Artificial Intelligence-Based RDA Quality Assessment Method and System

Abstract

This invention discloses an AI-based RDA quality evaluation method and system, relating to the fields of artificial intelligence and fintech. The method includes: connecting to the charging station data chain to retrieve multi-branch data on equipment operation, maintenance, and fund flow; evaluating the data integrity of each branch to obtain the completeness of each data; performing correlation matching evaluation on the multi-branch data based on the data integrity to obtain the matching degree of operation-funds, operation-maintenance, and maintenance-funds; and finally, conducting a comprehensive quality evaluation based on the above six indicators to obtain the RDA data quality evaluation result. It includes corresponding modules for data acquisition, integrity evaluation, correlation matching evaluation, and comprehensive evaluation. This invention, by combining integrity evaluation with correlation verification, constructs a multi-dimensional automated quality evaluation system, effectively solving the problem of underlying data quality assessment in RDA, and providing key technical support for the trusted encapsulation and value assessment of data assets.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and financial technology, specifically to an RDA quality evaluation method and system based on artificial intelligence. Background Technology

[0002] Registered Digital Assets (RDA) are a new type of financial tool that empowers the digital transformation of the real economy. Their core lies in encapsulating the operational data of physical assets using trusted technologies to create data assets that can be valued and traded. In the new energy vehicle industry, constructing RDAs from multiple data points related to charging station equipment operation, maintenance, and cash flow has become an important way to revitalize assets and attract investment.

[0003] However, the value realization of RDA is highly dependent on the quality of the underlying data assets. Currently, the following technical bottlenecks exist in the construction of charging station RDA: First, data integrity is difficult to guarantee. The charging station data chain involves multiple independent systems such as equipment, operation and maintenance, and funds. Data collection is easily lost due to network interruptions, equipment failures, or human oversights, leading to data fragmentation. Second, the reliability of data associations is insufficient. Existing technologies lack effective means to verify the logical consistency between multi-branch data. For example, whether equipment runtime matches power generation revenue, or whether operation and maintenance records correspond to equipment fault states. The lack or misalignment of such associations directly undermines the value foundation of RDA. Finally, there is a lack of a quantitative quality evaluation system. Current assessments of RDA data quality rely heavily on manual sampling and simple rule-based judgments, which suffer from strong subjectivity, low efficiency, and difficulty in comprehensive coverage, failing to meet the rigid demand of the financial sector for efficient, objective, automated, and reliable authentication of data assets. Summary of the Invention

[0004] This invention addresses the technical problems in existing technologies, such as difficulty in ensuring the integrity of underlying RDA data, insufficient verification of the credibility of correlations between multi-source data, and the lack of an automated quantitative evaluation system. It provides an artificial intelligence-based RDA quality evaluation method and system.

[0005] The technical solution of the present invention to solve the above-mentioned technical problems is as follows:

[0006] In a first aspect, the present invention provides an artificial intelligence-based method for evaluating the quality of RDA, comprising:

[0007] Connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data, and capital flow data;

[0008] The equipment operation data, equipment maintenance data, and capital flow data are evaluated for data integrity to obtain the integrity of operation data, maintenance data, and capital data.

[0009] Based on the completeness of the operational data, the completeness of the maintenance data, and the completeness of the financial data, the data association and matching evaluation of the equipment operation data, equipment maintenance data, and financial flow data are performed to obtain the operation-finance matching degree, operation-maintenance matching degree, and maintenance-finance matching degree.

[0010] The data quality of the charging station data chain is evaluated based on the completeness of the operational data, the completeness of the maintenance data, the completeness of the financial data, the matching degree between operation and funds, the matching degree between operation and maintenance, and the matching degree between maintenance and funds, to obtain the RDA data quality evaluation result.

[0011] Secondly, the present invention provides an RDA quality evaluation system based on artificial intelligence, comprising:

[0012] The data acquisition module is used to connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data and capital flow data.

[0013] The data integrity evaluation module is used to evaluate the data integrity of the equipment operation data, equipment maintenance data, and fund flow data respectively, and obtain the integrity of the operation data, maintenance data, and fund data.

[0014] The data association matching evaluation module is used to perform data association matching evaluation on the equipment operation data, equipment maintenance data and fund flow data based on the completeness of the operation data, the completeness of the maintenance data and the completeness of the fund data, and to obtain the operation-fund matching degree, operation-maintenance matching degree and maintenance-fund matching degree;

[0015] The data quality comprehensive evaluation module is used to evaluate the data quality of the charging station data chain based on the completeness of the operational data, the completeness of the maintenance data, the completeness of the financial data, the matching degree between operation and funds, the matching degree between operation and maintenance, and the matching degree between maintenance and funds, and obtain the RDA data quality evaluation result.

[0016] The beneficial effects of this invention are:

[0017] First, by independently evaluating the integrity of multi-branch data, the system achieves precise quantification of data quality across various dimensions of equipment operation, maintenance management, and fund flow, effectively solving the problem of incomplete data due to data fragmentation. Second, it innovatively introduces a data association matching verification mechanism based on integrity, which can intelligently identify and evaluate the logical consistency and credibility between different data chains, overcoming the technical challenge of insufficient multi-source data association verification. Third, by employing a multi-level, multi-indicator comprehensive evaluation model, it organically combines integrity and correlation measurements, constructing a comprehensive and objective automated evaluation system for RDA data quality. Finally, this invention provides a reliable data trust authentication tool for the issuance, trading, and financial supervision of charging station RDA, significantly improving the transparency and value stability of data assets, and has good application prospects and promotional value. Attached Figure Description

[0018] Figure 1 A flowchart illustrating the AI-based RDA quality evaluation method provided by this invention;

[0019] Figure 2 A schematic diagram of the structure of the AI-based RDA quality evaluation system provided by this invention.

[0020] In the attached diagram, the components represented by each label are as follows:

[0021] Data acquisition module 11, data integrity evaluation module 12, data association and matching evaluation module 13, and data quality comprehensive evaluation module 14. Detailed Implementation

[0022] 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, and 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.

[0023] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0024] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0025] Example 1, as Figure 1 As shown, this embodiment of the invention provides an artificial intelligence-based RDA quality evaluation method, including:

[0026] S10: Connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data, and capital flow data;

[0027] Specifically, the charging station data chain is connected, and multi-branch data is retrieved. This multi-branch data includes equipment operation data, equipment maintenance data, and fund flow data, including:

[0028] Receive the collection time window input from the user terminal;

[0029] Connect the charging station data chain and determine the equipment operation storage block, equipment maintenance storage block, and fund flow storage block in the charging station data chain;

[0030] Based on the collection time window, data is collected from the equipment operation storage block, the equipment maintenance storage block, and the fund flow storage block to obtain equipment operation data, equipment maintenance data, and fund flow data, which are used as the multi-branch data.

[0031] This step is the data input and preparation phase. Its core purpose is to establish a secure connection with the source storing the charging station's operational data and extract the structured data required for subsequent quality evaluation. Specifically, the charging station data chain refers to a distributed data storage system that records comprehensive operational data of the charging station, composed of IoT devices, databases, blockchain nodes, etc.; multi-branch data reflects the diversity of data sources and the differences in dimensions, which is the basis for multi-dimensional quality evaluation.

[0032] First, the system receives the data collection time window input from the user. This time window is a specific time interval parameter set according to the user's evaluation needs, such as one day, one week, one month, or one year. By defining the time range for data extraction, the system ensures the relevance and timeliness of the evaluation. This time window sets a unified time benchmark for all subsequent data collection and evaluation, guaranteeing the comparability and consistency of data collected from multiple independent data sources across the time dimension.

[0033] Secondly, the charging station data chain is connected, and within this chain, three storage blocks are identified: equipment operation storage, equipment maintenance storage, and fund flow storage. Specifically, a trusted connection is established between the charging station data chain and the target charging station's data infrastructure. Based on a pre-defined data map or metadata configuration, the physical or logical storage locations of these three types of data are precisely located within the complex data chain. These include: the equipment operation storage block, which stores the real-time status of charging piles, such as power, voltage, current, charging process records, and fault alarms in a database or table; the equipment maintenance storage block, which stores maintenance work orders, inspection records, spare parts replacements, and manual dispatch data in the maintenance management system database; and the fund flow storage block, which stores charging transaction orders, payment records, handling fees, and revenue sharing records in the financial system or payment platform database.

[0034] Furthermore, based on the collection time window, data is collected separately in the equipment operation storage block, equipment maintenance storage block, and fund flow storage block to obtain equipment operation data, equipment maintenance data, and fund flow data as multi-branch data. This realizes the transformation from scattered and heterogeneous data sources to structured datasets to be evaluated, which facilitates the provision of raw data input for subsequent integrity and correlation evaluations.

[0035] S20: Perform data integrity evaluation on the equipment operation data, equipment maintenance data, and capital flow data respectively, and obtain the integrity of operation data, maintenance data, and capital data;

[0036] Specifically, data integrity evaluations are performed on the equipment operation data, equipment maintenance data, and fund flow data respectively to obtain the integrity of operation data, maintenance data, and fund data, including:

[0037] Activate the data integrity evaluator, which includes an operational data evaluation unit, an operation and maintenance data evaluation unit, and a financial data evaluation unit;

[0038] The completeness of equipment operation data, equipment maintenance data, and capital flow data are evaluated by the operation data evaluation unit, maintenance data, and capital flow data, respectively, to obtain the completeness of operation data, maintenance data, and capital data.

[0039] Data integrity evaluation is performed on the raw multi-branch data extracted from the data chain. Integrity evaluation refers to the process of systematically detecting and quantifying the existence of quality defects such as missing values, outliers, format errors, and logical conflicts in the dataset. It is used for quantitative diagnosis and accurate assessment of the completeness and reliability of various types of data.

[0040] First, activate the data integrity evaluator, which refers to a pre-built and trained dedicated artificial intelligence evaluation model or set of models, encapsulating complex logic and algorithms for data integrity evaluation, and including a running data evaluation unit, an operational data evaluation unit, and a financial data evaluation unit.

[0041] Specifically, the construction steps of the data integrity evaluator include:

[0042] Collect historical charging station data, which includes historical equipment operation data, historical equipment maintenance data, and historical fund flow data. Divide the historical equipment operation data, historical equipment maintenance data, and historical fund flow data into events to construct sample equipment operation event datasets, sample equipment maintenance event datasets, and sample fund flow event datasets.

[0043] Data integrity is labeled for the sample equipment operation event dataset, sample equipment maintenance event dataset, and sample fund flow event dataset respectively, and sample operation event data integrity sets, sample maintenance event data integrity sets, and sample fund event data integrity sets are constructed.

[0044] Set up a first deep neural network, a second deep neural network, and a third deep neural network;

[0045] The first deep neural network is trained based on the sample device operation event dataset and the sample operation event data completeness set to generate an operation data evaluation unit;

[0046] The second deep neural network is trained based on the sample equipment operation and maintenance event dataset and the sample operation and maintenance event data completeness set to generate an operation and maintenance data evaluation unit.

[0047] The third deep neural network is trained based on the sample fund flow event dataset and the sample fund event data completeness set to generate a fund data evaluation unit;

[0048] The operational data evaluation unit, the maintenance data evaluation unit, and the financial data evaluation unit are combined to form the data integrity evaluator.

[0049] Specifically, the construction of the data integrity evaluator is based on a supervised learning model training and ensemble process. First, historical charging station data is collected, including historical equipment operation data, historical equipment maintenance data, and historical fund flow data. Each type of historical data is then segmented into events, such as by business event units like charging sessions, maintenance work orders, and transaction orders, constructing sample equipment operation event datasets, sample equipment maintenance event datasets, and sample fund flow event datasets with clear business semantics.

[0050] Secondly, the data integrity of the divided sample equipment operation event dataset, sample equipment maintenance event dataset, and sample fund flow event dataset is labeled, such as the data missing rate, outlier ratio, and logical consistency score of each event, to form the corresponding sample operation event data integrity set, sample maintenance event data integrity set, and sample fund event data integrity set, which serve as the ground truth labels for model training.

[0051] Furthermore, three structurally independent but functionally similar deep neural networks, such as LSTM networks and convolutional neural networks, are set up, denoted as the first deep neural network for running data; the second deep neural network for operation and maintenance data; and the third deep neural network for financial data.

[0052] Specifically, due to the significant temporal correlation and dynamic evolution characteristics of multi-branch data from charging stations, the current data state directly affects the quality of subsequent data. Furthermore, the data generation pattern dynamically changes with factors such as equipment aging and operational strategy adjustments, and may exhibit periodic patterns and long-term trend dependencies. For example, the LSTM (Long Short-Term Memory) network model is chosen as the core architecture for data integrity assessment. The LSTM network is a special type of recurrent neural network specifically designed for processing temporal data. It effectively captures long-term dependencies in the data. Its core achieves information flow control through memory cells and gating mechanisms: memory cells act as continuous information transmission channels, storing long-term patterns of data integrity features; during information transmission, gating units selectively regulate the flow: the forget gate determines which historical data quality-irrelevant information to discard; the input gate determines which new data features need to be stored in the memory cells; and the output gate controls the feature output of the current state.

[0053] Dedicated networks were constructed and trained for each of the three types of data, taking into account their characteristics:

[0054] First, a dataset of historical device operation events, such as time-series data divided by charging sessions, is input into a first deep neural network. The input layer of this network receives the sequence of operational features, including time-series parameters such as voltage, current, and power. The LSTM layer learns the continuity and abrupt change patterns of the device's operating state, such as steady-state characteristics during normal operation and abnormal patterns during faults. The fully connected layer integrates the time-series features, and the output layer outputs the completeness of the operational data through a sigmoid activation function. Using the labeled completeness of operational events as a label, the network parameters are optimized through backpropagation, focusing on capturing the correlation between missing sensor data and device state.

[0055] Secondly, the historical equipment maintenance event dataset, such as sequence data divided by maintenance work orders, is input into a second deep neural network. The input layer of this network receives maintenance feature sequences, including work order type, processing time, and spare parts replacement records. The LSTM layer learns the periodic and sudden patterns of maintenance activities, such as the differences between regular maintenance and emergency repairs. The fully connected layer extracts key maintenance logic features, and the output layer generates maintenance data completeness. During training, the maintenance event completeness label serves as a supervisory signal to optimize the network's ability to assess the completeness of work order records and process compliance.

[0056] Next, the historical fund flow event dataset, such as time-series data divided by transaction orders, is input into a third deep neural network. The input layer of this network receives fund feature sequences, including transaction amounts, payment channels, and settlement times. The LSTM layer learns the continuity and consistency of fund flows, such as the matching pattern between revenue and orders. The fully connected layer integrates financial time-series features, and the output layer generates the fund data completeness. During training, the accuracy of the network's judgment on transaction flow continuity and account balance is optimized using fund event completeness labels as the target.

[0057] The aforementioned network training all employs the mean squared error loss function. It learns the integrity characteristics of specific data patterns through the training set and monitors generalization performance using the validation set. Iterative optimization continues until the model converges, achieving a validation set accuracy of over 90%. This results in the formation of operational data evaluation units, maintenance data evaluation units, and financial data evaluation units, each used for different data types. Furthermore, these units are combined to form a data integrity evaluator.

[0058] Furthermore, through the aforementioned operational data evaluation unit, maintenance data evaluation unit, and financial data evaluation unit, the completeness of equipment operational data, equipment maintenance data, and financial flow data are evaluated respectively to obtain the completeness of operational data, maintenance data, and financial data, including:

[0059] Event identification is performed on the equipment operation data, equipment maintenance data, and fund flow data respectively to extract multiple equipment operation event data, multiple equipment maintenance event data, and multiple fund flow event data;

[0060] The multiple device operation event data are sequentially input into the operation data evaluation unit to obtain the completeness of multiple operation event data, and the average of the multiple operation event data completeness is calculated to obtain the operation data completeness.

[0061] The multiple device operation and maintenance event data are sequentially input into the operation and maintenance data evaluation unit to obtain the completeness of multiple operation and maintenance event data, and the average of the multiple operation and maintenance event data completeness is calculated to obtain the operation and maintenance data completeness.

[0062] The data of the multiple fund flow events are sequentially input into the fund data evaluation unit to obtain the completeness of the multiple fund event data. The mean of the completeness of the multiple fund event data is then calculated to obtain the fund data completeness.

[0063] First, event identification is performed on equipment operation data, equipment maintenance data, and fund flow data. Continuous, raw data is segmented into independent, semantically complete minimum business units—i.e., events—based on their business meaning. Specifically, event identification is performed on equipment operation data to extract multiple equipment operation event data, such as segmenting the data stream into independent charging session events, each event containing all data from the start to the end of a charging session; event identification is performed on equipment maintenance data to extract multiple equipment maintenance event data, such as segmenting the data into independent maintenance work order events or inspection record events; and event identification is performed on fund flow data to extract multiple fund flow event data, such as segmenting the data into transaction events corresponding to each charging session or each order.

[0064] Secondly, a dedicated evaluation unit is used to score the completeness of each individual event. Specifically, the extracted data from multiple device operation events are sequentially input into a trained operation data evaluation unit, i.e., the first deep neural network. This unit outputs a score for each charging session event, i.e., the completeness of the operation event data. The data from multiple device maintenance events are sequentially input into a maintenance data evaluation unit, i.e., the second deep neural network, to obtain the completeness of the maintenance event data for each work order or inspection record. The data from multiple fund transfer events are sequentially input into a fund data evaluation unit, i.e., the third deep neural network, to obtain the completeness of the fund event data for each transaction event.

[0065] Furthermore, the average of the fine-grained scores for each independent event is calculated and aggregated into a macro-level indicator that represents the overall data quality. The arithmetic mean of the data completeness of all operational events output by the operational data evaluation unit is calculated to obtain a comprehensive operational data completeness score, which represents the average completeness of all charging session data within the evaluation time window. Similarly, the average of the data completeness of all maintenance events is calculated to obtain a comprehensive maintenance data completeness score, representing the average record quality of all maintenance activities; and the average of the data completeness of all financial events is calculated to obtain a comprehensive financial data completeness score, representing the average completeness of all financial transaction data.

[0066] In summary, this step first decomposes the large dataset through event recognition, then conducts a refined evaluation by constructing a data integrity evaluator, and finally obtains a reliable overall index through statistical aggregation, which greatly enhances the accuracy and reliability of the evaluation results.

[0067] S30: Based on the completeness of the operation data, the completeness of the maintenance data, and the completeness of the financial data, perform data association matching evaluation on the equipment operation data, equipment maintenance data, and financial flow data to obtain the operation-finance matching degree, operation-maintenance matching degree, and maintenance-finance matching degree;

[0068] Specifically, based on the completeness of the operational data, the completeness of the maintenance data, and the completeness of the financial data, a data association and matching evaluation is performed on the equipment operational data, equipment maintenance data, and financial flow data to obtain the operation-finance matching degree, operation-maintenance matching degree, and maintenance-finance matching degree, including:

[0069] Based on the completeness of the operational data and the completeness of the financial data, a first matching benchmark data and a first matching verification data are determined from the equipment operational data and the financial flow data. The first matching benchmark data and the first matching verification data are then correlated and matched to obtain the operational-financial matching degree.

[0070] Based on the completeness of the operational data and the completeness of the maintenance data, a second matching benchmark data and a second matching verification data are determined from the equipment operational data and the equipment maintenance data. The second matching benchmark data and the second matching verification data are then correlated and matched to obtain the operational-maintenance matching degree.

[0071] Based on the completeness of the operation and maintenance data and the completeness of the financial data, a third matching benchmark data and a third matching verification data are determined from the equipment operation and maintenance data and the financial flow data. The third matching benchmark data and the third matching verification data are then correlated and matched to obtain the operation and maintenance-finance matching degree.

[0072] Since charging station operation is essentially a process of generating services through equipment operation, ensuring equipment maintenance and management, and ultimately realizing value through capital flow, there are mandatory business logic connections between each of these three aspects. Data verification from any single dimension cannot fully guarantee the credibility of the entire data chain. Therefore, this step aims to construct a cross-validated data credibility triangle through three sets of pairwise correlation matching evaluations.

[0073] Specifically, the operation-maintenance matching is used to verify the logical consistency between the physical state of equipment (such as failure or wear and tear) and operation and maintenance activities (such as repair and maintenance records). A missing match indicates a delayed operation and maintenance response or inaccurate records. The operation-fund matching is used to verify the causal correspondence between service supply (such as charging volume) and financial income (such as service fees). A missing match indicates revenue leakage or billing fraud risk. The maintenance-fund matching is used to verify the compliance between resource consumption (such as parts replacement and labor input) and cost expenditure. A missing match indicates resource waste or suspected financial fraud.

[0074] Meanwhile, the core of this step lies in the fact that the matching verification is not a simple comparison, but rather a dynamic selection of data branches with higher credibility as the matching benchmark based on the data completeness calculated in previous steps, while those with relatively lower credibility are used as the data to be verified. This dynamic selection mechanism significantly improves the accuracy and reliability of the association verification process, effectively avoiding evaluation distortion caused by inherent defects in the quality of a single data source.

[0075] Specifically, firstly, based on the completeness of the operational data and the completeness of the financial data, a first matching benchmark data and a first matching verification data are determined from the equipment operational data and the financial flow data. The first matching benchmark data and the first matching verification data are then correlated and matched to obtain the operational-financial matching degree, including:

[0076] The completeness of the operational data and the completeness of the financial data are compared. When the completeness of the operational data is greater than or equal to the completeness of the financial data, the equipment operational data is used as the first matching benchmark data and the financial flow data is used as the first matching verification data.

[0077] Based on the first matching benchmark data, the first matching test data is subjected to event association matching, and the first event matching degree is obtained according to the event association matching result.

[0078] Based on the event association matching results, the first matching benchmark data and the first matching test data are matched for event content data to obtain the first data matching degree.

[0079] The operation-fund matching degree is obtained based on the first event matching degree and the first data matching degree.

[0080] It should be noted that when matching equipment operation data with cash flow data, typical matching scenarios include charging operation records and charging revenue records, equipment electricity consumption records and electricity expense records, and equipment failure records and failure loss records. The verification content includes whether the charging amount and the charge amount match according to the electricity price, whether the charging time and the charge time correspond, and whether the equipment power consumption and the electricity expense are consistent.

[0081] Furthermore, the operation-fund matching specifically refers to the correlation verification of fund flows directly related to equipment operation services. Its scope is limited to the matching relationship between the fund flows generated by the operation of charging equipment, including service revenue (such as charging electricity fees and service fees) and its direct costs (such as electricity expenses), and the corresponding equipment operation records (such as charging sessions). Specifically, the events involved in the first matching benchmark data and the first matching verification data specifically refer to events related to this core business, namely, charging session events in the equipment operation data, and corresponding revenue records (such as orders) and direct electricity cost records in the fund flow data. It does not include fund flows related to equipment operation and maintenance management (such as spare parts procurement expenditures and maintenance labor costs). The matching relationship between this portion of the fund flows and the operation and maintenance data is handled separately by the operation and maintenance-fund matching step.

[0082] Specifically, firstly, the completeness values ​​of operational data and financial data are compared, and the baseline direction for data verification is dynamically established based on the completeness measurement results. When the operational data completeness value is greater than or equal to the financial data completeness value, the operational data is determined to be of relatively superior quality and is designated as the first matching baseline data, while the financial data is designated as the first matching data to be verified. Furthermore, when the financial data completeness value is greater than the operational data completeness value, the financial data is used as the first matching baseline data, and the operational data is used as the first matching data to be verified.

[0083] The core of this benchmark selection mechanism lies in adhering to the principle of prioritizing high-completeness data, using datasets with higher completeness scores as validation benchmarks, and constructing a credible reference system to test the consistency of relatively low-completeness datasets. This fundamentally avoids the problem of distorted validation results caused by using low-quality data as benchmarks, and significantly improves the accuracy and reliability of data association matching evaluation.

[0084] Secondly, based on the first matching benchmark data, event association matching is performed on the first matching test data, and the first event matching degree is obtained according to the event association matching result, including:

[0085] The total number of events is determined based on the intersection of the events of the first matching benchmark data and the first matching test data.

[0086] By using time identifiers and device identifiers, an event association relationship is established between the first matching benchmark data and the first matching data to be verified, an event association matching result is formed, and the number of matching events is obtained.

[0087] The matching degree of the first event is calculated based on the ratio of the number of matched events to the total number of events.

[0088] Specifically, first, the total number of events is determined. The total number of events is determined based on the union of events from the first matching benchmark dataset and the first matching verification dataset. This total number of events refers to the sum of all unique events related to the charging service business within the evaluation time window, in both the first matching benchmark data and the first matching verification data. For example, 100 charging session events are extracted from the equipment operation data, and 95 charging transaction events and 5 electricity bill settlement events are extracted from the fund flow data; all of these are fund events directly related to the service. If 5 of these fund events do not have corresponding session records in the operation data, then the total number of all unique events is 105. This total represents the complete set of all business activities related to the charging service that should theoretically be fully recorded.

[0089] Secondly, event associations are established and the number of matching events is obtained. Using time identifiers (such as charging start timestamps and transaction completion times) and device identifiers (such as unique charging pile codes) as the joint matching key, precise association matching is performed between event records in the first matching benchmark data and the first matching verification data. By traversing all event records in the first matching benchmark data, a corresponding record that simultaneously meets the time tolerance requirement and device identifier consistency is searched in the second matching verification data. The number of event pairs that successfully establish associations is recorded as the number of matching events.

[0090] Finally, the first event matching degree is calculated. The first event matching degree is calculated using the following formula: First event matching degree = Number of matched events / Total number of events × 100%. This quantitative indicator accurately reflects the degree of matching between event records of two types of data. For example, when there are 100 total events, and 80 of them are successfully matched with corresponding fund transfer records, the calculated first event matching degree is 80%. This value directly indicates that 20% of business activities have missing or inconsistent data records.

[0091] Further, based on the event association matching results, event content data matching is performed on the first matching benchmark data and the first matching test data to obtain a first data matching degree, including:

[0092] Extract event pairs with established event relationships from the event association matching results to determine the total number of content matching verification events;

[0093] Perform data content consistency verification on event pairs that have established event relationships, and obtain the data content consistency verification results;

[0094] The statistical data content consistency verification result is the number of event pairs that pass the data content consistency verification, which yields the number of content-matching events;

[0095] The first data matching degree is calculated based on the ratio of the number of content matching events to the total number of content matching verification events.

[0096] Specifically, firstly, event pairs with established event relationships are extracted from the event association matching results to determine the total number of content matching verification events. In other words, from the aforementioned event association matching results, event pairs with successfully established relationships are selected; the total number of these event pairs represents the total number of content matching verification events, indicating the total number of samples requiring further content consistency verification.

[0097] Furthermore, data content consistency verification is performed on event pairs that have established event relationships. This verification employs multi-dimensional and in-depth validation rules to conduct a comprehensive consistency check on the key business attributes of the associated event pairs, ensuring the complete reliability of the data at the numerical, temporal, and business logic levels. Specific verification methods include:

[0098] Numerical consistency verification compares whether key quantifiable numerical fields in an event pair are consistent or within acceptable tolerances. For example, in a charging session-payment order event pair, it verifies the matching of values ​​such as charging amount (kWh), unit price (yuan / kWh), and final amount (yuan) to ensure the accuracy of core business metering.

[0099] Timing consistency verification checks whether the timestamp records of business actions with a time sequence in an event pair conform to a predefined sequential logic. For example, it verifies whether the start time of a charging session is earlier than the payment completion time of the corresponding payment order to ensure the rationality of the business process sequence.

[0100] Logical consistency verification checks whether specific constraints are met between multiple data fields in an event pair based on predefined business rules. For example, it verifies whether (charging amount × unit price + service fee) equals the total order amount, or checks whether charging revenue cannot be generated when the device is in a faulty state, thereby uncovering deep-seated errors or fraud at the business rule level.

[0101] The above three verifications together constitute a rigorous data content consistency verification framework, which ensures data quality from three dimensions: numerical accuracy, temporal rationality, and logical compliance.

[0102] After performing the above verification on each event pair, a data content consistency verification result is generated, indicating either pass or fail. Subsequently, the number of event pairs that passed the verification is counted, yielding the number of content-matching events. This value reflects the number of reliable events where the data content is completely consistent across all related events.

[0103] Finally, the first data matching degree is calculated based on the ratio of the number of content-matching events to the total number of content-matching verification events. First data matching degree = (Number of content-matching events / Total number of content-matching verification events) × 100%. This ratio quantifies the degree of consistency between the data to be verified and the benchmark data at the data content level. For example, if 72 out of 80 matching event pairs have completely identical data content, the data matching degree is 90%.

[0104] A higher ratio indicates that the data to be verified not only has complete event records, but also highly accurate core business data; a lower ratio reveals a large number of events that match but the data does not match, such as discrepancies between charging amount and billing amount, which can effectively indicate data entry errors, system billing loopholes or potential fraud risks.

[0105] In summary, this step, based on the established event associations, further quantifies the consistency and accuracy of key business data between the first matching benchmark data and the first matching test data from the perspective of data content, and calculates the first data matching degree, which is used to verify whether the detailed data records in the matched event pairs match, providing key evidence of content consistency for assessing the credibility of data assets.

[0106] Finally, based on the first event matching degree and the first data matching degree, the operation-fund matching degree is obtained. Considering both event matching and data content matching dimensions, the operation-fund matching degree = first event matching degree × first data matching degree. A multiplicative operation is used instead of a weighted average to emphasize the close coupling between the existence of event records and the accuracy of data content. Defects in either dimension will significantly impact the final result, thus more sensitively revealing the overall vulnerability of the data chain. For example, if the event matching degree is 80% and the data matching degree is 90%, the product result is 72%, objectively reflecting that although most events have been recorded and most data is accurate, the overall data quality is significantly degraded due to the combined effect of the two factors.

[0107] Operation-fund matching is a normalized continuous quantitative indicator with a value range of [0%, 100%]. The higher the value of this indicator, the more complete, accurate, and reliable the overall data link from event triggering to fund recording. This indicator provides a direct and quantitative basis for RDA value assessment and risk identification.

[0108] Based on the same technical approach and calculation model, the operation-maintenance matching degree and the maintenance-funds matching degree are calculated respectively to comprehensively evaluate the correlation between the three sets of data. It should be noted that when matching equipment operation data with equipment maintenance data, typical matching scenarios include equipment fault records and fault handling records, equipment performance degradation and maintenance records, and equipment status changes and maintenance operation records. The verification content includes whether the fault occurrence time and repair response time are reasonable, whether the abnormal equipment status corresponds to the maintenance record, and whether the maintenance completion time matches the equipment recovery time. When matching equipment maintenance data with fund flow data, typical matching scenarios include maintenance operation records and maintenance expenses, spare parts replacement records and spare parts procurement costs, and manual maintenance records and labor expenses. The verification content includes whether maintenance items match the corresponding expenses, whether the quantity of spare parts used matches the procurement cost, and whether maintenance hours correspond to labor costs.

[0109] Finally, by calculating the operation-maintenance matching degree, operation-fund matching degree, and maintenance-fund matching degree using the above methods, the symmetry and comparability of the correlation between the three sets of data on equipment operation, maintenance management, and fund flow were evaluated. The three matching degree indicators together constitute a multi-dimensional cross-validation data credibility triangle, forming the core evidence chain supporting the comprehensive evaluation of RDA data quality.

[0110] S40: Evaluate the data quality of the charging station data chain based on the completeness of the operational data, the completeness of the maintenance data, the completeness of the financial data, the matching degree between operation and funds, the matching degree between operation and maintenance, and the matching degree between maintenance and funds, and obtain the RDA data quality evaluation result.

[0111] Operational data completeness, maintenance data completeness, and financial data completeness reflect the completeness and reliability within each data branch; operation-finance matching degree, operation-maintenance matching degree, and maintenance-finance matching degree reflect the correlation and consistency between different data branches. These six indicators are used as the data quality evaluation results of the charging station data chain to obtain the RDA data quality evaluation results.

[0112] In summary, the embodiments of this application have at least the following technical effects:

[0113] Compared to existing technologies, this application first constructs an independent integrity evaluation system for equipment operation data, equipment maintenance data, and fund flow data. Employing a deep learning-based data integrity evaluator, it achieves accurate detection and quantitative assessment of missing and anomalies in multi-source heterogeneous data, effectively solving the problems of low efficiency and incomplete coverage in traditional manual sampling methods, and significantly improving the accuracy and reliability of data quality assessment. Secondly, it innovatively proposes a data integrity-based association matching verification mechanism. By intelligently determining benchmark data and data to be verified, and sequentially performing event association matching and data content consistency verification, it can deeply explore the inherent logical relationships between multiple branches of data such as equipment, maintenance, and funds, overcoming the technical challenge of insufficient credibility verification of associations between multi-source data.

[0114] Furthermore, a multi-level, multi-indicator comprehensive evaluation model is adopted to organically integrate data integrity measurement and data correlation measurement, thereby constructing an automated quality evaluation system with both depth and breadth. This enables a comprehensive, objective, and intelligent assessment of the charging station data chain quality, providing a unified and quantifiable standard for data quality evaluation.

[0115] Finally, this invention not only provides key data trust authentication technology support for the issuance and trading of charging station RDAs, protecting the legitimate rights and interests of investors, but can also be widely applied to scenarios such as financial supervision and operational optimization, providing effective technical means for building a transparent and trustworthy digital economy environment, and has significant social benefits and promotion and application value.

[0116] Example 2, as Figure 2 As shown, based on the same inventive concept as the AI-based RDA quality evaluation method provided in Embodiment 1, this embodiment of the invention also provides an AI-based RDA quality evaluation system, including:

[0117] Data acquisition module 11 is used to connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data and capital flow data;

[0118] The data integrity evaluation module 12 is used to evaluate the data integrity of the equipment operation data, equipment maintenance data and fund flow data respectively, and obtain the integrity of the operation data, maintenance data and fund data.

[0119] The data association matching evaluation module 13 is used to perform data association matching evaluation on the equipment operation data, equipment maintenance data and fund flow data based on the completeness of the operation data, the completeness of the maintenance data and the completeness of the fund data, and to obtain the operation-fund matching degree, operation-maintenance matching degree and maintenance-fund matching degree;

[0120] The data quality comprehensive evaluation module 14 is used to evaluate the data quality of the charging station data chain based on the completeness of the operation data, the completeness of the maintenance data, the completeness of the financial data, the matching degree of operation-finance, the matching degree of operation-maintenance, and the matching degree of maintenance-finance, and to obtain the RDA data quality evaluation result.

[0121] Specifically, the data acquisition module 11 is used for:

[0122] Connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data, and fund flow data, including:

[0123] Receive the collection time window input from the user terminal;

[0124] Connect the charging station data chain and determine the equipment operation storage block, equipment maintenance storage block, and fund flow storage block in the charging station data chain;

[0125] Based on the collection time window, data is collected from the equipment operation storage block, the equipment maintenance storage block, and the fund flow storage block to obtain equipment operation data, equipment maintenance data, and fund flow data, which are used as the multi-branch data.

[0126] The data integrity evaluation module 12 is specifically used for:

[0127] The equipment operation data, equipment maintenance data, and cash flow data are each evaluated for data integrity to obtain the integrity of operation data, maintenance data, and cash data, including:

[0128] Activate the data integrity evaluator, which includes an operational data evaluation unit, an operation and maintenance data evaluation unit, and a financial data evaluation unit;

[0129] The completeness of equipment operation data, equipment maintenance data, and capital flow data are evaluated by the operation data evaluation unit, maintenance data, and capital flow data, respectively, to obtain the completeness of operation data, maintenance data, and capital data.

[0130] Specifically, the construction steps of the data integrity evaluator include:

[0131] Collect historical charging station data, which includes historical equipment operation data, historical equipment maintenance data, and historical fund flow data. Divide the historical equipment operation data, historical equipment maintenance data, and historical fund flow data into events to construct sample equipment operation event datasets, sample equipment maintenance event datasets, and sample fund flow event datasets.

[0132] Data integrity is labeled for the sample equipment operation event dataset, sample equipment maintenance event dataset, and sample fund flow event dataset respectively, and sample operation event data integrity sets, sample maintenance event data integrity sets, and sample fund event data integrity sets are constructed.

[0133] Set up a first deep neural network, a second deep neural network, and a third deep neural network;

[0134] The first deep neural network is trained based on the sample device operation event dataset and the sample operation event data completeness set to generate an operation data evaluation unit;

[0135] The second deep neural network is trained based on the sample equipment operation and maintenance event dataset and the sample operation and maintenance event data completeness set to generate an operation and maintenance data evaluation unit.

[0136] The third deep neural network is trained based on the sample fund flow event dataset and the sample fund event data completeness set to generate a fund data evaluation unit;

[0137] The operational data evaluation unit, the maintenance data evaluation unit, and the financial data evaluation unit are combined to form the data integrity evaluator.

[0138] Specifically, the completeness of equipment operation data, equipment maintenance data, and capital flow data are assessed through the operation data assessment unit, maintenance data assessment unit, and capital flow data, respectively, to obtain the completeness of operation data, maintenance data, and capital data, including:

[0139] Event identification is performed on the equipment operation data, equipment maintenance data, and fund flow data respectively to extract multiple equipment operation event data, multiple equipment maintenance event data, and multiple fund flow event data;

[0140] The multiple device operation event data are sequentially input into the operation data evaluation unit to obtain the completeness of multiple operation event data, and the average of the multiple operation event data completeness is calculated to obtain the operation data completeness.

[0141] The multiple device operation and maintenance event data are sequentially input into the operation and maintenance data evaluation unit to obtain the completeness of multiple operation and maintenance event data, and the average of the multiple operation and maintenance event data completeness is calculated to obtain the operation and maintenance data completeness.

[0142] The data of the multiple fund flow events are sequentially input into the fund data evaluation unit to obtain the completeness of the multiple fund event data. The mean of the completeness of the multiple fund event data is then calculated to obtain the fund data completeness.

[0143] The data association matching evaluation module 13 is specifically used for:

[0144] Based on the completeness of the operational data, the completeness of the maintenance data, and the completeness of the financial data, a data association and matching evaluation is performed on the equipment operational data, equipment maintenance data, and financial flow data to obtain the operation-finance matching degree, operation-maintenance matching degree, and maintenance-finance matching degree, including:

[0145] Based on the completeness of the operational data and the completeness of the financial data, a first matching benchmark data and a first matching verification data are determined from the equipment operational data and the financial flow data. The first matching benchmark data and the first matching verification data are then correlated and matched to obtain the operational-financial matching degree.

[0146] Based on the completeness of the operational data and the completeness of the maintenance data, a second matching benchmark data and a second matching verification data are determined from the equipment operational data and the equipment maintenance data. The second matching benchmark data and the second matching verification data are then correlated and matched to obtain the operational-maintenance matching degree.

[0147] Based on the completeness of the operation and maintenance data and the completeness of the financial data, a third matching benchmark data and a third matching verification data are determined from the equipment operation and maintenance data and the financial flow data. The third matching benchmark data and the third matching verification data are then correlated and matched to obtain the operation and maintenance-finance matching degree.

[0148] Specifically, based on the completeness of the operational data and the completeness of the financial data, a first matching benchmark data and a first matching verification data are determined from the equipment operational data and the financial flow data. The first matching benchmark data and the first matching verification data are then correlated and matched to obtain the operational-financial matching degree, including:

[0149] The completeness of the operational data and the completeness of the financial data are compared. When the completeness of the operational data is greater than or equal to the completeness of the financial data, the equipment operational data is used as the first matching benchmark data and the financial flow data is used as the first matching verification data.

[0150] Based on the first matching benchmark data, the first matching test data is subjected to event association matching, and the first event matching degree is obtained according to the event association matching result.

[0151] Based on the event association matching results, the first matching benchmark data and the first matching test data are matched for event content data to obtain the first data matching degree.

[0152] The operation-fund matching degree is obtained based on the first event matching degree and the first data matching degree.

[0153] Further, based on the first matching benchmark data, event association matching is performed on the first matching test data, and a first event matching degree is obtained according to the event association matching result, including:

[0154] The total number of events is determined based on the intersection of the events of the first matching benchmark data and the first matching test data.

[0155] By using time identifiers and device identifiers, an event association relationship is established between the first matching benchmark data and the first matching data to be verified, an event association matching result is formed, and the number of matching events is obtained.

[0156] The matching degree of the first event is calculated based on the ratio of the number of matched events to the total number of events.

[0157] Furthermore, based on the event association matching results, event content data matching is performed on the first matching benchmark data and the first matching test data to obtain a first data matching degree, including:

[0158] Extract event pairs with established event relationships from the event association matching results to determine the total number of content matching verification events;

[0159] Perform data content consistency verification on event pairs that have established event relationships, and obtain the data content consistency verification results;

[0160] The statistical data content consistency verification result is the number of event pairs that pass the data content consistency verification, which yields the number of content-matching events;

[0161] The first data matching degree is calculated based on the ratio of the number of content matching events to the total number of content matching verification events.

[0162] Specifically, the data quality comprehensive evaluation module 14 is used for:

[0163] The data quality of the charging station data chain is evaluated based on the completeness of the operational data, the completeness of the maintenance data, the completeness of the financial data, the matching degree between operation and funds, the matching degree between operation and maintenance, and the matching degree between maintenance and funds, to obtain the RDA data quality evaluation result.

[0164] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0165] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0166] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. An RDA quality evaluation method based on artificial intelligence, characterized in that, The method includes: Connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data, and capital flow data; The equipment operation data, equipment maintenance data, and capital flow data are evaluated for data integrity to obtain the integrity of operation data, maintenance data, and capital data. Based on the completeness of the operational data, the completeness of the maintenance data, and the completeness of the financial data, the data association and matching evaluation of the equipment operation data, equipment maintenance data, and financial flow data are performed to obtain the operation-finance matching degree, operation-maintenance matching degree, and maintenance-finance matching degree. The data quality of the charging station data chain is evaluated based on the completeness of the operational data, the completeness of the maintenance data, the completeness of the financial data, the matching degree between operation and funds, the matching degree between operation and maintenance, and the matching degree between maintenance and funds, to obtain the RDA data quality evaluation result; Specifically, based on the completeness of the operational data and the completeness of the financial data, a first matching benchmark data and a first matching verification data are determined from the equipment operational data and the financial flow data. The first matching benchmark data and the first matching verification data are then correlated and matched to obtain the operational-financial matching degree, including: The completeness of the operational data and the completeness of the financial data are compared. When the completeness of the operational data is greater than or equal to the completeness of the financial data, the equipment operational data is used as the first matching benchmark data and the financial flow data is used as the first matching verification data. Based on the first matching benchmark data, the first matching test data is subjected to event association matching, and the first event matching degree is obtained according to the event association matching result. Based on the event association matching results, the first matching benchmark data and the first matching test data are matched for event content data to obtain the first data matching degree. The operation-fund matching degree is obtained based on the first event matching degree and the first data matching degree; Based on the technical approach and calculation model that achieves the same operation-fund matching degree, the operation-maintenance matching degree and the maintenance-fund matching degree are calculated respectively. When matching equipment operation data with cash flow data, the verification content includes whether the charging amount and the charging amount match the electricity price, whether the charging time and the charging time correspond, and whether the equipment power consumption and the electricity expenditure are consistent. When matching equipment operation data with equipment maintenance data, the verification content includes whether the fault occurrence time and the maintenance response time are reasonable, whether the abnormal equipment status and the maintenance record correspond, and whether the maintenance completion time and the equipment recovery time match. When matching equipment maintenance data with cash flow data, the verification content includes whether the maintenance items and the corresponding expenses match, whether the number of spare parts used and the procurement cost are consistent, and whether the maintenance hours and labor costs correspond.

2. The method according to claim 1, characterized in that, Connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data, and fund flow data, including: Receive the collection time window input from the user terminal; Connect the charging station data chain and determine the equipment operation storage block, equipment maintenance storage block, and fund flow storage block in the charging station data chain; Based on the collection time window, data is collected from the equipment operation storage block, the equipment maintenance storage block, and the fund flow storage block to obtain equipment operation data, equipment maintenance data, and fund flow data, which are used as the multi-branch data.

3. The method according to claim 1, characterized in that, The equipment operation data, equipment maintenance data, and cash flow data are each evaluated for data integrity to obtain the integrity of operation data, maintenance data, and cash data, including: Activate the data integrity evaluator, which includes an operational data evaluation unit, an operation and maintenance data evaluation unit, and a financial data evaluation unit; The completeness of equipment operation data, equipment maintenance data, and capital flow data are evaluated by the operation data evaluation unit, maintenance data, and capital flow data, respectively, to obtain the completeness of operation data, maintenance data, and capital data.

4. The method according to claim 3, characterized in that, The construction steps of the data integrity evaluator include: Collect historical charging station data, which includes historical equipment operation data, historical equipment maintenance data, and historical fund flow data. Divide the historical equipment operation data, historical equipment maintenance data, and historical fund flow data into events to construct sample equipment operation event datasets, sample equipment maintenance event datasets, and sample fund flow event datasets. Data integrity is labeled for the sample equipment operation event dataset, sample equipment maintenance event dataset, and sample fund flow event dataset respectively, and sample operation event data integrity sets, sample maintenance event data integrity sets, and sample fund event data integrity sets are constructed. Set up a first deep neural network, a second deep neural network, and a third deep neural network; The first deep neural network is trained based on the sample device operation event dataset and the sample operation event data completeness set to generate an operation data evaluation unit; The second deep neural network is trained based on the sample equipment operation and maintenance event dataset and the sample operation and maintenance event data completeness set to generate an operation and maintenance data evaluation unit. The third deep neural network is trained based on the sample fund flow event dataset and the sample fund event data completeness set to generate a fund data evaluation unit; The operational data evaluation unit, the maintenance data evaluation unit, and the financial data evaluation unit are combined to form the data integrity evaluator.

5. The method according to claim 4, characterized in that, The completeness of equipment operation data, equipment maintenance data, and capital flow data are assessed through the operation data assessment unit, maintenance data assessment unit, and capital flow data, respectively, to obtain the completeness of operation data, maintenance data, and capital data, including: Event identification is performed on the equipment operation data, equipment maintenance data, and fund flow data respectively to extract multiple equipment operation event data, multiple equipment maintenance event data, and multiple fund flow event data; The multiple device operation event data are sequentially input into the operation data evaluation unit to obtain the completeness of multiple operation event data, and the mean of the multiple operation event data completeness is calculated to obtain the operation data completeness. The multiple device operation and maintenance event data are sequentially input into the operation and maintenance data evaluation unit to obtain the completeness of multiple operation and maintenance event data, and the average of the multiple operation and maintenance event data completeness is calculated to obtain the operation and maintenance data completeness. The data of the multiple fund flow events are sequentially input into the fund data evaluation unit to obtain the completeness of the multiple fund event data. The mean of the completeness of the multiple fund event data is then calculated to obtain the fund data completeness.

6. The method according to claim 1, characterized in that, Based on the first matching benchmark data, event association matching is performed on the first matching test data, and a first event matching degree is obtained according to the event association matching result, including: The total number of events is determined based on the intersection of the events of the first matching benchmark data and the first matching test data. By using time identifiers and device identifiers, an event association relationship is established between the first matching benchmark data and the first matching verification data to form an event association matching result and obtain the number of matching events. The matching degree of the first event is calculated based on the ratio of the number of matching events to the total number of events.

7. The method according to claim 1, characterized in that, Based on the event association matching results, event content data matching is performed on the first matching benchmark data and the first matching test data to obtain a first data matching degree, including: Extract event pairs with established event relationships from the event association matching results to determine the total number of content matching verification events; Perform data content consistency verification on event pairs that have established event relationships, and obtain the data content consistency verification results; The statistical data content consistency verification result is the number of event pairs that pass the data content consistency verification, which yields the number of content-matching events. The first data matching degree is calculated based on the ratio of the number of content matching events to the total number of content matching verification events.

8. An RDA quality evaluation system based on artificial intelligence, characterized in that, The method for performing the AI-based RDA quality assessment method according to any one of claims 1-7 includes: The data acquisition module is used to connect to the charging station data chain and retrieve multi-branch data, including equipment operation data, equipment maintenance data and capital flow data. The data integrity evaluation module is used to evaluate the data integrity of the equipment operation data, equipment maintenance data, and fund flow data respectively, and obtain the integrity of the operation data, maintenance data, and fund data. The data association matching evaluation module is used to perform data association matching evaluation on the equipment operation data, equipment maintenance data and fund flow data based on the completeness of the operation data, the completeness of the maintenance data and the completeness of the fund data, and to obtain the operation-fund matching degree, operation-maintenance matching degree and maintenance-fund matching degree; The data quality comprehensive evaluation module is used to evaluate the data quality of the charging station data chain based on the completeness of the operational data, the completeness of the maintenance data, the completeness of the financial data, the matching degree between operation and funds, the matching degree between operation and maintenance, and the matching degree between maintenance and funds, and obtain the RDA data quality evaluation result.

Images