Power grid financial process automation method based on rpa and blockchain technology

Abstract

The application discloses a power grid financial process automation method and system based on RPA and blockchain technology, belongs to the technical field of power grids, and through real-time collection of original data of interface calls in the RPA execution process, extraction of continuous response time abnormal fluctuation rate and upper chain transaction delay abnormal values, normalization processing, and evaluation of interface abnormal risk by using a weighted comprehensive scoring model, abnormal states are determined according to preset threshold values, and a multi-level abnormal processing mechanism is automatically executed when an abnormality is detected, including automatic retry, link switching, local caching and asynchronous compensation, further judgment of abnormal release effects, compensation of key information that is not successfully chained according to local caching or retry records, so that the real-time performance and completeness of financial data chaining are effectively improved, the risk of data loss caused by node failure is reduced, and the stability and automation level of the system are significantly enhanced.

Description

Technical Field

[0001] This invention relates to the field of power grid technology, and more specifically to a method and system for automating power grid financial processes based on RPA and blockchain technology. Background Technology

[0002] Power grid financial process automation based on RPA (Robotic Process Automation) and blockchain technology refers to using RPA tools to automatically execute various standardized and repetitive financial operation tasks in power grid enterprises, such as expense reimbursement review, account reconciliation, and invoice processing. At the same time, combined with the immutability and full traceability of blockchain, transparent management and trusted sharing of financial data can be achieved, thereby improving work efficiency, reducing human error rate, and enhancing the security and compliance of the power grid financial system.

[0003] The existing technology has the following shortcomings:

[0004] In the automation of power grid financial processes based on RPA and blockchain technologies, interface reliability and anomaly handling are critical issues. Because blockchain nodes or smart contract interfaces may experience brief failures under high load conditions, such as downtime or response timeouts, if the RPA process lacks robust anomaly detection and retry mechanisms, data may fail to be uploaded to the blockchain in a timely manner, leading to information loss or processing failures. For example, in the financial reimbursement process, approval records need to be automatically synchronized to the blockchain. However, if a node downtime goes undetected, some data may not be successfully written, ultimately revealing missing records during subsequent audits, affecting the integrity and traceability of financial data. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for automating power grid financial processes based on RPA and blockchain technology, in order to address the shortcomings of the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for automating power grid financial processes based on RPA and blockchain technology, comprising:

[0007] During RPA execution, relevant raw data from API calls are collected in real time.

[0008] Based on the collected raw data, the abnormal volatility of the continuous response time of the interface is extracted to characterize the stability of the interface link, and the abnormal value of the on-chain transaction delay is used to characterize the efficiency of blockchain packaging and confirmation.

[0009] The extracted abnormal volatility of continuous response time and abnormal values ​​of on-chain transaction delay are normalized, and the interface abnormal risk value is calculated using a weighted comprehensive scoring model.

[0010] Determine the interface exception risk value according to the preset risk threshold. When the interface exception risk value exceeds the threshold, it is determined that the interface has an exception;

[0011] When an interface exception is detected, automatically execute the exception handling mechanism;

[0012] Judge the effect of exception resolution, and compensate for the key information that has not been successfully chained according to the local cache or retry record.

[0013] Preferably, the relevant original data includes interface response time, return status code, transaction hash, write confirmation delay, and node load information.

[0014] Preferably, the extraction method of the continuous response time exception volatility is as follows: set a continuous time window N, collect the interface response time each time, and form a response time series: ; Calculate the average response time within the time window , and the expression is: ; In the formula, is the actual response time of the i-th interface call, calculate the standard deviation of the response time within the time window , and the expression is: ; Calculate the continuous response time exception volatility : .

[0015] Preferably, the extraction method of the chained transaction delay exception value is as follows: set a continuous sampling window, collect the delay time data of the last M chained transactions, and form a delay time series: , where represents the chained confirmation delay time of the i-th transaction. Sort the delay time data D from small to large, and calculate the first quartile Q1: that is, the delay value ranked at the 25% position, and the third quartile Q3, that is, the delay value ranked at the 75% position;

[0016] The interquartile range IQR represents the span of the middle 50% of the data, and the calculation formula is: IQR = Q3 - Q1; Determine the exception judgment boundary, and set the normal range as: Lower Bound = Q1 - 1.5 × IQR; Upper Bound = Q3 + 1.5 × IQR; LowerBound is the lower boundary, and Upper Bound is the upper boundary;

[0017] Judge each delay time : If < Lower Bound or > Upper Boundd, then it is determined that the transaction delay is abnormal, and calculate the average value of the delay times of all transaction delay abnormalities as the chained transaction delay exception value.

[0018] Preferably, the abnormal volatility of continuous response time and the abnormal value of on-chain transaction delay are normalized so that they are both between [0,1], and the abnormal risk value of the interface is calculated based on the abnormal volatility of continuous response time and the abnormal value of on-chain transaction delay after normalization.

[0019] Preferably, based on a preset risk threshold, the interface anomaly risk value is determined. When the interface anomaly risk value exceeds the threshold, it is determined that the interface is abnormal. Specifically, this includes:

[0020] The obtained interface anomaly risk value is compared with the preset risk threshold. If the interface anomaly risk value exceeds the preset risk threshold, the interface is determined to be abnormal; if the interface anomaly risk value does not exceed the preset risk threshold, the interface is normal.

[0021] Preferably, when an interface anomaly is detected, an anomaly handling mechanism is automatically executed. The anomaly handling mechanism includes: automatic retry, backup link switching, local cache delayed asynchronous uploading, or manual intervention notification.

[0022] Preferably, to determine the effectiveness of exception resolution, based on local cache or retry records, key information regarding unsuccessful on-chain data is compensated, specifically as follows:

[0023] Collect key performance metrics data of current API calls in real time, including API response time, status codes, transaction confirmation latency, and node load; summarize the performance data of the most recent K API calls to form a time-series feature set. ;

[0024] Intelligent determination of anomaly resolution status:

[0025] Use the model to calculate anomaly scores for the indicator feature F within the latest window;

[0026] Determine whether the anomaly score of the most recent M consecutive interface calls is lower than the set recovery threshold ϵ;

[0027] If the condition is met M times consecutively, the interface anomaly is determined to be resolved and the system returns to normal.

[0028] After confirming that the anomaly has been resolved, a compensation mechanism is initiated to process the backlog of unlisted data in the local cache or retry log.

[0029] This invention also provides an automated system for power grid financial processes based on RPA and blockchain technology, including a data acquisition module, a feature extraction module, a risk assessment module, an anomaly detection module, an anomaly handling module, and a compensation and data consistency module;

[0030] Data acquisition module: During RPA execution, it collects relevant raw data from interface calls in real time;

[0031] Feature extraction module: Based on the collected raw data, extract the abnormal volatility of the continuous response time of the interface to characterize the stability of the interface link, and extract the abnormal value of the on-chain transaction delay to characterize the efficiency of blockchain packaging and confirmation.

[0032] Risk assessment module: Normalizes the extracted abnormal volatility of continuous response time and abnormal values ​​of on-chain transaction delay, and uses a weighted comprehensive scoring model to calculate the interface abnormal risk value;

[0033] Anomaly detection module: Based on a preset risk threshold, the module determines the anomaly risk value of the interface. When the anomaly risk value of the interface exceeds the threshold, it is determined that the interface is abnormal.

[0034] Exception handling module: Automatically executes exception handling mechanism when an interface exception is detected;

[0035] Compensation and Data Consistency Module: Determines the effectiveness of anomaly resolution and compensates for key information that failed to be uploaded to the blockchain based on local cache or retry records.

[0036] The technical effects and advantages provided by the present invention in the above technical solution are as follows:

[0037] 1. This invention collects key performance data of interface calls in real time during RPA execution, extracts abnormal volatility in continuous response time and outliers in on-chain transaction delays, and employs a normalized and weighted comprehensive scoring model for interface anomaly risk assessment. This enables timely and accurate identification of interface link instability or blockchain packaging delay issues. Compared to existing technologies that rely solely on a single interface response for simple judgment, this invention integrates multi-dimensional features for dynamic monitoring, effectively improving the intelligence and accuracy of interface anomaly detection and reducing the risk of misjudgment and missed judgment.

[0038] 2. This invention employs a multi-layered anomaly handling mechanism, including automatic retries, backup link switching, asynchronous compensation via local caching, and manual intervention notifications, to ensure that critical information is ultimately uploaded to the blockchain even in the event of node failures or network congestion. Simultaneously, by introducing an anomaly resolution determination method based on intelligent performance data analysis, it ensures the timely initiation of compensation operations and the restoration of data integrity. This significantly improves the continuity, stability, and automation capabilities of the power grid financial processes, while reducing manual intervention and system maintenance costs. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings.

[0040] Figure 1 This is a mind map of the method of the present invention.

[0041] Figure 2 This is a mind map of the system modules of the present invention. Detailed Implementation

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0043] Example 1, please refer to Figure 1 As shown in this embodiment, the power grid financial process automation method based on RPA and blockchain technology includes:

[0044] During RPA execution, relevant raw data from API calls are collected in real time.

[0045] Based on the collected raw data, the abnormal volatility of the continuous response time of the interface is extracted to characterize the stability of the interface link, and the abnormal value of the on-chain transaction delay is used to characterize the efficiency of blockchain packaging and confirmation.

[0046] The extracted abnormal volatility of continuous response time and abnormal values ​​of on-chain transaction delay are normalized, and the interface abnormal risk value is calculated using a weighted comprehensive scoring model.

[0047] Based on a preset risk threshold, the abnormal risk value of the interface is determined. When the abnormal risk value of the interface exceeds the threshold, it is determined that the interface is abnormal.

[0048] When an interface anomaly is detected, the exception handling mechanism is automatically executed;

[0049] To determine the effectiveness of the exception resolution, key information about unsuccessful on-chain data is compensated based on local cache or retry records.

[0050] During RPA execution, real-time collection of raw data related to interface calls includes: Interface response time: This refers to the total time elapsed from sending a request to receiving a response when RPA calls a blockchain interface, typically measured in milliseconds (ms). Collection method: Timestamp recording logic is embedded in the RPA task script to record the request sending time (T1) and response receiving time (T2). Interface response time = T2 - T1. This is used to monitor for abnormal latency in interface communication and quickly identify issues such as link congestion or degraded node processing performance.

[0051] Return status code: This refers to the status identifier returned by the blockchain interface, typically including HTTP protocol status codes (such as 200, 400, 500) or smart contract execution statuses (such as Success / Failure, Error Code). Collection method: The status field is directly extracted from the interface's returned message and recorded. This is used to quickly identify whether the request was successfully processed and to pinpoint the type of failure (such as interface exception, permission denied, incorrect parameters, etc.).

[0052] Transaction hash: A unique transaction identifier returned by a blockchain system after successfully receiving a write request. It is typically a fixed-length encrypted string (such as the 64-bit hexadecimal hash value returned by Ethereum). Collection method: The transaction hash field is parsed from the data returned by the API and recorded in the RPA internal log or database. It is used for subsequent tracking of transaction status (confirmed / unconfirmed), locating specific on-chain data, and performing on-chain query verification operations.

[0053] Write confirmation latency: This refers to the time difference between when a data transaction is submitted to a blockchain node and when it is packaged into a new block and confirmed. Collection method: After the transaction hash is returned, RPA periodically queries the on-chain transaction status (e.g., via the eth_getTransactionReceipt interface); records the transaction submission time and the first confirmation completion time, and calculates the time difference between the two. This is used to assess the processing load and transaction congestion of the blockchain network, and to promptly identify slow block generation or node synchronization delays.

[0054] Node load information: refers to the current operating load of blockchain nodes, including system indicators such as CPU utilization, memory utilization, processing queue length, and mempool congestion.

[0055] Data collection methods: Node running status is collected periodically via monitoring interfaces or system APIs (such as the Node Metrics API provided by Hyperledger Fabric); or indirectly through RPA (e.g., node load information is included in the response header, or on-chain load monitoring smart contracts are reserved). This helps determine whether a node's interface response is slow or processing fails due to resource bottlenecks, providing early warnings of performance degradation risks at the system level.

[0056] Based on the collected raw data, abnormal volatility of continuous response time of the interface is extracted to characterize the stability of the interface link, and abnormal values ​​of on-chain transaction delay are extracted to characterize the efficiency of blockchain packaging and confirmation, specifically including:

[0057] The method for extracting abnormal volatility in continuous response time is as follows: A continuous time window is set (e.g., the most recent N API calls), and the response time of each API call is collected to form a response time series. Calculate the average response time within this time window. The expression is: In the formula, Calculate the standard deviation of the response time within the time window for the actual response time (in milliseconds, ms) of the i-th interface call. The expression is: ; Calculate the abnormal volatility of continuous response time : ; Calculated Compare with the set abnormal fluctuation threshold Rthreshold, if >Rthreshold determines that there are abnormal fluctuations in the interface link.

[0058] When the continuous response time exhibits high abnormal fluctuations, it indicates poor stability of the interface link. Specifically, if the interface response time fluctuates drastically within a short period, exhibiting significant uncertainty and discontinuity, it suggests that the link may be affected by factors such as network congestion, abnormal node load, and link jitter. This can easily lead to data transmission delays or request failures during RPA execution, reducing the overall system reliability and real-time performance.

[0059] When the abnormal fluctuation rate of continuous response time is low, it indicates that the interface link is relatively stable. At this time, the interface response time changes smoothly with small fluctuations, indicating that the link communication environment is stable, the node processing capacity is sufficient, and it can continuously provide consistent and fast response services, thereby ensuring the efficient execution of RPA processes and the timely on-chaining of financial data, and improving the overall business continuity and reliability of the system.

[0060] The method for extracting outliers from on-chain transaction delays is as follows:

[0061] Set a continuous sampling window to collect the latency data of the most recent M on-chain transactions, forming a latency time series: ,in, This represents the on-chain confirmation delay time for the i-th transaction. Calculate the first quartile Q1 and the third quartile Q3, and sort the delay time data D from smallest to largest. Calculate the first quartile Q1: the delay value at the 25th percentile, and the third quartile Q3: the delay value at the 75th percentile.

[0062] The interquartile range (IQR) represents the span of the middle 50% of the data, and is calculated as: IQR = Q3 − Q1; The outlier detection boundaries are determined, and the normal range is set as: Lower Bound = Q1 − 1.5 × IQR; Upper Bound = Q3 + 1.5 × IQR; Lower Bound is the lower boundary, and Upper Bound is the upper boundary.

[0063] For each delay time make a judgment: If <Lower Bound or >Upper Boundd, it is determined that the delay of this transaction is abnormal. Calculate the average value of the delay times of all transactions with abnormal delays as the abnormal value of the on-chain transaction delay.

[0064] When the abnormal value of the on-chain transaction delay is large, it indicates that the packaging and confirmation efficiency of the blockchain is low. Specifically, the time required for a transaction to be submitted and confirmed significantly exceeds the normal range, which may be due to insufficient processing capacity of blockchain nodes, network congestion, consensus mechanism delay, or a decrease in the block generation speed, reflecting a significant degradation in the transaction processing performance of the blockchain system at the current stage, and may thus affect the real-time nature and reliability of the upper-layer business processes.

[0065] When the abnormal value of the on-chain transaction delay is small, it indicates that the packaging and confirmation efficiency of the blockchain is high. At this time, transactions can be packaged and confirmed in a relatively short time, indicating that the blockchain nodes are operating normally, the network transmission is unobstructed, the block generation speed is stable, the overall transaction processing process is efficient, and it can effectively support the requirements of the RPA system for timely on-chain of financial data, ensuring the continuity and data consistency of the power grid financial automation process.

[0066] Normalize the extracted abnormal volatility of the continuous response time and the abnormal value of the on-chain transaction delay, and use a weighted comprehensive scoring model to calculate the interface abnormal risk value, specifically including:

[0067] Normalize the abnormal volatility of the continuous response time and the abnormal value of the on-chain transaction delay so that they are both within [0,1], and calculate the interface abnormal risk value based on the normalized abnormal volatility of the continuous response time and the abnormal value of the on-chain transaction delay.

[0068] For example, the present invention can calculate the interface abnormal risk value using the following formula, and the calculation expression is: ; where is the interface abnormal risk value, is the abnormal volatility of the continuous response time, is the abnormal value of the on-chain transaction delay, is the weight coefficient of the abnormal volatility of the continuous response time and the abnormal value of the on-chain transaction delay (which can be optimized according to experimental experience or machine learning), and are both greater than 0.

[0069] Judge the interface abnormal risk value according to the preset risk threshold. When the interface abnormal risk value exceeds the threshold, it is determined that the interface is abnormal;

[0070] The obtained interface anomaly risk value is compared with the preset risk threshold. If the interface anomaly risk value exceeds the preset risk threshold, the interface is determined to be abnormal; if the interface anomaly risk value does not exceed the preset risk threshold, the interface is normal.

[0071] When the interface anomaly risk value exceeds a preset threshold, the system automatically triggers an anomaly handling mechanism, which specifically includes the following strategies:

[0072] Automatic retry mechanism trigger conditions: Retry is initiated immediately after an interface call fails, a response times out, or an error status code is returned.

[0073] Implementation: Set the maximum number of retries Nretry and the retry interval Tinterval.

[0074] The exponential backoff algorithm is used to optimize the retry interval and prevent frequent resource consumption.

[0075] During retries, the original data request is resent to maintain idempotency of the request (avoiding duplicate uploads to the chain).

[0076] This addresses transient anomalies caused by short-term network jitter or temporary node overload, improving request success rates.

[0077] Backup link switching mechanism trigger conditions: After multiple consecutive failed retries, the primary link (primary node) is deemed unavailable, and backup switching is initiated.

[0078] Implementation: Multiple blockchain nodes are pre-configured (primary + backup). The system selects and switches to the optimal backup node based on node availability checks (e.g., heartbeat detection, health check API results). Connection parameters are automatically updated during the switchover process without interrupting the main RPA workflow.

[0079] Function: When the master node is unavailable, it can quickly restore the ability to write data, ensure that data is uploaded to the chain in a timely manner, and improve the fault tolerance and high availability of the system.

[0080] The local cache delayed asynchronous on-chain mechanism is triggered when the node is unavailable and the backup node is also unavailable, or when the interface is abnormal for a period of time exceeding the set threshold.

[0081] Implementation: Data that failed to be successfully uploaded to the blockchain (such as transaction records, contract vouchers, and approval records) is temporarily cached in a local database or message queue. A background asynchronous task continuously monitors the node status and automatically uploads all cached data to the blockchain in batches once the node recovers. Each cached data entry is accompanied by a timestamp and a unique identifier to ensure that the uploaded data is not lost or duplicated.

[0082] Function: To prevent data loss due to link unavailability, ensure uninterrupted process, and guarantee eventual data consistency and integrity.

[0083] The manual intervention notification mechanism is triggered when automatic retry, link switching, and local cache compensation all fail or when a serious anomaly occurs (such as the loss of multiple consecutive batches of data).

[0084] Implementation: Anomaly reports are automatically generated, including the anomaly type, occurrence time, list of affected data, and processing logs. These reports are pushed to system maintenance personnel or financial managers via email, SMS, and system notifications. Manual triggering of data retries, node switching, or direct on-chain operations is supported.

[0085] Function: When the system's automatic mechanisms are unable to resolve complex anomalies, timely manual intervention is introduced to prevent long-term data loss or system paralysis.

[0086] To determine the effectiveness of the exception resolution, based on local cache or retry records, compensate for key information regarding unsuccessful on-chain data, specifically:

[0087] Step 1: Collect key performance metrics data for the current API call in real time, including API response time, status code, transaction confirmation latency, node load, etc.; summarize the performance data over a recent period (e.g., the last K API calls) to form a time-series feature set. .

[0088] Step 2: Intelligently determine the anomaly resolution status, for example, by using the IsolationForest anomaly detection model based on a sliding window or the dynamic Z-Score method based on an adaptive threshold.

[0089] The specific steps are as follows:

[0090] Use the model to calculate anomaly scores for the indicator feature F within the latest window;

[0091] Determine whether the anomaly score of the most recent M consecutive interface calls is lower than the set recovery threshold ϵ;

[0092] If the condition is met M times consecutively, the interface anomaly is determined to be resolved and the system returns to normal.

[0093] Step 3: After confirming that the exception has been resolved, initiate the compensation mechanism to process the backlog of unlisted data in the local cache or retry record.

[0094] The compensation steps specifically include: retrieving all key information records marked as incomplete on-chain processing from the local cache database or message queue; sorting the data according to the original submission timestamp to ensure that the data is processed on the chain in the order of business operations; re-initiating the on-chain operation for each record and calling the blockchain interface to submit the compensation data; confirming each piece of compensation data to ensure successful writing and updating its status to complete on-chain processing.

[0095] Step 4: After all compensation is completed, perform a system-level data consistency check, including: verifying the consistency between cached data and on-chain data (e.g., transaction hash comparison); verifying the integrity of business system data (e.g., whether all reimbursement records and fund flows are closed-loop). If the check passes, clear the cache; if an anomaly is found, proceed to manual review or a second compensation process.

[0096] This embodiment provides a method for automating power grid financial processes based on RPA and blockchain technology. By collecting key performance data of interface calls in real time during RPA execution, it extracts abnormal fluctuations in continuous response time and abnormal values ​​of on-chain transaction delays, and employs a normalized and weighted comprehensive scoring model to accurately assess interface anomaly risks. Upon detecting an interface anomaly, the invention designs a multi-layered anomaly handling mechanism, including automatic retries, backup link switching, asynchronous on-chain storage with local cache delays, and manual intervention notifications. This ensures that critical information is intelligently compensated and uploaded to the blockchain after the anomaly is resolved through advanced algorithms. This method effectively improves the stability and reliability of data processing in the power grid financial system, reduces the risk of data loss due to interface anomalies, significantly enhances the level of automation, and reduces manual intervention and maintenance costs.

[0097] Example 2, please refer to Figure 2 As shown in the figure, the power grid financial process automation system based on RPA and blockchain technology described in this embodiment includes a data acquisition module, a feature extraction module, a risk assessment module, an anomaly detection module, an anomaly handling module, and a compensation and data consistency module.

[0098] Data acquisition module: During RPA execution, it collects relevant raw data from interface calls in real time;

[0099] Feature extraction module: Based on the collected raw data, extract the abnormal volatility of the continuous response time of the interface to characterize the stability of the interface link, and extract the abnormal value of the on-chain transaction delay to characterize the efficiency of blockchain packaging and confirmation.

[0100] Risk assessment module: Normalizes the extracted abnormal volatility of continuous response time and abnormal values ​​of on-chain transaction delay, and uses a weighted comprehensive scoring model to calculate the interface abnormal risk value;

[0101] Anomaly detection module: Based on a preset risk threshold, the module determines the anomaly risk value of the interface. When the anomaly risk value of the interface exceeds the threshold, it is determined that the interface is abnormal.

[0102] Exception handling module: Automatically executes exception handling mechanism when an interface exception is detected;

[0103] Compensation and Data Consistency Module: Determines the effectiveness of anomaly resolution and compensates for key information that failed to be uploaded to the blockchain based on local cache or retry records.

[0104] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0105] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0106] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0107] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A method for automating power grid financial processes based on RPA and blockchain technology, characterized by: include: During RPA execution, relevant raw data from API calls are collected in real time. Based on the collected raw data, the abnormal volatility of the continuous response time of the interface is extracted to characterize the stability of the interface link, and the abnormal value of the on-chain transaction delay is used to characterize the efficiency of blockchain packaging and confirmation. The extracted abnormal volatility of continuous response time and abnormal values ​​of on-chain transaction delay are normalized, and the interface abnormal risk value is calculated using a weighted comprehensive scoring model. Based on a preset risk threshold, the abnormal risk value of the interface is determined. When the abnormal risk value of the interface exceeds the threshold, it is determined that the interface is abnormal. When an interface anomaly is detected, the exception handling mechanism is automatically executed; To determine the effectiveness of the exception resolution, key information about unsuccessful on-chain data is compensated based on local cache or retry records.

2. The method for automating power grid financial processes based on RPA and blockchain technology according to claim 1, characterized in that: The relevant raw data includes interface response time, return status code, transaction hash, write confirmation delay, and node load information.

3. The method for automating power grid financial processes based on RPA and blockchain technology according to claim 1, characterized in that: The method for extracting abnormal volatility in continuous response time is as follows: A continuous time window N is set, and the response time of each interface is collected to form a response time series. ;Calculate the average response time within the time window The expression is: In the formula, Calculate the standard deviation of the response time within the time window, given the actual response time of the i-th interface call. The expression is: ; Calculate the abnormal volatility of continuous response time : .

4. The method for automating power grid financial processes based on RPA and blockchain technology according to claim 3, characterized in that: The method for extracting outliers in on-chain transaction delays is as follows: A continuous sampling window is set to collect the delay time data of the most recent M on-chain transactions, forming a delay time series. ,in, Let D represent the on-chain confirmation delay time of the i-th transaction. Sort the delay time data D from smallest to largest and calculate the first quartile Q1: the delay value at the 25th percentile, and the third quartile Q3: the delay value at the 75th percentile. The interquartile range (IQR) represents the span of the middle 50% of the data, calculated as: IQR = Q3 − Q1; The outlier detection boundaries are determined, and the normal range is set as: Lower Bound = Q1 − 1.5 × IQR; Upper Bound = Q3 + 1.5 × IQR; Lower Bound is the lower boundary, and Upper Bound is the upper boundary. For each delay time make a judgment: If <Lower Bound or >Upper Boundd, it is determined that the transaction delay is abnormal, and the average value of the delay times of all abnormally delayed transactions is calculated as the on-chain transaction delay anomaly value.

5. The method for automating power grid financial processes based on RPA and blockchain technology according to claim 4, characterized in that: The abnormal volatility of continuous response time and the abnormal value of on-chain transaction delay are normalized so that they are both between [0,1]. The abnormal risk value of the interface is calculated based on the abnormal volatility of continuous response time and the abnormal value of on-chain transaction delay after normalization.

6. The method for automating power grid financial processes based on RPA and blockchain technology according to claim 5, characterized in that: Based on a preset risk threshold, the interface anomaly risk value is determined. When the interface anomaly risk value exceeds the threshold, it is determined that the interface is abnormal, specifically including: The obtained interface anomaly risk value is compared with the preset risk threshold. If the interface anomaly risk value exceeds the preset risk threshold, the interface is determined to be abnormal; if the interface anomaly risk value does not exceed the preset risk threshold, the interface is normal.

7. The method for automating power grid financial processes based on RPA and blockchain technology according to claim 6, characterized in that: When an interface anomaly is detected, an anomaly handling mechanism is automatically executed. The anomaly handling mechanism includes: automatic retry, backup link switching, local cache delayed asynchronous uploading, or manual intervention notification.

8. The method for automating power grid financial processes based on RPA and blockchain technology according to claim 7, characterized in that: To determine the effectiveness of the exception resolution, based on local cache or retry records, compensate for key information regarding unsuccessful on-chain data, specifically: Collect key performance metrics data of current API calls in real time, including API response time, status codes, transaction confirmation latency, and node load; summarize the performance data of the most recent K API calls to form a time-series feature set. ; Intelligent determination of anomaly resolution status: Use the model to calculate anomaly scores for the indicator feature F within the latest window; Determine whether the anomaly score of the most recent M consecutive interface calls is lower than the set recovery threshold ϵ; If the condition is met M times consecutively, the interface anomaly is determined to be resolved and the system returns to normal. After confirming that the anomaly has been resolved, a compensation mechanism is initiated to process the backlog of unlisted data in the local cache or retry log.

9. A power grid financial process automation system based on RPA and blockchain technology, used to implement the power grid financial process automation method based on RPA and blockchain technology as described in any one of claims 1-8, characterized in that: It includes a data acquisition module, a feature extraction module, a risk assessment module, an anomaly detection module, an anomaly handling module, and a compensation and data consistency module; Data acquisition module: During RPA execution, it collects relevant raw data from interface calls in real time; Feature extraction module: Based on the collected raw data, extract the abnormal volatility of the continuous response time of the interface to characterize the stability of the interface link, and extract the abnormal value of the on-chain transaction delay to characterize the efficiency of blockchain packaging and confirmation. Risk assessment module: Normalizes the extracted abnormal volatility of continuous response time and abnormal values ​​of on-chain transaction delay, and uses a weighted comprehensive scoring model to calculate the interface abnormal risk value; Anomaly detection module: Based on a preset risk threshold, the module determines the anomaly risk value of the interface. When the anomaly risk value of the interface exceeds the threshold, it is determined that the interface is abnormal. Exception handling module: Automatically executes exception handling mechanism when an interface exception is detected; Compensation and Data Consistency Module: Determines the effectiveness of anomaly resolution and compensates for key information that failed to be uploaded to the blockchain based on local cache or retry records.

Images