A power grid robot process automation construction method and system based on RPA+AI

Abstract

The application discloses a power grid robot process automation construction method and system based on RPA+AI, relates to the technical field of power automation and artificial intelligence, and comprises the following steps: capturing non-standardized scene data to be acquired; utilizing OCR file detection technology, identifying the data type based on pre-constructed power grid text data detection rules, and determining a business mode; utilizing a decision tree model to analyze the captured non-standardized scene data and the determined business mode, and determining specific operation steps of an automation process; applying a man-machine interaction mechanism, simultaneously collecting state information of task execution and feedback opinions of users by an RPA background, forming a closed-loop feedback, and performing decision optimization; the application significantly improves the capability of power grid business in response to non-standardized scenes, improves the stability of power grid business processes, simultaneously establishes a man-machine interaction mechanism, realizes secondary data verification, and improves the accuracy and interaction capability of power grid business processes.

Description

Technical Field

[0001] This invention relates to the fields of power automation and artificial intelligence technology, and in particular to a method and system for constructing automated power grid robotic processes based on RPA+AI. Background Technology

[0002] With the continuous expansion of the power grid and the increasing complexity of its operations, power grid companies need to process massive amounts of data and information, including equipment condition monitoring, fault diagnosis, load forecasting, and power dispatching. These data processing tasks are typically repetitive and standardized, but manual processing is not only inefficient but also prone to errors. Therefore, power grid companies urgently need a method to automate this data processing to improve efficiency and accuracy.

[0003] Existing power grid process automation systems suffer from problems such as low system stability, low data accuracy, low autonomy, and poor interactivity. Low system stability arises because the method involves information interaction between multiple systems (distribution automation system, control cloud system, etc.), and the actual business scenarios are complex, with a large amount of non-standardized data and complex operational steps. Applying traditional robotic process automation (RPA)... RPA (Robotic Process Automation) technology, while handling related tasks, carries risks such as operational steps not meeting actual task requirements, resulting in low stability. It also lacks verification mechanisms, leading to low data accuracy: Traditional RPA lacks data verification mechanisms, posing risks such as data collection errors causing data source errors in business systems; data loss or errors during multi-system information exchange; data processing anomalies due to non-standardized scenarios; and the inability of existing RPA technologies to perform manual verification of critical data during process execution, leading to data parsing anomalies and system operational risks. Furthermore, it suffers from low autonomy and inefficiency: Traditional RPA can only execute fixed operations according to pre-set steps or times, lacking autonomous monitoring, resulting in rigid process handling and low efficiency. Finally, it lacks interaction and controllability: The absence of human-computer interaction mechanisms prevents manual supplementation of data that RPA cannot automatically acquire, and it lacks information awareness and status control during RPA task execution. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is: how to improve the stability, data accuracy, autonomy and interactivity of the power grid robotic process automation system to meet the needs of power grid dispatch for non-standardized scenarios, high accuracy and dynamic response.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a method for automating the construction of power grid robot processes based on RPA+AI, including:

[0008] Two methods, web page data crawling and file data crawling, are used to crawl non-standardized scenario data.

[0009] Based on the captured non-standardized scenario data, OCR file detection technology is used to identify data types and determine business models based on pre-built power grid text data detection rules;

[0010] The decision tree model is used to analyze the captured non-standardized scenario data and the identified business model, and the specific operation steps of the automation process are determined based on the analysis results.

[0011] Based on the specific operational steps of the automated process, a human-computer interaction mechanism is applied. In steps where manual data supplementation or verification is required, proactive interaction is carried out. At the same time, the RPA backend collects task execution status information and user feedback to form a closed-loop feedback and optimize decision-making.

[0012] As a preferred solution for a power grid robotic process automation construction method based on RPA+AI, wherein:

[0013] The webpage data crawling includes:

[0014] Determine the pages that need to be accessed in the power grid business scenario and the target elements to be crawled;

[0015] The RPA robot sends an access request based on the target page to obtain the target page's HTML document;

[0016] The obtained HTML document is loaded and the webpage style is parsed to obtain the element nodes and style information of the target page, and the position and size of the element to be obtained or manipulated are calculated.

[0017] Based on the calculated element position and size information, the RPA robot uses a browser plugin to locate the element and then crawls the located target element.

[0018] As a preferred solution for a power grid robotic process automation construction method based on RPA+AI, wherein:

[0019] The file data capture includes:

[0020] Identify the file types of power grid business documents;

[0021] Based on the identified file type, different tool libraries are invoked to extract the file data content.

[0022] As a preferred solution for a power grid robotic process automation construction method based on RPA+AI, wherein:

[0023] The method of using OCR file detection technology to identify data types based on pre-built power grid text data detection rules includes:

[0024] Obtain the image and text files of the power grid services executed by RPA;

[0025] The image text file is decoded and characters are extracted using OCR recognition technology.

[0026] Based on the decoded file content and extracted character sequences, a language model is applied for text recognition.

[0027] Data extraction is performed on the identified content.

[0028] As a preferred solution for a power grid robotic process automation construction method based on RPA+AI, wherein:

[0029] The method of using OCR file detection technology to identify data types based on pre-built power grid text data detection rules also includes:

[0030] Collect power grid business text data, including power grid specification documents, power grid rules and regulations, and power grid business data;

[0031] Based on actual power grid business application scenarios, texts containing interference information, sensitive information, and key information features are artificially generated to form a text dataset for algorithm training.

[0032] The text dataset is preprocessed to remove noise and convert the text data into a uniform format.

[0033] Based on the specific needs of power grid operations, the preprocessed text dataset is divided into different categories and classification features are extracted.

[0034] The extracted classification features are manually or semi-automatically labeled to distinguish normal text from text containing inappropriate, sensitive, or critical information, and then labeled accordingly.

[0035] Based on the extracted features and labels, a text detection model is constructed;

[0036] Based on the text detection model, key discrimination rules or feature combinations are extracted to form a specific rule set for filtering interference information, detecting sensitive information, and extracting key information.

[0037] The rule set is integrated into the RPA system to detect newly generated text data in real time; at the same time, the system's performance is continuously monitored, and the rules are adjusted or the text detection model is retrained according to the actual situation.

[0038] As a preferred solution for a power grid robotic process automation construction method based on RPA+AI, wherein:

[0039] The process of using a decision tree model to analyze the captured non-standardized scenario data and the determined business model, and determining the specific operational steps of the automated process based on the analysis results, includes:

[0040] Based on the captured non-standardized scenario data and the defined business model, we analyze the repetitive tasks and data processing requirements in the power grid business and construct a decision tree model.

[0041] Based on the decision tree model, anomaly detection or fault prediction is performed on the RPA execution status.

[0042] A data interaction channel is used to exchange business data and status information during the actual execution of the RPA robot with the results predicted and detected by the decision tree model, and the prediction mechanism of the decision tree model and the data processing mechanism of the RPA robot are coordinated.

[0043] Based on the actual execution of business operations, analyze the problems encountered in the practical application of RPA, and collect user feedback through the RPA robot backend to optimize the operation process and decision-making of the RPA robot for the problems existing in the actual business process.

[0044] As a preferred solution for a power grid robotic process automation construction method based on RPA+AI, wherein:

[0045] The specific operational steps based on the automated process utilize human-computer interaction mechanisms, including:

[0046] Perceive contextual information during the execution of the RPA robot;

[0047] Based on the context information, the execution content of the preceding and following operation steps is obtained;

[0048] The content is interactively evaluated to determine whether manual intervention is required.

[0049] If interaction with the user is required, the RPA robot will proactively display a pop-up window or issue a voice prompt;

[0050] For special data types, a secondary judgment is added. If special data or security information is detected, the RPA robot will pop up a prompt window again, requiring the dispatcher to confirm the information a second time.

[0051] Secondly, embodiments of the present invention provide an automated construction system for power grid robotic processes based on RPA+AI, including:

[0052] The data crawling module is used to crawl non-standardized scenario data using two methods: web page data crawling and file data crawling.

[0053] The detection and recognition module is used to identify data types and determine business models based on the captured non-standardized scene data, using OCR file detection technology and pre-built power grid text data detection rules.

[0054] The process establishment module is used to analyze the captured non-standardized scenario data and the determined business model using a decision tree model, and determine the specific operation steps of the automated process based on the analysis results.

[0055] The human-computer interaction module is used to apply human-computer interaction mechanisms to specific operation steps based on automated processes. It allows for proactive interaction in steps that require manual data supplementation or verification. At the same time, the RPA backend collects task execution status information and user feedback to form a closed-loop feedback loop for decision optimization.

[0056] Thirdly, embodiments of the present invention provide a computing device, including:

[0057] Memory and processor;

[0058] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the one or more programs are executed by the one or more processors, the one or more processors implement the RPA+AI-based power grid robot process automation construction method as described in any embodiment of the present invention.

[0059] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the aforementioned method for constructing automated power grid robot processes based on RPA+AI.

[0060] The beneficial effects of this invention are as follows: By introducing AI technology, this invention achieves pattern recognition of business scenarios, customizes rules according to the scenario, and automatically executes process tasks, assisting dispatchers in completing daily tedious tasks, reducing their workload, improving work efficiency, and solving the problems of complex power grid application scenarios and the strong limitations of traditional RPA, thus improving the stability of power grid operation. It employs machine learning algorithms to establish an adaptive learning model for the RPA robot, adding feedback data based on the actual needs and feedback of power grid regulation to form a closed-loop feedback, enabling RPA self-learning and decision optimization. Furthermore, it applies OCR recognition technology to identify data types, confirm business patterns, and select the optimal operation process through a decision model, thereby optimizing process execution efficiency. By applying the controller and actuator mechanism of an intelligent robot, it solves problems such as key data supplementation, secondary data verification, and execution permission confirmation through proactive human-machine interaction during execution, improving system interactivity and data accuracy. Finally, it develops a real-time event perception mechanism to monitor the process execution status in real time, identify critical moments requiring dispatcher intervention, and ensure timely dispatcher response, achieving efficient decision-making and operation. Attached Figure Description

[0061] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is an overall flowchart of the power grid robot process automation construction method based on RPA+AI described in the first embodiment of the present invention;

[0063] Figure 2 This is a flowchart of the text detection rule construction process of the power grid robot process automation construction method based on RPA+AI described in the first embodiment of the present invention;

[0064] Figure 3 This is a schematic diagram of the anomaly detection process of the RPA+AI-based power grid robot process automation construction method described in the first embodiment of the present invention;

[0065] Figure 4 This is a schematic diagram of the human-computer interaction process of the power grid robot process automation construction method based on RPA+AI described in the first embodiment of the present invention;

[0066] Figure 5 This is a schematic diagram of the background customized operation process in a simulation example of the power grid robot process automation construction method based on RPA+AI described in the second embodiment of the present invention. Detailed Implementation

[0067] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0068] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0069] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0070] Example 1

[0071] Reference Figures 1-4 This is the first embodiment of the present invention, which provides a method for automating the construction of power grid robotic processes based on RPA+AI, including:

[0072] S1: Use two methods, web page data crawling and file data crawling, to crawl the non-standardized scenario data to be obtained;

[0073] Specifically, common non-standardized data includes: chart data generated from web pages (line charts, bar charts, real-time comparison charts, etc.), on-site power grid equipment image data, nameplate images of power equipment, equipment monitoring information under multiple systems, and specified data in power grid data text, etc., which are business data under unconventional conditions.

[0074] In this embodiment of the application, web page data crawling includes:

[0075] The RPA robot determines the pages required for access in the power grid business scenario and the target elements to be crawled. It then sends a request to the target page through a browser and obtains the HTML document of the accessed page. The RPA robot parses the obtained HTML document, loads and parses the webpage style, obtains the node and style information of the target page elements, and calculates the position and size of the elements to be retrieved or manipulated. Using the webpage style information, it locates the page position of the target data, identifies and confirms the text data to be crawled, locates the element position on the webpage, and completes the data crawling by calling a browser plugin.

[0076] For example, the page to be accessed and the target element to be crawled in the power grid business scenario are determined. The page to be accessed is the Fujian Intelligent Control Cloud Platform (dcloud.fj.sgcc.com.cn) and the target data to be crawled is the daily electricity consumption of the provincial and municipal coal-fired installed capacity. The RPA robot sends a request to the target page through the browser and obtains a response, thereby obtaining the HTML document of the Fujian Intelligent Control Cloud Platform.

[0077] The RPA robot parses the HTML document obtained from the Fujian Intelligent Control Cloud Platform, loads and parses the webpage's CSS styles, obtains the node and style information of the target page elements, and calculates the position and size of the elements to be retrieved or manipulated.

[0078] By using the CSS style information of the webpage, the target data is located on the page. Then, the text data to be crawled is identified and confirmed through OCR. The element position on the webpage is located, and finally, the RPA robot completes the data crawling by calling the browser plugin.

[0079] In this embodiment of the application, file data capture includes:

[0080] File type identification is performed using an RPA robot; different tool libraries are invoked based on the file type to assist in data extraction;

[0081] Specifically, for Excel or Word files, the Apache POI or OpenXML utility classes are used; for PDF files, libraries such as PDFBox and iText are used; and for non-text formats such as images, OCR recognition technology is employed.

[0082] For example, it is necessary to filter the downloaded power generation plan document to obtain pumped storage type units and sum the data from 0:15 to 0:00 the next day. At this time, the RPA robot determines that the file type is an Excel file, calls the utility class, obtains the data in the Excel file, filters and sums the target data, and exports the summed data to generate a new document.

[0083] S2: Based on the captured non-standardized scene data, OCR file detection technology is used to identify data types and determine business models based on pre-built power grid text data detection rules;

[0084] In this embodiment of the application, identifying data types based on pre-built power grid text data detection rules includes:

[0085] Obtain the image and text files of the power grid services executed by RPA;

[0086] The image text file is decoded and characters are extracted using OCR recognition technology.

[0087] Based on the decoded file content and extracted character sequences, a language model is applied for text recognition.

[0088] Data extraction is performed on the identified content.

[0089] Collect power grid business text data, including power grid specification documents, power grid rules and regulations, and power grid business data;

[0090] Based on actual power grid business application scenarios, texts containing interference information, sensitive information, and key information features are artificially generated to form a text dataset for algorithm training.

[0091] The text dataset is preprocessed to remove noise and convert the text data into a uniform format.

[0092] Based on the specific needs of power grid operations, the preprocessed text dataset is divided into different categories and classification features are extracted.

[0093] The extracted classification features are manually or semi-automatically labeled to distinguish normal text from text containing inappropriate, sensitive, or critical information, and then labeled accordingly.

[0094] Based on the extracted features and labels, a text detection model is constructed;

[0095] Based on the text detection model, key discrimination rules or feature combinations are extracted to form a specific rule set for filtering interference information, detecting sensitive information, and extracting key information.

[0096] The rule set is integrated into the RPA system to detect newly generated text data in real time; at the same time, the system's performance is continuously monitored, and the rules are adjusted or the text detection model is retrained according to the actual situation.

[0097] For example, collecting power grid business text data mainly includes power grid specification documents (such as safety operation manuals, maintenance guidelines, technical standards, etc.), power grid rules and regulations (laws and regulations, internal company policies, etc.), and power grid business data (daily reports, weekly reports, accident contingency plans, etc.).

[0098] Based on actual power grid business application scenarios, some texts containing features such as interference information, sensitive information (which need to be manually checked or filled in), and key information are artificially generated for algorithm training.

[0099] The text dataset is preprocessed to remove noise from the text, such as extra spaces, non-text characters, repeated sentences or paragraphs, and the text data is converted into a uniform format.

[0100] Based on the specific needs of power grid operations, the preprocessed text data is divided into different categories, such as technical documents, policy documents, safety guidelines, and daily report data.

[0101] Classification features, such as keywords (wind power, hydropower, thermal power, and other power grid business-related terms), phrases, sentence structure, and semantic features, are extracted from the preprocessed text data.

[0102] S3: Utilize decision tree models to analyze the captured non-standardized scenario data and determined business models, and determine the specific operational steps of the automated process based on the analysis results;

[0103] In this embodiment of the application, based on the captured non-standardized scenario data and the determined business model, the repetitive tasks and data processing requirements in the power grid business are analyzed, and a decision tree model is constructed.

[0104] Based on the decision tree model, anomaly detection or fault prediction is performed on the RPA execution status.

[0105] A data interaction channel is used to exchange business data and status information during the actual execution of the RPA robot with the results predicted and detected by the decision tree model, and the prediction mechanism of the decision tree model and the data processing mechanism of the RPA robot are coordinated.

[0106] Based on the actual execution of business operations, analyze the problems encountered in the practical application of RPA, and collect user feedback through the RPA robot backend to optimize the operation process and decision-making of the RPA robot for the problems existing in the actual business process.

[0107] S4: Based on the specific operation steps of the automated process, a human-computer interaction mechanism is applied. In steps where manual data supplementation or verification is required, proactive interaction is carried out. At the same time, the RPA backend collects task execution status information and user feedback to form a closed-loop feedback and optimize decision-making.

[0108] In this embodiment of the application, the application of human-computer interaction mechanisms in the specific operation steps based on the automated process includes:

[0109] Perceive contextual information during the execution of the RPA robot;

[0110] Based on the context information, the execution content of the preceding and following operation steps is obtained;

[0111] The content is interactively evaluated to determine whether manual intervention is required.

[0112] If interaction with the user is required, the RPA robot will proactively display a pop-up window or issue a voice prompt;

[0113] For special data types, a secondary judgment is added. If special data or security information is detected, the RPA robot will pop up a prompt window again, requiring the dispatcher to confirm the information a second time.

[0114] The above is an illustrative scheme of the power grid robot process automation construction method based on RPA+AI in this embodiment. It should be noted that the technical solution of the power grid robot process automation construction system based on RPA+AI and the technical solution of the power grid robot process automation construction method based on RPA+AI described above belong to the same concept. Details not described in detail in the technical solution of the power grid robot process automation construction system based on RPA+AI in this embodiment can be found in the description of the technical solution of the power grid robot process automation construction method based on RPA+AI described above.

[0115] This embodiment of the power grid robotic process automation construction system based on RPA+AI includes:

[0116] The data crawling module is used to crawl non-standardized scenario data using two methods: web page data crawling and file data crawling.

[0117] The detection and recognition module is used to identify data types and determine business models based on the captured non-standardized scene data, using OCR file detection technology and pre-built power grid text data detection rules.

[0118] The process establishment module is used to analyze the captured non-standardized scenario data and the determined business model using a decision tree model, and determine the specific operation steps of the automated process based on the analysis results.

[0119] The human-computer interaction module is used to apply human-computer interaction mechanisms to specific operation steps based on automated processes. It allows for proactive interaction in steps that require manual data supplementation or verification. At the same time, the RPA backend collects task execution status information and user feedback to form a closed-loop feedback loop for decision optimization.

[0120] This embodiment also provides a computing device applicable to the construction method of power grid robot process automation based on RPA+AI, including:

[0121] The system includes a memory and a processor. The memory stores computer-executable instructions, and the processor executes these instructions to realize the RPA+AI-based power grid robot process automation construction method proposed in the above embodiments.

[0122] This embodiment also provides a storage medium on which a computer program is stored. When the program is executed by a processor, it implements the RPA+AI-based power grid robot process automation construction method proposed in the above embodiment.

[0123] The storage medium proposed in this embodiment belongs to the same inventive concept as the RPA+AI-based power grid robot process automation construction method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0124] Example 2

[0125] Reference Figure 5 As an embodiment of the present invention, a method for automating the construction of power grid robot processes based on RPA+AI is provided. To verify the beneficial effects of the present invention, a simulation experiment is conducted for scientific demonstration.

[0126] Business objective: Generate the required information from the online contingency plan preparation function of the control cloud, and then generate the corresponding list using RPA and input it into the Guofen Cloud fault contingency plan joint preparation system.

[0127] Target webpage: Guofen Cloud Fault Contingency Plan Joint Compilation System

[0128] (http: / / faulthand.dcloud.gf.dc.sgcc.com.cn / pages / homePage / index.html#)

[0129] Target data: Information on the emergency response plan for major maintenance accidents compiled by Fujian Province in the XXth week of 202X.

[0130] Business steps:

[0131] 1. Jointly developed system for contingency plans for Guofen Cloud;

[0132] 2. Click the "New" button to bring up the following pop-up window, which contains various information.

[0133] 3. Select the correct contingency plan, click the corresponding Excel sheet, and retrieve the contingency plan information;

[0134] 4. In the new contingency plan window, fill in the contingency plan information;

[0135] 5. After completing the form, click "Please have the dispatcher confirm that everything is correct" (a pop-up window will appear). Once the dispatcher confirms that everything is correct;

[0136] 6. Automatically click to issue the compilation task.

[0137] Detailed operating procedures:

[0138] (1) Based on the above operation steps, the backend sets up operation steps for the business scenario, such as Figure 5 The diagram shows the customized operation process in the backend.

[0139] (2) Read the information from the national emergency response plan;

[0140] (3) Read the risk analysis content of the risk warning form of the corresponding plan, fill in the content of the power grid mode change (i.e., if XX trips) into the initial operation mode of the power grid, and generate an Excel table.

[0141] (4) Open the RPA robot task interface and execute the national survey plan upload task.

[0142] (5) The RPA robot automatically logs into the Guofen Cloud Fault Contingency Plan Joint Compilation System according to the business process.

[0143] (6) (First Human-Machine Verification) After the RPA robot automatically completes the login, a prompt window pops up for the user to confirm the contingency plan time. After the user confirms, the RPA robot will automatically execute the subsequent contingency plan selection process and call the API interface provided by Guofen Cloud to generate an Excel file containing contingency plan information.

[0144] (7) (Second Human-Machine Verification) The RPA robot automatically opens the Excel file containing the contingency plan information and pops up a prompt window. After the user reviews the information in the table, the user confirms the information in the pop-up window. After the dispatcher confirms the information, the RPA robot will extract the contingency plan information from the Excel table and automatically fill it in.

[0145] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A power grid robot process automation (RPA) + artificial intelligence (AI) based construction method, characterized in that, The application relates to a method for realizing automatic optimization of an RPA (Robotic Process Automation) robot based on a decision tree model. The method comprises the following steps: The non-standardized scene data to be acquired is acquired by using webpage data acquisition and file data acquisition; Based on the acquired non-standardized scene data, the data type is identified by using an OCR file detection technology and based on pre-constructed power grid text data detection rules, and a business mode is determined; The acquired non-standardized scene data and the determined business mode are analyzed by using a decision tree model, and specific operation steps of an automatic process are determined according to an analysis result; A man-machine interaction mechanism is applied based on the specific operation steps of the automatic process, active interaction is carried out in steps requiring manual data supplement or manual data checking, meanwhile, state information of task execution and feedback opinions of users are collected by the RPA background, a closed-loop feedback is formed, and decision optimization is carried out; The method comprises the following steps: Based on the acquired non-standardized scene data and the determined business mode, repetitive tasks and data processing requirements in power grid business are analyzed, and a decision tree model is constructed; Based on the decision tree model, RPA execution states are abnormally detected or fault predicted; Data interaction channels are adopted to interact business data and state information in an actual execution process of the RPA robot with result information predicted and detected by the decision tree model, and the decision tree model prediction mechanism and the data processing mechanism of the RPA robot are cooperated; According to actual business execution conditions, RPA actual application problems are analyzed, feedback opinions of users are collected by the RPA robot background, and operation processes and decisions of the RPA robot are optimized according to problems existing in actual business processes; The method comprises the following steps: Context information in an execution process of the RPA robot is perceived; Based on the context information, execution contents of previous and subsequent operation steps are acquired; The contents are interactively judged to judge whether manual operation steps are required; If interaction with users is required, the RPA robot initiatively pops up a window or issues a voice prompt; Secondary judgment is added to special data types, if special data or safety information is detected, the RPA robot pops up a prompt window again to require dispatch personnel to perform secondary information confirmation; The method comprises the following steps: Power grid business text data, including power grid specification documents, power grid rules and regulations and power grid business data, are collected; According to actual power grid business application scenes, part of texts containing interference information, sensitive information and key information features are artificially generated to form a text data set for algorithm training; The text data set is preprocessed to remove noise in the text, and the text data is converted into a unified format; According to specific requirements of the power grid business, the preprocessed text data set is divided into different categories and classification features are extracted; The extracted classification features are manually or semi-automatically labeled to distinguish normal texts and texts containing bad, sensitive or key information, and corresponding labels are marked. According to the extracted features and labels, a text detection model is constructed; Based on the text detection model, key discriminant rules or feature combinations are extracted to form a specific rule set for interference information filtering, sensitive information detection and key information extraction; The rule set is integrated into the RPA system to detect newly generated text data in real time; at the same time, the performance of the system is continuously monitored, and the rules are adjusted or the text detection model is retrained according to the actual situation.

2. The RPA+AI based power grid robotic process automation building method of claim 1, wherein, The webpage data crawling includes: Determine the pages to be accessed and the target elements to be crawled required by the power grid business scenario; An access request is sent based on the target page using an RPA robot to obtain the target page HTML document; The obtained HTML document is loaded and the webpage style is parsed to obtain the element nodes and style information of the target page, and the element position and size to be obtained or operated are calculated; Based on the calculated element position and size information, the RPA robot calls the browser plug-in to locate the target element and crawl the target element.

3. The RPA+AI based power grid robotic process automation building method of claim 2, wherein, The file data crawling includes: Identify the file type of the power grid business file; According to the identified file type, different tool libraries are called to extract the file data content.

4. The RPA+AI based power grid robotic process automation building method of claim 3, wherein, The OCR file detection technology is used to identify the data type based on the pre-constructed power grid text data detection rules, including: Obtain the image text file of the power grid business executed by the RPA; Apply OCR recognition technology to decode and extract characters from the image text file; Based on the decoded file content and extracted character sequence, apply a language model for text recognition; Extract data from the recognized content.

5. A system employing the RPA+AI based power grid robotic process automation construction method according to any one of claims 1-4, characterized in that, It includes: A data crawling module for crawling non-standardized scenario data to be obtained by using webpage data crawling and file data crawling; A detection and recognition module for identifying the data type based on the pre-constructed power grid text data detection rules using the OCR file detection technology based on the crawled non-standardized scenario data to determine the business mode; A process establishment module for analyzing the crawled non-standardized scenario data and the determined business mode using a decision tree model, and determining the specific operation steps of the automation process according to the analysis result; A human-computer interaction module for actively interacting in steps requiring manual data supplement or manual data verification based on the specific operation steps of the automation process, while the RPA background collects state information and user feedback to form a closed-loop feedback for decision optimization.

6. A computing device, comprising: a memory and a processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions, which realize the steps of the power grid robot process automation construction method based on RPA+AI in any one of claims 1 to 4 when executed by the processor. 7.A computer readable storage medium storing computer executable instructions which, when executed by a processor, implement the steps of the method for constructing a power grid robot process automation based on RPA+AI according to any one of claims 1 to 4.

Images