Prediction method and device for maintenance project cost in power grid and electronic equipment

Abstract

The invention discloses a method and a device for predicting maintenance project cost in a power grid, and electronic equipment. The method comprises the steps that target influence factors in a maintenance project to be subjected to cost prediction in a power grid are acquired, the target influence factors are used for representing target working procedures in the maintenance project, and the target working procedures are used for representing working procedures with the execution times larger than the preset times in basic working procedures forming the maintenance project; and the project cost of the maintenance project is predicted based on the target influence factors by using a target maintenance cost calculation model to obtain a cost prediction result of the maintenance project, and the target maintenance cost calculation model is used for representing a machine learning model for cost prediction by using different types of kernel functions as base models. According to the invention, the technical problem of low efficiency of predicting the cost of the maintenance project in the power grid in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of power grid project cost, and in particular to a method, device and electronic equipment for predicting the cost of a maintenance project in a power grid. Background Art

[0002] According to the characteristics and scale of the project, the overall cost of the maintenance project in the power grid is predicted through analysis and calculation. That is, the cost of the maintenance project in the power grid is predicted, which can help project managers make budget plans in advance, ensure the smooth progress of the project and control costs.

[0003] At present, the cost prediction of maintenance projects in power grids in related technologies is usually based on manual calculation and prediction, or most existing studies use a single model to estimate the cost of engineering projects, which cannot give full play to the advantages of various calculation models and avoid the defects of a single model, resulting in low efficiency in predicting the cost of maintenance projects in power grids.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a method, device and electronic device for predicting the cost of maintenance projects in a power grid, so as to at least solve the technical problem of low efficiency in predicting the cost of maintenance projects in a power grid in the related art.

[0006] According to one aspect of an embodiment of the present invention, there is provided a method for predicting the cost of a maintenance project in a power grid, comprising: obtaining target influencing factors in the maintenance project to be cost predicted in the power grid, wherein the target influencing factors are used to represent the target process in the maintenance project, and the target process is used to represent the process whose execution times are greater than a preset number of basic processes constituting the maintenance project; predicting the engineering cost of the maintenance project based on the target influencing factors using a target maintenance cost calculation model to obtain a cost prediction result of the maintenance project, wherein the target maintenance cost calculation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction.

[0007] Optionally, the method also includes: obtaining historical maintenance statistical data of the maintenance project, wherein the historical maintenance statistical data at least includes historical maintenance engineering quantity data and historical maintenance settlement data; constructing a training data set based on the historical maintenance statistical data, wherein the input vectors of the training sample pairs in the training data set are generated by the historical maintenance engineering quantity data, and the output vectors of the training sample pairs are generated by the historical maintenance settlement data; and constructing a target maintenance cost estimation model based on the training data set and a preset statistical machine learning algorithm.

[0008] Optionally, based on the training data set and the preset statistical machine learning algorithm, a target maintenance cost estimation model is constructed, including: constructing a target error term between the cost prediction result and the output vector in the training data set; based on the target error term, constructing a first objective function for cost prediction; based on the training data set and the preset statistical machine learning algorithm, constructing a first constraint condition for cost prediction; based on the first constraint condition and the first objective function, constructing a target maintenance cost estimation model.

[0009] Optionally, based on the first constraint condition and the first objective function, a target maintenance cost estimation model is constructed, including: solving the first objective function based on the first constraint condition to obtain target parameters for constructing the model; constructing multiple target regression functions based on different types of kernel functions and target parameters; constructing a target maintenance cost estimation model based on multiple target regression functions and a target weight matrix, wherein the target weight matrix is ​​used to represent a matrix composed of weight coefficients of the prediction results based on multiple target regression functions in the prediction results of the target maintenance cost estimation model.

[0010] Optionally, the method also includes: based on multiple objective regression functions, performing cost forecasting on historical maintenance engineering quantity data to obtain a forecast result matrix; based on the forecast result matrix and historical maintenance settlement data, constructing a second objective function for measuring the forecast results; constructing constraints that need to be satisfied by each weight coefficient in the target weight matrix to obtain a second constraint; solving the second objective function based on the second constraint to obtain a target weight matrix.

[0011] Optionally, the target maintenance cost calculation model is used to predict the engineering cost of the maintenance project based on target influencing factors to obtain a cost prediction result of the maintenance project, including: using multiple target regression functions in the target maintenance cost calculation model to predict the engineering cost of the maintenance project based on target influencing factors to obtain multiple initial cost prediction results; based on the target weight matrix in the target maintenance cost calculation model, weighted summation of multiple initial cost prediction results is performed to obtain a cost prediction result.

[0012] According to another aspect of an embodiment of the present invention, there is also provided a device for predicting the cost of a maintenance project in a power grid, comprising: an acquisition module for acquiring target influencing factors in a maintenance project to be cost predicted in the power grid, wherein the target influencing factors are used to represent target processes in the maintenance project, and the target processes are used to represent processes in the basic processes constituting the maintenance project that are executed more than a preset number of times; a prediction module for predicting the engineering cost of the maintenance project based on the target influencing factors using a target maintenance cost calculation model to obtain a cost prediction result of the maintenance project, wherein the target maintenance cost calculation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction.

[0013] According to another aspect of an embodiment of the present invention, there is further provided an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the method in each embodiment of the present invention is executed when the program is running.

[0014] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is provided. The computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0015] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0016] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0017] According to another aspect of the embodiments of the present invention, a computer program is further provided. When the computer program is executed by a processor, the methods in the embodiments of the present invention are implemented.

[0018] In an embodiment of the present invention, a method for predicting the cost of a maintenance project in a power grid is provided, comprising: obtaining target influencing factors in the maintenance project to be cost predicted in the power grid, wherein the target influencing factors are used to represent the target process in the maintenance project, and the target process is used to represent the process whose execution times are greater than a preset number of basic processes constituting the maintenance project; using a target maintenance cost estimation model to predict the engineering cost of the maintenance project based on the target influencing factors to obtain a cost prediction result of the maintenance project, wherein the target maintenance cost estimation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction. It is easy to notice that in the process of predicting the cost of maintenance projects in the power grid, the present application can determine the target influencing factors in the maintenance projects. The target influencing factors will have a greater impact on the cost of the entire project, providing a strong basis and reference for subsequent cost predictions; the present application adopts a target maintenance cost calculation model to predict the engineering cost of the maintenance project based on the target influencing factors. The target maintenance cost calculation model uses different types of kernel functions as a base model. The target maintenance cost calculation model using multiple different types of kernel functions is used to calculate the engineering cost of the maintenance project. It can give full play to the advantages of multiple types of calculation models and avoid the defects of a single model. The target maintenance cost calculation model can output prediction results, that is, the cost prediction results of the maintenance project. The cost prediction results can help project managers make decisions and evaluate the costs and risks of the project, so as to better plan and manage the project, thereby solving the technical problem of low efficiency in predicting the cost of maintenance projects in the power grid in related technologies. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0020] Figure 1 is a flow chart of a method for predicting the cost of a maintenance project in a power grid according to an embodiment of the present invention;

[0021] Figure 2 is a flow chart of an optional method for predicting the cost of a maintenance project in a power grid according to an embodiment of the present invention;

[0022] Figure 3 is a schematic diagram of a process structure of an optional maintenance item according to an embodiment of the present invention;

[0023] Figure 4 is a schematic diagram of an optional process cost composition according to an embodiment of the present invention;

[0024] Figure 5It is a flow chart of optimizing the target maintenance cost estimation model by using an ant colony optimization algorithm according to an embodiment of the present invention;

[0025] Figure 6 It is a schematic diagram of a device for predicting the cost of maintenance items in a power grid according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] According to one aspect of an embodiment of the present invention, a method for predicting the cost of maintenance projects in a power grid is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in an order different from that shown here.

[0029] Figure 1 FIG. 1 is a flow chart of a method for predicting the cost of a maintenance project in a power grid according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0030] Step S102, obtaining target influencing factors in a maintenance project to be cost predicted in a power grid.

[0031] The target influencing factor is used to represent the target process in the maintenance project, and the target process is used to represent the process whose execution times are greater than the preset times among the basic processes constituting the maintenance project.

[0032] The above-mentioned maintenance projects may refer to construction projects for regular inspection, maintenance and repair of power equipment and facilities. The maintenance projects may include various links such as transmission lines, substations, and distribution equipment, and may include cleaning, testing, repair, and replacement of equipment. Through regular maintenance projects, equipment problems can be discovered and resolved in a timely manner, extending the service life of the equipment and ensuring the normal operation of the power system. Maintenance projects can be determined based on actual needs and are not limited here.

[0033] The above-mentioned cost forecast can refer to the forecast of the total cost required for the project through analysis and evaluation of the costs of materials, labor, equipment, etc. involved in the maintenance project. Cost forecast can help the power grid management department make reasonable budget arrangements to ensure the smooth progress of the project and control costs. At the same time, it can also discover possible risks and problems in advance so as to make timely adjustments and solutions. Through cost forecasting, the maintenance projects in the power grid can be better managed and controlled to ensure the smooth and efficient completion of the project.

[0034] The above-mentioned basic process can refer to the basic constituent unit of the engineering project obtained by dividing the maintenance project. The determination and division of the basic process can be determined according to actual needs and is not limited here. The process can be used as the smallest constituent unit of the engineering project, and the sum of all process costs can form the total project cost.

[0035] The above-mentioned preset number of times may refer to a preset value for determining the target influencing factor. For example, the preset number of times may be three times, five times, etc. The preset number of times may be determined according to actual needs and is not limited here.

[0036] The above-mentioned target influencing factors may refer to the processes that need to be executed more than the preset number of times after the maintenance projects are divided. Different maintenance projects may determine different target influencing factors. For example, in different transformer maintenance projects, similar projects may include a large number of similar processes, such as concrete pole assembly, tension corner tower conductor suspension and insulator installation in tower projects. The cost of such processes accounts for most of the project cost and is the key process of the project. Therefore, they can be determined as the target influencing factors that determine the total cost of the project. The determination of the target influencing factors may be based on actual needs and is not limited here.

[0037] In an optional embodiment, in the process of predicting the cost of a maintenance project in a power grid, the maintenance project to be predicted can be determined, and then the target influencing factors in the maintenance project, that is, the target process, can be identified. The target process refers to the process whose execution times are greater than the preset times in the basic processes that constitute the maintenance project. The target influencing factors may have a greater impact on the cost of the entire project. Specifically, the specific maintenance project for which cost prediction is required can be determined first, which may include information such as the scope of the project, work content, and execution cycle. For the maintenance project, all basic processes can be listed, which may include inspection, maintenance, replacement, installation and other work contents. According to the actual situation of the project, the number of times each basic process actually needs to be executed can be counted, and processes whose execution times are greater than the preset times can be screened out. These processes are the target influencing factors, that is, the target processes, which facilitates the subsequent in-depth analysis of the target processes to understand the degree of their impact on the cost of the entire project, the reasons for the possible cost increase, etc. Through the above steps, the target influencing factors in the maintenance project can be effectively identified, providing a strong basis and reference for subsequent cost prediction.

[0038] Step S104, using the target maintenance cost estimation model to predict the engineering cost of the maintenance project based on the target influencing factors, to obtain a cost prediction result of the maintenance project.

[0039] Among them, the target maintenance cost estimation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction.

[0040] The above-mentioned cost prediction result may refer to the engineering cost of the maintenance project obtained through prediction, or may be the total cost of the maintenance project, etc.

[0041] In an optional embodiment, the target maintenance cost estimation model can be a model that predicts the engineering cost of the maintenance project based on the target influencing factors. During the use of the model, the target influencing factors of the maintenance project can be input, and the target influencing factors will be passed into the model as input features. The target maintenance cost estimation model uses different types of kernel functions as base models, such as linear kernel functions, polynomial kernel functions, radial basis functions, etc. These kernel functions will transform the input features and map them into a high-dimensional space for calculation. By performing kernel transformation on the input features, the target maintenance cost estimation model can better capture the nonlinear relationship between the input features, thereby improving the accuracy of the prediction. The target maintenance cost estimation model can generate a prediction result based on the combination of the input features and the kernel function, that is, the engineering cost of the maintenance project. The target maintenance cost estimation model can output a prediction result, that is, the cost prediction result of the maintenance project. The cost prediction result can help project managers make decisions and evaluate the costs and risks of the project, thereby better planning and managing the project.

[0042] In an embodiment of the present invention, a method for predicting the cost of a maintenance project in a power grid is provided, comprising: obtaining target influencing factors in the maintenance project to be cost predicted in the power grid, wherein the target influencing factors are used to represent the target process in the maintenance project, and the target process is used to represent the process whose execution times are greater than a preset number of basic processes constituting the maintenance project; using a target maintenance cost estimation model to predict the engineering cost of the maintenance project based on the target influencing factors to obtain a cost prediction result of the maintenance project, wherein the target maintenance cost estimation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction. It is easy to notice that in the process of predicting the cost of maintenance projects in the power grid, the present application can determine the target influencing factors in the maintenance projects. The target influencing factors will have a greater impact on the cost of the entire project, providing a strong basis and reference for subsequent cost predictions; the present application adopts a target maintenance cost calculation model to predict the engineering cost of the maintenance project based on the target influencing factors. The target maintenance cost calculation model uses different types of kernel functions as a base model. The target maintenance cost calculation model using multiple different types of kernel functions is used to calculate the engineering cost of the maintenance project. It can give full play to the advantages of multiple types of calculation models and avoid the defects of a single model. The target maintenance cost calculation model can output prediction results, that is, the cost prediction results of the maintenance project. The cost prediction results can help project managers make decisions and evaluate the costs and risks of the project, so as to better plan and manage the project, thereby solving the technical problem of low efficiency in predicting the cost of maintenance projects in the power grid in related technologies.

[0043] Optionally, the method also includes: obtaining historical maintenance statistical data of the maintenance project, wherein the historical maintenance statistical data at least includes historical maintenance engineering quantity data and historical maintenance settlement data; constructing a training data set based on the historical maintenance statistical data, wherein the input vectors of the training sample pairs in the training data set are generated by the historical maintenance engineering quantity data, and the output vectors of the training sample pairs are generated by the historical maintenance settlement data; and constructing a target maintenance cost estimation model based on the training data set and a preset statistical machine learning algorithm.

[0044] The above-mentioned historical maintenance engineering quantity data may refer to the engineering quantity statistical data of maintenance projects carried out in the power grid within a certain period of time in the past, which may include but is not limited to the type, scale, construction hours, human resource investment, etc. of the project. The historical maintenance engineering quantity data can help predict the size of the engineering quantity of future maintenance projects, thereby determining the project budget and resource allocation.

[0045] The above-mentioned historical maintenance settlement data may refer to the settlement data of maintenance projects completed in the power grid within a certain period of time in the past, including but not limited to the actual cost of the project, cost analysis, etc. The historical maintenance settlement data can help predict the cost of future maintenance projects.

[0046] The above-mentioned preset statistical machine learning algorithm may refer to a pre-set machine learning algorithm in the statistical field. The preset statistical machine learning algorithm may be a machine learning algorithm such as a support vector machine (SVM), and may also be determined according to actual needs, which is not limited here.

[0047] In an optional embodiment, historical maintenance statistical data of maintenance projects can be obtained, historical data of past maintenance projects in the power grid can be collected, which may include engineering quantity data and settlement data of historical maintenance projects, which can be obtained from the database of the power grid company, or obtained through investigation and recording. The historical maintenance statistical data can be converted into a training data set, wherein the input vector of each training sample pair is generated by the historical maintenance engineering quantity data, and the output vector is generated by the historical maintenance settlement data. Data processing tools can be used to process and convert data, and a machine learning algorithm suitable for predicting the cost of maintenance projects can be selected. The selected machine learning algorithm and training data set can be used to construct a target maintenance cost estimation model, and a machine learning library can be used to implement model training and evaluation. The constructed model can also be evaluated, and methods such as cross-validation can be used to evaluate the performance of the model. The model can be optimized and adjusted according to the evaluation results to improve the accuracy and generalization ability of the model. Through the above steps, the construction and optimization process of the target maintenance cost prediction model based on historical maintenance statistical data can be realized, which can help power grid companies more accurately predict and manage the cost of maintenance projects and improve the efficiency and cost control capabilities of the projects.

[0048] When the above-mentioned preset statistical machine learning algorithm adopts a support vector machine, the target maintenance cost estimation model can be constructed as a least squares support vector machine. Specifically, a support vector machine is a machine learning method based on statistical learning theory, which can achieve better generalization performance by minimizing structural risk rather than empirical risk. The least squares support vector machine (LSSVM) is an improvement and extension of the standard SVM. LSSVM transforms the inequality constraints of the standard SVM into equality constraints, uses the loss function and the square of the error as the empirical loss of the training sample, and transforms the quadratic programming problem of the SVM into the solution of a linear equation system, which reduces the computational complexity and greatly facilitates the solution process. It overcomes the abnormal regression caused by the roughness of the data set and the large volatility of the data set, and can effectively avoid the local optimal problems that appear in methods such as BP neural network. The training data set can be expressed as T={(x i ,y i ),i=1,2,…,N}, where N is the number of training sample pairs, x i It can represent the input vector, y i can represent the output vector.

[0049] Optionally, based on the training data set and the preset statistical machine learning algorithm, a target maintenance cost estimation model is constructed, including: constructing a target error term between the cost prediction result and the output vector in the training data set; based on the target error term, constructing a first objective function for cost prediction; based on the training data set and the preset statistical machine learning algorithm, constructing a first constraint condition for cost prediction; based on the first constraint condition and the first objective function, constructing a target maintenance cost estimation model.

[0050] In an optional embodiment, in the process of predicting the cost of maintenance projects in the power grid, a target error term between the cost prediction result and the output vector in the training data set can be constructed, that is, the difference between the actual prediction result and the actual cost in the training data set is calculated. The target error term can be obtained by calculating the mean square error or other error measurement indicators between the predicted value and the actual value; based on the target error term, a first objective function for cost prediction can be constructed, and minimizing the target error term can be selected as the objective function. In machine learning algorithms, a loss function is usually used to measure the difference between the predicted result and the actual value, and the loss function is minimized by an optimization algorithm; based on the training data set and a preset statistical machine learning algorithm, a first constraint condition for cost prediction can be constructed, and a suitable model and algorithm can be selected according to specific problems and data characteristics, and the parameters and constraints of the model can be set to ensure that the model can converge to a reasonable prediction result during the training process; finally, based on the first constraint condition and the first objective function, a target The maintenance cost estimation model can specifically collect and organize data sets related to the cost of maintenance projects, including characteristic variables of maintenance projects and corresponding cost data, perform preprocessing operations such as missing value processing, outlier processing and feature engineering on the data sets to ensure the quality and availability of the data, divide the data sets into training sets and test sets, train the model with the training set, and evaluate the performance of the model with the test set, select appropriate machine learning algorithms such as linear regression, decision tree, support vector machine, etc. according to the characteristics of the specific problem and the characteristics of the data set, use the training set to train the selected machine learning algorithm, optimize the objective function, minimize the loss function, use the test set to evaluate the performance of the model, compare the error between the predicted results and the actual values, adjust the model parameters to improve the prediction accuracy, apply the trained model to new maintenance project data, perform cost prediction, and make decisions and plans based on the prediction results. Through the above steps, an effective target maintenance cost estimation model can be constructed to accurately predict and plan the cost of maintenance projects in the power grid.

[0051] This application can use the following regression function to obtain the corresponding predicted value, that is, the cost prediction value of each maintenance project:

[0052] ;

[0053] The training data set can be expressed as T={(x i ,y i ),i=1,2,…,N}, where N is the number of training sample pairs, x i It can represent the input vector, y i Can represent the output vector; is a nonlinear function that maps the input vector to a high-dimensional feature space. It can be determined according to actual needs and is not limited here; w is a weight vector; b is a bias parameter, w and b can be determined according to actual needs and are not limited here.

[0054] Then, we can construct the target error term between the cost prediction result and the output vector in the training data set, that is, we can get the predicted value and the output vector y i The error term ε i , the optimization problem can be expressed as follows, that is, the first objective function and the first constraint condition can be obtained as follows:

[0055] ;

[0056] Among them, C is the penalty factor; ε i is the above-mentioned target error term. The above-mentioned first objective function and first constraint condition can also be determined according to actual needs and are not limited here.

[0057] Optionally, based on the first constraint condition and the first objective function, a target maintenance cost estimation model is constructed, including: solving the first objective function based on the first constraint condition to obtain target parameters for constructing the model; constructing multiple target regression functions based on different types of kernel functions and target parameters; constructing a target maintenance cost estimation model based on multiple target regression functions and a target weight matrix, wherein the target weight matrix is ​​used to represent a matrix composed of weight coefficients of the prediction results based on multiple target regression functions in the prediction results of the target maintenance cost estimation model.

[0058] The above-mentioned target parameters may refer to parameters to be solved for constructing a target maintenance cost estimation model. The target parameters may be determined according to the different target maintenance cost estimation models and actual needs, and are not limited here.

[0059] In an optional embodiment, in the process of predicting the cost of maintenance projects in the power grid, the first objective function can be solved based on the first constraint condition to obtain the target parameters for building the model, and the target parameters for building the model can be solved based on the actual data and the constraint condition through data analysis and mathematical modeling methods; multiple target regression functions can be constructed based on different types of kernel functions and target parameters, and appropriate kernel functions can be selected, such as linear kernel functions, polynomial kernel functions, Gaussian kernel functions, etc., and corresponding parameters, and then multiple target regression functions can be constructed through machine learning algorithms, such as support vector machines, kernel ridge regression, etc.; multiple target regression functions can be constructed based on multiple target regression functions and target parameters. A target maintenance cost estimation model is constructed by combining multiple target regression functions and target weight matrices. The target weight matrix is ​​used to represent the weight coefficients of the prediction results of different target regression functions in the prediction model to achieve a weighted combination of the prediction results. Through the above steps, a target maintenance cost estimation model that comprehensively considers multiple target regression functions and weight matrices can be constructed, thereby achieving accurate prediction of the cost of maintenance projects in the power grid. Different types of kernel functions and target parameters can be fully utilized to improve the accuracy and robustness of the prediction model, providing strong support for the budget and management of power grid maintenance projects.

[0060] Optionally, in order to solve the above first objective function, the Lagrangian function can be introduced as follows:

[0061] ;

[0062] Among them, β i is the Lagrange multiplier, which can be determined according to actual needs and is not limited here.

[0063] By solving the partial derivatives of the Lagrangian function, we can get the following partial derivative equation:

[0064] ;

[0065] By eliminating w and ε in the above partial derivative equation i , we can get the following linear equation:

[0066] ;

[0067] in, is a kernel function that satisfies the Mercer condition.

[0068] Target parameters b and β i It can be obtained by solving the above linear equation, and then the regression function of LSSVM can be expressed as follows:

[0069] ;

[0070] The target maintenance cost estimation model can be set with different types of kernel functions. Optionally, the kernel function can be set as follows:

[0071] The radial basis function can be expressed as follows:

[0072] ;

[0073] The linear kernel function can be expressed as follows:

[0074] ;

[0075] The polynomial kernel function can be expressed as follows:

[0076] ;

[0077] in: is the kernel function width, which can be determined according to actual needs and is not limited here; different types of kernel functions can also be determined according to actual needs and are not limited here.

[0078] Based on the above kernel functions and the regression function of LSSVM, multiple target regression functions can be obtained.

[0079] Optionally, the method also includes: based on multiple objective regression functions, performing cost forecasting on historical maintenance engineering quantity data to obtain a forecast result matrix; based on the forecast result matrix and historical maintenance settlement data, constructing a second objective function for measuring the forecast results; constructing constraints that need to be satisfied by each weight coefficient in the target weight matrix to obtain a second constraint; solving the second objective function based on the second constraint to obtain a target weight matrix.

[0080] In an optional embodiment, in the process of predicting the cost of maintenance projects in the power grid, a machine learning algorithm or a statistical method can be used to construct multiple target regression functions to obtain a prediction result matrix; a second target function for measuring the accuracy of the prediction results can be constructed based on the prediction result matrix and historical maintenance settlement data, which can be an error function or other evaluation index; and then the constraints that need to be satisfied by each weight coefficient in the target weight matrix can be determined, such as the sum of the weights being 1; the second target function can be solved based on the second constraint to obtain an optimal target weight matrix for evaluating the accuracy and reliability of the prediction results. Through the above steps, the prediction of the cost of maintenance projects in the power grid can be achieved, and by constructing multiple target regression functions and target weight matrices, the prediction accuracy and credibility can be improved, providing a reference for the budget and decision-making of power grid maintenance projects.

[0081] The present application can establish different prediction base models based on LSSVM of radial basis function, linear kernel function and polynomial kernel function respectively, and the prediction result of the virtual prediction point, that is, the prediction result matrix can be expressed as:

[0082] ;

[0083] in, It can represent the prediction result of the sth prediction point of the lth base model; N s is the number of predicted points.

[0084] The weight matrix corresponding to V, that is, the target weight matrix, can be expressed as:

[0085] ;

[0086] Among them, m l It can represent the weight corresponding to the lth base model.

[0087] The objective function of the optimal virtual prediction, that is, the second objective function and the second constraint condition can be expressed as:

[0088] ;

[0089] Among them, u s It can be the actual value of the predicted point s, and the weight vector M can be the decision variable of this optimization problem, so the sum of its weights can be 1.

[0090] In view of the above optimization problem, the present invention can adopt the differential evolution algorithm to solve it. According to the optimal weight M obtained from the optimization problem, the target weight matrix can be obtained, which is convenient for subsequent combination with the actual prediction values ​​obtained by different base models to obtain the final prediction result of the maintenance cost.

[0091] Optionally, the target maintenance cost calculation model is used to predict the engineering cost of the maintenance project based on target influencing factors to obtain a cost prediction result of the maintenance project, including: using multiple target regression functions in the target maintenance cost calculation model to predict the engineering cost of the maintenance project based on target influencing factors to obtain multiple initial cost prediction results; based on the target weight matrix in the target maintenance cost calculation model, weighted summation of multiple initial cost prediction results is performed to obtain a cost prediction result.

[0092] In an optional embodiment, in the process of predicting the cost of maintenance projects in the power grid, multiple target regression functions in the target maintenance cost estimation model can be used for prediction. Specifically, multiple target regression functions are established based on the collected data related to the maintenance projects. The degree of influence and correlation of each target influencing factor on the project cost can be determined through regression analysis; the established multiple target regression functions can be used to predict the project cost of the maintenance project to obtain multiple initial cost prediction results; the importance and weight of each target can be determined based on the target weight matrix in the target maintenance cost estimation model, and the multiple initial cost prediction results can be weighted and summed up based on the target weight matrix to obtain the final cost prediction result. Through the above steps, using the multiple target regression functions and target weight matrix in the target maintenance cost estimation model, the project cost of the maintenance project can be effectively predicted, the prediction accuracy and reliability can be improved, and managers can better understand the cost influencing factors and carry out cost control and optimization in a targeted manner.

[0093] This application combines the prediction results of different base models to construct a comprehensive prediction model based on the idea of ​​virtual optimal prediction. By taking the prediction results of different cost prediction base models as virtual prediction points, the weight coefficients of each single prediction base model can be adjusted with the goal of minimizing the sum of squares of prediction errors, and the prediction values ​​of different base models at actual prediction points can be combined to obtain a comprehensive prediction result for cost estimation.

[0094] Optionally, the present application may also use an ant colony optimization algorithm to optimize the hyperparameters of LSSVM. The ant colony optimization (ACO) algorithm may be a simulated evolutionary algorithm based on the principle of bionics, that is, a global search algorithm with the advantages of good positive information feedback mechanism, distributed computing, and greedy heuristic search. The present application may regard the parameter selection process of the LSSVM model as a parameter combination optimization problem, construct an ant colony optimization objective function according to the parameter combination, and then use the ACO algorithm to seek its minimum value, and finally output the optimal combination value of the hyperparameters.

[0095] Specifically, the minimum mean square error (MSE) can be used as the ant colony optimization objective function F of the optimization problem. The optimal value F is found through ACO iteration to obtain the optimal parameter combination set , the ant colony optimization objective function is established as follows:

[0096] ;

[0097] In the formula, is the actual value of the maintenance cost of project i; is the predicted value of the maintenance cost of project i; N is the total number of test project samples; and are the minimum and maximum values ​​of parameter C respectively; and The parameters are The minimum and maximum values.

[0098] The idea of ​​parameter optimization of LSSVM is to iteratively search for a set of parameters through ACO. , so that the F value is minimized. The dynamic random method is used to guide the ant colony to perform global search. After multiple iterations, the optimal prediction model of ACO optimized LSSVM, namely ACO-LSSVM, can be obtained. The specific steps can be as follows:

[0099] Parameter initialization, you can set the number of ants to M, the pheromone volatility coefficient , Maximum number of cycles Randomly select a set of parameter sequences as the initial position X of the ant i , the pheromone concentration of ant individual i can be calculated by the following formula.

[0100] ;

[0101] For path transfer, p ants can be randomly selected from the ant colony M, and the ant with the smallest F value (the one with the largest pheromone) is selected as the head ant. , whose location is , and the other ants in the colony move to The location is searched globally in large steps:

[0102] ;

[0103] in, It is an adjustable coefficient and can be determined according to actual needs.

[0104] The head ant generated in the last iteration , perform a local fine search in its adjacent area according to the following two formulas:

[0105] ;

[0106] ;

[0107] in, For the head ant The position after the local search; .

[0108] h is the dynamic search step size, and the update formula can be as follows:

[0109] ;

[0110] in, and are all initial constants and can be taken as , is the current iteration number; is the maximum number of iterations.

[0111] Pheromone update, during the iteration process, the concentration of ant colony pheromone at each location will be updated, and the update rule can be as follows:

[0112] ;

[0113] in, is the pheromone volatility coefficient.

[0114] Check the iteration termination conditions. If the convergence conditions are met, terminate the iteration and output the optimal parameter combination. ; Otherwise, return to the above path transfer step and continue.

[0115] As users' requirements for power supply quality increase, the number of maintenance projects for power grid equipment increases year by year, which makes power companies more and more strict in controlling project funds. Traditional engineering cost methods lack guidance and credibility in actual engineering applications and are insufficient to meet the needs of modern engineering. In the modern power industry, the engineering cost system is developing in the direction of high speed, large scale and intelligence to meet the requirements of modern engineering projects. Power grid equipment maintenance projects are characterized by periodicity and high repetitiveness, and the workload of engineering cost is large. The accuracy and efficiency of manual cost work are low, and it may cause waste of funds, bringing certain losses to power grid companies. With the rapid development of computer technology and network technology, the use of intelligent technical tools to solve the problem of engineering cost prediction is of great significance to improving the efficiency of engineering cost work and reducing waste of funds.

[0116] At present, domestic and foreign researchers have conducted a lot of research on engineering cost estimation technology in combination with computer technology, especially intelligent engineering cost estimation models based on heuristic algorithms and machine learning have received widespread attention. For example, cost estimation models are constructed based on support vector machines, BP neural networks, deep belief networks, etc. Existing studies mostly rely on subjective experience for the selection of machine learning hyperparameters, and the accuracy of the model cannot be guaranteed; in addition, most existing studies use a single model to estimate the cost of engineering projects, which cannot give full play to the advantages of various estimation models and avoid the defects of a single model. The present invention combines the prediction results of different base models to construct a comprehensive prediction model framework based on the idea of ​​virtual optimal prediction, so as to optimize the prediction results and make the prediction results of the estimation model more practical in engineering practice.

[0117] The following describes a comprehensive cost prediction model for 10kV transformer maintenance project based on ant colony optimization algorithm and multi-core least squares support vector machine. The model can include the following steps: historical 10kV transformer maintenance project settlement data can be collected, and the bill of quantities and cost structure can be deeply analyzed to identify the key influencing factors that determine the project cost, and on this basis, a project cost estimation model can be constructed; historical project data and general project rules can be combined. 10kV transformer maintenance projects can include foundation projects, pole tower projects, line engineering, pole-mounted substation and distribution equipment, and demolition projects. Each project contains several sub-projects, and the sub-projects are composed of multiple processes. Among them, the costs of foundation projects, pole tower projects, and line engineering account for a large proportion, usually accounting for more than 90% of the total project cost.

[0118] A process can be the smallest component unit of an engineering project. All process costs are aggregated to form the total project cost. According to the cost calculation method, the process can be divided into a one-time process and a quota process. A one-time process may only include a one-time cost. The cost of a quota process is compiled according to the national or regional quota, including direct costs, indirect costs, profits, and price differences in the compilation base period. The cost structure of a quota process is relatively complex, requiring production project managers to master the professional knowledge of power engineering cost quotas, and omissions may occur when compiling project budgets, resulting in inaccurate project budgets. In addition, power grid equipment maintenance projects are usually characterized by periodicity and high repetitive content. Although the engineering cost is labor-intensive, there are more mechanical and repetitive tasks. By building a cost estimation model based on machine learning, some manual cost work can be replaced to improve cost accuracy and work efficiency.

[0119] As the preliminary work and basic data support for building the estimation model, this application combines historical engineering data samples to identify the key influencing factors that determine the total project cost through numerical statistics and expert experience. In different transformer maintenance projects, similar projects usually contain a large number of similar processes, such as concrete pole assembly, tension corner tower conductor suspension and insulator installation in pole tower projects, and the cost of such processes accounts for most of the project cost and is the key process of the project, that is, the key influencing factor that determines the total project cost. The key processes of each major project can be shown in Table 1 below.

[0120] Table 1

[0121]

[0122] Figure 2 is a flow chart of an optional method for predicting the cost of a maintenance project in a power grid according to an embodiment of the present invention, such as Figure 2As shown, step S202, obtain historical maintenance statistical data of the maintenance project, construct a training data set based on the historical maintenance statistical data, and construct a target maintenance cost estimation model based on the training data set and a preset statistical machine learning algorithm; step S204, obtain the target influencing factors in the maintenance project to be cost predicted in the power grid; step S206, use multiple target regression functions in the target maintenance cost estimation model to predict the engineering cost of the maintenance project based on the target influencing factors, and obtain multiple initial cost prediction results; step S208, based on the target weight matrix in the target maintenance cost estimation model, perform weighted summation on multiple initial cost prediction results to obtain a cost prediction result.

[0123] Figure 3 is a schematic diagram of a process structure of an optional maintenance item according to an embodiment of the present invention, such as Figure 3 As shown in the figure, the transformer maintenance production project can be divided into three projects: foundation engineering, tower engineering and wire laying engineering; the project can be further divided into sub-projects. The foundation engineering can be divided into the on-site transportation of foundation engineering materials and foundation masonry. The tower engineering can be divided into the on-site transportation of tower engineering materials and tower assembly. The wire laying engineering can be divided into the on-site transportation of wire laying engineering materials and conductor laying. The sub-projects can be divided into processes. The on-site transportation of foundation engineering materials can be divided into manpower, vehicle and ship transportation. The foundation masonry can be divided into concrete mixing and pouring, steel bar processing and Production and manual excavation and backfilling of pole pits, tower pits and wire pulling pits, on-site transportation of pole tower engineering materials can be divided into manpower, vehicles and ship transportation, pole tower assembly can be divided into line re-survey and sub-pit, concrete pole assembly and tension corner pole tower conductor hanging and insulator installation, on-site transportation of line engineering materials can be divided into manpower, vehicles and ship transportation, conductor erection can be divided into conductor erection below 10kv, crossarm installation below 10kv and straight pole tower insulator suspension installation, the process composition of the maintenance project can also be determined according to actual needs and is not limited here.

[0124] Figure 4 is a schematic diagram of an optional process cost composition according to an embodiment of the present invention, such as Figure 4As shown, taking the quota-based process costs as an example, the quota-based process costs may include indirect costs, price differences in the preparation base period, profits and direct costs, among which indirect costs include regulatory fees and enterprise management fees, regulatory fees include social insurance premiums and housing provident funds, price differences in the preparation base period include labor price difference costs, machinery price difference costs, price difference costs of priced materials supplied by Party A and price difference costs of priced materials supplied by Party B, direct costs include direct engineering costs and measures costs, direct engineering costs include labor costs, construction machinery usage fees, priced materials supplied by Party A, priced materials supplied by Party B, non-priced materials supplied by Party A and non-priced materials supplied by Party B, measures costs include additional costs for construction in winter and rainy seasons, additional costs for nighttime construction, fees for the use of construction tools and equipment, additional costs for construction in special areas, temporary facilities fees, safe and civilized construction fees, additional costs for multiple entries and exits and construction organization relocation fees, and the process cost composition can also be determined according to actual needs, which is not limited here.

[0125] Figure 5 is an optional flow chart of optimizing the target maintenance cost estimation model by using the ant colony optimization algorithm according to an embodiment of the present invention, such as Figure 5 As shown, step S502 is to initialize the target maintenance cost estimation model and the ant colony optimization algorithm parameters; step S504 is to initialize the ant colony and perform the ant colony search operation; step S506 is to calculate the ant colony optimization objective function using the target maintenance cost estimation model training samples; step S508 is to determine whether the termination conditions are met to output the search optimal parameters.

[0126] The technical solution proposed in this application can analyze the project composition and cost composition of the transformer maintenance project, and identify the key influencing factors that determine the total project cost based on the characteristics of historical engineering data, and build a machine learning-based engineering cost estimation model framework based on this; LSSVM models based on radial basis functions, linear kernel functions and polynomial kernel functions can be constructed to establish different estimation base models, and the ant colony optimization algorithm can be used to optimize the LSSVM hyperparameters; the prediction results of different base models can be combined to construct a comprehensive prediction model based on the idea of ​​virtual optimal prediction, and the optimal prediction result can be obtained by optimizing the weights of different base models.

[0127] This application optimizes the hyperparameters of LSSVM through ACO, thereby effectively improving the prediction accuracy of the LSSVM prediction model; this application uses LSSVM prediction models with multiple different types of kernel functions as base models to calculate the project costs, giving full play to the advantages of various calculation models and avoiding the defects of a single model; the comprehensive prediction framework based on the virtual optimal prediction idea proposed in the present invention solves the weights of different base models, effectively combines the prediction results of different base models, obtains more stable and excellent prediction performance, and optimizes the prediction results of project costs.

[0128] According to another aspect of an embodiment of the present invention, a device for predicting the cost of maintenance projects in a power grid is also provided. The device can execute the method for predicting the cost of maintenance projects in a power grid of the above embodiment. The specific implementation method and preferred application scenario are the same as those of the above embodiment and will not be repeated here.

[0129] Figure 6 is a schematic diagram of a device for predicting the cost of a maintenance project in a power grid according to an embodiment of the present application, such as Figure 6 As shown, the device includes the following: an acquisition module 602 and a prediction module 604 .

[0130] Among them, the acquisition module 602 is used to obtain the target influencing factors in the maintenance project to be cost predicted in the power grid, wherein the target influencing factors are used to represent the target process in the maintenance project, and the target process is used to represent the process whose execution times are greater than the preset times in the basic processes constituting the maintenance project; the prediction module 604 is used to use the target maintenance cost calculation model to predict the engineering cost of the maintenance project based on the target influencing factors, and obtain the cost prediction result of the maintenance project, wherein the target maintenance cost calculation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction.

[0131] Among them, the above-mentioned device also includes a construction module, which is used to obtain historical maintenance statistical data of the maintenance project, wherein the historical maintenance statistical data at least includes historical maintenance engineering quantity data and historical maintenance settlement data; construct a training data set based on the historical maintenance statistical data, wherein the input vector of the training sample pair in the training data set is generated by the historical maintenance engineering quantity data, and the output vector of the training sample pair is generated by the historical maintenance settlement data; based on the training data set and a preset statistical machine learning algorithm, a target maintenance cost estimation model is constructed.

[0132] Among them, the construction module is also used to construct a target error term between the cost prediction result and the output vector in the training data set; based on the target error term, a first objective function for cost prediction is constructed; based on the training data set and a preset statistical machine learning algorithm, a first constraint condition for cost prediction is constructed; based on the first constraint condition and the first objective function, a target maintenance cost estimation model is constructed.

[0133] Among them, the construction module is also used to solve the first objective function based on the first constraint condition to obtain the target parameters of the construction model; construct multiple target regression functions based on different types of kernel functions and target parameters; construct a target maintenance cost estimation model based on multiple target regression functions and target weight matrices, wherein the target weight matrix is ​​used to represent a matrix composed of weight coefficients of the prediction results based on multiple target regression functions in the prediction results of the target maintenance cost estimation model.

[0134] Among them, the construction module is also used to make cost predictions for historical maintenance engineering quantity data based on multiple objective regression functions to obtain a prediction result matrix; based on the prediction result matrix and historical maintenance settlement data, a second objective function for measuring the prediction results is constructed; the constraints that need to be satisfied by each weight coefficient in the target weight matrix are constructed to obtain a second constraint; based on the second constraint, the second objective function is solved to obtain a target weight matrix.

[0135] Among them, the prediction module is also used to utilize multiple target regression functions in the target maintenance cost estimation model to predict the engineering cost of the maintenance project based on the target influencing factors, and obtain multiple initial cost prediction results; based on the target weight matrix in the target maintenance cost estimation model, multiple initial cost prediction results are weighted and summed to obtain the cost prediction result.

[0136] An embodiment of the present application further provides an electronic device, comprising: a memory storing an executable program; and a processor for running the program, wherein the method in each embodiment of the present invention is executed when the program is running.

[0137] The above-mentioned memory may refer to a device inside a computer for storing data and programs, and may include memory, hard disk, etc., wherein the memory may be used to temporarily store running programs and data, the hard disk may be used to store programs and data for a long time, and the memory may be used to enable the computer to read and write data, and execute programs; the above-mentioned processor may be responsible for executing instructions in computer programs and performing data processing, and may be responsible for controlling and executing various operations, including arithmetic operations, logical operations, data transmission, etc.

[0138] An embodiment of the present application further provides a computer-readable storage medium, which includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the methods in various embodiments of the present invention.

[0139] The above-mentioned computer storage medium may refer to a medium in a computer memory used to store certain discontinuous physical quantities. Computer storage media mainly include semiconductors, magnetic cores, magnetic drums, magnetic tapes, laser disks, etc.; the stored program included in the computer-readable storage medium may be a set of instructions that can be recognized and executed by a computer, running on an electronic computer, and is an information tool that meets certain needs of people.

[0140] An embodiment of the present application further provides a computer program product, including a computer program, which implements the methods in various embodiments of the present invention when executed by a processor.

[0141] The above-mentioned computer program product may refer to a software program that has been written, tested and released and can be run on a computer or other device. The computer program product may include an application, an operating system, tool software, etc., which is used to implement specific functions or solve specific problems.

[0142] An embodiment of the present application further provides a computer program product, including a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium is used to store a computer program, and when the computer program is executed by a processor, the method in each embodiment of the present invention is implemented.

[0143] The above-mentioned non-volatile computer-readable storage medium may refer to a medium for storing data. The non-volatile computer-readable storage medium can keep the data from being lost when the power is off, and can be used to store long-term data, such as operating systems, applications, and user files. The non-volatile storage medium may include hard disk drives, solid-state drives, optical disks, and flash memory storage devices, etc.

[0144] The embodiments of the present application further provide a computer program, which implements the methods in the above-mentioned embodiments of the present invention when executed by a processor.

[0145] The above-mentioned computer program may refer to a collection of instructions used to tell a computer to perform a specific task or operation. A computer program may be written by a programmer using a specific programming language and may include algorithms, data structures, logic, and control flows. A computer program may be used for a variety of purposes, including application software, operating systems, and the like.

[0146] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0148] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0149] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0150] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc. Various media that can store program codes.

[0151] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for predicting the cost of maintenance projects in power grids. It is characterized in that include: Obtaining target influencing factors in a maintenance project to be cost-predicted in a power grid, wherein the target influencing factors are used to represent target processes in the maintenance project, and the target processes are used to represent processes whose execution times are greater than a preset number of times among basic processes constituting the maintenance project; The target maintenance cost estimation model is used to predict the engineering cost of the maintenance project based on the target influencing factors to obtain the cost prediction result of the maintenance project, wherein the target maintenance cost estimation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction.

2. The method for predicting the cost of maintenance projects in a power grid according to claim 1, It is characterized in that The method further comprises: Obtaining historical maintenance statistical data of the maintenance project, wherein the historical maintenance statistical data at least includes historical maintenance engineering quantity data and historical maintenance settlement data; Constructing a training data set based on the historical maintenance statistical data, wherein the input vectors of the training sample pairs in the training data set are generated by the historical maintenance engineering quantity data, and the output vectors of the training sample pairs are generated by the historical maintenance settlement data; Based on the training data set and a preset statistical machine learning algorithm, the target maintenance cost estimation model is constructed.

3. The method for predicting the cost of maintenance projects in a power grid according to claim 2, It is characterized in that Based on the training data set and the preset statistical machine learning algorithm, the target maintenance cost estimation model is constructed, including: Constructing a target error term between the cost prediction result and the output vector in the training data set; Based on the target error term, construct a first objective function for cost prediction; Based on the training data set and the preset statistical machine learning algorithm, construct a first constraint condition for cost prediction; Based on the first constraint condition and the first objective function, the target maintenance cost estimation model is constructed.

4. The method for predicting the cost of maintenance projects in a power grid according to claim 3, It is characterized in that Based on the first constraint condition and the first objective function, constructing the target maintenance cost estimation model includes: Solving the first objective function based on the first constraint condition to obtain target parameters for constructing a model; Based on the different types of kernel functions and the target parameters, construct a plurality of target regression functions; Based on the multiple target regression functions and the target weight matrix, the target maintenance cost estimation model is constructed, wherein the target weight matrix is ​​used to represent a matrix composed of weight coefficients of the prediction results based on the multiple target regression functions in the prediction results of the target maintenance cost estimation model.

5. The method for predicting the cost of maintenance projects in a power grid according to claim 4, It is characterized in that The method further comprises: Based on the multiple target regression functions, the cost of the historical maintenance engineering quantity data is predicted to obtain a prediction result matrix; Based on the prediction result matrix and the historical maintenance settlement data, construct a second objective function for measuring the prediction result; Constructing the constraint conditions that each weight coefficient in the target weight matrix needs to satisfy to obtain a second constraint condition; The second objective function is solved based on the second constraint condition to obtain the objective weight matrix.

6. The method for predicting the cost of maintenance projects in a power grid according to claim 1, It is characterized in that The target maintenance cost estimation model is used to predict the engineering cost of the maintenance project based on the target influencing factors to obtain the cost prediction result of the maintenance project, including: Using multiple target regression functions in the target maintenance cost estimation model, based on the target influencing factors, the engineering cost of the maintenance project is predicted to obtain multiple initial cost prediction results; Based on the target weight matrix in the target maintenance cost estimation model, the multiple initial cost prediction results are weighted and summed to obtain the cost prediction result.

7. A device for predicting the cost of maintenance projects in power grids. It is characterized in that include: An acquisition module, used for acquiring target influencing factors in a maintenance project to be cost predicted in a power grid, wherein the target influencing factors are used to represent target processes in the maintenance project, and the target processes are used to represent processes whose execution times are greater than a preset number of times among the basic processes constituting the maintenance project; A prediction module is used to predict the engineering cost of the maintenance project based on the target influencing factors using a target maintenance cost estimation model to obtain a cost prediction result of the maintenance project, wherein the target maintenance cost estimation model is used to represent a machine learning model that uses different types of kernel functions as base models for cost prediction.

8. An electronic device, It is characterized in that include: A memory storing an executable program; A processor is used to run the program, wherein when the program is run, the method for predicting the cost of maintenance projects in the power grid as described in any one of claims 1 to 6 is executed.

9. A computer-readable storage medium, It is characterized in that The computer-readable storage medium includes a stored executable program, wherein when the executable program is executed, the device where the storage medium is located is controlled to execute the method for predicting the cost of maintenance projects in a power grid as described in any one of claims 1 to 6.

10. A computer program product, It is characterized in that It comprises a computer program, which, when executed by a processor, implements the method for predicting the cost of maintenance projects in a power grid according to any one of claims 1 to 6.

Images