Power transmission and transformation project settlement method based on data collaboration and intelligent quota matching

Abstract

The invention provides a power transmission and transformation project settlement method based on data collaboration and intelligent quota matching, and the method comprises the steps: obtaining multi-dimensional data of a to-be-settled electric power project, and carrying out the preprocessing of the multi-dimensional data, and obtaining target multi-dimensional data; sequentially carrying out multi-dimensional division and dynamic fusion processing on the target multi-dimensional data to obtain a trusted project amount; the dynamic fusion processing comprises the step of performing dynamic fusion on the target multi-dimensional data by utilizing a weight evaluation model; extracting a target feature vector of the trusted project amount, performing settlement scheme matching on the target feature vector and a standard quota library, and obtaining a target settlement scheme in combination with the environment adjustment coefficient; and when the target settlement scheme accords with the preset conditions of grading auditing and settlement scheme evidence integrity evaluation, implementing the target settlement scheme. According to the scheme, the precision problem of an existing settlement method in the aspects of data consistency, change management, quota matching, risk control and the like is solved.

Description

Technical Field

[0001] This embodiment relates to the interdisciplinary field of power engineering management and information technology, and in particular to a settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching. Background Technology

[0002] With the expansion of the power grid and the advancement of the construction of new power systems, the number of power transmission and transformation projects has surged, leading to increasingly complex and sophisticated requirements for settlement management.

[0003] Traditional settlement models are increasingly revealing systemic shortcomings in modern engineering practice. First, settlement data sources are scattered, with inconsistent standards and formats across design models, construction records, supervision and acceptance documents, and material ledgers, leading to time-consuming and labor-intensive verification of quantities and frequent errors. Second, change order management lacks standardized procedures, resulting in problems such as non-standard change descriptions, incomplete supporting documentation, and unclear review authority, becoming a major source of settlement disputes. Third, quota matching relies excessively on manual experience, making it difficult to guarantee accuracy and consistency in the face of complex pricing rules and diverse environmental factors. In addition, risk control mechanisms are weak, with issues such as deviations in quantities and errors in cost accounting often only emerging at the settlement stage, leading to high rectification costs.

[0004] Although digital technologies such as 3D design and smart construction sites have been initially applied, existing technologies have not yet achieved deep integration with settlement operations, and information barriers still exist between various systems. These factors collectively restrict the improvement of settlement efficiency and quality, and there is an urgent need to establish a new settlement methodology system that is data-driven, intelligent in decision-making, and fully controllable. Summary of the Invention

[0005] This application provides a settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching, in order to solve the accuracy problems of existing settlement methods in terms of data consistency, change management, quota matching and risk control.

[0006] The method includes:

[0007] Acquire multi-dimensional data of the power project to be settled, and preprocess the multi-dimensional data to obtain target multi-dimensional data;

[0008] The target multi-dimensional data is sequentially divided into multiple dimensions and dynamically fused to obtain reliable engineering quantities; the dynamic fusion process includes dynamically fusing the target multi-dimensional data using a weight evaluation model;

[0009] Extract the target feature vector of the credible engineering quantity, match the target feature vector with the standard quota library for settlement schemes, and combine the environmental adjustment coefficient to obtain the target settlement scheme;

[0010] Once the target settlement plan meets the preset conditions for tiered review and assessment of the completeness of settlement plan evidence, the target settlement plan will be implemented.

[0011] Preferably, the process of implementing the target settlement scheme further includes:

[0012] When data changes occur in the multi-dimensional data of the power project to be settled, a new settlement plan is generated;

[0013] The regenerated settlement plan is compared with the target settlement plan to determine the deviation. If the regenerated settlement plan meets the deviation requirements, the regenerated settlement plan is implemented.

[0014] Preferably, the preprocessing includes:

[0015] Noise reduction and consistency verification are performed on all data in the multi-dimensional data in sequence.

[0016] Preferably, the multi-dimensional data includes three-dimensional design data, smart construction site perception data, construction ledger data, material management data, and supervision and acceptance data.

[0017] Preferably, the training process of the weight evaluation model includes:

[0018] Extract historical engineering case training samples from the historical engineering case library, and extract the training feature vectors of the engineering training samples;

[0019] Obtain historical settlement data, and use the historical settlement data and the training feature vector to train the weight evaluation model; the trained weight evaluation model has weight configuration.

[0020] Preferably, the step of dynamically fusing the target multi-dimensional data using a weighted evaluation model includes:

[0021] Import the target multi-dimensional data, after multi-dimensional segmentation, into the weight evaluation model;

[0022] By using the weight configuration, weights are assigned to each data point in the target multi-dimensional data that has been divided into multiple dimensions, resulting in several weighted data points.

[0023] The reliable engineering quantity is obtained by fusing all the weighted data.

[0024] Preferably, the engineering training samples include voltage level, professional classification, equipment parameters, construction method, and environmental label.

[0025] Preferably, the step of comparing the regenerated settlement scheme with the target settlement scheme includes:

[0026] Analyze the differences between the regenerated settlement scheme and the target settlement scheme to obtain difference information;

[0027] Based on the discrepancy information, the discrepancy items are sent to the corresponding responsible parties; the discrepancy items are the data items in the multi-dimensional data that have undergone data changes.

[0028] The responsible party shall review the discrepancies.

[0029] Preferably, when the regenerated settlement plan does not meet the deviation requirements, the confidence level of the regenerated settlement plan is calculated based on historical engineering cases and historical settlement data, and when the confidence level of the regenerated settlement plan meets the confidence level threshold, the engineering case corresponding to the regenerated settlement plan and the settlement plan are uploaded to the historical engineering case library.

[0030] Preferably, the noise reduction process includes:

[0031] Wavelet decomposition is performed on all data in the multi-dimensional data to obtain the first preprocessed data;

[0032] The first preprocessed data is subjected to noise filtering using an adaptive threshold to obtain the second preprocessed data;

[0033] The high-frequency coefficients of the second preprocessed data are quantized using a soft thresholding method to obtain quantized high-frequency coefficients;

[0034] The quantized high-frequency coefficients and the low-frequency coefficients in the second preprocessed data are reconstructed by wavelet inverse transform to obtain denoised data.

[0035] As described above, this application provides a settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching. The method includes acquiring multi-dimensional data of the power project to be settled, preprocessing the multi-dimensional data to obtain target multi-dimensional data, sequentially performing multi-dimensional division and dynamic fusion processing on the target multi-dimensional data to obtain reliable engineering quantities, and using a weighted evaluation model to dynamically fuse the target multi-dimensional data. The method also includes extracting target feature vectors from the reliable engineering quantities, matching these target feature vectors with a standard quota library to obtain a settlement scheme, and combining this with an environmental adjustment coefficient to obtain a target settlement scheme. Once the target settlement scheme meets the preset conditions for hierarchical review and settlement scheme evidence integrity assessment, the target settlement scheme is implemented. This application solves the accuracy problems of existing settlement methods in terms of data consistency, change management, quota matching, and risk control through the above solution. Attached Figure Description

[0036] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a flowchart illustrating a settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching, as proposed in this application.

[0038] Figure 2 Flowchart for adding new project settlement items;

[0039] Figure 3 This is a flowchart illustrating the training method of the weight evaluation model in the power transmission and transformation engineering settlement method based on data collaboration and intelligent quota matching, as described in this application.

[0040] Figure 4 This is a flowchart illustrating the dynamic integration of data collaboration and intelligent quota matching in the power transmission and transformation engineering settlement method proposed in this application.

[0041] Figure 5 This is a flowchart of the noise reduction process in a power transmission and transformation engineering settlement method based on data collaboration and intelligent quota matching, as described in this application. Detailed Implementation

[0042] The technical solutions of this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this embodiment, and not all embodiments. Based on the embodiments of this embodiment, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the protection scope of this embodiment.

[0043] It should be noted that the brief descriptions of terms in this application are only for the convenience of understanding the embodiments described below, and are not intended to limit the embodiments of this application. Unless otherwise stated, these terms should be understood in their ordinary and common meaning.

[0044] It should be noted that, in this application, the terms "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or illustration. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design solutions. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0045] Figure 1 This is a flowchart illustrating a settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching, as proposed in this application.

[0046] See Figure 1 As can be seen, this embodiment provides a settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching, the method including:

[0047] S100: Obtain multi-dimensional data of the power project to be settled, and preprocess the multi-dimensional data to obtain target multi-dimensional data.

[0048] Specifically, in this embodiment, 3D design data, smart construction site perception data, construction ledger data, material management data, and supervision and acceptance data are collected, and noise reduction processing and consistency verification are performed on the multi-source data.

[0049] Among them, see Figure 5 It can be seen that the noise reduction process includes:

[0050] S110, perform wavelet decomposition on all data in the multi-dimensional data to obtain the first preprocessed data;

[0051] S120, noise is filtered out from the first preprocessed data using an adaptive threshold to obtain the second preprocessed data;

[0052] S130, the high-frequency coefficients of the second preprocessed data are quantized using a soft thresholding method to obtain quantized high-frequency coefficients;

[0053] S140, the quantized high-frequency coefficients and the low-frequency coefficients in the second preprocessed data are reconstructed by wavelet inverse transform to obtain noise-reduced data.

[0054] The specific data collected includes component geometric dimensions, spatial coordinates, and material specifications; construction site location and image data of key processes; construction logs, acceptance records of sub-projects and items, acceptance records of concealed works, construction plans and technical briefing documents; equipment purchase contracts, factory certificates of conformity, on-site retest reports, requisition forms, and inventory records; supervision logs, on-site supervision records, supervision notices and replies, and project payment records.

[0055] In power transmission and transformation projects, IoT sensor data is susceptible to interference from strong electromagnetic fields, signal transmission packet loss, and transient sensor failures, resulting in pulse-like or high-frequency random noise. This noise manifests as "abrupt components" at the signal level.

[0056] Considering the characteristics of power transmission and transformation engineering data, which often contain non-stationary and abrupt components, the Daubechies (dbN) series of wavelets is preferred as the basis function. This series of wavelets can effectively capture the local features of the signal. Specifically, db4 or db5 is recommended for smoothly varying signals; db2 or db3 is recommended for signals containing many abrupt components. This embodiment uses the db4 wavelet as an example for illustration.

[0057] Step 1: Perform wavelet decomposition on the original data:

[0058]

[0059] Where a is the scaling parameter, which controls the scaling of the wavelet function; b is the translation parameter, which controls the position of the wavelet function; ψ(t) is the wavelet basis function; and x(t) is the original time series data.

[0060] Step 2: Apply an adaptive threshold to filter out noise from the processed data.

[0061]

[0062] Where σ is the noise standard deviation, which characterizes the fluctuation intensity caused by the above interference sources in the monitoring signal; MAD is the absolute median deviation; and α is an adjustment coefficient that is dynamically adjusted according to the data quality, with a range of [0.1, 0.5].

[0063] Step 3: After performing wavelet calculation on the signal and calculating the adaptive threshold T, the high-frequency coefficients obtained from the decomposition need to be quantized using a soft thresholding method.

[0064]

[0065] Where W(a,b) are the original coefficients obtained from wavelet decomposition; These are the estimated coefficients after threshold processing; T is the adaptive threshold obtained from the above calculation; sign(·) is the sign coefficient, used to preserve the positive or negative sign of the coefficients.

[0066] Step 4: Utilize the coefficients after thresholding. The denoised signal is reconstructed using inverse wavelet transform, along with the unprocessed low-frequency approximation coefficients.

[0067]

[0068] Among them, C ψ The wavelet admissibility constant, the value of which depends on the selected wavelet basis function ψ(t).

[0069] To visually demonstrate and verify the effectiveness of this noise reduction method, a simulation is conducted using raw time-series data collected by a GIS temperature sensor in an actual construction site. The unit is °C, the sampling frequency is 1 time / hour, and there are a total of 8 data points:

[0070] x(t)=[20.0, 20.1, 20.3, 19.5, 20.2, 20.1, 19.8, 20.0];

[0071] The fourth data point, 19.5, may be a mutation component caused by transient interference.

[0072] (1) The time series data is generally smooth. Using the db4 wavelet basis function, the above signal is decomposed into approximation coefficients and detail coefficients:

[0073] W(a,b)=[0.02,-0.01,0.05,-0.45,0.03,0.01,-0.02,0.01];

[0074] Among them, the absolute value of the fourth data point, -0.45, is much larger than that of the other data points, corresponding to the data point of 19.5 above;

[0075] (2) Calculate the adaptive threshold:

[0076]

[0077] We arrive at T≈0.312, where the MAD calculation rule is: first calculate the absolute value of the coefficients, then sort them and take the median;

[0078] (3) Perform soft threshold quantization on high-frequency coefficients:

[0079]

[0080] (4) Process the high-frequency detail coefficients The denoised signal is reconstructed by performing an inverse wavelet transform with the unprocessed low-frequency approximation coefficients.

[0081]

[0082] The reconstructed time series data is as follows:

[0083]

[0084] Further improvements to the above methods, the standardization of measurement standards for power transmission and transformation project settlement data, include: after cleaning the collected raw data, addressing the inconsistencies caused by differences in measurement tools, units of measurement, calculation rules, and statistical rules across different data sources. This necessitates the construction of a systematic measurement conversion and calibration engine. Its core is a multi-level conversion model that ensures all input data is standardized into a unified and comparable data format before entering the core calculation. The mathematical expression of this multi-dimensional measurement conversion model is:

[0085] V std =k s ·k u ·k m ·V src ;

[0086] Where, ks This is a scale conversion factor used to handle scale differences caused by variations in measurement accuracy or statistical dimensions. Its calculation is based on material properties and construction techniques.

[0087]

[0088] Where, α i β is the standard coefficient of the i-th type of loss; i Adjustment factors for on-site working conditions.

[0089] k u These are unit conversion factors, based on the unified conversion system of the International System of Units (SI):

[0090]

[0091] Among them, U t U is the target unit of measurement. s γ is the original unit of measurement, and γ is the environmental correction factor.

[0092] k m These are correction factors for measurement methods, used to correct for systematic errors in different measurement methods.

[0093]

[0094] Among them, w j R represents the weight of the j-th measurement method; j This represents the relative error coefficient between the various measurement methods.

[0095] To visually demonstrate and verify the effectiveness of this noise reduction method, a simulation was conducted using the concrete pouring volume of the foundation section in a power transmission project:

[0096] The theoretical volume of the design model is 50 cubic meters. The delivery note from the commercial concrete plant records a volume of 51.5 cubic meters. The construction unit's settlement statement reports a completed volume of 52.8 cubic meters, which includes operational losses and over-pouring.

[0097] (1) Normal operation loss (i=1) α1=1.5%, on-site operation standard β1=1; Over-excavation of foundation causing over-irrigation (i=2) α2=1%, according to the foundation over-excavation record confirmed by the supervisor, β2=1.5:

[0098]

[0099] To obtain k s =1.03, reasonable dosage 50 × 1.03 = 51.5m 3 <52.8, the amount of work settled by the construction unit has exceeded reasonable expectations;

[0100] (2) All data in this project are in the same unit, which is m. 3 There is no need to standardize the units of measurement;

[0101] (3) The construction unit estimated the project quantity (j=1) w1=0.4, and the estimation tended to be overestimated, with a relative error coefficient R1=1.03; the supervisor's spot check (j=2) w2=0.6, R2=0.995:

[0102]

[0103] To obtain k m =1.009;

[0104] (4) Calculate the data after the metering conversion process.

[0105] V std =k s ·k u ·k m ·V src ;

[0106] V std =54.88m 3 The above calculation results have significant errors compared to the designed engineering quantity and reasonable loss amount. The construction party needs to provide a reasonable explanation and key evidence for this.

[0107] Subsequently, a structured conversion coefficient database was established based on the "Construction Engineering Quantity List Pricing Specification", "Substation Engineering Quantity Calculation Specification" and "Transmission Line Engineering Quantity List Pricing Specification". The principles for determining each coefficient include: the material loss coefficient is based on the "Construction Engineering Material Consumption Quota"; and the conversion of each data unit quantity follows the "International System of Units and its Application".

[0108] Based on the aforementioned database, a coefficient optimization model based on machine learning is constructed:

[0109]

[0110] Where, k o The optimized conversion coefficients are used for standardization of measurement standards; k b λ represents the initial baseline transformation coefficients, determined based on specifications or historical data; λ is the learning rate, a hyperparameter controlling the step size for each update, determining the speed of model adjustment; V a,i V represents the historical project's actual engineering quantity for the i-th data point, derived from on-site measurements or a reliable source; e,i The estimated workload for the historical project at the i-th data point is calculated using the current conversion factor; N is the sample size.

[0111] The optimization coefficients k obtained after training oThese pre-trained coefficients will be used as new stability coefficients in the settlement data processing of subsequent new engineering projects. During the settlement process of the current project, the system will consistently use these pre-trained coefficients for standardized measurement, and will no longer dynamically modify these coefficients using the current project's data, thus ensuring the consistency and stability of the processing logic. Furthermore, after the current project's settlement is completed and finalized, its data can be tagged and stored in the historical database as a data reserve for future model retraining and version upgrades.

[0112] The method further includes:

[0113] S200, the target multi-dimensional data is sequentially divided into multiple dimensions and dynamically fused to obtain reliable engineering quantities; the dynamic fusion process includes dynamically fusing the target multi-dimensional data using a weight evaluation model;

[0114] S300, extract the target feature vector of the credible engineering quantity, match the target feature vector with the standard quota library for settlement scheme, and obtain the target settlement scheme by combining the environmental adjustment coefficient.

[0115] Specifically, in this embodiment, based on the verified data, the project is divided into multiple dimensions according to component type, engineering characteristics and environmental factors, and the engineering quantities reported by each data source are dynamically integrated using a weighted evaluation model to generate reliable engineering quantities.

[0116] See Figure 3 It can be seen that the training process of the weight evaluation model includes:

[0117] S010, Extract historical engineering case training samples from the historical engineering case library, and extract the training feature vector of the engineering training samples;

[0118] S020, acquire historical settlement data, and train the weight evaluation model using the historical settlement data and the training feature vector; the trained weight evaluation model has weight configuration.

[0119] See Figure 4 It can be seen that, further, in some embodiments, the step of dynamically fusing the target multi-dimensional data using a weighted evaluation model includes:

[0120] S210, import the target multi-dimensional data that has been divided into multiple dimensions into the weight evaluation model;

[0121] S220, Using the weight configuration, assign weights to each data in the target multi-dimensional data that has been divided into multiple dimensions, to obtain several weighted data;

[0122] S230, all the weight data are fused to obtain the reliable engineering quantity.

[0123] Specifically, in this embodiment, based on further improvements to the above method, the dynamic reliability assessment of power transmission and transformation project settlement data includes: dynamically calculating the weight of each data source in the final "reliable engineering quantity" by quantitatively assessing the inherent quality, equipment status, and timeliness of each data source, ensuring maximum reliability of the settlement basis. To assess the weight of each data source, a weight assessment model is constructed:

[0124]

[0125] Among them, w i R represents the reliability of the data source at the current moment. i To assess the health status of data acquisition equipment, this indicator is typically a value between 0 and 1, calculated based on equipment records, calibration logs, and real-time diagnostic status (such as battery voltage and signal strength).

[0126] R i =w bat ·S bat +w sig ·S sig +w cal ·S cal ;

[0127] Among them, S bat Rate the power supply status:

[0128]

[0129] Where V is the real-time power supply voltage, V norm The rated voltage, V min This is the minimum operating voltage for the equipment;

[0130] S sig Score the signal strength:

[0131] Where RSSI stands for Received Signal Strength Indicator. max Ideal signal strength;

[0132] S cal The calibration status score decays linearly based on the time since the last calibration:

[0133]

[0134] Where, ΔT cal T is the time since the last calibration. valid For calibration validity period;

[0135] wbat w sig w cal The weights of each indicator are such that their sum equals 1.

[0136] R max It is the highest reliability score among all similar devices in the system, used for normalization; Δt is the time interval from data acquisition to the current moment; τ is the decay time constant, set according to the evolution rate of the data type; Q i The data quality factor for the i-th data source is based on the historical performance of the data source.

[0137]

[0138] Where, σ i 2 This is the variance of the error between the reported data and the final verified value of the data source across multiple completed historical projects. A smaller variance indicates a more stable and reliable data source, and a higher quality factor Q. i Calculations and updates are performed during system initialization or periodic maintenance, and the system remains relatively stable over a settlement cycle. Q i The closer Q is to 1 (highest quality), the greater the variance, and the higher Q's value. i A value close to 0 indicates that the data source is noisy and unreliable.

[0139] Based on the above weighted evaluation model, the final reliable engineering quantity is synthesized, and its uncertainty is evaluated:

[0140]

[0141] Among them, V trust For the final credible engineering quantity; V i Report quantities for all data sources; δ conf Confidence interval correction value:

[0142]

[0143] Where k is the coverage factor, which takes different values ​​depending on the confidence level.

[0144] Further improvements to the above methods include three-dimensional visualization and discrepancy analysis of multi-source settlement data for power transmission and transformation projects. This involves setting up spatial deviation location and quantification algorithms to automatically identify and sort key deviation components in the project.

[0145]

[0146] Among them, P dev Ω represents the location of the largest deviation; Ω is the three-dimensional model space of the entire power transmission and transformation project, containing the set of all components to be analyzed; V trust(p) represents the reliable engineering quantity at spatial location p; V design (p) represents the engineering quantity of the design model at spatial location point p.

[0147] Based on the above deviation identification, the deviation analysis results are visualized:

[0148] (1) Components with deviation > 10% are marked with a red warning, and a work order with the highest priority is generated. The alarm is pushed to the project manager and relevant personnel for timely response and handling.

[0149] (2) Components with deviations of 5% < ≤ 10% are given an orange warning, a high-priority work order is generated, and they are discussed in detail in the regular meeting, and a rectification plan is formulated.

[0150] (3) Components with deviations of 2% to 5% will be given a yellow warning, recorded in the system, included in the periodic tracking list, and followed up by relevant departments.

[0151] (4) Components with a deviation of ≤2% are marked as green normal status, which is considered normal fluctuation and does not require special treatment.

[0152] The method further includes:

[0153] S400: When the target settlement plan meets the preset conditions for hierarchical review and settlement plan evidence integrity assessment, the target settlement plan is implemented.

[0154] Specifically, in this embodiment, based on further improvements to the above method, the hierarchical review and evidence scoring of power transmission and transformation engineering change orders includes: using structured representation of the change object, implementing hierarchical review paths, and combining Bayesian methods to calculate compliance scores, thereby achieving intelligent review and process optimization of change orders.

[0155] Figure 2 This is a flowchart for adding new project settlement items.

[0156] See Figure 2 It is understood that, furthermore, in some embodiments, the process of implementing the target settlement scheme also includes:

[0157] S500, when data changes occur in the multi-dimensional data of the power project to be settled, a settlement plan is regenerated;

[0158] S600: The regenerated settlement plan is compared with the target settlement plan to determine the deviation. If the regenerated settlement plan meets the deviation requirements, the regenerated settlement plan is implemented.

[0159] To achieve automated workflow and intelligent analysis of change management, this embodiment transforms unstructured change requests into readable and computable structured data objects, including the following core fields: change identifier, in the format of "project code-change sequence number"; change description; change amount; scope of impact, divided into local, system-level, and global according to the "Change Management of Power Transmission and Transformation Engineering"; change type, such as design change, construction process change, material substitution, etc.; change supporting materials; change submission time; and relevant responsible persons for the change.

[0160] Further improvements to the above method include designing a tiered review path for changes, based on the structured representation of the data and according to the change amount threshold and scope of impact, dividing the review path into three levels:

[0161] (1) Changes with a change amount ≤ 10,000 yuan and a localized impact are considered minor changes. On-site verification of quantity and price is required. Essential milestones include on-site confirmation by the supervisor and matching of the work quantity with the change. If the completion rate of this milestone is C... minor =1, then the review is approved.

[0162] (2) Changes with a change amount between RMB 10,000 and RMB 100,000, or changes affecting the system level, are classified as medium-level changes. These require additional professional confirmation steps, including confirmation by the supervisor and technical review and verification of quantity and price by professional technicians. Completion score:

[0163]

[0164] If C medium If the value is 1, then it passes; otherwise, manual intervention is required.

[0165] (3) Changes exceeding 100,000 yuan or affecting the entire system are considered major changes. Design approval documents and budget adjustment plans must be submitted. Essential steps include confirmation by the supervisor, technical review by professional technicians, approval from the design unit, budget adjustment documents, and project manager approval. Completion score:

[0166]

[0167] If C major =1, then pass; otherwise, return for correction.

[0168] The above completion rate calculation formula is:

[0169]

[0170] Among them, S i Let N represent the state of the i-th node (1 indicates completion, 0 indicates incomplete), and N represent the total number of nodes required for the current level change.

[0171] Further improvements to the above methods include the following for calculating compliance scores for changes in power transmission and transformation projects: using a Bayesian method to calculate compliance scores based on historical change approval rates and the credibility of on-site evidence, in order to quantify the probability of approval.

[0172] The formula for calculating the historical change approval rate is as follows:

[0173]

[0174] Similar changes are matched by change type, change amount, and scope of impact.

[0175] Formula for calculating the credibility of on-site evidence:

[0176] E = w1·E completeness +w2·E clarity +w3·E authority ;

[0177] Among them, E completeness The evidence completeness score is calculated using the following formula: E clarity E is the evidence clarity score, quantified by indicators such as image resolution and text readability, ranging from [0, 1]; authority This is a score representing the authority of the evidence, based on the level of authority of the source of the evidence; w i Each score has a weight, and its value is dynamically adjusted according to the project type.

[0178] Based on the above pass rate and credibility calculations, the compliance of the change is scored:

[0179]

[0180] The likelihood functions L(E|compliant) and L(E|non-compliant) are trained based on historical data:

[0181]

[0182] Where, μ c and The mean and variance of the credibility of historical compliance changes; μ r and The mean and variance of the credibility of historical non-compliant changes.

[0183] Based on the compliance score and review completion rate, the following procedure will be followed to determine the changes to power transmission and transformation projects:

[0184] (1) If the compliance score S < 0.5 or the audit completion rate C < 1, the system will automatically generate a list of items to be returned for correction and specify the deadline for supplementary certification.

[0185] (2) If the compliance score S < 0.8 and the audit completion rate C = 1, the change shall be sent to a professional manual review for expert review.

[0186] (3) If the compliance score S≥0.8 and the audit completion rate C=1, the system will automatically match the quota library and generate settlement items.

[0187] Further improvements to the above methods lead to the design of an intelligent matching and measurement rule application system for settlement quotas in power transmission and transformation projects. This system involves constructing a feature set for project items, associating it with environmental influencing factors, and using a machine learning model to achieve accurate and automatic quota matching. The final output is a structured recommendation result.

[0188] Each project item is characterized in a standardized and structured manner. The project item feature set includes the following core elements: project voltage level, which directly affects the selection of quota sub-items; professional letter category, which is divided according to the "Standard for Budget Compilation and Calculation of Power Grid Engineering Construction"; equipment model / section parameters, to refine quota selection; unit of measurement, which must be consistent with the unit in the quota library; construction method of each project item; project operation environment label, used to associate environmental adjustment coefficient; and final reliable project quantity.

[0189] The above feature set is stored and computed in vector form in the system:

[0190]

[0191] Among them, the environmental adjustment coefficient associated with the engineering operation environment label is dynamically matched through linkage with external databases: data such as slope and elevation of the engineering operation site are obtained from the GIS system to match the "terrain enhancement coefficient"; historical and forecast data are obtained from publicly available information from meteorological departments to match the "weather construction coefficient"; and salinity and air pollution levels are obtained from environmental monitoring departments to match the "corrosion prevention enhancement coefficient". Power transmission and transformation engineering construction is often affected by multiple environmental factors mentioned above. The comprehensive adjustment coefficient is calculated as follows:

[0192]

[0193] Where, β q The adjustment rate for the qth environmental factor is determined primarily based on various official documents issued by the China Electric Power Engineering Cost and Quota Management Center.

[0194] Further improvements to the above method include similarity matching between the feature vectors of power transmission and transformation projects and standard quotas, which involves calculating the matching degree between the feature vectors of project items and the feature vectors of quota sub-items in the quota library.

[0195]

[0196] Where K is the feature dimension; wk The weight of the Kth feature is automatically matched according to the "project type-voltage level" combination of the current project. Among them, the weight coefficient of component type and material specification is automatically increased for substation construction projects, while the weight ratio of environmental features such as terrain slope and crossing type is significantly increased for transmission line projects, and the dominant role of equipment model and cross-sectional parameters is highlighted for high-voltage cable laying projects, so as to reflect the different importance of different features to quota selection.

[0197] To assess the reliability of the matching results, a confidence level is calculated for the above matching degree:

[0198]

[0199] in, and These are the feature vectors of the quota sub-items with the highest and second highest matching degrees, respectively. The closer the confidence level Conf is to 1, the higher the reliability of the matching result.

[0200] When the unit of measurement for an engineering item is not entirely consistent with the unit in the quota library, a conversion function should be established:

[0201] V quota =g(V trust );

[0202] The function g operates by converting units such as length and weight based on the equipment model parameter library.

[0203] Based on further improvements to the above methods, the output of the quota intelligent matching and measurement rule system includes: recommended quota number, which is the quota sub-item code with the highest similarity among each project item; key parameters that are easy to match after being processed by the caliber conversion function; environmental adjustment coefficient; confidence score of matching the feature vector of each project item with the credible quantification; and possible abnormal information prompts.

[0204] Further improvements to the above methods include: establishing a systematic mechanism for identifying, tracking, attributing, and accumulating experience in power transmission and transformation engineering settlement discrepancies, effectively addressing the problems of low efficiency in discrepancy handling and the inability to effectively transfer advanced experience in the traditional settlement model for power transmission and transformation engineering.

[0205] Step 1: The system automatically compares the data of this settlement version with the previous benchmark version, identifying differences in quantities, costs, and related quota applications. For each settlement item, it calculates the absolute difference and relative difference rate of the quantities.

[0206] ΔV i =|V current,i -V baseline,i |;

[0207]

[0208] Where i represents each settlement entry; ΔV i The absolute difference between the two versions of settlement data for each settlement item; δ i V represents the relative difference rate between the two versions of settlement data for each settlement item. current,i This represents the current version's project size; V baseline,i This is the engineering quantity data from the previous baseline version.

[0209] Step 2: After clarifying the differences, automatically package the differences according to the professional classification and responsibility attribution rule matrix, and send them to the corresponding responsible parties in a timely manner.

[0210] Further improvements to the above methods include tracing the audit results of power transmission and transformation project settlements by constructing a complete version sequence for each settlement item.

[0211] EntryHistory=[Ver1,Ver2,...,Ver n ];

[0212] Each version Ver t Including t as the version generation time; V t This refers to the project volume for this version; P t This version's pricing parameters include key parameters such as quota number, unit price, and environmental adjustment factor; E t For settlement evidence, this includes related change order numbers, on-site video footage, and supervisor confirmation records; O t Those responsible for modifying or confirming this version.

[0213] Calculate the hash values ​​for each version to ensure that the settlement data is immutable:

[0214] Hash(Ver t ) = SHA-256(Hash(Ver) t-1 Content(Ver) t ));

[0215] Where || represents string concatenation, Content(Ver t ) is the core content of version t.

[0216] Further improvements to the above methods, including the identification of settlement issues in power transmission and transformation projects and the accumulation of relevant experience, include:

[0217] (1) Mining historical discrepancy data, extracting the problem feature vector for each closed discrepancy item:

[0218]

[0219] The information for each dimension is quantified using the following method: p 专业 One-hot encoding is used; p 责任方 Based on the output of the responsibility attribution rule matrix; p 环境因素 By associating the aforementioned environmental adjustment coefficients, environmental feature labels are extracted; p 构件类型 Identify specific component categories based on a component coding system; p 偏差类型 Keyword extraction is performed on the problem description using natural language processing. After constructing a bag-of-words model, the weight vector is calculated using the TF-IDF algorithm.

[0220] Step 1: Extract keywords from the relevant problem descriptions to grasp the core elements of the problem, specifically why what materials were used in the component, what problems occurred, and what consequences resulted. Remove duplicates to create a unique bag of terms containing a large number of engineering terms.

[0221] Step 2: For any newly emerging or historically existing problem, generate a vector based on the bag-of-words. The generation method is as follows: Check each word in the bag-of-words against the specific problem description; if a keyword is in the bag-of-words and appears in the specific problem description, mark the corresponding position as "1"; otherwise, if a word in the bag-of-words does not appear in the problem description, mark it as "0". Thus, each problem described in natural language is transformed into a fixed-length numerical sequence consisting of a series of "1"s and "0"s. This sequence precisely characterizes "which keywords a problem contains," while temporarily ignoring the order and grammatical relationships of these words.

[0222] Step 3: Evaluate the weight of each keyword, converting the "0"s and "1"s in the above sequence into meaningful weight scores. The weight of a keyword is directly proportional to the number of times it appears in the current question and inversely proportional to the number of times it appears in all questions. The formula for calculating the weight of each keyword is:

[0223] TF-IDF = TF × IDF;

[0224] Where TF stands for term frequency, which is the frequency with which the keyword appears in the question description:

[0225]

[0226] IDF stands for Inverse Document Frequency, which represents how rare a keyword is in the bag of words.

[0227]

[0228] The impact of the amount is normalized:

[0229]

[0230] Where, ΔC i ΔC represents the absolute value of the cost impact of this difference item. max This represents the maximum cost impact of similar differences in the current project.

[0231] Frequency of occurrence refers to the relative frequency of this problem pattern in historical projects:

[0232]

[0233] (2) For settlement processing solutions without problems, they are encapsulated as structured knowledge items, including a list of key evidence points, a standard processing action sequence, an applicable condition matrix, and an effect evaluation index.

[0234] The evidence checklist outlines the evidentiary materials that must be collected to resolve such issues, along with their required attributes.

[0235] E req ={e1,e2,...,e n};

[0236] Among them, each piece of evidence e i It includes the type of evidence, required fields, and format requirements.

[0237] The standard processing sequence is the standard operating procedure for solving this type of problem:

[0238] A seq =[a1,a2,...,a m ];

[0239] Among them, each processing action a i It includes the action type, the role to be executed, the parameter requirements, and the expected output.

[0240] The applicability condition matrix clarifies the scope of application for this knowledge item:

[0241] C apply =[c min,1 ,c max,1 ]×[c min,2 ,c max,2 ]×…×[c min,k ,c max,k ];

[0242] Each dimension corresponds to a key feature of the feature vector.

[0243] The effectiveness evaluation metric is the historical application effectiveness of this knowledge item:

[0244] P eff = {Number of successful applications, average resolution time, average user rating};

[0245] (3) Intelligent knowledge push is achieved through similarity calculation, and the knowledge base is continuously optimized through feedback mechanism.

[0246] Step 1: Match the feature vector of the new settlement project with the feature vector of historical questions in the knowledge base, and calculate the multi-dimensional similarity:

[0247]

[0248] in, For the feature vector of the new settlement item, w is the feature vector of the historical problem. k denoted as the weight of each feature k.

[0249] Step 2: Calculate the confidence level for each push:

[0250]

[0251] Where Δt is the time interval since the last successful application, T half The half-life of knowledge.

[0252] Step 3: Dynamically update the utility value of each knowledge item based on usage feedback:

[0253]

[0254] Among them, R i The user is rated, τ is the decay time constant, and λ is the learning rate.

[0255] (4) When going beyond matching a single problem, construct a knowledge graph of settlement problems.

[0256] Step 1: Construct a triplet relationship network: causal relationship, problem A leads to problem B; concurrency relationship, problem A and problem B occur simultaneously; disposal dependency relationship, solving problem A is a prerequisite for solving problem B.

[0257] Step 2: When a specific problem is identified, related problems are discovered by traversing the graph:

[0258] P related ={p j |Path(p i ,p j )≤D max};

[0259] Wherein, Path(p i ,p j D represents the shortest path length between two nodes. max This represents the maximum association depth.

[0260] Further improvements to the above methods include the following: the risk classification and closed-loop completion determination for power transmission and transformation engineering settlement include: constructing a differential risk classification system based on multi-dimensional quantitative rules and a clear dual verification closed-loop mechanism to achieve accurate identification of settlement process risks and efficient convergence of the handling process.

[0261] Based on the aforementioned reliable quantity fusion and settlement version comparison results, each identified discrepancy item is subject to risk quantification and classification. The system simultaneously calculates the relative quantity discrepancy rate and the absolute value of the cost impact for each discrepancy item.

[0262]

[0263] Where, δ i V represents the relative difference rate of the quantities of the differential items; c,i For the i-th settlement item, the quantity of work declared or confirmed by the system in the current settlement version is derived from the aforementioned reliable quantity of work; V b,i For the i-th settlement entry, the quantity of the previous settlement version used as the comparison benchmark.

[0264] ΔC i =|C c,i -C b,i |;

[0265] Where, ΔC i C represents the absolute value of the impact of the difference item costs; c,i C represents the total cost of the current settlement version in the i-th settlement entry; b,i For the i-th settlement entry, the total cost of the previous settlement version used as the comparison benchmark.

[0266] Based on the calculation of the relative difference rate of the above-mentioned quantities and the absolute value of the cost impact, the difference items are divided and classified for handling, as shown in the table below.

[0267] Table 1

[0268]

[0269] Based on the above calculations of the relative difference rate of project quantities and the absolute value of cost impact, the difference closure judgment is initiated: each difference item must pass both model validation and rule validation simultaneously before it can be determined that the closure is complete.

[0270] Model validation focuses on the stability of the data state. This difference term must show a continuous decrease in the relative difference rate of the workload over two consecutive periods, and the δ value for the current period must be... i ≤2%;

[0271] Rule validation includes the following necessary conditions:

[0272] (1) The relative difference rate δ of the quantity of work in the current version for this difference itemi ≤2%;

[0273] (2) The final cost change resulting from the handling of this difference item has been approved, and the relative deviation from the latest estimated cost is ≤2%:

[0274]

[0275] Among them, C f,i For the approved final cost changes, C c,i This is the latest estimated cost;

[0276] (3) All supporting evidence documents necessary for verifying all such differences have been uploaded, and the completeness of the evidence chain is ≥95%;

[0277] (4) In the process of handling discrepancies, all pre-set responsible party nodes have completed their signatures, with a signature completion rate of 100%.

[0278] The system will automatically update the status of the difference item to "closed loop" only if all the conditions for model validation and rule validation mentioned above are met.

[0279] For ease of explanation, the above description has been provided in conjunction with specific embodiments. However, the discussion in some embodiments is not intended to be exhaustive or to limit the embodiments to the specific forms disclosed above. Various modifications and variations can be obtained based on the above teachings. The selection and description of the above embodiments are for the purpose of better explaining the contents of this disclosure, thereby enabling those skilled in the art to better utilize the embodiments.

Claims

1. A settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching, characterized in that, The method includes: Acquire multi-dimensional data of the power project to be settled, and preprocess the multi-dimensional data to obtain target multi-dimensional data; The target multi-dimensional data is sequentially divided into multiple dimensions and dynamically fused to obtain reliable engineering quantities; the dynamic fusion process includes dynamically fusing the target multi-dimensional data using a weight evaluation model; Extract the target feature vector of the credible engineering quantity, match the target feature vector with the standard quota library for settlement schemes, and combine the environmental adjustment coefficient to obtain the target settlement scheme; Once the target settlement plan meets the preset conditions for tiered review and assessment of the completeness of settlement plan evidence, the target settlement plan will be implemented.

2. The settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 1, characterized in that, The process of implementing the target settlement scheme also includes: When data changes occur in the multi-dimensional data of the power project to be settled, a new settlement plan is generated; The regenerated settlement plan is compared with the target settlement plan to determine the deviation. If the regenerated settlement plan meets the deviation requirements, the regenerated settlement plan is implemented.

3. The settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 1, characterized in that, The preprocessing includes: Noise reduction and consistency verification are performed on all data in the multi-dimensional data in sequence.

4. The settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 1, characterized in that, The multi-dimensional data includes 3D design data, smart construction site perception data, construction ledger data, material management data, and supervision and acceptance data.

5. The settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 1, characterized in that, The training process of the weight evaluation model includes: Extract historical engineering case training samples from the historical engineering case library, and extract the training feature vectors of the engineering training samples; Obtain historical settlement data, and use the historical settlement data and the training feature vector to train the weight evaluation model; the trained weight evaluation model has weight configuration.

6. The settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 5, characterized in that, The step of dynamically fusing the multi-dimensional data of the target using a weighted evaluation model includes: Import the target multi-dimensional data, after multi-dimensional segmentation, into the weight evaluation model; By using the weight configuration, weights are assigned to each data point in the target multi-dimensional data that has been divided into multiple dimensions, resulting in several weighted data points. The reliable engineering quantity is obtained by fusing all the weighted data.

7. The settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 5, characterized in that, The engineering training samples include voltage levels, professional categories, equipment parameters, construction methods, and environmental labels.

8. The settlement method for power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 2, characterized in that, The step of comparing the regenerated settlement scheme with the target settlement scheme includes: Analyze the differences between the regenerated settlement scheme and the target settlement scheme to obtain difference information; Based on the discrepancy information, the discrepancy items are sent to the corresponding responsible parties; the discrepancy items are the data items in the multi-dimensional data that have undergone data changes. The responsible party shall review the discrepancies.

9. A method for settling power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 2, characterized in that, When the regenerated settlement plan does not meet the deviation requirements, the confidence level of the regenerated settlement plan is calculated based on historical engineering cases and historical settlement data. When the confidence level of the regenerated settlement plan meets the confidence level threshold, the engineering case corresponding to the regenerated settlement plan and the settlement plan are uploaded to the historical engineering case library.

10. A method for settling power transmission and transformation projects based on data collaboration and intelligent quota matching according to claim 3, characterized in that, The noise reduction process includes: Wavelet decomposition is performed on all data in the multi-dimensional data to obtain the first preprocessed data; The first preprocessed data is subjected to noise filtering using an adaptive threshold to obtain the second preprocessed data; The high-frequency coefficients of the second preprocessed data are quantized using a soft thresholding method to obtain quantized high-frequency coefficients; The quantized high-frequency coefficients and the low-frequency coefficients in the second preprocessed data are reconstructed by wavelet inverse transform to obtain denoised data.

Images