Power transmission and transformation project special cost standard determination method based on multi-attribute clustering

Abstract

The invention discloses a power transmission and transformation project special cost standard determination method based on multi-attribute clustering, and relates to the technical field of power transmission and transformation, and the method comprises the following steps: classifying power transmission and transformation projects, and dividing different types of power transmission and transformation project special cost types; acquiring historical data of the power transmission and transformation project, and processing missing values and abnormal values according to the historical data; identifying influence factors of different types of power transmission and transformation project special expenses, and respectively judging the correlation degree between the fluctuation change trend of different expense types and the influence factors in combination with historical data; formulating a special cost standard according to the degree of association between the fluctuation change trend and the influence factors of the special cost of the power transmission and transformation project; and constructing a cost dynamic updating mechanism, and dynamically updating and revising the cost standard in combination with the difference between the actual project and the standard. Through deep analysis of historical data, an adversarial network technology is adopted to carry out abnormal value interpolation, so that the quality and integrity of the data are improved, and the scientificity of cost calculation is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of project data statistical analysis, and more specifically, to a method for determining special cost standards for power transmission and transformation projects based on multi-attribute clustering. Background Art

[0002] In the field of power transmission and transformation projects, the standardization of special expenses is a key link to ensure the economy and controllability of the project. As cost audits become stricter and stricter, higher requirements are placed on cost control. As an important part of the cost, the accuracy of special expenses has an important impact on investment pricing. With the continuous growth of electricity demand and the rapid development of technology, traditional cost standard formulation methods have been difficult to adapt to the complex and changing market environment. These traditional methods often rely on experience or simple statistical models, and it is difficult to fully consider the comprehensive effects of project characteristics, environmental impacts and other variables, resulting in insufficient accuracy in cost forecasting, which in turn affects project decision-making and resource allocation.

[0003] At present, the cost structure of power transmission and transformation projects mainly includes direct costs, labor costs, material costs, business costs and indirect costs, etc. These costs are affected by many factors, such as market fluctuations, seasonal changes, project scale and construction location. The interaction of these factors has formed a complex cost structure, making the scientific determination of cost standards particularly important. Due to the different characteristics of different types of projects, the use of a unified cost standard often cannot effectively reflect the actual cost, which in turn leads to budget overruns and waste of resources.

[0004] In order to improve the accuracy of cost forecasting, modern data analysis technology has been gradually introduced. Multi-attribute cluster analysis, as a data mining technology, can classify data according to multiple features, thereby revealing the cost characteristics and influencing factors of different project types. However, traditional clustering methods may face the problems of dimensional disaster and information loss when processing high-dimensional data, making it difficult to effectively extract useful information. In addition, the dynamic nature of costs also requires that cost standards have the ability to be updated in real time to adapt to changes in the market environment.

[0005] In this context, the technology for determining special cost standards for power transmission and transformation projects based on multi-attribute clustering came into being. This technology can comprehensively evaluate the fluctuation trends and influencing factors of various costs by classifying project types and combining historical data and real-time data analysis, providing strong support for the scientific determination of cost standards. At the same time, by building a dynamic update mechanism, the timeliness and adaptability of cost standards can be ensured.

[0006] In addition, the rapid development of artificial intelligence technology in recent years has also provided new ideas for cost forecasting. As a powerful nonlinear model, neural networks can handle complex input-output relationships and achieve efficient cost forecasting by learning potential patterns in large amounts of data. By combining genetic algorithms to optimize neural network models, the optimal solution can be found in parameter configuration to improve the model's prediction accuracy and stability. In the forecasting of indirect costs, the ARIMA model has become a popular analysis tool because of its powerful time series analysis capabilities and its ability to effectively capture trends and seasonal changes in data. This model provides a scientific basis for future budgeting and cost control by extracting patterns from historical data.

[0007] In summary, the current formulation of power transmission and transformation project cost standards urgently needs to adopt a diversified and systematic approach, combining data analysis and artificial intelligence technology to cope with complex project characteristics and market demands. The special cost standard determination technology based on multi-attribute clustering is a beneficial supplement to the traditional method, aiming to improve the scientificity and accuracy of cost standard formulation, thereby providing support for the sustainable development of power transmission and transformation projects.

[0008] The above disclosed technical solutions have at least the following technical problems: the traditional clustering method may face the problem of dimensional disaster and information loss when processing high-dimensional data, and it is difficult to effectively extract useful information. At the same time, the fee standard does not have the ability to be updated in real time and cannot adapt to changes in the market environment. In view of the above problems, the present invention proposes a solution. Summary of the invention

[0009] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for determining special fee standards for power transmission and transformation projects based on multi-attribute clustering. Through the ARIMA model, the problem of dimensional disaster and information loss that traditional clustering methods may face when processing high-dimensional data is solved, and the problem of difficulty in effectively extracting useful information is solved.

[0010] To achieve the above object, the present invention provides the following technical solutions: A method for determining special expense standards for power transmission and transformation projects based on multi-attribute clustering includes the following steps: classifying power transmission and transformation projects, and dividing special expense categories for power transmission and transformation projects of different categories; obtaining historical data of power transmission and transformation projects, and processing missing values ​​and abnormal values ​​based on the historical data; identifying influencing factors of special expenses for different types of power transmission and transformation projects, and combining statistical analysis of historical data to determine the correlation between the fluctuation trend of different expense types and the influencing factors; formulating special expense standards based on the correlation between the fluctuation trend and the influencing factors of special expenses for power transmission and transformation projects; and constructing a dynamic expense update mechanism to dynamically update and revise the expense standards based on the differences between actual projects and standards.

[0011] In a preferred embodiment, the missing value and outlier processing based on historical data is specifically as follows: obtaining the historical data of power transmission and transformation projects and the influencing factors of special expenses of power transmission and transformation projects, and detecting the outliers in the historical data and influencing factors by statistical methods; designing and constructing a generative adversarial network suitable for the historical data and influencing factors of power transmission and transformation projects, including the structure and parameter settings of the generator and the discriminator; using normal data to train the adversarial network, inputting the detected outliers into the trained generator, and replacing the outliers with the generated new data.

[0012] In a preferred embodiment, the detected outliers are input into a trained generator, and the generated new data are used to replace the outliers, specifically: a reconstructed sample for adversarial purposes is generated by a generator; a distance between the adversarial sample and the original sample is distinguished by a discriminator, wherein the original sample is any sample data of the historical cost data and the influencing factor data; a reconstruction loss function is used to reduce the distance between the reconstructed sample and the original sample, and a discriminant loss function is used to maximize the output probability of the discriminator; a filling loss function is generated according to the reconstruction loss function and the discriminant loss function, and the filling loss function is minimized based on the reconstructed sample and the original sample to generate missing values ​​for interpolation.

[0013] In a preferred embodiment, the degree of correlation between the fluctuation trend of different types of expenses and the influencing factors is determined by an improved anti-entropy-grey correlation analysis method, specifically: calculating the entropy value of each of the multiple key influencing factors; converting the entropy value of each key influencing factor into an anti-entropy value through an anti-entropy formula; generating an objective weight based on the anti-entropy value, and calculating the combined weight based on the objective weight and the subjective weight; calculating the correlation coefficient between each comparison sequence and the reference sequence based on the combined weight, selecting special expense data as the reference sequence, and taking multiple key influencing factors as multiple comparison sequences; calculating the correlation between the reference sequence and each comparison sequence based on the correlation coefficient as the correlation between the expense input and each key influencing factor.

[0014] In a preferred embodiment, the special cost standard is formulated according to the correlation between the fluctuation trend and the influencing factors of the special cost of the power transmission and transformation project, and the special cost of the power transmission and transformation project includes labor costs, material costs, business costs, indirect costs and direct costs; wherein, the cost standards of labor costs, material costs and business costs are formulated, specifically: each cost standard is confirmed by a neural network model, Genetic algorithm is used to optimize the neural network model to obtain the cost standards of labor costs, material costs, and business expenses; the standards for indirect costs are obtained by analyzing using time series analysis and the autoregressive integrated moving average (ARIMA) model; the standards for indirect costs are determined by using the multivariate linear regression method.

[0015] In a preferred embodiment, the cost standards for labor costs, material costs, and business costs are formulated as follows: based on the design characteristics of the bp neural network, the mutual influence relationship between the various factors is not considered, and the main influencing factors are selected as the input layer in the cost input; the target value is selected by the method of processing data in the horizontal and vertical directions as the output target value; a model is established based on a genetic algorithm and the neural network model is optimized; based on MATLAB software, the existing data on the maintenance and operation costs of substation equipment are used as samples, and the GA-BP neural network is used for training; the training data results are verified, and the K-fold cross-validation method is adopted to determine the accuracy of the model, wherein when the model results are tested, the reliability and validity of the model are considered, the reliability refers to the reliability of the model, and the validity refers to the effectiveness of the model.

[0016] In a preferred embodiment, the genetic algorithm-based model is established and the neural network model is optimized, specifically: the population is initialized using GA to obtain the initial weights and thresholds of the BP neural network; the individuals after the population is initialized are substituted into the BP neural network to iteratively search for the optimal initial weights and thresholds.

[0017] In a preferred embodiment, the detected outliers are input into a trained generator and replaced with generated new data. The specific steps are: generating a filling loss function based on a reconstruction loss function and a discriminant loss function; minimizing the filling loss function based on the reconstructed samples and the original samples to generate missing values ​​for interpolation.

[0018] The technical effects and advantages of the method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering of the present invention are as follows: 1. The present invention realizes the precise management of the power transmission and transformation project budget through the cost standard determination technology based on multi-attribute clustering. First, the method can classify according to the project type, effectively identify the special needs of different types of projects, and avoid a one-size-fits-all cost standard. Secondly, through in-depth analysis of historical data and the use of adversarial network technology for outlier interpolation, the quality and integrity of the data are improved, ensuring the scientific nature of cost calculation.

[0019] 2. The present invention, by adopting an improved neural network model, captures the complex nonlinear relationship that affects these costs in the prediction of labor costs, material costs and business costs, and provides more accurate prediction results, thereby providing real-time support for project decision-making. In addition, the process of optimizing the neural network using a genetic algorithm can explore the optimal solution in a multidimensional parameter space, further improving the accuracy and generalization ability of the prediction. For the calculation of indirect costs, the fluctuation pattern of indirect costs can be effectively identified by analyzing time series data through the ARIMA model, ensuring the dynamic adaptability of the budget. Through the multiple linear regression method, multiple influencing factors such as material cost, labor cost and equipment use can be considered simultaneously, thereby establishing an accurate prediction model. This method can deeply analyze the linear relationship between the factors, identify key cost drivers, and provide a scientific basis for project management. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 The present invention is a schematic structural diagram of a method for determining special cost standards for power transmission and transformation projects based on multi-attribute clustering. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Embodiment 1, Figure 1 The present invention provides a method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering, which includes the following steps: S1, classify the power transmission and transformation projects, and divide the special expenses of different types of power transmission and transformation projects into categories; According to the project type, the project is divided into different types such as substation, converter station, overhead transmission line, cable transmission line project, etc., and the next step of cost standard determination is carried out accordingly, which has the following benefits: Improve the accuracy of cost management: By subdividing projects into different types, more accurate cost standards can be formulated for each type of project. For example, substation and converter station projects may involve a large amount of equipment purchase and installation and commissioning costs, while overhead transmission line and cable transmission line projects may focus more on material costs and construction costs. By refining the cost standards, the cost of each type of project can be estimated more accurately, thereby improving the accuracy of cost management.

[0023] Promote the controllability of project costs: Formulating cost standards for different types of projects helps to achieve controllability of project costs. During the project implementation process, various costs can be monitored and adjusted in real time according to the cost standards to ensure that the project costs are always controlled within a reasonable range. This helps to avoid cost overruns and waste and improve the economic benefits of the project.

[0024] Optimize resource allocation: Classifying projects into different types and determining corresponding cost standards can help optimize resource allocation. Before the implementation of the project, the required resources can be reasonably planned and allocated according to the cost standards to ensure that all resources are fully utilized. This helps to improve the implementation efficiency and quality of the project, while reducing resource waste and costs.

[0025] Enhance budget transparency: By establishing clear cost standards, budget transparency can be enhanced. During the project budget formulation process, the cost structure and budget amount of each type of project can be clearly listed, making the project budget clearer and more transparent. This helps to strengthen the supervision and management of project budgets and prevent budget overspending and misuse of funds.

[0026] Provide a basis for decision-making: Clear cost standards can provide an important basis for decision-making in engineering projects. In the process of project decision-making, cost-benefit analysis can be conducted on different plans based on cost standards to select the best plan. This helps to improve the decision-making level and efficiency of engineering projects and ensure the smooth implementation and completion of engineering projects.

[0027] Promoting the standardization of project management: Formulating cost standards for different types of projects will help promote the standardization of project management. By formulating unified cost standards and management processes, the implementation and management of engineering projects can be standardized, and the management level and efficiency of engineering projects can be improved. This will help promote the standardization and normalization of engineering projects.

[0028] To sum up, dividing the projects into different types such as substations, converter stations, overhead transmission lines, cable transmission line projects, etc. according to the project type, and carrying out the next step of determining the cost standards accordingly, has multiple benefits such as improving the accuracy of cost management, promoting the controllability of project costs, optimizing resource allocation, enhancing budget transparency, providing a basis for decision-making, and promoting the standardization of project management.

[0029] S2, obtain historical data of power transmission and transformation projects, and process missing values ​​and outliers based on the historical data; In the process of determining the special fee standards for power transmission and transformation projects, the data collected is crucial. The following are the main types of data that need to be collected: (1) Material costs: The unit price and historical price change records of various building materials (such as steel, cement, insulation materials, etc.).

[0030] Material procurement quantity and its specifications.

[0031] (2) Labor costs: Wage standards for workers in different positions.

[0032] Working hours and overtime for various types of workers.

[0033] (3) Machinery costs: Rental costs and operating costs of machinery and equipment.

[0034] Frequency of use and maintenance records of mechanical equipment.

[0035] (4) Administrative expenses: Salaries, training and other administrative costs for the project management team.

[0036] Purchase and maintenance costs of related management software.

[0037] (5) Project scale and construction location: The total investment amount, construction area and construction period of the project.

[0038] Geographic information of the construction site and its impact on construction.

[0039] (6) Environmental impact data: Environmental protection requirements and standards in the project area.

[0040] Relevant data for environmental impact assessment, such as pollutant emissions, noise control, etc.

[0041] (7) Historical data: Cost data and actual expenditure of similar projects in the past.

[0042] Analysis results of cost change trends and influencing factors.

[0043] (8) External economic data: Local economic conditions, market conditions and industry standards.

[0044] Changes in policies and regulations and their potential impact on project costs.

[0045] Through the comprehensive analysis of these data, the key factors affecting fee fluctuations can be more accurately identified, providing a solid foundation for the subsequent determination of fee standards. In the process of data collection and processing, it is crucial to ensure the accuracy and completeness of the data so that the subsequent analysis and modeling work can achieve effective results.

[0046] The historical data and the related influencing factor data are interpolated for outliers through an adversarial network. The adversarial network usually refers to a generative adversarial network, which consists of a generator and a discriminator. The generator attempts to generate realistic data to deceive the discriminator, while the discriminator strives to distinguish between real data and data generated by the generator. Through the continuous confrontation and optimization of the two, high-quality data is finally generated.

[0047] The historical data and the related influencing factor data are processed by an adversarial network for outlier interpolation, and the specific steps are as follows: Data preparation: Collect and organize historical data and relevant influencing factor data to ensure the accuracy and completeness of the data.

[0048] Constructing adversarial networks: Designing and constructing generative adversarial networks that are suitable for data characteristics, including the structure and parameter settings of the generator and discriminator.

[0049] Training the adversarial network: Use normal data to train the adversarial network, so that the generator can learn the distribution characteristics of the data, and the discriminator can continuously improve its ability to distinguish between real data and generated data.

[0050] Outlier detection: Detect outliers in historical data and related influencing factor data through statistical methods.

[0051] Outlier interpolation: The detected outliers are input into the trained generator, which will generate new data similar to the normal data distribution to replace the outliers.

[0052] Specifically, performing outlier interpolation processing on historical cost input data and related influencing factor data through an adversarial network further includes: generating a reconstructed sample for adversarial use through a generator; distinguishing the distance between the adversarial sample and the original sample through a discriminator, wherein the original sample is any sample data of the historical cost data and the related influencing factor data; and reducing the distance between the reconstructed sample G(z) and the original sample x using a reconstruction loss function, wherein the reconstruction loss function is represented by the following formula: ; The discriminant loss function is used to maximize the output probability of the discriminator to ensure that the reconstructed sample G(z) conforms to the distribution of the original sample. The discriminant loss function is expressed by the following formula: ; A filling loss function is generated according to the reconstruction loss function and the discriminant loss function, and the filling loss function is minimized based on the reconstructed samples and the original samples to generate missing values ​​for interpolation. The filling loss function and the missing value are respectively expressed by the following formulas: ; ; Where x represents the original sample; m is a mask matrix used to indicate the missing parts in the historical cost input data and related influencing factor data; G(z) represents the reconstructed sample, Lr(z) represents the distance between the generated sample and the original sample; Ld(z) represents the discriminant loss function, which is used to evaluate the output probability of the discriminator; D(G(z)) is the output of the discriminator, which represents the probability that the reconstructed sample G(z) is discriminated as the original sample; Limputation(z) is a comprehensive loss function composed of the reconstruction loss function and the discriminant loss function; λ is a hyperparameter used to control the ratio between the reconstruction loss and the discriminant loss; ximputed represents the padded data sample; x⋅m represents the non-missing part in the original sample; (1−m)⋅G(z) represents the corresponding missing part in the reconstructed sample; Indicates filling the missing values ​​of the original sample.

[0053] S3, identify the influencing factors of special expenses of different types of power transmission and transformation projects, and combine the statistical analysis of historical data to determine the correlation between the fluctuation trend of different types of expenses and the influencing factors; The determination of the correlation between the fluctuation trend of different expense types and the influencing factors is specifically as follows: The improved anti-entropy-grey correlation analysis method is used to judge the correlation between cost input and each key influencing factor, further including: Calculating an entropy value for each of the multiple key influencing factors; The entropy value of each key influencing factor is converted into an anti-entropy value by using an anti-entropy formula, wherein the higher the anti-entropy value, the more certain the information provided by the key influencing factor and the greater the difference, so as to be given a greater weight; Generate an objective weight based on the anti-entropy value, and calculate a combined weight based on the objective weight and the subjective weight; Based on the combined weight, a correlation coefficient between each comparison sequence and a reference sequence is calculated, wherein the cost data is selected as the reference sequence and the multiple key influencing factors are selected as multiple comparison sequences; and based on the correlation coefficient, a correlation degree between the reference sequence and each comparison sequence is calculated as a correlation degree between the cost input and each key influencing factor.

[0054] Specifically, the improved anti-entropy-grey correlation analysis method is used to determine the correlation between the cost scale and each key influencing factor, further including: the matrix of key influencing factors is expressed by the following formula: for: ; in, For the The object's The value of a key influencing factor, selected from multiple key influencing factors The key influencing factors Regions, Annual data; based on the entropy values ​​of multiple key influencing factors, the anti-entropy of the following key influencing factors is calculated: ; ; in, For the Objects under The entropy value of each key influencing factor is calculated; the anti-entropy of each key influencing factor is normalized to obtain the objective weight of the corresponding anti-entropy: ; in, For the The objective weights of the key influencing factors are calculated; based on the objective weights and subjective weights, the following combined weights are calculated: ; in, To evaluate the key influencing factors The combined weight of , Respectively The objective weight and subjective weight of the key influencing factors are calculated by the following formula: : ; ; ; ; in, is the correlation between cost input and key influencing factor j; is the correlation coefficient between the cost input and the key influencing factor j determined based on the combined weight; is the residual degree of each key influencing factor r; is the relative entropy, which represents the ratio of the anti-entropy of the key influencing factor r to the maximum entropy; , are the anti-entropy and maximum entropy of the key influencing factors respectively; b(r) is the proportional weight of other key influencing factors r.

[0055] The cost types of special expenses analyzed by the system include direct costs, labor costs, material costs, business costs, and indirect costs.

[0056] The relevant factors for each cost are finally determined as follows: (1) Labor costs: Construction environment: such as the impact of high altitude, high temperature or low temperature areas on workers' labor intensity.

[0057] Project type: Demand for skilled workers for complex projects such as tunnels or bridges.

[0058] Labor supply and demand: A shortage or surplus of workers in a particular area affects wage levels.

[0059] Safety requirements: Safety training and equipment costs for high-risk construction environments.

[0060] Seasonal impact: The impact of construction season (such as rainy season) on work efficiency.

[0061] (2) Material costs: Source of raw materials: transportation distance and cost of material procurement location.

[0062] Market price fluctuations: Changes in the price of a particular material in different markets.

[0063] Environmental impact standards: Increased material costs to comply with environmental regulations.

[0064] Supply chain stability: supplier's delivery capability and its stability.

[0065] Quality standard requirements: Inspection and certification costs required for high-standard materials.

[0066] (3) Business expenses: Market competition status: pricing strategies of similar service providers.

[0067] Degree of service customization: additional costs caused by customer-specific needs.

[0068] Contract Terms: The cost sharing and liability provisions set out in a contract.

[0069] Project size: The difference in management costs between large projects and small projects.

[0070] Industry regulations: Additional costs caused by industry regulation and compliance requirements.

[0071] (4) Direct costs: Construction technology: the direct cost difference of using different construction technologies.

[0072] Equipment efficiency: Additional expenses caused by equipment failure or inefficiency.

[0073] Resource consumption: the consumption of materials and manpower by different construction methods.

[0074] Design complexity: Increased costs due to design changes.

[0075] Construction schedule arrangement: increase in direct costs caused by construction delays.

[0076] (5) Indirect costs: Administrative costs: Staffing and related expenses of the project management team.

[0077] Facility maintenance costs: management and maintenance of office space and temporary facilities.

[0078] Insurance costs: various insurance expenses involved in the project.

[0079] On-site support services: costs of temporary services such as medical care and catering.

[0080] Information Technology Support: Costs of IT infrastructure and support required for the project.

[0081] S4. Formulate special cost standards based on the correlation between the fluctuation trend and the factors affecting the special costs of power transmission and transformation projects; (1) Labor costs, material costs, and business expenses These costs are usually affected by multiple factors, such as market fluctuations, seasonal demand, and project size, and involve multiple decision variables. Therefore, an improved neural network method is adopted to determine the cost standards for the above costs. Neural networks are good at handling complex nonlinear relationships, which enables them to effectively capture the subtle connections between influencing factors when analyzing multidimensional data related to costs. Secondly, the self-learning ability of neural networks means that it can continuously optimize its own model as new data is input, thereby improving the accuracy of predictions. In addition, through deep learning technology, neural networks can identify patterns hidden in the data, adapt to market changes, and reflect real-time cost fluctuation trends. In short, neural networks provide a flexible, dynamic, and accurate solution for cost forecasting.

[0082] This patent uses genetic algorithms to optimize neural network models. The principle of genetic algorithms (GA) to optimize neural networks simulates the natural selection process and finds the best parameter configuration through a series of steps. First, a group of neural network individuals are randomly generated and their performance is evaluated by the fitness function. Then, excellent individuals are selected and crossover and mutation are performed to generate new individuals. These new individuals will replace the original population to form a new population. By repeatedly iterating this process, GA can explore in a complex parameter space and gradually approach the optimal solution, thereby improving the prediction accuracy and generalization ability of the neural network.

[0083] 1) Design of input and output layers. According to the design characteristics of BP neural network, without considering the mutual influence relationship between various factors, the main influencing factors can be selected as the input layer and output layer cost investment values ​​in the cost investment.

[0084] 2) The target value is selected by processing data in horizontal and vertical directions as the output target value.

[0085] 3) Model establishment.

[0086] Genetic algorithm part: First, use GA to initialize the population to obtain the initial weights and thresholds of the BP neural network, and then substitute the individuals in the initialized population into the BP neural network to iterate and find the optimal initial weights and thresholds.

[0087] ① Encode the population of real number strings consisting of the weights between the input layer, hidden layer and output layer, and the thresholds of the hidden layer and the output layer.

[0088] ②The fitness value F of an individual is composed of the absolute value of the error between its expected output value and the predicted output value:

[0089] In the formula is the number of output nodes; For the The expected output value of each node; For the The expected output value of each node.

[0090] ③ According to the individual fitness results calculated by the above formula, select better individuals to form a new population. The calculation method is:

[0091]

[0092] in, For the The fitness value of a group is calculated by the formula: is the group size, for The probability of a group being selected.

[0093] ④ Randomly select two fathers from the population and , with a certain probability The real number crossing method of the points is:

[0094] in, is a random number between 0 and 1.

[0095] ⑤ Through probability calculation, randomly select an individual from the new population To mutate, The point variation modes are:

[0096] In the formula , are the upper and lower bounds of the individual, respectively; This is the number of iterations at this time; is the maximum number of evolutions; is a random number ranging from 0 to 1.

[0097] ⑥ Check whether the obtained individuals meet the optimal standards, and re-check the fitness of individuals that cannot meet the standards; if they meet the standards, the optimal individuals are used as the weights and thresholds of the BP neural network, repeat the learning and training times, and adjust the neural network parameters to the set accuracy range or until the learning times are met.

[0098] Neural network part: After obtaining the initial weights and thresholds of the BP neural network, the steps for analyzing its internal structure model are as follows.

[0099] ①Initialize the network.

[0100]

[0101] In the formula Input vector for neurons and influencing factors for costs; is the number of nodes in the input layer; is the neuron output vector, is the cost input scale, is the number of nodes in the output layer.

[0102] ②Calculate the hidden layer output. Hidden layer output value The calculation method is:

[0103] In the formula is the number of hidden layer nodes; is the weight between the input layer and the hidden layer; is the threshold value; is the activation function.

[0104] ③Calculate the output of the output layer. Output layer output value The calculation method is: In the formula is the weight between the hidden layer and the output layer; is the threshold value.

[0105]

[0106] ④ Calculate the error and verify. Error Represented as predicted output With the expected output The difference is calculated as:

[0107] set up is the maximum allowable error value, if , then adjust the BP neural network parameters and re-train until all training data meet the specified error standard.

[0108] ⑤Update weights and thresholds. Update based on the error you are looking for and :

[0109]

[0110] in, is the learning rate, between 0 and 1. The threshold is corrected to

[0111]

[0112] 4) Model calculation. The model calculation uses MATLAB software. During the calculation process, the existing data of substation equipment maintenance and operation costs are used as samples, and the GA-BP neural network is used for training. In order to avoid overfitting and improve the accuracy of the GA-BP neural network model, the K-fold cross-validation method is used for the training data to determine the accuracy of the model.

[0113] 5) Prediction effect evaluation indicators. When testing the results of the model, it is necessary to consider the reliability and validity of the model. Reliability refers to the reliability of the model, which is manifested in the reproducibility and stability of the model results; validity refers to the effectiveness of the model, which is manifested in the degree of consistency between the model results and the content to be examined.

[0114] This paper mainly adopts Evaluate the impact of K-fold cross validation on model accuracy. The calculation formula is:

[0115] The mean absolute percentage error and root mean square error (RMSE) are used to measure the validity of the network, that is, the prediction effect. The calculation formula is:

[0116]

[0117] in, is the actual output of the model, is the expected output. In order to test the reliability of the model, this paper takes the , The standard deviation is used as the final test indicator.

[0118] (2) Indirect costs Indirect costs are usually affected by seasonal and cyclical changes and show obvious time series characteristics. Therefore, time series analysis and autoregressive integrated moving average (ARIMA) model are used to analyze and predict indirect costs.

[0119] The ARIMA model (Autoregressive Integrated Moving Average) is widely used in time series data analysis, especially in the economic and financial fields. Its structure consists of three parts: autoregression (AR), difference (I) and moving average (MA), which enables it to flexibly capture trends, seasonality and cyclical changes in data. In the forecast of indirect costs, the ARIMA model is particularly suitable because indirect costs are often affected by external environment and internal management changes, showing complex time series characteristics.

[0120] First, the autoregressive component can use the values ​​of previous moments to predict the current value and capture the inherent correlation of the data. Secondly, the moving average component further smoothes the data by weighted averaging the historical forecast errors to reduce the interference of noise on the forecast results. Through differential processing, ARIMA can transform non-stationary time series into stationary series to ensure the effectiveness of model parameters. Since changes in indirect costs often show cyclical characteristics, the ARIMA model can effectively adapt to these characteristics through its flexibility, thereby providing accurate budget support and decision-making basis. Combined with the analysis of historical data, the ARIMA model can not only help identify the fluctuation pattern of indirect costs, but also provide strong support for future cost control and resource allocation.

[0121] The Auto-Regressive and Moving Average (ARMA) model is a combination of the Auto-Regressive model (AR) and the Moving Average method (MA).

[0122] The autoregressive model assumes that in the time series, the current variable With the past The future correlation of variables consists of a linear combination of past values ​​of the variables plus a random error term. Order autoregressive model In , the general form of the variables is:

[0123] In the formula, Indicates the degree of correlation between the current variable and its historical value; is independent and identically distributed white noise, is a random error term, and satisfies , .

[0124] The moving average model considers the current variable of the time series It is the memory of past noise interference and can be expressed as a linear combination of white noise sequences with random interference. Moving average model It can be expressed as:

[0125] In the formula, Indicates the degree of correlation between variables and random interference, It is also a white noise sequence.

[0126] The autoregressive moving average model ARMA is a combination of the two. It is the system's memory of its past state and the noise that enters the system. In a time series, the current variable can be expressed as a linear combination of historical variables plus a white noise sequence. It is represented by an order moving average. , Autoregressive Moving Average Model Generally expressed as:

[0127] In the formula, is the moving average parameter, which determines the extent to which the value of the time series at the previous moment affects the current value. is a white noise sequence.

[0128] (3) Direct costs For direct costs, the cost standard is determined using the multiple linear regression method. The cost structure of direct costs is affected by multiple factors, including material costs, labor costs, and equipment usage. There may be complex linear relationships between these factors. Through multiple linear regression, the joint effects of these variables on direct costs can be considered at the same time to establish a more accurate prediction model. In addition, multiple linear regression can provide contribution analysis of each factor, which helps to identify key cost drivers, thereby providing a scientific basis for cost control and decision-making.

[0129] Multiple linear regression analysis is a statistical method that predicts and explains the changes in dependent variables by establishing a linear relationship model between independent variables and dependent variables. This method can simultaneously consider the impact of multiple factors on indirect costs, thereby providing a scientific basis for the formulation of cost standards.

[0130] Regression Analysis refers to the use of statistical data to analyze the quantitative change patterns of each variable, and to reflect and describe this relationship in the form of a regression equation. If there are multiple independent variables involved in the regression analysis, it is a multiple linear regression analysis, and the result is called a multiple linear regression equation. Due to the diversity of factors affecting the cost of power transmission and transformation projects, multiple linear regression analysis is often used in factor analysis.

[0131] The mathematical model of multiple linear regression is:

[0132] In the above two formulas, Refers to the part of y that is changed by other random factors; that is The regression coefficient is the partial derivative of y with respect to x, so it is also called the partial regression coefficient.

[0133] The steps of direct cost analysis based on multiple linear regression are as follows: (1) Determine the independent variable in the regression equation , which is the influencing factor. The dependent variable y is the cost level. Collect sample data.

[0134] (2) Estimate the parameters in the model under certain statistical fitting criteria and obtain the regression equation.

[0135] (3) Test the regression coefficient and analyze the results.

[0136] Finally, the forecast results are superimposed to form a scientific determination of the special consulting fee standard.

[0137] S5, build a dynamic cost update mechanism, dynamically update and revise cost standards based on actual project and standard differences.

[0138] The first task in establishing a dynamic fee update mechanism is to clarify its goal, which is to ensure the accuracy and rationality of fees to adapt to market changes and business needs. The mechanism should follow the principles of fairness, justice, transparency, timeliness and rationality to ensure that all relevant parties can understand and accept the fee update.

[0139] Comprehensive information collection: Establish comprehensive information collection channels, including market price monitoring, cost analysis, business demand changes, etc., to ensure the accuracy and timeliness of information.

[0140] Analyze data regularly: Analyze the collected data regularly to identify trends and reasons for cost changes and provide a basis for cost updates.

[0141] Clear trigger conditions: Set clear trigger conditions for fee updates, such as market price fluctuations exceeding a certain range, business volume changes reaching a certain proportion, etc.

[0142] Contractual provisions: Clearly stipulate in the contract the triggering conditions and adjustment methods for fee updates to avoid disputes.

[0143] Develop process: Develop a detailed expense update process, including application, review, approval, execution and feedback, to ensure a smooth and efficient process.

[0144] Standardized operations: Clarify the operational specifications for cost updates, such as calculation method, adjustment range, etc., to ensure the rationality and accuracy of the updates.

[0145] Timely communication: During the cost update process, timely communication should be conducted with relevant parties to explain the reasons and basis for the update and ensure transparency and sharing of information.

[0146] Collect feedback: Collect opinions and suggestions from all parties on fee updates, and continuously improve and optimize the mechanism.

[0147] Establish a supervision mechanism: Establish a supervision mechanism for the dynamic update mechanism of expenses to ensure the implementation and effectiveness of the mechanism.

[0148] Regular evaluation: Regularly evaluate the dynamic fee update mechanism, identify problems and make timely improvements to ensure continuous optimization and improvement of the mechanism.

[0149] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0150] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0151] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0152] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0153] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0154] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for determining special cost standards for power transmission and transformation projects based on multi-attribute clustering, characterized in that: The following steps are involved: Classify power transmission and transformation projects, and divide the special expenses of different types of power transmission and transformation projects into categories; Obtain historical data of power transmission and transformation projects, and process missing values ​​and outliers based on the historical data; Identify the influencing factors of special expenses for different types of power transmission and transformation projects, and combine the statistical analysis of historical data to determine the correlation between the fluctuation trend of different types of expenses and the influencing factors; Formulate special expense standards based on the correlation between fluctuation trends and factors affecting special expenses of power transmission and transformation projects; Establish a dynamic cost update mechanism, and dynamically update and revise cost standards based on actual project and standard differences.

2. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 1 is characterized in that: The missing value and outlier processing according to historical data is specifically as follows: Obtain historical data of power transmission and transformation projects and influencing factors of special costs of power transmission and transformation projects, and detect abnormal values ​​in historical data and influencing factors through statistical methods; Design and construct a generative adversarial network that is suitable for the historical data and influencing factors of power transmission and transformation projects, including the structure and parameter settings of the generator and discriminator; The adversarial network is trained using normal data, the detected outliers are input into the trained generator, and the outliers are replaced with the generated new data.

3. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 2 is characterized in that: The detected outliers are input into the trained generator and replaced with the generated new data, specifically: Generate adversarial reconstruction samples through the generator; Distinguish the distance between the adversarial sample and the original sample by a discriminator, wherein the original sample is any sample data of the historical cost data and the influencing factor data; The reconstruction loss function is used to reduce the distance between the reconstructed sample and the original sample, and the discriminant loss function is used to maximize the output probability of the discriminator; A filling loss function is generated according to the reconstruction loss function and the discrimination loss function, and the filling loss function is minimized based on the reconstructed samples and the original samples to generate missing values ​​for interpolation.

4. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 3 is characterized in that: The determination of the correlation between the fluctuation trend of different expense types and the influencing factors is made by using the improved anti-entropy-grey correlation analysis method, specifically: Calculating an entropy value for each of the multiple key influencing factors; The entropy value of each key influencing factor is converted into an anti-entropy value by using an anti-entropy formula; Generate an objective weight based on the anti-entropy value, and calculate a combined weight based on the objective weight and the subjective weight; Calculate the correlation coefficient between each comparison sequence and the reference sequence based on the combined weight, select the special expense data as the reference sequence, and use multiple key influencing factors as multiple comparison sequences; The correlation between the reference sequence and each comparison sequence is calculated based on the correlation coefficient as the correlation between the cost input and each key influencing factor.

5. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 4 is characterized in that: According to the correlation between the fluctuation trend and the factors affecting the special expenses of power transmission and transformation projects, special expense standards are formulated. The special expenses of power transmission and transformation projects include labor costs, material costs, business costs, indirect costs and direct costs; Among them, the cost standards for labor costs, material costs, and business costs are formulated as follows: The neural network model is used to confirm various cost standards. Genetic algorithms are used to optimize the neural network model to obtain the cost standards for labor costs, material costs, and business costs; The standards for indirect costs are as follows: The analysis was obtained by using time series analysis and autoregressive integrated moving average (ARIMA) model; The standards for indirect costs are as follows: The cost standard was determined using the multivariate linear regression method.

6. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 5 is characterized in that: The cost standards for labor costs, material costs, and business costs are as follows: According to the design characteristics of BP neural network, the mutual influence relationship between various factors is not considered, and the main influencing factors in cost input are selected as the input layer; The target value is selected by processing data in horizontal and vertical directions as the output target value; Establish models based on genetic algorithms and optimize neural network models; Based on MATLAB software, the existing data of substation equipment maintenance and operation costs are used as samples and trained using GA-BP neural network; The training data results are verified and the K-fold cross-validation method is adopted to determine the accuracy of the model. When the model results are tested, the reliability and validity of the model are considered. The reliability refers to the reliability of the model, and the validity refers to the effectiveness of the model.

7. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 6 is characterized in that: The genetic algorithm-based model is established and the neural network model is optimized, specifically: Use GA to initialize the population and get the initial weights and thresholds of the BP neural network; After the population is initialized, the individuals are substituted into the BP neural network to iteratively find the optimal initial weights and thresholds.

8. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 3 is characterized in that: The detected outliers are input into the trained generator and replaced with the generated new data. The specific steps are as follows: Generate a filling loss function based on the reconstruction loss function and the discriminant loss function; Minimize the filling loss function based on the reconstructed samples and the original samples to generate missing values ​​for interpolation; Among them, the reconstruction loss function is: ； The discriminant loss function is: ； The filling loss function and missing values ​​are: ； ； Where x represents the original sample; m is a mask matrix used to indicate the missing parts in the historical cost input data and related influencing factor data; G(z) represents the reconstructed sample, Lr(z) represents the distance between the generated sample and the original sample; Ld(z) represents the discriminant loss function, which is used to evaluate the output probability of the discriminator; D(G(z)) is the output of the discriminator, which represents the probability that the reconstructed sample G(z) is discriminated as the original sample; Limputation(z) is a comprehensive loss function composed of the reconstruction loss function and the discriminant loss function; λ is a hyperparameter used to control the ratio between the reconstruction loss and the discriminant loss; ximputed represents the padded data sample; x⋅m represents the non-missing part in the original sample; (1−m)⋅G(z) represents the corresponding missing part in the reconstructed sample; Indicates filling the missing values ​​of the original sample.

9. The method for determining the special cost standard of power transmission and transformation projects based on multi-attribute clustering according to claim 8, characterized in that: The correlation degree between the fluctuation trend and the influencing factors is expressed by the following formula: ； ； ； ； in, is the correlation between cost input and key influencing factor j; is the correlation coefficient between the cost input and the key influencing factor j determined based on the combined weight; is the residual degree of each key influencing factor r; is the relative entropy, which represents the ratio of the anti-entropy of the key influencing factor r to the maximum entropy; , are the anti-entropy and maximum entropy of the key influencing factors respectively; b(r) is the proportional weight of other key influencing factors r.

10. The method for determining special cost standards for power transmission and transformation projects based on multi-attribute clustering according to claim 5, characterized in that: The training data results are verified by using the K-fold cross-validation method to determine the accuracy of the model, specifically: use Evaluate the impact of K-fold cross validation on model accuracy. The calculation formula is: The mean absolute percentage error and root mean square error are used to measure the validity of the network, that is, the prediction effect. The calculation formula is: in, is the actual output of the model, is the expected output. In order to test the reliability of the model, we take , The standard deviation is used as the final test indicator.

Images