Multi-dimensional lean-oriented power grid enterprise purchase demand intelligent review management system

Abstract

The invention discloses a power grid enterprise purchase demand intelligent review management system for multi-dimensional lean. The system comprises a multi-mode purchase compliance verification module, a power grid violation clause identification module, a purchase file logic verification module, a power demand compliance research and judgment module, a power grid purchase review management module and a system cooperative control module. The multi-modal purchase compliance verification module processes text, table and image purchase data, the power grid violation clause identification module compares clauses with a violation clause library, the purchase file logic verification module analyzes clause logic association, the power demand compliance research and judgment module integrates data for comprehensive research and judgment, the power grid purchase review management module manages and controls the process and archived data, and the power grid procurement management module manages and controls the power grid procurement data. And the system cooperative control module realizes module cooperation. According to the invention, the violation identification accuracy and the logic examination capability are improved, the examination process is standardized, the procurement demand examination lean and intelligent demands of power grid enterprises are met, and the digital transformation of procurement management is supported.

Description

Technical Field

[0001] This invention relates to the field of procurement review technology for power grid enterprises, and in particular to an intelligent review and management system for procurement needs of power grid enterprises that is oriented towards multi-dimensional lean management. Background Technology

[0002] As the procurement scale of power grid companies continues to expand, the types of materials, technical parameters, and compliance standards involved in procurement needs are becoming increasingly complex. Procurement documents often include multimodal data such as text, tables, and images, and must comply with specific compliance clauses and logical norms of the power grid industry. Traditional procurement requirement review relies on manual work. When faced with massive amounts of data, this not only consumes a lot of manpower and time, but also requires reviewers to be proficient in identifying power grid procurement violation clauses and verifying the logical consistency of procurement documents. Furthermore, it requires the integration of multi-dimensional compliance standards for comprehensive judgment. Currently, power grid companies have increasingly higher requirements for the refinement and intelligence of procurement requirement review. There is an urgent need for an integrated system that can cover multimodal data processing, violation clause identification, logical verification, compliance judgment, and process management to meet the needs of efficiency, accuracy, and standardization in procurement requirement review, and support the digital transformation of power grid companies' procurement management.

[0003] Existing technologies have two significant drawbacks in the field of power grid enterprise procurement requirement review: First, existing review technologies struggle to effectively handle multimodal data in procurement requirements, often focusing only on compliance verification of single text data. They fail to deeply extract and verify the logical parameters in tables or the technical information in images, leading to the easy omission of compliance risks in multimodal data and hindering comprehensive procurement compliance verification. Second, existing technologies lack a systematic review capability for the logical consistency of procurement documents. They cannot efficiently link the identification results of non-compliant clauses with the logical relationships between various clauses in the procurement documents, nor can they integrate compliance assessment results for closed-loop control of the review process. This results in logical gaps and process disconnects during the review process, failing to provide full-process, integrated technical support for power grid enterprise procurement requirement review. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of existing technologies, this invention provides an intelligent review and management system for procurement needs of power grid enterprises that is oriented towards multi-dimensional lean manufacturing.

[0005] The technical solution adopted in this invention is an intelligent review and management system for power grid enterprise procurement needs, oriented towards multi-dimensional lean manufacturing. This system includes: a multimodal procurement compliance verification module, a power grid violation clause identification module, a procurement document logic verification module, a power demand compliance analysis module, a power grid procurement review and management module, and a system collaborative control module. The multimodal procurement compliance verification module receives raw power grid procurement needs data and performs feature extraction and preliminary compliance verification on text, table, and image-based procurement data using a multimodal procurement compliance deep verification model. The verified data is then transmitted to the power grid violation clause identification module. The power grid violation clause identification module calls a power grid procurement violation clause identification algorithm to match and identify violation features in the received data, and outputs the violation clause identification result to the procurement document logic verification module. The procurement document logic verification module employs a self-consistent logic verification model for procurement documents. It verifies the logical relationships between various clauses in the procurement document based on the results of violation clause identification, generates a logic verification report, and sends it to the power demand compliance assessment module. The power demand compliance assessment module, through the power procurement demand compliance assessment platform, integrates the logic verification report with power grid procurement demand parameters to comprehensively assess the compliance of procurement demands and outputs the assessment results to the power grid procurement review management module. Based on the parameters of the intelligent review management of power grid enterprise procurement demands, the power grid procurement review management module controls the review process and archives data for the assessment results, feeding the control data back to the system collaborative control module. The system collaborative control module receives data and results transmitted from each module, performs data interaction and collaborative work between different modules, and ensures the overall operation of the system.

[0006] Furthermore, the expression for the multimodal procurement compliance depth verification model in the multimodal procurement compliance verification module is as follows: For the multimodal procurement compliance verification results, To purchase text data, For procurement form data, To procure image data, For the number of data modalities, For the first Weighting coefficients for modal data These are the coefficients for feature extraction from text, table, and image data, respectively. For text feature extraction functions, Parameters for extracting text features For table feature extraction functions, Parameters for extracting table features. For image feature extraction functions, For image feature extraction parameters, These are the benchmark parameters for compliance in power grid procurement.

[0007] Furthermore, the algorithm expression for identifying power grid procurement violation clauses in the power grid violation clause identification module is as follows: , For the results of identifying the violations, For the set of procurement terms to be identified, For the first Purchase terms to be identified This is a collection of regulations related to power grid violations. For the first Grid violation clauses For the number of procurement terms to be identified, The number of violations of power grid regulations, For the function of calculating clause similarity, For the first Article to be identified and Article The matching weight of each violation clause, The threshold for determining violations. These are characteristic parameters of the power grid procurement terms.

[0008] Furthermore, the expression for the self-consistent verification model of the procurement document logic in the procurement document logic verification module is as follows: , To verify the logical consistency of the results, For the collection of terms and conditions of the procurement documents, For the first Terms and conditions of the procurement documents For the first Terms and conditions of the procurement documents For a set of logical verification rules, For the first Logical verification rules, For the first Logical verification rules, For the number of terms in the procurement documents, The number of logical verification rules. A function to calculate the logical relationship between clauses and rules. For logical verification coefficients, These are the logical characteristic parameters of the power grid procurement documents.

[0009] Furthermore, the calculation expression for the compliance assessment of power procurement demand in the power demand compliance assessment module is as follows: , Based on the results of the demand compliance assessment, This is the parameter set for the logic verification report. For the first Item logic verification report parameters, For the first Additional logic verification parameters, The set of parameters for identifying violations. For the first Parameters for identifying non-compliant clauses For the first Additional violation identification parameters, This is a set of parameters for power grid procurement requirements. For the first Parameters for power grid procurement requirements, For the first Additional procurement requirement parameters, To calibrate the number of parameters, To supplement the number of parameters, As a benchmark parameter for power grid demand assessment, This is a compliance assessment coefficient.

[0010] Furthermore, the intelligent review and management calculation expression for power grid enterprise procurement needs in the power grid procurement review and management module is as follows: , In order to review the management results, For the set of parameters of compliance assessment results, For the first Parameters for compliance assessment of item calibration For the first Additional compliance assessment parameters To review the set of process parameters, For the first Parameters for item calibration and review process. For the first Additional review process parameters, For data archiving parameter set, For the first Item calibration data archiving parameters, For the first Additional data archiving parameters, To calibrate the number of management parameters, To supplement the number of management parameters, To review the management coefficients, These serve as benchmark parameters for power grid procurement review.

[0011] Furthermore, the procurement document logic verification module includes a logic association extraction unit, a clause conflict detection unit, and a logic integrity verification unit. The logic association extraction unit receives the violation clause identification results output by the power grid violation clause identification module, and extracts the logic association information between clauses by analyzing the reference relationships, conditional relationships, and constraint relationships between clauses in conjunction with the content of each clause in the procurement document. Based on the extracted logic association information, the clause conflict detection unit compares the procurement requirements, technical parameters, and compliance standards involved in different clauses to identify contradictory, repetitive, and conflicting content between clauses. The logic integrity verification unit checks whether there are logical gaps, missing labeling clauses, and incoherent processes in the procurement document according to the specification requirements of the power grid procurement document, and generates a logic integrity verification result.

[0012] Furthermore, the power demand compliance assessment module includes a demand parameter parsing unit, a compliance standard matching unit, a comprehensive assessment calculation unit, and an assessment result output unit. The demand parameter parsing unit receives a logic verification report generated by the procurement document logic verification module, and decomposes and analyzes the procurement demand quantity, technical indicators, budget amount, and delivery cycle parameters involved in the report to obtain standardized demand parameters. The compliance standard matching unit compares the standardized demand parameters with various standards in the power grid procurement compliance standard library to determine the compliance basis and standard threshold corresponding to the demand parameters. The comprehensive assessment calculation unit combines the matched compliance standards and uses the algorithm of the power procurement demand compliance assessment platform to quantitatively calculate the compliance of the demand parameters. The assessment result output unit organizes the results obtained from the comprehensive assessment calculation into a structured report and transmits it to the power grid procurement review and management module.

[0013] Furthermore, the power grid procurement review management module includes a review process scheduling unit, a real-time data monitoring unit, a review result archiving unit, and a historical data query unit. The review process scheduling unit receives the assessment results output by the power demand compliance assessment module, allocates review tasks to corresponding review nodes, and schedules review personnel and resources according to the process specifications of intelligent review management of power grid enterprise procurement needs. The real-time data monitoring unit tracks data transmission, parameter modification, and review operations during the review process in real time, and records calibration data and time nodes during the review process. The review result archiving unit organizes the results, relevant supporting materials, and process data after the review is completed according to preset classification standards and stores them in the corresponding database. The historical data query unit receives external query requests, retrieves the corresponding historical review data from the database according to the query conditions in the request, and feeds it back to the querying party.

[0014] A multi-dimensional lean intelligent review and management system for power grid enterprise procurement needs includes the following steps: S1. Receiving procurement need-related data submitted by power grid enterprises, classifying the data into three types: text, tables, and images, and transmitting it to the multi-modal procurement compliance verification module; S2. The multi-modal procurement compliance verification module calls the multi-modal procurement compliance deep verification model to extract features from the three types of data, fuse the extracted features, perform preliminary compliance verification, and output preliminary verification results; S3. The power grid violation clause identification module receives the preliminary verification results, calls the power grid procurement violation clause identification algorithm, compares the clause content in the preliminary verification results with the power grid violation clause database, identifies the violation clauses, and generates a list of violation clauses; S4. Procurement... The procurement document logic verification module receives a list of non-compliant clauses, uses a self-consistent logic verification model for procurement documents to analyze the relationship between each clause in the procurement document and the list of non-compliant clauses, checks the logical relationships between clauses, and generates a logic verification report. The S5 power demand compliance assessment module receives the logic verification report, integrates the report data with basic parameters of power grid procurement needs through the power procurement demand compliance assessment platform, performs multi-dimensional calculations on the compliance of procurement needs, and obtains the compliance assessment results. The S6 power grid procurement review and management module receives the compliance assessment results, reviews, controls, and archives the assessment results according to the intelligent review and management process for power grid enterprise procurement needs, and the system collaborative control module integrates the data from each module to enable collaborative operation of different modules, completing the intelligent review of procurement needs.

[0015] Beneficial Effects: This invention proposes an intelligent review and management system for procurement needs of power grid enterprises, oriented towards multi-dimensional lean management. It utilizes a multi-modal procurement compliance verification module to process text, tables, and image data across all types, improving the coverage of multi-source data; a power grid violation clause identification module to accurately locate violations, improving the accuracy of violation identification; a procurement document logic verification module to identify logical problems in clauses, improving the rigor of document logic; a power demand compliance assessment module to conduct comprehensive analysis, improving the comprehensiveness of compliance assessment; a power grid procurement review and management module to standardize processes and archived data, improving the standardization of the review process; and a system collaborative control module to achieve linkage between various links, improving overall operational synergy. Ultimately, this significantly reduces manual intervention, substantially improving review efficiency and accuracy, while effectively meeting the needs of lean and intelligent procurement review, and strongly supporting the digital transformation of procurement management. The system utilizes a multimodal procurement compliance verification module to extract features from different types of data and conduct compliance verification, covering the review of table parameter logic and image technology information, thereby improving multimodal data processing capabilities and avoiding omissions of compliance risks. The system also uses a procurement document logic verification module to link the identification results of non-compliant clauses with the logic of procurement document clauses. Combined with the power demand compliance assessment module and the power grid procurement review management module, the system integrates assessment results and achieves closed-loop control of the review process, improving the systematic nature of logical review, eliminating logical gaps and process disconnects, and providing integrated technical support for the entire process. Attached Figure Description

[0016] Figure 1 This is a diagram showing the system module composition of the present invention; Figure 2 This is a flowchart of the system operation steps of the present invention. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] like Figure 1 As shown, a smart review and management system for procurement needs of power grid enterprises with a focus on multi-dimensional lean manufacturing includes: a multi-modal procurement compliance verification module, a power grid violation clause identification module, a procurement document logic verification module, a power demand compliance analysis module, a power grid procurement review and management module, and a system collaborative control module. The multimodal procurement compliance verification module receives the original data of power grid procurement requirements, performs feature extraction and preliminary compliance verification on text, table, and image procurement data through the multimodal procurement compliance deep verification model, and transmits the verified data to the power grid violation clause identification module. Specifically, the multimodal procurement compliance verification module, as the starting point of the system's data processing, receives raw data on power grid procurement requirements. Data types include text-based procurement specifications, tabular material parameter lists, and image-based technical drawings. It supports processing a maximum of 500MB of data per batch, with a data transfer rate of no less than 10MB / s, and is compatible with more than 10 common formats such as PDF, Excel, JPG, and PNG. This module relies on a multimodal procurement compliance deep verification model. For text data, it uses semantic understanding-based feature extraction to extract keywords, clause descriptions, and other labeling information, with a text recognition accuracy set at no less than 98.5%. For tabular data, it uses row-column association analysis to verify parameter formats, numerical ranges, and other content, with the verification error controlled within ±0.1%. For image data, it uses feature point matching to identify technical parameters, specification markings, and other information in the drawings, with an image recognition response time of no more than 3 seconds. During implementation, the module first parses the received data in terms of format, then initiates the corresponding extraction process according to the data type. After integrating the three types of data characteristics, it compares the data with the basic compliance parameter library for power grid procurement. The parameter library includes more than 200 basic compliance indicators such as voltage level, material model, and procurement cycle. Finally, it outputs preliminary verification results including compliant items, items to be verified, and non-compliant items, providing standardized data support for subsequent modules. The implementation of this module can reduce manual data preprocessing time by more than 60% and improve the initial compliance rate of procurement data to more than 90%.

[0019] The power grid violation clause identification module calls the power grid procurement violation clause identification algorithm to match and identify the violation features of the procurement clauses involved in the received data, and outputs the violation clause identification results to the procurement document logic verification module. Specifically, the power grid violation clause identification module receives the preliminary verification results output by the multimodal procurement compliance verification module. The module has a built-in power grid procurement violation clause library, containing over 800 violation clauses extracted from documents such as the State Grid Procurement Management Measures and local power procurement standards. The clauses are categorized into 12 types, including violations of qualification requirements, price setting, and performance requirements. The clauses are updated monthly to ensure their timeliness. This module uses a power grid procurement violation clause identification algorithm. During implementation, the clause content in the preliminary verification results is first segmented into words with an accuracy of over 99%. Then, the segmented results are matched with the clauses in the violation clause library. Matching dimensions include expression similarity, constraint consistency, and applicability scenario fit. The expression similarity threshold is set at 85%, constraint consistency requires matching at least eight calibration conditions, and applicability scenario fit must correspond to a specific procurement category, such as transmission equipment procurement or distribution material procurement. The module's processing response time is controlled within 5 seconds, and it can process no fewer than 200 clauses at a time. The output results include the violation clause number, violation type, original text of the violation, and the clause on which the violation is based. It also includes a violation severity classification, divided into three levels: minor violation, general violation, and serious violation. The classification is set according to the impact of the violation clause on the procurement process and the difficulty of rectification. Minor violations must be rectified within 24 hours, general violations must be rectified within 48 hours, and serious violations require suspension of the procurement process and special review. The implementation of this module can improve the accuracy rate of violation clause identification to over 97%, reducing missed and false judgments.

[0020] The procurement document logic verification module adopts a procurement document logic self-consistency verification model, and verifies the logical correlation between various clauses in the procurement document by combining the results of violation clause identification, generates a logic verification report and sends it to the power demand compliance assessment module. Specifically, the procurement document logic verification module receives the violation clause identification results output by the power grid violation clause identification module. The module has a built-in procurement document logic rule library, which includes six major categories: clause reference logic, parameter association logic, and process connection logic. Each category contains 50-80 specific rules, such as "technical parameter clauses must match material model clauses" and "delivery cycle clauses must coordinate with payment method clauses." Rules can be customized according to the power grid company's procurement process, with a configuration response time of no more than 10 minutes. This module adopts a procurement document logic self-consistency verification model. During implementation, it first establishes an association mapping between the violation clauses in the violation clause identification results and other clauses in the procurement document. The mapping covers at least 95% of the total number of clause associations. Then, it verifies the logical relationships between clauses item by item according to the logic rule library. Verification dimensions include forward logic consistency, reverse logic non-conflict, and supplementary logic no omission. Forward logic consistency requires that there are no contradictions in the labeling requirements between clauses; reverse logic non-conflict requires that contradictory expressions between clauses be excluded; and supplementary logic no omission requires that key process node clauses are not missing. The module can process up to 100 pages of procurement documents at a time, with a logic verification completion time of no more than 8 seconds. The output logic verification report includes the percentage of logically compliant items, the location and content of logical conflicts, and conflict rectification suggestions. The percentage of logically compliant items is calculated as the ratio of the number of compliant clauses to the total number of clauses. Conflict rectification suggestions must explicitly cite the corresponding rules in the logic rule base. Implementing this module can increase the logical compliance rate of procurement documents to over 92%, avoiding procurement disputes caused by logical issues in clauses.

[0021] The power demand compliance assessment module integrates the logic verification report and power grid procurement demand parameters through the power procurement demand compliance assessment platform to comprehensively assess the compliance of procurement demands and output the assessment results to the power grid procurement review and management module. Specifically, the power demand compliance assessment module receives the logic verification report generated by the procurement document logic verification module. This module relies on the power procurement demand compliance assessment platform, which integrates three major calibration resources: a power grid procurement compliance standard library, a historical assessment case library, and a parameter calculation model library. The compliance standard library includes over 300 compliance standards issued by the National Energy Administration and power grid companies; the historical assessment case library stores over 10,000 procurement demand assessment cases from the past five years; and the parameter calculation model library includes eight types of calculation models, such as demand rationality assessment and compliance risk quantification. During implementation, this module first decomposes the logic verification report, extracting over 50 calibration parameters from the procurement requirements, including material quantity, technical indicators, budget amount, delivery cycle, and quality requirements. The parameter extraction completeness is set to no less than 99%. Then, the decomposed parameters are compared with the compliance standard library to determine the compliance standard threshold for each parameter, such as the budget amount needing to be within 15% of the total annual procurement budget and the delivery cycle needing to meet the reasonable range of material production cycle + transportation cycle. Finally, models from the parameter calculation model library are called to quantify the parameter compliance. The dimensions include parameter compliance rate, risk occurrence rate, and reasonable deviation. The parameter compliance rate is the ratio of the number of compliant parameters to the total number of parameters. The risk occurrence rate is the percentage of parameters with compliance risks. The reasonable deviation must be controlled within ±5%. Finally, the calculation results are integrated to generate an assessment result including compliance score, risk point list, and rectification suggestions. The compliance score is based on a 100-point scale, with 80 points or above considered compliant, 60-80 points indicating rectification needed, and below 60 points considered non-compliant. Implementing this module can improve the efficiency of compliance assessment of procurement needs by more than 70%, ensuring the objectivity and accuracy of the assessment results.

[0022] The power grid procurement review and management module, based on the parameters of the intelligent review and management of power grid enterprise procurement needs, controls the review process and archives the data of the assessment results, and feeds the control data back to the system collaborative control module. Specifically, the power grid procurement review and management module receives the assessment results output by the power demand compliance assessment module. The module is based on the intelligent review and management parameters of the power grid enterprise's procurement needs. The parameters include the number of review nodes, the configuration standards for review personnel, the review time limit requirements, and the data archiving format. The review nodes are set as three benchmark nodes: initial review of needs, compliance review, and final review. Each node is configured with no less than two review personnel. The time limit for the initial review of needs is no more than 4 hours, the time limit for the compliance review is no more than 8 hours, and the time limit for the final review is no more than 12 hours. The data archiving adopts XML format for easy subsequent query and retrieval. This module first determines the review level based on the assessment results. If the assessment results are compliant, a regular review process is initiated; if rectification is required, a special review process is initiated; and if non-compliant, a rejection and resubmission process is initiated. Review tasks are then assigned according to the review level. The regular review process only requires passing three calibration nodes, the special review process requires an additional technical assessment node, and the rejection and resubmission process requires feedback on specific rectification opinions. During the review process, the reviewers' operations, review opinions, and modification traces are recorded in real time, with the records retained for at least five years. After the review is completed, the review results, supporting materials, and process records are archived. The archived data storage capacity supports a maximum of 100MB of archived data for a single procurement requirement. A unique file number is generated after archiving for easy traceability. This module can increase the standardization rate of the procurement review process to over 95%, shorten the review time by 50%, and achieve full lifecycle management of review data.

[0023] The system collaborative control module receives data and results transmitted from each module, performs data interaction and collaborative work between different modules, and ensures the overall operation of the system.

[0024] Specifically, the system's collaborative control module receives data and results from the multimodal procurement compliance verification module, the power grid violation clause identification module, the procurement document logic verification module, the power demand compliance assessment module, and the power grid procurement review and management module. The module incorporates a data interaction protocol, collaborative control rules, and a fault handling mechanism. The data interaction protocol uses TCP / IP to ensure data transmission stability, with a packet loss rate controlled within 0.1%. The collaborative control rules include the module startup sequence, data flow path, and result feedback mechanism. The module startup sequence strictly follows the order of "data verification - violation identification - logic verification - compliance assessment - review and management." The data flow path uses a targeted transmission method to avoid data mistransmission or omission. The fault handling mechanism includes module fault alarms, data backup and recovery, and process breakpoint resumption. The module fault alarm response time is no more than 1 second. Data backup uses a real-time dual backup method, ensuring 100% consistency between backup data and original data. Process breakpoint resumption allows execution to continue from the breakpoint after fault repair, without restarting the entire process. This module monitors the real-time operating status of each module during implementation. Monitoring metrics include CPU utilization, memory usage, and data processing progress. The CPU utilization threshold is set to no more than 80%, and the memory usage threshold is set to no more than 70%. Resource allocation is adjusted based on the data processing progress of each module, prioritizing computing resources for modules with longer processing times. Simultaneously, it receives the output results from each module, integrates them according to a preset format, and generates a comprehensive system operation report. The report includes key indicators such as processing time, processing results, and compliance rate for each module. Implementing this module can improve the overall system stability to over 99.5% and increase data interaction efficiency between modules by 40%, ensuring efficient and continuous system operation.

[0025] Preferably, the expression for the multimodal procurement compliance depth verification model in the multimodal procurement compliance verification module is: For the multimodal procurement compliance verification results, To purchase text data, For procurement form data, To procure image data, For the number of data modalities, For the first Weighting coefficients for modal data These are the coefficients for feature extraction from text, table, and image data, respectively. For text feature extraction functions, Parameters for extracting text features For table feature extraction functions, Parameters for extracting table features. For image feature extraction functions, For image feature extraction parameters, These are the benchmark parameters for compliance in power grid procurement.

[0026] Specifically, the multimodal procurement compliance deep verification model in the multimodal procurement compliance verification module first determines the number of data modalities during implementation. In a single processing run, this number is typically set to 3, corresponding to text, tables, and images, respectively. The weight coefficients for each data type are allocated based on data importance: text data typically has a weight coefficient of 0.4, table data 0.35, and image data 0.25, ensuring a higher weight for the labeled data. The feature extraction coefficients for text, table, and image data are set to 0.5, 0.3, and 0.2, respectively, with the coefficient values ​​determined based on the impact of each data feature on compliance verification. Text feature extraction parameters include semantic window size and keyword matching threshold; the window size is set to 5-8 characters, and the matching threshold is no less than 90%. Table feature extraction parameters include row and column matching accuracy and numerical error range; the matching accuracy reaches 99%, and the error range is ±0.5%. Image feature extraction parameters include the number of feature points and matching similarity; the number of feature points is no less than 50, and the matching similarity is no less than 85%. The compliance benchmark parameters for power grid procurement include over 200 indicators such as voltage level standard range, material model coding rules, and procurement cycle limits. For example, the voltage level must be within the 10kV-500kV range, and the procurement cycle must not exceed 180 days. During implementation, the module first extracts various data features according to the above parameters, then calculates and merges these features using weights and coefficients, and finally combines them with the compliance benchmark parameters to obtain the verification results. This model can improve the accuracy of multimodal data compliance verification to over 98%, reducing the limitations of single data type verification and ensuring comprehensive compliance screening of procurement data.

[0027] Preferably, the algorithm expression for identifying power grid procurement violation clauses in the power grid violation clause identification module is as follows: , For the results of identifying the violations, For the set of procurement terms to be identified, For the first Purchase terms to be identified This is a collection of regulations related to power grid violations. For the first Grid violation clauses For the number of procurement terms to be identified, The number of violations of power grid regulations, For the function of calculating clause similarity, For the first Article to be identified and Article The matching weight of each violation clause, The threshold for determining violations. These are characteristic parameters of the power grid procurement terms.

[0028] Specifically, the algorithm for identifying power grid procurement violation clauses in the power grid violation clause identification module determines the number of procurement clauses to be identified based on the size of a single procurement document, typically 50-150 clauses. The total number of power grid violation clauses is fixed at over 800, consistent with the module's built-in clause library. The clause similarity calculation function needs to set similarity grading standards, divided into perfect match (above 95%), high match (85%-94%), general match (70%-84%), and low match (below 70%). Only clauses with high or higher matching are included in subsequent calculations. The matching weight between the clause to be identified and the violation clause is allocated according to the clause type: qualification requirement clauses have a weight of 0.3, price setting clauses have a weight of 0.25, performance requirement clauses have a weight of 0.2, and other types have a weight of 0.25, ensuring that key violation types have higher weights. The violation threshold is set at 80 points. This score is calculated using matching weights and similarity. For example, if a clause to be identified highly matches a violation clause and falls under the qualification requirement category, a score of 85 points is considered a violation. The characteristic parameters of power grid procurement clauses include the standardization of clause expression and the completeness of constraints. Standardization must conform to the power grid procurement document writing standards, and completeness must include at least five defined constraints. During implementation, the clauses are first segmented into words with an accuracy of over 99%. Then, similarity and matching weights are calculated. Finally, the identification result is obtained by combining the judgment threshold and characteristic parameters. This algorithm can improve the efficiency of violation clause identification by more than 60%, while controlling the false positive rate to within 3%, ensuring accurate identification of violation clauses.

[0029] Preferably, the expression for the self-consistent logical verification model of the procurement documents in the procurement document logical verification module is as follows: , To verify the logical consistency of the results, For the collection of terms and conditions of the procurement documents, For the first Terms and conditions of the procurement documents For the first Terms and conditions of the procurement documents For a set of logical verification rules, For the first Logical verification rules, For the first Logical verification rules, For the number of terms in the procurement documents, The number of logical verification rules. A function to calculate the logical relationship between clauses and rules. For logical verification coefficients, These are the logical characteristic parameters of the power grid procurement documents.

[0030] Specifically, the procurement document logic self-consistency verification model in the procurement document logic verification module determines the number of clauses in the procurement document based on the document's page count. For documents under 100 pages, the number of clauses is generally 100-200. The number of logic verification rules is fixed at 360-480, covering six major categories of logic rules. The function for calculating the logical correlation between clauses and rules needs to set correlation levels, categorized as strong correlation (above 90%), medium correlation (75%-89%), and weak correlation (below 75%). Only strong and medium correlations are included in the calculation. The logic verification coefficient is set according to the rule's importance: 0.3 for clause reference logic, 0.28 for parameter association logic, 0.22 for process connection logic, and 0.2 for other categories, ensuring higher weighting for calibrated logic rules. The logical characteristic parameters of the power grid procurement documents include clause logical coherence and rule adaptability. Coherence requires no logical gaps between clauses, and adaptability must conform to the power grid procurement document logic specifications. During implementation, the association mapping between clauses and non-compliant clauses is first established, with a mapping coverage of no less than 95%. Then, the logical correlation between clauses and rules is calculated. The verification result is obtained by multiplying the correlation and using the verification coefficient and feature parameters. The implementation of this model can improve the identification rate of logical conflicts in procurement documents to over 95% and the logical compliance rate by 15-20 percentage points, avoiding procurement execution deviations caused by logical problems.

[0031] Preferably, the calculation expression for the compliance assessment of power procurement demand in the power demand compliance assessment module is as follows: , Based on the results of the demand compliance assessment, This is the parameter set for the logic verification report. For the first Item logic verification report parameters, For the first Additional logic verification parameters, The set of parameters for identifying violations. For the first Parameters for identifying non-compliant clauses For the first Additional violation identification parameters, This is a set of parameters for power grid procurement requirements. For the first Parameters for power grid procurement requirements, For the first Additional procurement requirement parameters, To calibrate the number of parameters, To supplement the number of parameters, As a benchmark parameter for power grid demand assessment, This is a compliance assessment coefficient.

[0032] Specifically, the electricity demand compliance assessment module's electricity procurement demand compliance assessment calculation uses 30 calibration parameters during implementation, including key parameters such as material quantity, technical indicators, and budget amount, and 20 supplementary parameters, including auxiliary parameters such as delivery location and acceptance standards. In the logic verification report parameters, the calibration parameters must have a numerical precision of two decimal places, and the supplementary parameters must be 100% complete. In the violation clause identification parameters, the calibration parameters must include the violation type and severity, and the supplementary parameters must include rectification suggestions. In the power grid procurement demand parameters, the calibration parameters must comply with power grid material procurement standards, and the supplementary parameters must meet project-specific requirements. The power grid demand assessment benchmark parameters include over 300 indicators such as the upper limit of budget percentage (15% of the annual budget) and the upper limit of delivery cycle (180 days). Compliance assessment coefficients are set according to the assessment dimensions: 0.4 for technical compliance, 0.3 for budget compliance, and 0.3 for cycle compliance. During implementation, various parameters are first broken down and integrated, and then calculated according to the calibration and supplementary parameters. The judgment results are obtained by combining the benchmark parameters and the judgment coefficients. The implementation of this calculation method can improve the quantitative accuracy of demand compliance judgment to more than 90%, and the degree of consistency between the judgment results and the actual compliance situation reaches 97%, providing a scientific quantitative basis for the compliance of procurement needs.

[0033] Preferably, the intelligent review and management calculation expression for power grid enterprise procurement needs in the power grid procurement review and management module is as follows: , In order to review the management results, For the set of parameters of compliance assessment results, For the first Parameters for compliance assessment of item calibration For the first Additional compliance assessment parameters To review the set of process parameters, For the first Parameters for item calibration and review process. For the first Additional review process parameters, For data archiving parameter set, For the first Item calibration data archiving parameters, For the first Additional data archiving parameters, To calibrate the number of management parameters, To supplement the number of management parameters, To review the management coefficients, These serve as benchmark parameters for power grid procurement review.

[0034] Specifically, the intelligent review and management calculation for power grid enterprise procurement needs in the power grid procurement review management module is implemented with 25 calibrated management parameters, including key parameters such as review node completion time and reviewer qualification level, and 15 supplementary management parameters, including auxiliary parameters such as review opinion completeness and supporting material quantity. In the compliance assessment result parameters, the calibrated parameters must include compliance score and risk level, and the supplementary parameters must include the number of risk points; in the review process parameters, the calibrated parameters must include the review time for each node, and the supplementary parameters must include the number of process adjustments; in the data archiving parameters, the calibrated parameters must include archiving completeness and storage format, and the supplementary parameters must include the number of backups. The review management coefficient is set according to the management stage: 0.35 for the process scheduling stage, 0.3 for the data monitoring stage, and 0.35 for the archiving query stage; the power grid procurement review benchmark parameters include indicators such as the upper limit of review time (24 hours) and the archiving completeness standard (100%). During implementation, various management parameters are first extracted and categorized for statistical analysis. Then, weighted scores are calculated based on calibrated and supplementary parameters. The management results are obtained by combining management coefficients and benchmark parameters. This calculation method can increase the standardization rate of review management to over 98%, control the deviation rate of the review process to within 2%, and simultaneously achieve precision in review data management, providing technical support for the whole process control of procurement review.

[0035] Preferably, the procurement document logic verification module includes a logic association extraction unit, a clause conflict detection unit, and a logic integrity verification unit. The logic association extraction unit receives the violation clause identification results output by the power grid violation clause identification module, and extracts the logic association information between clauses by analyzing the reference relationships, conditional relationships, and constraint relationships between clauses in combination with the content of each clause in the procurement document. The clause conflict detection unit, based on the extracted logic association information, compares the procurement requirements, technical parameters, and compliance standards involved in different clauses to identify contradictory, repetitive, and conflicting content between clauses. The logic integrity verification unit checks whether there are logical gaps, missing labeling clauses, and incoherent processes in the procurement document according to the specification requirements of the power grid procurement document, and generates a logic integrity verification result.

[0036] Specifically, the procurement document logic verification module includes a logic association extraction unit, a clause conflict detection unit, and a logic integrity verification unit. During implementation, the logic association extraction unit first receives the violation clause identification results output by the power grid violation clause identification module. These results include the violation clause number, type, and content. The unit must complete the receiving and parsing of the results within 5 seconds. Subsequently, it analyzes the reference relationships (such as cross-references of clause numbers), conditional relationships (such as "if clause A is met, then clause B must be executed"), and constraint relationships (such as "the parameter of clause C must not exceed the limit of clause D") between clauses in the procurement document. The coverage rate of extracted logic association information must reach at least 98% of the total number of clauses. The association information is stored in a structured data format for easy retrieval by subsequent units. The clause conflict detection unit, based on the extracted logic association information, compares the procurement requirements (such as qualification thresholds), technical parameters (such as voltage level ranges), and compliance standards (such as acceptance criteria) of different clauses. The conflict detection response time must not exceed 8 seconds, the number of clauses that can be processed in a single instance must not be less than 200, and the detection accuracy must reach at least 97%, ensuring that no contradictory, duplicate, or conflicting content is missed. The detection results must be labeled with the conflict clause number, conflict type, and a summary of the conflict content. The logical integrity verification unit checks the procurement documents for logical gaps (such as missing process steps), missing key clauses (such as payment method clauses not mentioned), and inconsistencies in the process (such as no connection between delivery cycle and acceptance cycle) in accordance with the requirements of the power grid procurement document specifications (including the State Grid Procurement Document Writing Guidelines, local power procurement specifications, etc.). The verification completion time is controlled within 10 seconds. The output logical integrity verification results must include the percentage of complete items (must be above 90%), a list of missing items, and supplementary suggestions. Through the collaboration of the three units, this module can improve the comprehensiveness of the logical review of procurement documents to over 95%, reducing the procurement risks caused by logical problems.

[0037] Preferably, the power demand compliance assessment module includes a demand parameter parsing unit, a compliance standard matching unit, a comprehensive assessment calculation unit, and an assessment result output unit. The demand parameter parsing unit receives a logic verification report generated by the procurement document logic verification module, and decomposes and analyzes the procurement demand quantity, technical indicators, budget amount, and delivery cycle parameters involved in the report to obtain standardized demand parameters. The compliance standard matching unit compares the standardized demand parameters with various standards in the power grid procurement compliance standard library to determine the compliance basis and standard threshold corresponding to the demand parameters. The comprehensive assessment calculation unit combines the matched compliance standards and uses the algorithm of the power procurement demand compliance assessment platform to quantitatively calculate the compliance of the demand parameters. The assessment result output unit organizes the results obtained from the comprehensive assessment calculation into a structured report and transmits it to the power grid procurement review and management module.

[0038] Specifically, the power demand compliance assessment module includes a demand parameter analysis unit, a compliance standard matching unit, a comprehensive assessment calculation unit, and an assessment result output unit. During implementation, the demand parameter analysis unit receives a logic verification report generated by the procurement document logic verification module. The report includes logical compliance items, conflict items, and rectification suggestions. The unit must complete report analysis within 6 seconds, breaking down and analyzing parameters such as procurement demand quantity (e.g., number of equipment units), technical indicators (e.g., equipment power, insulation class), budget amount (accurate to yuan), and delivery cycle (accurate to days). The parameter breakdown accuracy must be above 99%, and the standardized demand parameters must conform to the power grid procurement parameter coding specifications for easy subsequent matching. The compliance standard matching unit compares the standardized demand parameters with the power grid procurement compliance standard library (including over 300 national and industry standards). Matching dimensions include parameter name, parameter range, and parameter unit. The matching response time must not exceed 7 seconds, and the matching accuracy must be above 98%. It determines the compliance basis (e.g., standard number, standard clause) and standard threshold (e.g., budget amount must not exceed 15% of the annual procurement budget) for each parameter, and the matching results must form a parameter-standard correspondence table. The comprehensive assessment and calculation unit, combining matching compliance standards, invokes the algorithm of the power procurement demand compliance assessment platform to quantify the compliance of demand parameters. The calculation dimensions include parameter compliance rate (number of compliant parameters / total number of parameters) and risk occurrence rate (number of risky parameters / total number of parameters). The calculation completion time is controlled within 9 seconds, and the quantitative results must be retained to two decimal places. The assessment result output unit organizes the calculation results into a structured report, which includes a compliance score (out of 100, with 80 points or above considered compliant) and a risk point list (including risk parameters and risk levels). The report generation time is no more than 3 seconds, and it is then transmitted to the power grid procurement review management module. This module, through the collaboration of four units, can improve the efficiency of demand compliance assessment by more than 70%, ensuring the objectivity and accuracy of the assessment results.

[0039] Preferably, the power grid procurement review management module includes a review process scheduling unit, a real-time data monitoring unit, a review result archiving unit, and a historical data query unit. The review process scheduling unit receives the assessment results output by the power demand compliance assessment module, allocates review tasks to corresponding review nodes, and schedules review personnel and resources according to the process specifications of intelligent review management of power grid enterprise procurement needs. The real-time data monitoring unit tracks data transmission, parameter modification, and review operations during the review process in real time, and records calibration data and time nodes during the review process. The review result archiving unit organizes the results, relevant supporting materials, and process data after the review is completed according to preset classification standards and stores them in the corresponding database. The historical data query unit receives external query requests, retrieves the corresponding historical review data from the database according to the query conditions in the request, and provides feedback to the querying party.

[0040] Specifically, the power grid procurement review management module includes a review process scheduling unit, a real-time data monitoring unit, a review result archiving unit, and a historical data query unit. During implementation, the review process scheduling unit receives the assessment results output by the power demand compliance assessment module. These results include compliance scores and risk levels. The unit must receive the results within 4 seconds. Based on the power grid enterprise procurement demand intelligent review management process specifications (including 3 calibrated review nodes: initial demand review, compliance review, and final review), review tasks are assigned to the corresponding review nodes. Each node must be assigned at least two reviewers with relevant qualifications (such as intermediate or higher procurement review qualifications). Simultaneously, resources such as the parameter library and historical case library required for the review are scheduled. The task allocation accuracy must reach 100% to avoid misassignment. The real-time data monitoring unit tracks data transmission (such as parameter modification records), parameter modifications (such as budget adjustments), and review operations (such as submission of review opinions) during the review process in real time. The monitoring frequency is set to once per second, recording key data (such as parameter values ​​before and after modification) and time nodes (accurate to the second). The monitoring data storage duration is no less than 5 years to ensure the traceability of the review process. The review results archiving unit organizes the completed review results (such as review conclusions and compliance scores), supporting materials (such as compliance standards), and process data (such as review opinion records) according to preset classification standards (such as classification by procurement project number and review date). The archiving completion time is no more than 10 seconds, and the storage capacity supports a maximum of 100MB of archived data for a single procurement requirement. The archived data is stored with dual backups, and the backup consistency reaches 100%. The historical data query unit receives external query requests (including query conditions: project number, review date, reviewer). The query response time is no more than 3 seconds. It retrieves the corresponding historical review data from the database, with a retrieval accuracy rate of over 99%, and then provides feedback to the querying party. This module, operating through four units, can improve the standardization rate of the review process to over 95%, shorten the review time by 50%, and realize full lifecycle management of review data.

[0041] The multimodal procurement compliance deep verification model is a model used in this invention to process the preliminary verification of compliance of multiple types of data in power grid procurement. It can perform feature extraction and compliance screening on three types of procurement data: text, tables, and images, solving the problem of incomplete verification of traditional single data types. Its implementation process requires first determining the number of data modalities (usually 3), assigning weight coefficients (text 0.4, tables 0.35, images 0.25) and feature extraction coefficients (text 0.5, tables 0.3, images 0.2) to each type of data, and then setting specific extraction parameters: for text data, the semantic window size (5-8 characters) and keyword matching threshold (not less than 90%) need to be set; for table data, the row and column matching accuracy (99%) and numerical error range (±0.5%) need to be set; and for image data, the number of feature points (not less than 50) and matching similarity (not less than 85%) need to be set. Finally, combined with power grid procurement compliance benchmark parameters (such as voltage level 10kV-500kV, procurement cycle not exceeding 180 days), the preliminary verification results are output after feature fusion. This model transforms multimodal procurement data into standardized compliance verification data, filters out obviously non-compliant information, provides data support for subsequent modules, improves the accuracy of multimodal data compliance verification to over 98%, reduces manual preprocessing time by over 60%, fills the technical gap in simultaneous verification of multiple types of data in power grid procurement, and ensures the comprehensiveness and efficiency of the initial screening of procurement data.

[0042] The power grid procurement violation clause identification algorithm of this invention is used to accurately identify violations in procurement clauses. Relying on a built-in database of over 800 violation clauses, it can specifically match violation types such as qualification requirements, price settings, and performance requirements in power grid procurement scenarios. In implementation, the number of clauses to be identified in a single processing run (50-150 clauses) must first be determined. The clauses are then segmented (with an accuracy of over 99%). A similarity calculation function (divided into four levels: complete match (over 95%), high match (85%-94%), etc.) and matching weights (qualification requirements 0.3, price settings 0.25, etc.) are used to calculate the matching score. Combined with a violation judgment threshold (80 points) and power grid procurement clause feature parameters (expression standardization, constraint completeness), the algorithm determines whether a clause is in violation and marks the violation type and basis. This algorithm accurately identifies non-compliant content in procurement terms and generates a list of non-compliant terms, avoiding oversights and misjudgments that can occur with manual identification. It improves the efficiency of identifying non-compliant terms by more than 60%, and controls the misjudgment rate to within 3%. This ensures that power grid procurement terms comply with national and industry standards, reduces procurement disputes and legal risks caused by non-compliant terms, and lays the foundation for the compliance of procurement documents.

[0043] The procurement document logical self-consistency verification model of this invention is used to review the logical correlation between various clauses in procurement documents. It covers 360-480 logical rules across six major categories, including clause citation, parameter correlation, and process connection, resolving logical gaps and conflicts in procurement documents. The implementation process first establishes a correlation mapping between procurement clauses and non-compliant clauses (coverage no less than 95%). Then, it analyzes the matching degree between clauses and rules using a logical correlation calculation function (divided into three levels: strong correlation (above 90%), medium correlation (75%-89%), etc.). Finally, it combines logical verification coefficients (0.3 for clause citation, 0.28 for parameter correlation, etc.) with logical characteristic parameters of power grid procurement documents (coherence, adaptability) to calculate the proportion of logically compliant items, identify logical conflict locations, and generate rectification suggestions. This model ensures that all clauses in the procurement document are logically coherent, conflict-free, and complete, outputting a logical verification report. It increases the logical conflict identification rate of procurement documents to over 95%, improves the logical compliance rate by 15-20 percentage points, avoids procurement execution deviations caused by logical contradictions in clauses, ensures the technical rigor of procurement documents, and supports subsequent compliance assessment work.

[0044] The power procurement demand compliance assessment platform of this invention is a platform that integrates multi-dimensional data to conduct comprehensive compliance analysis of procurement demands. It integrates three major resources: a compliance standard library (more than 300 national and industry standards), a historical case library (more than 10,000 cases in the past 5 years), and a parameter calculation model library (8 types of calculation models). In implementation, the logic verification report is first decomposed by the demand parameter parsing unit to extract 30 calibration parameters (material quantity, technical indicators, etc.) and 20 supplementary parameters (delivery location, acceptance standards, etc.), with an accuracy rate of over 99%. Then, the compliance standard matching unit compares the parameters with the standard library to determine the compliance basis and threshold (e.g., budget not exceeding 15% of the annual budget). Subsequently, the calculation model library is called to quantify the calculation from three dimensions: parameter compliance rate, risk occurrence rate, and reasonable deviation (within ±5%). Finally, a compliance score out of 100 (compliance score above 80) and a risk point list are generated. The platform scientifically and quantitatively assesses the compliance of procurement needs, outputs structured assessment results, improves the efficiency of compliance assessment by more than 70%, and achieves a 97% consistency between the assessment results and the actual compliance situation. It provides objective quantitative basis for power grid procurement review, avoids subjective judgment bias, promotes the transformation of procurement need review from "experience-based judgment" to "data-driven", and supports the lean management of procurement.

[0045] like Figure 2As shown, a multi-dimensional lean intelligent review and management system for power grid enterprise procurement needs is described. The system operates through the following steps: S1. Receiving procurement need-related data submitted by the power grid enterprise, classifying the data into three types—text, tables, and images—and transmitting it to the multi-modal procurement compliance verification module; S2. The multi-modal procurement compliance verification module calls the multi-modal procurement compliance deep verification model to extract features from the three types of data, fuses the extracted features, performs preliminary compliance verification, and outputs preliminary verification results; S3. The power grid violation clause identification module receives the preliminary verification results, calls the power grid procurement violation clause identification algorithm, compares the clause content in the preliminary verification results with the power grid violation clause database, identifies the violation clauses, and generates a list of violation clauses; S4. The procurement document logic verification module receives a list of non-compliant clauses, uses a self-consistent logic verification model for procurement documents to analyze the relationship between each clause in the procurement document and the list of non-compliant clauses, checks the logical relationships between clauses, and generates a logic verification report. The S5 power demand compliance assessment module receives the logic verification report, integrates the report data with basic parameters of power grid procurement needs through the power procurement demand compliance assessment platform, performs multi-dimensional calculations on the compliance of procurement needs, and obtains the compliance assessment results. The S6 power grid procurement review and management module receives the compliance assessment results, reviews, controls, and archives the assessment results according to the intelligent review and management process for power grid enterprise procurement needs, and the system collaborative control module integrates the data from each module to enable collaborative operation of different modules, completing the intelligent review of procurement needs.

[0046] This intelligent review and management system for power grid enterprise procurement needs, designed for multi-dimensional lean operations, employs multi-modal data processing. Its multi-modal procurement compliance verification module specifically processes text, table, and image-based procurement data, eliminating reliance on a single data type and comprehensively mining procurement information from various data types. It boasts high accuracy in violation identification; the power grid violation clause identification module specifically compares procurement clauses with a power grid violation clause database, avoiding oversights during manual identification. It possesses strong logical review capabilities; the procurement document logic verification module deeply analyzes the relationships between clauses, ensuring logical coherence in procurement documents. It offers more comprehensive compliance assessment; the power demand compliance assessment module integrates multi-dimensional data for comprehensive analysis, rather than relying on a single dimension. It features standardized process management; the power grid procurement review and management module integrates review process control and data archiving. Furthermore, it ensures efficient collaboration among modules; the system's collaborative control module breaks down data barriers between modules, improving overall operational efficiency.

[0047] This system addresses the challenge of handling multimodal data with existing technologies. Its multimodal procurement compliance verification module extracts features from different data types and performs compliance checks. It not only processes text data but also conducts in-depth reviews of parameter logic in tables and technical information in images, filling gaps in multimodal data processing and preventing oversights of compliance risks. Furthermore, addressing the lack of systematic logical review capabilities in existing technologies, the system uses a procurement document logic verification module to link the identification results of violation clauses with the logical structure of procurement document clauses. Simultaneously, it combines the comprehensive analysis results from the power demand compliance assessment module with the process control of the power grid procurement review management module to construct a complete system from logic verification to process closure. This eliminates logical gaps and process disconnects in the review process, providing end-to-end technical support for power grid companies' procurement demand review.

[0048] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," "link," and "fix" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal communication between two components. Those skilled in the art will understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0049] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A smart review and management system for procurement requirements of power grid enterprises, characterized in that: include: The system comprises a multimodal procurement compliance verification module, a power grid violation clause identification module, a procurement document logic verification module, a power demand compliance analysis module, a power grid procurement review and management module, and a system collaborative control module. The multimodal procurement compliance verification module receives raw power grid procurement demand data and performs feature extraction and preliminary compliance verification on text, table, and image-based procurement data using a multimodal procurement compliance deep verification model. The verified data is then transmitted to the power grid violation clause identification module. This module invokes a power grid procurement violation clause identification algorithm to match and identify violation features in the received data, and outputs the violation clause identification results to the procurement document logic verification module. The procurement document logic verification module employs a self-consistent logic verification model for procurement documents. It verifies the logical relationships between clauses in the procurement documents based on the results of violation clause identification, generates a logic verification report, and sends it to the power demand compliance assessment module. The power demand compliance assessment module, through the power procurement demand compliance assessment platform, integrates the logic verification report with power grid procurement demand parameters to comprehensively assess the compliance of procurement demands and outputs the assessment results to the power grid procurement review management module. Based on the parameters of the power grid enterprise's intelligent review management of procurement demands, the power grid procurement review management module controls the review process and archives data for the assessment results, feeding the control data back to the system collaborative control module. The system collaborative control module receives data and results transmitted from each module, performs data interaction and collaborative work between different modules, and ensures the overall operation of the system.

2. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The expression for the multimodal procurement compliance depth verification model in the multimodal procurement compliance verification module is as follows: For the multimodal procurement compliance verification results, To purchase text data, For procurement form data, To purchase image data, For the number of data modalities, For the first Weighting coefficients for modal data These are the coefficients for feature extraction from text, table, and image data, respectively. For text feature extraction functions, Parameters for extracting text features For table feature extraction functions, Parameters for extracting table features. For image feature extraction functions, For image feature extraction parameters, These are the benchmark parameters for compliance in power grid procurement.

3. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The algorithm expression for identifying power grid procurement violation clauses in the power grid violation clause identification module is as follows: , For the results of identifying the violations, For the set of procurement terms to be identified, For the first Purchase terms to be identified This is a collection of regulations related to power grid violations. For the first Grid violation clauses For the number of procurement terms to be identified, The number of violations of power grid regulations, For the function of calculating clause similarity, For the first Article to be identified and Article The matching weight of each violation clause, The threshold for determining violations. These are characteristic parameters of the power grid procurement terms.

4. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The expression for the self-consistent logical verification model of the procurement documents in the procurement document logical verification module is as follows: , To verify the logical consistency of the results, For the collection of terms and conditions of the procurement documents, For the first Terms and conditions of the procurement documents For the first Terms and conditions of the procurement documents For a set of logical verification rules, For the first Logical verification rules, For the first Logical verification rules, For the number of terms in the procurement documents, The number of logical verification rules. A function to calculate the logical relationship between clauses and rules. For logical verification coefficients, These are the logical characteristic parameters of the power grid procurement documents.

5. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The calculation expression for the compliance assessment of electricity procurement demand in the electricity demand compliance assessment module is as follows: , Based on the results of the demand compliance assessment, This is the parameter set for the logic verification report. For the first Item logic verification report parameters, For the first Additional logic verification parameters, The set of parameters for identifying violations. For the first Parameters for identifying non-compliant clauses For the first Additional violation identification parameters, This is a set of parameters for power grid procurement requirements. For the first Parameters for power grid procurement requirements, For the first Additional procurement requirement parameters, To calibrate the number of parameters, To supplement the number of parameters, As a benchmark parameter for power grid demand assessment, This is a compliance assessment coefficient.

6. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The intelligent review and management calculation expression for power grid enterprise procurement needs in the power grid procurement review and management module is as follows: , In order to review the management results, For the set of parameters of compliance assessment results, For the first Parameters for compliance assessment of item calibration For the first Additional compliance assessment parameters To review the set of process parameters, For the first Parameters for item calibration and review process. For the first Additional review process parameters, For data archiving parameter set, For the first Item calibration data archiving parameters, For the first Additional data archiving parameters, To calibrate the number of management parameters, To supplement the number of management parameters, To review the management coefficients, These serve as benchmark parameters for power grid procurement review.

7. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The procurement document logic verification module includes a logic association extraction unit, a clause conflict detection unit, and a logic integrity verification unit. The logic association extraction unit receives the violation clause identification results output by the power grid violation clause identification module, and extracts the logic association information between clauses by analyzing the reference relationships, conditional relationships, and constraint relationships between clauses in combination with the content of each clause in the procurement document. The clause conflict detection unit, based on the extracted logic association information, compares the procurement requirements, technical parameters, and compliance standards involved in different clauses to identify contradictory, repetitive, and conflicting content between clauses. The logic integrity verification unit checks whether there are logical gaps, missing labeling clauses, and incoherent processes in the procurement document according to the specification requirements of the power grid procurement document, and generates a logic integrity verification result.

8. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The electricity demand compliance assessment module includes a demand parameter parsing unit, a compliance standard matching unit, a comprehensive assessment calculation unit, and an assessment result output unit. The demand parameter parsing unit receives the logic verification report generated by the procurement document logic verification module, and decomposes and analyzes the procurement demand quantity, technical indicators, budget amount, and delivery cycle parameters involved in the report to obtain standardized demand parameters. The compliance standard matching unit compares the standardized demand parameters with various standards in the power grid procurement compliance standard library to determine the compliance basis and standard threshold corresponding to the demand parameters. The comprehensive judgment and calculation unit combines the matched compliance standards and uses the algorithm of the power procurement demand compliance judgment platform to quantitatively calculate the compliance of the demand parameters. The judgment result output unit organizes the results obtained from the comprehensive judgment and calculation into a structured report and transmits it to the power grid procurement review and management module.

9. The intelligent review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in claim 1, is characterized in that... The power grid procurement review management module includes a review process scheduling unit, a real-time data monitoring unit, a review result archiving unit, and a historical data query unit. The review process scheduling unit receives the assessment results output by the power demand compliance assessment module, allocates review tasks to corresponding review nodes, and schedules review personnel and resources according to the process specifications of intelligent review management for power grid enterprise procurement needs. The real-time data monitoring unit tracks data transmission, parameter modifications, and review operations during the review process in real time, recording calibration data and time nodes. The review result archiving unit organizes the completed review results, relevant supporting materials, and process data according to preset classification standards and stores them in the corresponding database. The historical data query unit receives external query requests, retrieves corresponding historical review data from the database based on the query conditions in the request, and provides feedback to the querying party.

10. A smart review and management system for procurement requirements of power grid enterprises oriented towards multi-dimensional lean manufacturing, as described in any one of claims 1-9, characterized in that, The system operates in the following steps: S1. Receive procurement request data submitted by the power grid company, classify the data into three types: text, tables, and images, and transmit it to the multimodal procurement compliance verification module; S2. The multimodal procurement compliance verification module calls the multimodal procurement compliance deep verification model to extract features from the three types of data received, fuse the extracted features, perform preliminary compliance verification, and output the preliminary verification results; S3. The power grid violation clause identification module receives the preliminary verification results, calls the power grid procurement violation clause identification algorithm, compares the clause content in the preliminary verification results with the power grid violation clause database, identifies the violation clauses, and generates a list of violation clauses; S4. The procurement document logic verification module receives the violation... The terms and conditions list uses a logical self-consistency verification model for procurement documents to analyze the relationship between each clause in the procurement documents and the list of non-compliant clauses, check the logical relationships between clauses, and generate a logical verification report. The S5 power demand compliance assessment module receives the logical verification report and, through the power procurement demand compliance assessment platform, integrates the report data with basic parameters of power grid procurement demands to perform multi-dimensional calculations on the compliance of procurement demands, obtaining compliance assessment results. The S6 power grid procurement review and management module receives the compliance assessment results and, according to the intelligent review and management process for power grid enterprise procurement demands, reviews, controls, and archives the assessment results. The system collaborative control module integrates data from various modules, enabling collaborative operation of different modules to complete the intelligent review of procurement demands.

Images