Land cable project cost evaluation and examination method based on artificial intelligence

Abstract

The invention relates to the technical field of cable engineering, in particular to a land cable engineering cost evaluation and examination method based on artificial intelligence, which comprises the following steps: S1, constructing a land cable engineering cost structured model based on historical cost data and technical and economic expert experience; when in use, the system is applied to intelligent cost review work of a power transmission cable line project, a professional review standard is provided for technical and economic review work of the cable project, comprehensive and efficient review of cost, price, cost, materials, indexes and the like is carried out, automatic project cost review and auxiliary opinion output are realized, and the project cost review efficiency is improved. The complexity and time consumption of a cost review business process are reduced, the review quality efficiency of cable engineering is improved, a matching rule is determined through a fuzzy comprehensive evaluation method, the extraction accuracy and integrity of key cost data are greatly improved, manual misjudgment caused by field ambiguity and format non-uniformity is avoided, and the efficiency of cable engineering is improved. And the problem that the quota is not matched due to regional differences and design scheme changes can be solved.

Description

Technical Field

[0001] This invention relates to the field of cable engineering technology, and specifically to an artificial intelligence-based method for cost assessment and review of onshore cable engineering projects. Background Technology

[0002] With the continuous growth of investment in cable transmission line projects in Shanghai, the demand for improving the quality of deliverables in the engineering design phase is becoming increasingly urgent. Currently, the review and verification of cost estimates for cable projects still heavily relies on personnel experience. Due to the complexity of irregular structures such as manholes and cable trenches in cable projects, reviewers lack quantitative analysis and verification methods, resulting in high costs and low efficiency in the review process. This encroaches on a significant amount of time for optimizing and adjusting cost estimates, leading to problems such as incomplete coverage of some cost estimates and high time costs. In addition, while technical and economic personnel at the Shanghai Institute of Economic Research have accumulated a series of indicators and experiences in the verification of cost estimates over a long period, their personal experience has not been effectively used to create value, resulting in the loss of digital assets and failing to meet the requirements of precise cost control. To fully utilize and leverage the value of cost data and gradually build a leading position in infrastructure technical and economic expertise, the Shanghai Institute of Economic Research has decided to develop an intelligent review and verification tool for cable projects. This tool aims to comprehensively improve the accuracy and response speed of cable project cost reviews, effectively utilize and accumulate digital assets, promote digital transformation, and improve the quality and efficiency of technical and economic work.

[0003] The patent with publication number CN105321026A describes in its specification a method for risk assessment and management of power transmission and transformation project costs. The method includes the following steps: 1) Receiving historical sample data of the power transmission and transformation project and setting a weighting coefficient n based on the risk level; 2) Initializing the population size of the chaotic particle swarm to establish a chaotic particle swarm model, and optimizing the parameters of the chaotic particle swarm model according to the chaotic particle swarm optimization algorithm; 3) Determining the iteration number of the chaotic particle swarm model and the optimal value m of the chaotic factor based on the historical sample data and the optimized chaotic particle swarm model; 4) Determining the parameters: learning factors C1, C2, population size N, and evolution number N; 5) Operating the particles according to equations 1 and 2; if the maximum number of generations is reached or a satisfactory solution is obtained, the optimization process ends; otherwise, return to step 4); 6) Receiving actual sample data of the power transmission and transformation project and generating a cost assessment result for the power transmission and transformation project based on the actual sample data. The beneficial effect is that it not only allows investors to conduct feasibility studies in the early stages of project construction... While the aforementioned technology can accurately estimate the cost of new construction projects and assist budget reviewers in conducting reasonable and rapid cost reviews during the preliminary design phase, achieving the goal of solving the learning difficulties of small sample data and realizing dynamic risk-cost correlation assessment to adapt to market changes by combining chaotic particle swarm optimization and principal component analysis, the complex structure of cable engineering makes it difficult for manual reviewers to quickly identify and locate abnormal data in cost estimates. Traditional manual verification relies on manual checking, which is prone to inaccurate quantities and prices, affecting the professionalism of the review. The inconsistent recording methods and incomplete information in manual review lead to high communication costs between reviewers and compilers, and the records are difficult to digitize, affecting the continuity and traceability of subsequent review work. At the same time, the review indicators and typical problems accumulated by technical and economic personnel exist in an unstructured form, lacking a standardized problem classification system and structured database storage, making it difficult to reuse historical experience in the system and providing data support for the review of new projects, resulting in duplication of work and loss of experience.

[0004] In conclusion, developing an artificial intelligence-based method for cost assessment and review of land-based cable engineering remains a critical issue that urgently needs to be addressed in the field of cable engineering technology. Summary of the Invention

[0005] The purpose of this invention is to address the problems in existing technologies, such as the difficulty in quickly identifying and locating abnormal data in cost estimates due to the complexity of cable engineering structures and the challenges of manual review. Traditional manual verification relies on manual checking, which is prone to inaccurate quantities and prices, affecting the professionalism of the review. The inconsistent recording methods and incomplete information in manual review lead to high communication costs between reviewers and compilers, and the records are difficult to preserve as digital assets, affecting the continuity and traceability of subsequent review work. At the same time, the review indicators and typical problems accumulated by technical and economic personnel exist in unstructured form, lacking a standardized problem classification system and structured database storage. This makes it difficult to reuse historical experience in the system, failing to provide data support for the review of new projects, resulting in duplication of work and loss of experience.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This invention provides an artificial intelligence-based method for cost assessment and review of onshore cable engineering projects, comprising the following steps: S1. Based on historical cost data and the experience of technical and economic experts, construct a structured cost model for land-based cable engineering. S2. Based on the structured cost model of the land cable project, and according to the preset extraction criteria, output key data; S3. Based on the verification rules for land cable engineering, construct a dynamically updated auxiliary rule library for engineering verification and formulate automatic verification rules; S4. Import the project file to be reviewed, call the project verification auxiliary rule library, perform verification through multiple verification methods, and automatically generate a verification report.

[0007] Furthermore, in step S1, the method for constructing a structured cost model for land-based cable engineering based on historical cost data and the experience of technical and economic experts is as follows: The historical cost data is analyzed and categorized in three dimensions: by cost document type (including but not limited to estimates, preliminary estimates, and budgets), differences in data granularity are identified; and by content dimension (including but not limited to review issues and handling opinions), frequently occurring issues are extracted and associated data are extracted. The expression is: In the formula, Indicates the number of statistics The total number of times such high-frequency problems appear in historical cost data. This indicates the total number of historical data entries. This represents the index variable for iterating through each historical data record. The indicator function indicates the judgment of the first... Does this data belong to the first... High-frequency problems, Used to distinguish different high-frequency problems Indicate the problem With the question The strength of the association, This indicates that problems occurred simultaneously. and Number of data entries They represent the first Class, No. The total number of occurrences of high-frequency problems.

[0008] Furthermore, in step S1, the method for constructing a structured cost model for land-based cable engineering based on historical cost data and the experience of technical and economic experts is as follows: Based on the scope, including but not limited to industry quotas and local regulations, regional rule labels are assigned, and label diversity is evaluated using the entropy method. The expression is: In the formula, This indicates the diversity of regional rule labels. This indicates the total number of regional rule labels participating in the evaluation. The index variable representing the iteration over each label is used to calculate the probability and entropy contribution of each label individually. Indicates the first One tag, Indicates the first The probability of a label appearing among all labels It is a logarithmic function. Based on the three-dimensional classification analysis, a structured model of land cable engineering cost is constructed, which covers basic engineering information including but not limited to voltage level, laying method, quantity and price data including but not limited to list quantity, quota, main material price, other costs including but not limited to design fee, supervision fee, etc., and cost indicators including but not limited to unit length cost, main material proportion. At the same time, a cascading relationship is established between list item → associated quota → main material equipment. The data table is designed using a relational database, and cross-table association is realized through primary key / foreign key, and flexible query and validation of SQL semantics is supported.

[0009] Furthermore, in step S2, based on the structured cost model of the land-based cable project and according to preset extraction criteria, the method for outputting key data is as follows: The extraction criteria are formulated through a two-dimensional layer: a business rules layer and a technical logic layer. The business rules layer, based on the mapping relationship between the integrated feasibility study / preliminary design specification and the cost document, is expressed as follows: In the formula, The matrix element represents the first... The first feasibility study parameter and the second The mapping strength of each cost field, For the first Feasibility study parameters for each project The value, For the first The cost field of a project The value, The cosine similarity function is used. Taking samples from historical projects ,exist When, determine the parameters With fields The mapping relationship is used to output a mapping table, which locates the storage fields of parameters including but not limited to transport distance and soil ratio.

[0010] Furthermore, in step S2, based on the structured cost model of the land-based cable project and according to preset extraction criteria, the method for outputting key data is as follows: Based on pre-planned standards, including but not limited to the "Standards for Budget Preparation and Calculation of Power Grid Engineering Construction," a matching rule for typical modules and quota items is established. In land cable engineering, considering various factors such as different laying methods (including but not limited to direct burial, duct laying, and cable trench laying), different voltage levels (including but not limited to 10kV, 35kV, and 110kV), and different cable types (including but not limited to cross-linked polyethylene insulated cables and oil-paper insulated cables), different laying methods, material usage, voltage levels, and cable types are defined as a typical module. Each typical module corresponds to a quota item, and the matching degree is calculated using the fuzzy comprehensive evaluation method. ,expression: In the formula, Typical modules With quota items The degree of matching, These are typical module feature vectors. This is the feature vector for the quota sub-item. For the first The weights of each feature For the first The membership function for each feature uses Gaussian membership. The summation symbol indicates that from arrive In the cumulative summation, the technical logic layer parses the cost estimation documents using Python's Pandas library, matches key data fields using regular expressions, outputs a structured intermediate table, and performs null / extreme value checks on the extracted key data.

[0011] Furthermore, in step S3, a dynamically updated auxiliary rule base for engineering verification is constructed by combining the verification rules for land-based cable projects. The method for formulating automatic verification rules is as follows: The aforementioned engineering verification auxiliary rule library, based on industry pricing standards including but not limited to the "Construction Engineering Quantity List Pricing Standard" and regional management implementation rules, establishes a material price library and a benchmark fee template library, and connects to the government information price platform. The price series is updated quarterly using an index-weighted moving average, expressed as: In the formula, The first term represents the result after exponentially weighted moving average processing. The final price of a material in the period. This indicates that the price was obtained from the government information price platform. The material in the first The original market price of the period, Indicates the first The material in the first The price after the period has been processed by the index-weighted moving average. The learning rate range At the same time, differentiate between the fee structures for municipal and power grid projects.

[0012] Furthermore, in step S3, a dynamically updated auxiliary rule base for engineering verification is constructed by combining the verification rules for land-based cable projects. The method for formulating automatic verification rules is as follows: Based on typical regional design schemes, pre-planning standard project classification, and soil quality business relationships, a quota-materials-typical module association library is constructed, expressed as: In the formula, Indicates quota Supplies Typical modules The correlation strength values ​​among the three This is an indicator function that returns 1 if the condition is met and 0 if the condition is not met. It is used to determine the first... Does the quota / material / module in the data match the target value? Indicates the first Typical modules in historical data, Indicates the first Materials in historical data, Indicates the first Typical modules in historical data, These are the target matching values ​​for quotas, materials, and typical modules, used to filter combinations that meet the criteria from historical data. This indicates the total number of historical data entries. The dynamic time warping distance is used to measure the mismatch between quotas, materials, and modules in the time / process dimension. This indicates that the DTW distance is subjected to exponential decay processing. Based on the historical review data in the historical cost data, high-frequency problems are analyzed through association rule mining algorithms to form a typical historical problem library. Based on cost indicators of similar projects in the past 3 years, including but not limited to cost per unit length and main material ratio, the threshold is dynamically adjusted using the moving average method. If the threshold is exceeded, an intelligent warning is triggered. At the same time, a rule priority mechanism is established: industry standard rules > regional detailed rules > historical problem rules. A rule editing platform is also developed to support technical and economic experts to add / modify rules online.

[0013] Further, in step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The imported project files are parsed using Apache Tika to identify metadata including but not limited to project name, voltage level, and cost stage. Unidentified project files are automatically rejected with a warning. Multiple verification methods are executed sequentially. The first layer involves a material semantic matching algorithm to clean, integrate, and standardize the material data in the design drawings and cost documents. An LSTM network is used to parse the semantics of material names, combined with an XGBoost regression model to predict price ranges. ,expression: In the formula, It is the mathematical symbol for taking the minimum value. The summation symbol comes from Start to For each item Accumulate. This represents the total number of samples. The Huber loss function is used to measure the model's predicted values. and the true value The error between them It is the first The true price of each sample Indicates the first The price predicted from a sample Indicates the first The regularization values ​​of each base learner Indicates the first Individual base learners, It is the penalty coefficient for the number of leaf nodes. It is the total number of leaf nodes in the current base learner. It is the weighted squared penalty coefficient. The summation symbol comes from arrive right Accumulate. This represents the total number of parameters participating in another regularization. Indicates the first The output weights of each leaf node It is a scaling factor that calculates the semantic similarity of materials between the design drawings and the cost documents. An early warning is triggered when the similarity is less than 0.85. The 0.85 is set based on the statistical analysis of historical review data.

[0014] Further, in step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The process sequentially executes multiple verification methods. The second layer verifies the appropriateness of quota application and the consistency of unit quantities based on the quota association library. The appropriateness of quota application is determined by calculating the matching score between engineering features and quota sub-items, expressed as: In the formula, Indicate engineering features Quota Sub-items The degree of matching, This represents a vector composed of the specific features of the project currently under review. This represents a vector composed of the standard features of the quota sub-item. This represents the total number of features involved in the matching calculation. Indicates the first The importance of each feature in the matching rule, Indicate engineering features Quota Sub-items In the The degree of similarity in each feature Indicate engineering features The Middle A specific feature, Indicates the characteristics of quota sub-items The Middle A specific feature, in If the quota application is suspected to be inappropriate, the quota item corresponding to the second highest score will be recommended. The consistency of the unit quantity is calculated based on the conversion relationship between quota unit and design quantity. If the quantity deviation rate exceeds 5%, an early warning will be triggered, the reason for the deviation will be displayed simultaneously, the cost index identification algorithm will be called to extract the indicators of the project to be reviewed, and the regional benchmark indicators will be compared. If the comparison result exceeds 15%, an anomaly will be marked.

[0015] Further, in step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The process sequentially executes multiple verification methods. In the third layer, an ELM (Extreme Learning Machine) model is set up to extract abnormal feature vectors from a typical historical problem database. These abnormal feature vectors include, but are not limited to, quota numbers, material deviation rates, and indicators exceeding thresholds. The input layer of the ELM model loads these abnormal feature vectors. The expression for the number of hidden layer nodes in the abnormal feature vectors is as follows: In the formula, This represents the number of hidden layer nodes in the ELM (Extreme Learning Machine) model. For input dimensions, For the output dimension, according to The number of hidden layer nodes is dynamically optimized. The abnormal feature vector output layer calculates the abnormal probability value of the current project. When the probability exceeds 0.7, it is marked as abnormal. The 0.7 is determined by ROC curve analysis. The verification report includes abnormal location, reasons for additions and subtractions, and correction suggestions. It automatically generates an Excel format quantity and price comparison table and a Word format rectification suggestion report.

[0016] Beneficial effects Compared with known public technologies, the technical solution provided by this invention has the following beneficial effects: When in use, this invention is applied to the intelligent cost review of power transmission cable line projects, providing professional review standards for the technical and economic review of cable projects, conducting comprehensive and efficient verification of cost quantity, price, fees, materials, indicators, etc., realizing automated review of project costs and auxiliary output of opinions, reducing the complexity and time consumption of cost review business processes, and improving the quality and efficiency of cable project review.

[0017] When in use, this invention uses fuzzy comprehensive evaluation to legally match rules, which can significantly improve the accuracy and completeness of extracting key cost data, avoid human error caused by field ambiguity and inconsistent formats, and help solve the problem of incompatible quotas caused by regional differences and changes in design schemes. Through dynamic and real-time verification and early warning of material prices, fee templates and engineering features, it can effectively cope with market price fluctuations and regional engineering differences, and help identify quota and material matching problems in the engineering process, thereby improving verification accuracy. Attached Figure Description

[0018] Figure 1 This is a flowchart of an artificial intelligence-based method for cost assessment and review of land-based cable engineering projects according to the present invention. Detailed Implementation

[0019] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0020] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but includes other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] The present invention will now be described in further detail with reference to the accompanying drawings: Example: like Figure 1 As shown, this invention provides an artificial intelligence-based method for cost assessment and review of land-based cable engineering projects, comprising the following steps: S1. Based on historical cost data and the experience of technical and economic experts, construct a structured cost model for land-based cable engineering. Furthermore, in step S1, the method for constructing a structured cost model for land-based cable engineering based on historical cost data and the experience of technical and economic experts is as follows: The historical cost data is analyzed and categorized in three dimensions: by cost document type (including but not limited to estimates, preliminary estimates, and budgets), differences in data granularity are identified; and by content dimension (including but not limited to review issues and handling opinions), frequently occurring issues are extracted and associated data are extracted. The expression is: In the formula, Indicates the number of statistics The total number of times such high-frequency problems appear in historical cost data. This indicates the total number of historical data entries. This represents the index variable for iterating through each historical data record. The indicator function indicates the judgment of the first... Does this data belong to the first... High-frequency problems, Used to distinguish different high-frequency problems Indicate the problem With the question The strength of the association, This indicates that problems occurred simultaneously. and Number of data entries They represent the first Class, No. The total number of occurrences of high-frequency problems.

[0022] Furthermore, in step S1, the method for constructing a structured cost model for land-based cable engineering based on historical cost data and the experience of technical and economic experts is as follows: Based on the scope, including but not limited to industry quotas and local regulations, regional rule labels are assigned, and label diversity is evaluated using the entropy method. The expression is: In the formula, This indicates the diversity of regional rule labels. This indicates the total number of regional rule labels participating in the evaluation. The index variable representing the iteration over each label is used to calculate the probability and entropy contribution of each label individually. Indicates the first One tag, Indicates the first The probability of a label appearing among all labels It is a logarithmic function. Based on the three-dimensional classification analysis, a structured model of land cable engineering cost is constructed, which covers basic engineering information including but not limited to voltage level, laying method, quantity and price data including but not limited to list quantity, quota, main material price, other costs including but not limited to design fee, supervision fee, etc., and cost indicators including but not limited to unit length cost, main material proportion. At the same time, a cascading relationship is established between list item → associated quota → main material equipment. The data table is designed using a relational database, and cross-table association is realized through primary key / foreign key, and flexible query and validation of SQL semantics is supported.

[0023] In this embodiment, historical cost data is first analyzed in three dimensions. The data granularity is sorted out according to cost document types such as estimates and preliminary estimates. Using formulas that include indicator functions and correlation strength calculations, high-frequency problem-related data are extracted from content dimensions such as review issues. Then, regional rule labels are marked according to the coverage scope such as industry quotas. The diversity of labels is evaluated using the entropy method. Based on these analyses, a structured cost model for land cable engineering is constructed, covering basic engineering information such as voltage level, quantity and price data such as bill of quantities, other costs such as design fees, and indicators such as unit length cost. At the same time, a cascading relationship between bill of quantities items, related quotas, and main materials and equipment is established. Tables are designed using a relational database, and cross-table relationships are supported through primary keys / foreign keys. SQL query verification is supported, which facilitates the improvement of the structure and standardization level of engineering cost data.

[0024] S2. Based on the structured cost model of the land cable project, and according to the preset extraction criteria, output key data; Furthermore, in step S2, based on the structured cost model of the land-based cable project and according to preset extraction criteria, the method for outputting key data is as follows: The extraction criteria are formulated through a two-dimensional layer: a business rules layer and a technical logic layer. The business rules layer, based on the mapping relationship between the integrated feasibility study / preliminary design specification and the cost document, is expressed as follows: In the formula, The matrix element represents the first... The first feasibility study parameter and the second The mapping strength of each cost field, For the first Feasibility study parameters for each project The value, For the first The cost field of a project The value, The cosine similarity function is used. Taking samples from historical projects ,exist When, determine the parameters With fields The mapping relationship is used to output a mapping table, which locates the storage fields of parameters including but not limited to transport distance and soil ratio.

[0025] Furthermore, in step S2, based on the structured cost model of the land-based cable project and according to preset extraction criteria, the method for outputting key data is as follows: Based on pre-planned standards, including but not limited to the "Standards for Budget Preparation and Calculation of Power Grid Engineering Construction," a matching rule for typical modules and quota items is established. In land cable engineering, considering various factors such as different laying methods (including but not limited to direct burial, duct laying, and cable trench laying), different voltage levels (including but not limited to 10kV, 35kV, and 110kV), and different cable types (including but not limited to cross-linked polyethylene insulated cables and oil-paper insulated cables), different laying methods, material usage, voltage levels, and cable types are defined as a typical module. Each typical module corresponds to a quota item, and the matching degree is calculated using the fuzzy comprehensive evaluation method. ,expression: In the formula, Typical modules With quota items The degree of matching, These are typical module feature vectors. This is the feature vector for the quota sub-item. For the first The weights of each feature For the first The membership function for each feature uses Gaussian membership. The summation symbol indicates that from arrive In the cumulative summation, the technical logic layer parses the cost estimation documents using Python's Pandas library, matches key data fields using regular expressions, outputs a structured intermediate table, and performs null / extreme value checks on the extracted key data.

[0026] In this embodiment, extraction standards are constructed through both business and technical logic dimensions. The business rule layer, based on the mapping relationship between the feasibility study preliminary design and the cost document, uses a matrix containing cosine similarity to consider the mapping strength between feasibility study parameters and cost fields, adapts the sample size to historical data, and stores parameters such as location and transportation distance. At the same time, in accordance with the pre-planned standards, typical modules are defined by combining laying methods, voltage levels, and cable types, and associated with quota items. Through fuzzy comprehensive evaluation, legal matching rules are established, which can significantly improve the accuracy and completeness of key cost data extraction, avoid human misjudgment caused by field ambiguity and inconsistent formats, and help solve the problem of quota mismatch caused by regional differences and design changes.

[0027] S3. Based on the verification rules for land cable engineering, construct a dynamically updated auxiliary rule library for engineering verification and formulate automatic verification rules; Furthermore, in step S3, a dynamically updated auxiliary rule base for engineering verification is constructed by combining the verification rules for land-based cable projects. The method for formulating automatic verification rules is as follows: The aforementioned engineering verification auxiliary rule library, based on industry pricing standards including but not limited to the "Construction Engineering Quantity List Pricing Standard" and regional management implementation rules, establishes a material price library and a benchmark fee template library, and connects to the government information price platform. The price series is updated quarterly using an index-weighted moving average, expressed as: In the formula, The first term represents the result after exponentially weighted moving average processing. The final price of a material in the period. This indicates that the price was obtained from the government information price platform. The material in the first The original market price of the period, Indicates the first The material in the first The price after the period has been processed by the index-weighted moving average. The learning rate range At the same time, differentiate between the fee structures for municipal and power grid projects.

[0028] Furthermore, in step S3, a dynamically updated auxiliary rule base for engineering verification is constructed by combining the verification rules for land-based cable projects. The method for formulating automatic verification rules is as follows: Based on typical regional design schemes, pre-planning standard project classification, and soil quality business relationships, a quota-materials-typical module association library is constructed, expressed as: In the formula, Indicates quota Supplies Typical modules The correlation strength values ​​among the three This is an indicator function that returns 1 if the condition is met and 0 if the condition is not met. It is used to determine the first... Does the quota / material / module in the data match the target value? Indicates the first Typical modules in historical data, Indicates the first Materials in historical data, Indicates the first Typical modules in historical data, These are the target matching values ​​for quotas, materials, and typical modules, used to filter combinations that meet the criteria from historical data. This indicates the total number of historical data entries. The dynamic time warping distance is used to measure the mismatch between quotas, materials, and modules in the time / process dimension. This indicates that the DTW distance is subjected to exponential decay processing. Based on historical review data in the historical cost data, high-frequency issues are analyzed through association rule mining algorithms to form a typical historical issue database. Based on cost indicators of similar projects in the past three years, including but not limited to cost per unit length and main material ratio, a moving average method is used to dynamically adjust thresholds. Exceeding the threshold triggers intelligent warnings. At the same time, a rule priority mechanism is established: industry standard rules > regional detailed rules > historical issue rules. A rule editing platform is further developed to support online addition / modification of rules by technical and economic experts. The rules are automatically verified for the rationality of basic parameters, the rationality of the calculation principles of construction, installation, and demolition engineering costs, the correctness of price difference calculation principles, and whether the quota application is correct. Appropriateness rules, consistency rules for quotas and unit quantities of materials, consistency rules for material quantities, reasonableness rules for the source of material prices, rules for applying the latest local information prices for local materials, authenticity rules for quota base prices, reasonableness rules for other cost calculation principles, indicator review rules, rules for comparison at different stages, pre-review and post-review comparisons, and comparisons with typical projects, rules for comparison with control line indicators, and rules for comparison with typical projects; reasonableness rules for basic parameters, based on the imported data from the integrated design specification of feasibility study and preliminary design, identifying basic parameters such as transportation distance, soil ratio, and delivery distance, verifying the corresponding parameters of the project to be reviewed, and providing identification reminders; rationality rules for the calculation principles of construction, installation, and demolition engineering costs. Based on industry pricing standards and regional management implementation rules, a benchmark fee template library is established. Using this library, fee tables in the cost engineering documents of projects under review are compared and reviewed, with alerts provided for any anomalies. Rules for the correctness of price difference calculation principles are established, verifying the accuracy of adjustment coefficients in the cost engineering documents of projects under review based on adjustment coefficients published by the quota station. Rules for the appropriateness of quota application are established, constructing typical modules based on typical regional design schemes. A quota relationship library is established by analyzing the relationships between typical modules, materials, and quotas. A quota relationship library is also established based on the analysis of the relationship between pre-planned standard project classification and quotas, and based on the business relationship between engineering soil quality and quotas. The relational database, combined with the design data submission form, analyzes and reviews the application of quotas in the cost engineering documents of the project under review, and provides alerts for any anomalies. For the consistency rules of quotas and material unit quantities, based on the imported engineering analysis, it checks whether the quantities of typical modules, materials, and quota units in the cost engineering documents of the project under review are consistent, and provides alerts for any anomalies. For the consistency rules of material quantities, based on the imported data from the integrated feasibility study and preliminary design specifications, it identifies the codes, names, specifications, etc., of major materials, extracts material information from the cost engineering documents of the project under review, compares it with the design deliverables—material list, verifies whether the material quantities are consistent, and provides alerts for any anomalies.The rules for the reasonableness of material price sources stipulate that local materials shall be subject to the latest local information price rules. Relying on the material price database, the material price information in the cost engineering documents of the project to be reviewed shall be analyzed and compared with the selected information price and the agreed inventory price, and an alert shall be given for any abnormal content. The rules for the authenticity of quota base prices shall be verified for custom quotas, reference quotas, and replacement quotas, and an alert shall be given for any abnormal content. The details are precise down to the specifics of personnel, materials, and machinery; the reasonableness rules for calculating other costs are established based on industry pricing standards and regional management implementation rules, creating a benchmark fee template library. This library is used to compare and review the fee tables in the cost engineering documents of projects under review, providing alerts for any anomalies; the indicator review rules utilize statistical analysis of historical reviewed projects to construct regional technical evaluation indicators, analyze the cost engineering documents of projects under review, extract key indicators such as earthwork volume, foundation concrete volume, and unit project cost, and compare them with the regional benchmark indicators, providing alerts for any significant anomalies; rules for comparison at different stages, pre-review and post-review comparisons, and comparisons with typical projects are established, defining the verification scope: project investment (static investment, construction costs, installation costs, equipment purchase costs, etc.). Other expenses, site construction costs), project information, fee calculation rules, and project quantities are automatically verified using the following rules: Table 1 cost comparison provides alerts for anomalies (by building a related factor database and automatically analyzing the causes of deviations), supporting drill-down to Table 3 to view specific reasons; Rules for comparison with control line indicators analyze component information in the cost engineering documents of the project under review, establish a mapping rule library between these components and control lines, and based on the component mapping results (entire cost, static investment), compare with control line benchmark data to identify the comprehensive unit price of abnormal components; Rules for comparison with typical projects analyze component information in the cost engineering documents of the project under review, establish a mapping rule library between these components and typical projects, and based on the component mapping results (component unit price), compare with typical design benchmark data to identify the component unit price of abnormal components.

[0029] In this embodiment, based on industry pricing standards, such as the "Construction Engineering Quantity List Pricing Standard" and regional detailed rules, a material price database and a benchmark fee template database are established. This database is connected to the government information price platform, and prices are updated quarterly using index-weighted moving averages. Simultaneously, fees are differentiated between municipal and power grid projects. Furthermore, based on typical regional designs and pre-planning standards, a database linking quotas, materials, and typical modules is constructed. High-frequency issues are extracted from historical cost data to create a database of typical historical issues. Threshold warnings are adjusted using moving averages based on similar project indicators from the past three years. Rule priorities are set: industry standards > regional detailed rules > historical issues. An editing platform is developed to support expert rule modification, facilitating dynamic and real-time verification and warning of material prices, fee templates, and project characteristics. This effectively addresses market price fluctuations and regional project differences, helps identify quota and material matching issues in the project process, and improves verification accuracy. The rule editing platform supports continuous optimization and iteration by technical and economic experts, ensuring the continuous evolution and adaptability of the project verification auxiliary rule database.

[0030] S4. Import the project file to be reviewed, call the project verification auxiliary rule library, perform verification through multiple verification methods, and automatically generate a verification report; Further, in step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The imported project files are parsed using Apache Tika to identify metadata including but not limited to project name, voltage level, and cost stage. Unidentified project files are automatically rejected with a warning. Multiple verification methods are executed sequentially. The first layer involves a material semantic matching algorithm to clean, integrate, and standardize the material data in the design drawings and cost documents. An LSTM network is used to parse the semantics of material names, combined with an XGBoost regression model to predict price ranges. ,expression: In the formula, It is the mathematical symbol for taking the minimum value. The summation symbol comes from Start to For each item Accumulate. This represents the total number of samples. The Huber loss function is used to measure the model's predicted values. and the true value The error between them It is the first The true price of each sample Indicates the first The price predicted from a sample Indicates the first The regularization values ​​of each base learner Indicates the first Individual base learners, It is the penalty coefficient for the number of leaf nodes. It is the total number of leaf nodes in the current base learner. It is the weighted squared penalty coefficient. The summation symbol comes from arrive right Accumulate. This represents the total number of parameters participating in another regularization. Indicates the first The output weights of each leaf node It is a scaling factor that calculates the semantic similarity of materials between the design drawings and the cost documents. An early warning is triggered when the similarity is less than 0.85. The 0.85 is set based on the statistical analysis of historical review data.

[0031] Further, in step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The process sequentially executes multiple verification methods. The second layer verifies the appropriateness of quota application and the consistency of unit quantities based on the quota association library. The appropriateness of quota application is determined by calculating the matching score between engineering features and quota sub-items, expressed as: In the formula, Indicate engineering features Quota Sub-items The degree of matching, This represents a vector composed of the specific features of the project currently under review. This represents a vector composed of the standard features of the quota sub-item. This represents the total number of features involved in the matching calculation. Indicates the first The importance of each feature in the matching rule, Indicate engineering features Quota Sub-items In the The degree of similarity in each feature Indicate engineering features The Middle A specific feature, Indicates the characteristics of quota sub-items The Middle A specific feature, in If the quota application is suspected to be inappropriate, the quota item corresponding to the second highest score will be recommended. The consistency of the unit quantity is calculated based on the conversion relationship between quota unit and design quantity. If the quantity deviation rate exceeds 5%, an early warning will be triggered, the reason for the deviation will be displayed simultaneously, the cost index identification algorithm will be called to extract the indicators of the project to be reviewed, and the regional benchmark indicators will be compared. If the comparison result exceeds 15%, an anomaly will be marked.

[0032] Further, in step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The process sequentially executes multiple verification methods. In the third layer, an ELM (Extreme Learning Machine) model is set up to extract abnormal feature vectors from a typical historical problem database. These abnormal feature vectors include, but are not limited to, quota numbers, material deviation rates, and indicators exceeding thresholds. The input layer of the ELM model loads these abnormal feature vectors. The expression for the number of hidden layer nodes in the abnormal feature vectors is as follows: In the formula, This represents the number of hidden layer nodes in the ELM (Extreme Learning Machine) model. For input dimensions, For the output dimension, according to The number of hidden layer nodes is dynamically optimized. The abnormal feature vector output layer calculates the abnormal probability value of the current project. When the probability exceeds 0.7, it is marked as abnormal. The 0.7 is determined by ROC curve analysis. The verification report includes abnormal location, reasons for additions and subtractions, and correction suggestions. It automatically generates an Excel format quantity and price comparison table and a Word format rectification suggestion report.

[0033] In this embodiment, to import the project files to be reviewed and to conduct verification and report generation, Apache Tika is first used to parse the files and identify metadata such as the project name. If metadata cannot be identified, the file is rejected. Then, a three-layer verification is performed sequentially. The first layer uses a material semantic analysis and matching algorithm to clean and integrate material data. After processing by an LSTM network and an XGBoost regression model, semantic similarity is calculated, and a warning is issued if the similarity is below 0.85. The second layer relies on a quota association library to calculate the matching score between project features and quota items, calculate the quantity deviation rate, and compare cost indicators to verify the application of quotas and the consistency of unit quantities. The third layer uses an ELM (Extreme Learning Machine) model to extract abnormal feature vectors from a typical historical problem database, dynamically optimize the number of hidden layer nodes, calculate the probability of anomalies (a probability exceeding 0.7 is considered abnormal, and 0.7 is determined by ROC curve analysis), and finally generate a verification report containing anomaly location and other information. It outputs an Excel quantity-price comparison table and a Word rectification suggestion report, which facilitates the automation of the entire process from file parsing to verification, early warning, and report output. This helps to ensure the rationality of materials and prices from multiple dimensions, reduce missed reviews and erroneous reviews, and enhance dynamic adaptation and intelligent recognition capabilities by matching industry and regional benchmarks with historical data.

[0034] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for cost assessment and review of onshore cable engineering based on artificial intelligence, characterized in that, Includes the following steps: S1. Based on historical cost data and the experience of technical and economic experts, construct a structured cost model for land-based cable engineering. S2. Based on the structured cost model of the land cable project, and according to the preset extraction criteria, output key data; S3. Based on the verification rules for land cable engineering, construct a dynamically updated auxiliary rule library for engineering verification and formulate automatic verification rules; S4. Import the project file to be reviewed, call the project verification auxiliary rule library, perform verification through multiple verification methods, and automatically generate a verification report.

2. The method for cost assessment and review of land-based cable engineering based on artificial intelligence according to claim 1, characterized in that, In step S1, the method for constructing a structured cost model for land-based cable engineering based on historical cost data and the experience of technical and economic experts is as follows: The historical cost data is analyzed and categorized in three dimensions: by cost document type (including but not limited to estimates, preliminary estimates, and budgets), differences in data granularity are identified; and by content dimension (including but not limited to review issues and handling opinions), frequently occurring issues are extracted and associated data are extracted. The expression is: In the formula, Indicates the number of statistics The total number of times such high-frequency problems appear in historical cost data. This indicates the total number of historical data entries. This represents the index variable for iterating through each historical data record. The indicator function indicates the judgment of the first... Does this data belong to the first... High-frequency problems, Used to distinguish different high-frequency problems Indicate the problem With the question The strength of the association, This indicates that problems occurred simultaneously. and Number of data entries They represent the first Class, No. The total number of occurrences of high-frequency problems.

3. The method for cost assessment and review of land-based cable engineering based on artificial intelligence according to claim 2, characterized in that, In step S1, the method for constructing a structured cost model for land-based cable engineering based on historical cost data and the experience of technical and economic experts is as follows: Based on the scope, including but not limited to industry quotas and local regulations, regional rule labels are assigned, and label diversity is evaluated using the entropy method. The expression is: In the formula, This indicates the diversity of regional rule labels. This indicates the total number of regional rule labels participating in the evaluation. The index variable representing the iteration over each label is used to calculate the probability and entropy contribution of each label individually. Indicates the first One tag, Indicates the first The probability of a label appearing among all labels It is a logarithmic function. Based on the three-dimensional classification analysis, a structured model of land cable engineering cost is constructed, which covers basic engineering information including but not limited to voltage level, laying method, quantity and price data including but not limited to list quantity, quota, main material price, other costs including but not limited to design fee, supervision fee, etc., and cost indicators including but not limited to unit length cost, main material proportion. At the same time, a cascading relationship is established between list item → associated quota → main material equipment. The data table is designed using a relational database, and cross-table association is realized through primary key / foreign key, and flexible query and validation of SQL semantics is supported.

4. The method for cost assessment and review of land-based cable engineering based on artificial intelligence as described in claim 3, characterized in that, In step S2, based on the structured cost model of the land-based cable project, and according to the preset extraction criteria, the method for outputting key data is as follows: The extraction criteria are formulated through a two-dimensional layer: a business rules layer and a technical logic layer. The business rules layer, based on the mapping relationship between the integrated feasibility study / preliminary design specification and the cost document, is expressed as follows: In the formula, The matrix element represents the first... The first feasibility study parameter and the second The mapping strength of each cost field, For the first Feasibility study parameters for each project The value, For the first The cost field of a project The value, The cosine similarity function is used. Taking samples from historical projects ,exist When, determine the parameters With fields The mapping relationship is used to output a mapping table, which locates the storage fields of parameters including but not limited to transport distance and soil ratio.

5. The method for cost assessment and review of land-based cable engineering based on artificial intelligence according to claim 4, characterized in that, In step S2, based on the structured cost model of the land-based cable project, and according to the preset extraction criteria, the method for outputting key data is as follows: Based on pre-planned standards, including but not limited to the "Standards for Budget Preparation and Calculation of Power Grid Engineering Construction," a matching rule for typical modules and quota items is established. In land cable engineering, considering various factors such as different laying methods (including but not limited to direct burial, duct laying, and cable trench laying), different voltage levels (including but not limited to 10kV, 35kV, and 110kV), and different cable types (including but not limited to cross-linked polyethylene insulated cables and oil-paper insulated cables), different laying methods, material usage, voltage levels, and cable types are defined as a typical module. Each typical module corresponds to a quota item, and the matching degree is calculated using the fuzzy comprehensive evaluation method. ,expression: In the formula, Typical modules With quota items The degree of matching, These are typical module feature vectors. This is the feature vector for the quota sub-item. For the first The weights of each feature For the first The membership function for each feature uses Gaussian membership. The summation symbol indicates that from arrive In the cumulative summation, the technical logic layer parses the cost estimation documents using Python's Pandas library, matches key data fields using regular expressions, outputs a structured intermediate table, and performs null / extreme value checks on the extracted key data.

6. The method for cost assessment and review of land-based cable engineering based on artificial intelligence according to claim 5, characterized in that, In step S3, a dynamically updated auxiliary rule base for engineering verification is constructed by combining the verification rules for land-based cable projects. The method for formulating automatic verification rules is as follows: The aforementioned engineering verification auxiliary rule library, based on industry pricing standards including but not limited to the "Construction Engineering Quantity List Pricing Standard" and regional management implementation rules, establishes a material price library and a benchmark fee template library, and connects to the government information price platform. The price series is updated quarterly using an index-weighted moving average, expressed as: In the formula, The first term represents the result after exponentially weighted moving average processing. The final price of a material in the period. This indicates that the price was obtained from the government information price platform. The material in the first The original market price of the period, Indicates the first The material in the first The price after the period has been processed by the index-weighted moving average. The learning rate range At the same time, differentiate between the fee structures for municipal and power grid projects.

7. The method for cost assessment and review of land-based cable engineering based on artificial intelligence according to claim 6, characterized in that, In step S3, a dynamically updated auxiliary rule base for engineering verification is constructed by combining the verification rules for land-based cable projects. The method for formulating automatic verification rules is as follows: Based on typical regional design schemes, pre-planning standard project classification, and soil quality business relationships, a quota-materials-typical module association library is constructed, expressed as: In the formula, Indicates quota Supplies Typical modules The correlation strength values ​​among the three This is an indicator function that returns 1 if the condition is met and 0 if the condition is not met. It is used to determine the first... Does the quota / material / module in the data match the target value? Indicates the first Typical modules in historical data, Indicates the first Materials in historical data, Indicates the first Typical modules in historical data, These are the target matching values ​​for quotas, materials, and typical modules, used to filter combinations that meet the criteria from historical data. This indicates the total number of historical data entries. The dynamic time warping distance is used to measure the mismatch between quotas, materials, and modules in the time / process dimension. This indicates that the DTW distance is subjected to exponential decay processing. Based on the historical review data in the historical cost data, high-frequency problems are analyzed through association rule mining algorithms to form a typical historical problem library. Based on cost indicators of similar projects in the past 3 years, including but not limited to cost per unit length and main material ratio, the threshold is dynamically adjusted using the moving average method. If the threshold is exceeded, an intelligent warning is triggered. At the same time, a rule priority mechanism is established: industry standard rules > regional detailed rules > historical problem rules. A rule editing platform is also developed to support technical and economic experts to add / modify rules online.

8. The method for cost assessment and review of land-based cable engineering based on artificial intelligence according to claim 7, characterized in that, In step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The imported project files are parsed using Apache Tika to identify metadata including but not limited to project name, voltage level, and cost stage. Unidentified project files are automatically rejected with a warning. Multiple verification methods are executed sequentially. The first layer involves a material semantic matching algorithm to clean, integrate, and standardize the material data in the design drawings and cost documents. An LSTM network is used to parse the semantics of material names, combined with an XGBoost regression model to predict price ranges. ,expression: In the formula, It is the mathematical symbol for taking the minimum value. The summation symbol comes from Start to For each item Accumulate. This represents the total number of samples. The Huber loss function is used to measure the model's predicted values. and the true value The error between them It is the first The true price of each sample Indicates the first The price predicted from a sample Indicates the first The regularization values ​​of each base learner Indicates the first Individual base learners, It is the penalty coefficient for the number of leaf nodes. It is the total number of leaf nodes in the current base learner. It is the weighted squared penalty coefficient. The summation symbol comes from arrive right Accumulate. This represents the total number of parameters participating in another regularization. Indicates the first The output weights of each leaf node It is a scaling factor that calculates the semantic similarity of materials between the design drawings and the cost documents. An early warning is triggered when the similarity is less than 0.

85. The 0.85 is set based on the statistical analysis of historical review data.

9. The method for cost assessment and review of land-based cable engineering based on artificial intelligence as described in claim 8, characterized in that, In step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The process sequentially executes multiple verification methods. The second layer verifies the appropriateness of quota application and the consistency of unit quantities based on the quota association library. The appropriateness of quota application is determined by calculating the matching score between engineering features and quota sub-items, expressed as: In the formula, Indicate engineering features Quota Sub-items The degree of matching, This represents a vector composed of the specific features of the project currently under review. This represents a vector composed of the standard features of the quota sub-item. This represents the total number of features involved in the matching calculation. Indicates the first The importance of each feature in the matching rule, Indicate engineering features Quota Sub-items In the The degree of similarity in each feature Indicate engineering features The Middle A specific feature, Indicates the characteristics of quota sub-items The Middle A specific feature, in If the quota application is suspected to be inappropriate, the quota item corresponding to the second highest score will be recommended. The consistency of the unit quantity is calculated based on the conversion relationship between quota unit and design quantity. If the quantity deviation rate exceeds 5%, an early warning will be triggered, the reason for the deviation will be displayed simultaneously, the cost index identification algorithm will be called to extract the indicators of the project to be reviewed, and the regional benchmark indicators will be compared. If the comparison result exceeds 15%, an anomaly will be marked.

10. The method for cost assessment and review of land-based cable engineering based on artificial intelligence as described in claim 8, characterized in that, In step S4, the method for importing the project file to be reviewed, calling the project verification auxiliary rule base, performing verification through multiple verification methods, and automatically generating a verification report is as follows: The process sequentially executes multiple verification methods. In the third layer, an ELM (Extreme Learning Machine) model is set up to extract abnormal feature vectors from a typical historical problem database. These abnormal feature vectors include, but are not limited to, quota numbers, material deviation rates, and indicators exceeding thresholds. The input layer of the ELM model loads these abnormal feature vectors. The expression for the number of hidden layer nodes in the abnormal feature vectors is as follows: In the formula, This represents the number of hidden layer nodes in the ELM (Extreme Learning Machine) model. For input dimensions, For the output dimension, according to The number of hidden layer nodes is dynamically optimized. The abnormal feature vector output layer calculates the abnormal probability value of the current project. When the probability exceeds 0.7, it is marked as abnormal. The 0.7 is determined by ROC curve analysis. The verification report includes abnormal location, reasons for additions and subtractions, and correction suggestions. It automatically generates an Excel format quantity and price comparison table and a Word format rectification suggestion report.

Images