve the problem of new installation and replacement of low-voltage batch users, solves a large number of system operation process duplication and complex processes, and helps marketing teams improve quality and efficiency. In terms of AI technology, it focuses on the application and analysis of artificial intelligence in the power industry, including intelligent machine inspection, intelligent equipment monitoring, intelligent writing and other fields, to help the grassroots carry out management decision-making analysis and support the improvement of lean management of front-line teams.
However, with the continuous increase of power business systems, the cross-system operation of report statistics is becoming more and more obvious, and the application requirements of RPA+AI in the office field are becoming increasingly prominent. It is an important digital technology means for the digital transformation of the power industry in the future to centralize RPA and AI technologies and make use of their respective technical advantages, based on the application of RPA+AI digital employee technology, to make traditional power services develop in the direction of more automation, intelligence, and intelligence.
02 Based on"rpa+ai"Convergence technology model design and application.
2.1 Design and application of ** monitoring status patrol model based on RPA+AI mode.
2.1.1 Problem analysis.
In the management of power production, on the basis of monitoring the production and operation environment of the substation, the substation monitoring also provides functions such as remote guidance of on-site operations and remote equipment fault judgment, which is the "clairvoyance" of substation operation and maintenance personnel in the production of power grids. With the in-depth construction of the high-definition inspection system of the power grid, the monitoring system of the power grid has gradually expanded, and the number of surveillance cameras as the basic monitoring unit has shown a rapid growth trend. The daily inspection is "visible" and the critical moment "can be broadcast" has become the internal requirements of the current monitoring system. However, the rapid increase in the scale of monitoring and the development of the current personnel operation and maintenance system are seriously mismatched, and the current management ability of the grassroots team is difficult to pay real-time and accurate attention to the operation status of each monitoring unit in the huge monitoring system, and the shortcomings in the rate and availability of the monitoring system are gradually highlighted. At present, most of the most advanced monitoring platforms in the power grid industry do not have perfect automatic inspection, classification and statistics functions, and due to the difference in system platforms, it is not only difficult to customize personalized data interfaces, but also difficult to achieve unified operation status research and judgment of multiple platforms, resulting in the fact that the status of the massive cameras in the monitoring platform can only be manually judged, and there are problems such as cross-platform cumbersome operation, repetitive click viewing, serious visual fatigue, and low efficiency.
2.1.2 Solution.
Based on the current status of maintenance and management of the state of the power grid monitoring system, through the use of RPA+AI fusion technology and the integration of artificial intelligence auxiliary tools of machine vision, the fault inspection model of the monitoring unit of the monitoring platform is constructed, which realizes the automatic inspection of the operation status of all monitoring units including cameras, automatically classifies and counts the situation of the monitoring unit module, reduces manual repetitive operations, improves the accuracy of inspection fault judgment, and automatically pushes the inspection result report to the designated location, greatly improving the inspection efficiency.
*The fault patrol model of the monitoring unit includes RPA module, AI module, report analysis module, etc. In terms of the process design of the model, through the development and configuration of the RPA process, the manual operation is simulated to click on the surveillance camera, the current camera page is stored, and the time and name on the camera are extracted through the AI technology middle platform image processing technology, and the camera is judged by time. If the device is **, the display time is successfully obtained, and the site information of the monitoring device is associated with the name, and the template is synchronously written. If the equipment device is not ** or the monitoring image is abnormal, it is impossible to obtain the ** time status information, judge the camera offline or faulty, and write the fault information such as no network and no frequency picture on the screen; The AI module judges whether there is ambiguity and position shift of the monitoring device through the ambiguity and comparison judgment algorithm model of the AI technology platform, and outputs the abnormal results to the abnormal report; According to the statistical results of the report, the statistical analysis module counts the number of faulty devices, the distribution of sites, the total rate, the rate of each site and other indicators.
Figure 1 **Fault patrol architecture model.
2.2 Design of fault defect identification model for substation equipment based on RPA+AI.
2.2.1 Problem analysis.
The demand for accurate and high-definition equipment remote monitoring has vigorously promoted the construction of high-definition inspection system. In order to accurately obtain the instrument readings, equipment operation status, temperature and other parameters of station-end substation equipment, the high-definition monitoring system adopts a multi-point and customized positioning and deployment method during construction. Although the high-definition monitoring system with wide coverage and perfect monitoring points has been established, the intelligent identification function of the operating status of the terminal equipment is currently in the process of gradual evolution, and the current information is mainly judged one by one by manually on the ** stream, ** and other information. After remotely obtaining insulator cracks, insulator cracks, foreign bodies hanging suspended solids, respirator silicone discoloration, switchgear pressure plate closure, switch cabinet pressure plate separation, bird's nest and other primary equipment operation image data through the high-definition ** monitoring system, the substation operation and maintenance personnel combine their own operation and maintenance experience to judge the ** or ** data one by one, manually mark the abnormal information after finding the abnormality, and carry out report statistics according to the classification of hidden defects, and finally feedback to the front-line operation and maintenance team of the substation for on-site confirmation. This inspection work has problems such as high work experience requirements, repetitive operations, long working hours, and high visual fatigue, and the accuracy is low. Based on the current working status of the high-definition inspection system, the efficient, accurate and intelligent identification of hidden defects of equipment has become an urgent need for power grid substation operation and maintenance personnel.
2.2.2 Solution.
In order to improve the accuracy of the identification of hidden defects of station-side equipment, shorten the time of manual research and judgment, and improve the quality and efficiency of substation operation and maintenance, technicians built a fault defect identification model for substation equipment based on RPA+AI fusion technology, which solved the problem of manual inspection of substation equipment fault defects by front-line operation and maintenance personnel without changing the original system architecture.
Figure 2 Substation equipment fault defect model design.
The intelligent identification model of equipment hidden defects includes RPA module, AI module, statistical analysis module, etc., and the RPA module is customized through the standard customization of the inspection process, combined with the information taken by the high-definition monitoring device of the substation at a fixed point at a fixed time, classifying and storing the equipment, automatically transmitting it to the AI platform in batches for calculation, labeling the results returned by calculation, and classifying and storing the defective equipment according to the site and equipment type; The AI module includes 25 kinds of hidden defect identification algorithms for equipment, such as insulator cracks, insulator cracks, foreign bodies hanging suspended solids, respirator silicone discoloration, frame ladder unlocking, respirator silicone discoloration, etc., which realizes the intelligent judgment of equipment hidden defects. The statistical analysis module classifies and summarizes the result data of the AI module and feeds back to the on-duty monitoring personnel.
03 Based on"rpa+ai"Convergence technology application effectiveness analysis.
3.1**Monitor the operation status and analyze the effectiveness of the inspection
The application of the monitoring status inspection model based on RPA+AI technology solves the painful problem that it is difficult to accurately control the operation status of a large number of monitoring device units in real time, and solidifies the functions of automatic statistical fault inspection report to meet the business needs of the front-line team.
In the traditional work mode, workers log in to various types of monitoring platforms every day, and the average time spent on system switching is 05 hours, click on the monitoring device one by one, manually study and judge the operation status of the monitoring device, the average single monitoring device takes 30 seconds, the total number of various types of devices is about 2,000, and it takes 33 hours to complete a single inspection task33 hours. After using RPA+AI fusion technology inspection model, the inspection of the monitoring system has evolved into a background programmed execution, no longer requiring manual intervention and operation, and the staff can view the overall inspection results of the monitoring platform at a fixed point in time every day, so as to find and deal with any hidden dangers or defects in the monitoring system in a more timely manner.
To sum up, the application of this achievement in a single city alone can save 11,774 hours of labor per year, and the annual labor cost can be saved by more than 380,000 yuan according to the calculation of 500 yuan of labor. If this achievement is rolled out, it will produce considerable economic benefits.
3.2. Analysis of the benefits of intelligent identification of defects and defects of substation equipment.
The intelligent identification model of hidden defects of substation equipment based on RPA+AI fusion technology makes full use of the advantages of RPA technology middle platform and artificial intelligence platform, and combines the two technologies to realize the intelligent identification of hidden defects of station-side primary equipment by substation operation and maintenance personnel, and solves the problem that it is difficult for substation operation and maintenance personnel to efficiently and accurately remotely study and judge the hidden defects of station-side primary equipment.
Under the traditional working mode, the manual judgment of the hidden defects of the station-end equipment needs to be manually identified according to the data type, which has problems such as slow judgment speed and easy error. At present, 1,500 pieces of station-side equipment data are collected remotely, and it takes 500 minutes to complete a research and judgment of hidden equipment defects and 3,041 hours of annual manual research and judgment based on the calculation of 20 seconds for a single single material.
The intelligent identification application of hidden defects of substation equipment based on RPA+AI fusion technology completely frees substation operation and maintenance personnel from the heavy work of image data research and judgment, and the staff only needs to view the inspection results under the formulated directory and selectively check the intelligent identification results. More importantly, the application of this achievement is convenient for finding the hidden defects of the operation of station-end equipment in a more timely and efficient manner, and can arrange the power grid maintenance plan in a more targeted manner, which has a very important guarantee role in supporting the safe operation of the power grid.
To sum up, the application of the results in only a single city can save 380 days of labor per year, and the annual labor cost can be saved by more than 190,000 yuan according to the calculation of 500 yuan of labor. If this achievement is rolled out, it will produce considerable economic benefits.
04 Conclusion.
In the process of digital transformation of the power industry, the focus is on the pain points such as front-line cross-system, repetitiveness, low efficiency, and error-proneness, and empowering traditional businesses through new digital technologies to improve the digital operation and maintenance level of grassroots employees.
This paper focuses on the long-standing pain and difficult problems such as batch repetitive process operation, cross-system operation, and artificial visual judgment in the first-class inspection of the grass-roots team substation, and uses RPA and AI technology to construct the architecture model design of the fault inspection and equipment defect judgment of the substation surveillance camera, and builds the camera inspection model and the substation identification model without changing the original information system, and realizes the automatic inspection of the substation camera and high-definition. It greatly reduces the repetitive operation and manual judgment of the on-duty monitoring personnel, saves a lot of labor time, solves the daily inspection problem of the front-line operation and maintenance team, and provides a useful reference for the application of other application scenarios in the field of substation in the later stage.
In the future, power companies will explore more businesses, more applications, and more scenarios, deeply integrate the reasoning, judgment, and decision-making capabilities of artificial intelligence technology with RPA technology, realize the cognitive ability of RPA process automation, build intelligent process robots, comprehensively empower power business, and promote the digital transformation of all businesses of power enterprises.
References. 1] Zhou Juan, Zhou Jiechong, Ni Yu, Wang Huihua, et al. Research and application of RPA technology for automated operation and maintenance[J].Computer Programming Skills and Maintenance, 2022, (03): 132-134
2] Yuan Xiaohan. Application of machine automation technology in marketing business[J].Volkswagen Electricity,2022,37(02):46-47
3] Hu Quangui, Xie Ke, Ren Lingling, et al. Analysis on the application of artificial intelligence in the power industry[J].Electric Power Information and Communication Technology,2021,19(1):73-80.
4] Liu Ziwen. RPA+AI leads the robot era of intelligent office[J].Shanghai Informatization,2020,17(8):31-33
5] Liu Haitao, Duan Jing, Wang Yanhua, Gu Wei, et al. Application analysis and architecture design of digital employees in power industry based on RPA+AI[J].Power Information and Communication Technology. 2022，20（04）：88-93
6] Liang Qianyun, Pan Kejia, et al. State Grid Sichuan Electric Power Company Technology Middle Platform Xiao Oh RPA Service Platform Application***R], Chengdu: State Grid Sichuan Electric Power Company, 2021
7] Zou Shifu, Zhang Jie, et al. State Grid Sichuan Electric Power Company Technology Middle Platform Artificial Intelligence Platform Application***r], Chengdu: State Grid Sichuan Electric Power Company, 2021
8] Wang Shu. Research on enterprise cost decision-making system based on "RPA+AI"[D], Chongqing: Chongqing University of Technology, 2021
9] Yin Yanhe, Jia Ziran, Zeng Yi, et al. Application of image recognition technology in substation inspection[J].Telecom power supply technology. 2020 ，37（03）：141-142.
10] Fried. Research and practice of artificial intelligence bid evaluation auxiliary tools for machine vision[J].Chinese tenders. 2019，（46）：34-38.
11] Zhang Jiafeng. Research on image recognition technology in artificial intelligence[J].Electronic Technology and Software Engineering. 2018，25（23）：250-251.
12] Zhou Zhiyong. On the development and application practice of image processing and image recognition technology[J].Network security technology and application. 2020，（06）：144-145.
13] Xiao turned red. Research on the application of computer technology in graphics image processing[J].The world of digital communication. 2021，（10）：188-189.
14] Deng Zhenming. Research on intelligent operation inspection and monitoring system of substation and its image recognition technology[D].Guangzhou:South China University of Technology,2020
15] Zheng Minying. Technical transformation and application of intelligent inspection in substation[J].Rural electrification. 2020，（11）.47-48.
Article information**:If there is any infringement, please contact to delete.