Electric power infrastructure project evaluation model construction method and system for risk evaluation

Abstract

The invention relates to an electric power infrastructure project assessment model construction method and system for risk assessment. The method comprises the steps of obtaining historical risk data and real-time monitoring data of a to-be-assessed electric power infrastructure project; preprocessing the historical risk data and the real-time monitoring data to obtain a standardized data set; performing feature engineering processing on the standardized data set, extracting statistical features, time sequence features and associated features, and performing screening to obtain key risk features; constructing feature vectors according to the key risk features; performing parameter initialization on a preset machine learning algorithm according to the dimension of the feature vector, and constructing an initial risk assessment model; and training and optimizing the initial risk assessment model by using the standardized data set to obtain a target risk assessment model which is used for performing risk assessment on the real-time monitoring data of the to-be-assessed power infrastructure project and outputting a risk assessment result. Compared with the prior art, the method improves the comprehensiveness and flexibility of project risk assessment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and in particular to a power infrastructure project evaluation model construction method and system for risk assessment. BACKGROUND

[0002] Power infrastructure projects (such as transmission line projects, substation construction, and new energy power plant development) have characteristics such as large investment scale, long construction period, high technical complexity, and many environmental impact factors. Risk management is always the core of project management. Traditional power infrastructure project risk assessment methods mainly rely on manual experience judgment and static statistical models, such as expert scoring, analytic hierarchy process (AHP), or fuzzy comprehensive evaluation methods for qualitative or semi-quantitative assessment. However, with the deepening of digital transformation of the power industry and the increasing demand for refined project management, the existing technology has exposed many limitations, which are embodied in the following aspects: Risk feature extraction is the core of model evaluation, but the existing technology has obvious shortcomings in feature engineering. Traditional methods rely more on manual experience to construct simple statistical features (such as mean, variance), while ignoring time series features (such as trend, period, autocorrelation) and correlation features (such as the correlation between equipment status and construction progress, and the correlation between environmental parameters and risk events). For example, the delay of construction progress may be related to consecutive rainfall weather, but the existing technology does not capture this dynamic correlation through time series analysis; equipment operation state anomalies may be related to high load operation in a specific construction phase, but the existing feature engineering does not mine this cross-dimensional correlation. In addition, the feature selection process relies more on subjective judgment or a single algorithm (such as only using Pearson correlation coefficient for screening), and lacks a multi-algorithm fusion screening mechanism (such as combining random forest importance score and L1 regularization), resulting in missing key risk features or retaining redundant features, affecting the generalization ability and evaluation accuracy of the model. SUMMARY

[0003] The purpose of the present application is to overcome the above-mentioned defects of the existing feature engineering not mining cross-dimensional correlation, and to provide a power infrastructure project evaluation model construction method and system for risk assessment.

[0004] The purpose of the present application can be achieved by the following technical solutions: A power infrastructure project evaluation model construction method for risk assessment, comprising: obtaining historical risk data and real-time monitoring data of the power infrastructure project to be evaluated; preprocessing the historical risk data and real-time monitoring data to obtain a standardized data set; perform feature engineering on the standardized dataset, extract statistical features, time series features and correlation features, and perform screening to obtain key risk features; construct a feature vector according to the key risk features; perform parameter initialization on a preset machine learning algorithm according to the dimension of the feature vector, and construct an initial risk assessment model; train and optimize the initial risk assessment model using the standardized dataset to obtain a target risk assessment model for real-time monitoring data of the power infrastructure project to be evaluated, and output a risk assessment result.

[0005] Further, the statistical features include mean, variance, maximum value, minimum value and change rate; The time series features include trend features, periodic features and autocorrelation features; The correlation features are extracted based on calculating the Pearson correlation coefficient, and the correlation features include the correlation between device status and construction progress, and the correlation between environmental parameters and risk events; Sort the key risk features according to a preset dimension to construct the feature vector.

[0006] Further, the calculation expression of the Pearson correlation coefficient is: In the formula, is the Pearson correlation coefficient between feature X and feature Y, is the covariance of feature and feature , and and are the standard deviations of feature and feature , respectively.

[0007] Further, the screening process of the key risk features includes: Calculate the Pearson correlation coefficient between each statistical feature, time series feature and correlation feature and the risk event, and select the features with an absolute value of the Pearson correlation coefficient greater than a preset correlation threshold as first features; Calculate the importance scores of each statistical feature, time series feature and correlation feature through a random forest algorithm, and select a plurality of features with higher importance scores as second features; Perform sparse processing on each statistical feature, time series feature and correlation feature through an L1 regularization logistic regression model to select a plurality of features as third features; Perform intersection or union processing on the first features, second features and third features to obtain the key risk features; The expression of the loss function of the L1 regularization logistic regression model is: In the formula, is a loss function, is an original logistic regression loss term, is a regularization coefficient, is a model parameter L1 norm of the model parameter by adjusting the regularization coefficient so that the parameters corresponding to the unimportant features tend to zero, thereby realizing feature selection.

[0008] Further, the construction process of the initial risk assessment model comprises: According to the input dimension of the feature vector, determine the input requirement of the machine learning algorithm; According to the input requirement of the machine learning algorithm, select a machine learning algorithm from a random forest model, an XGBoost model, a LightGBM model and a neural network model; Based on the input dimension of the feature vector and the parameter space of the selected machine learning algorithm, initialize the model parameters of the machine learning algorithm to obtain an initial risk assessment model.

[0009] Further, the initial risk assessment model is trained and optimized using the standardized data set to obtain a target risk assessment model, comprising: Divide the standardized data set into a training data set and a validation data set; Train the initial risk assessment model based on the training data set; Input the validation data set into the trained initial risk assessment model for evaluation, adjust the hyperparameters of the initial risk assessment model according to the evaluation result, until the preset training termination condition is reached, and obtain the target risk assessment model.

[0010] Further, the real-time monitoring data of the power infrastructure project to be evaluated is risk assessed, and a risk level is output, comprising: According to the real-time monitoring data of the power infrastructure project to be evaluated, obtain real-time feature data, which includes equipment status, construction progress and environmental parameters; Convert the real-time feature data into a feature vector format and input it into the target risk assessment model; Output a risk score through the target risk assessment model, which is a probability value of a risk event occurring in the project; According to a preset risk level division rule, map the risk score to different risk levels; The contribution of each feature to the risk score is calculated based on the SHAP value, and the risk factor explanation is output to indicate the key features that lead to the risk. The expression for calculating the SHAP value is: In the formula, To explain the model, For the total number of features, As the baseline value, For the first The SHAP value of each feature.

[0011] Furthermore, the method also includes: Construct at least two initial risk assessment models with different algorithm types to obtain multiple target risk assessment models; The outputs of each target risk assessment model are fused using a weighted average method to obtain a comprehensive risk score, which serves as the final risk assessment result. The formula for calculating the comprehensive risk score is as follows: In the formula, For comprehensive risk scoring, For the first The output score of the target risk assessment model For the first The weights of each target risk assessment model, and satisfying Weight The F1-Score of each target risk assessment model is determined based on the validation set.

[0012] Furthermore, the preprocessing includes: Data cleaning is performed on historical risk data and real-time monitoring data. This data cleaning includes removing duplicate data, filling in missing values, and correcting outliers. The cleaned data is then normalized. The normalized data is time-aligned to obtain the standardized dataset.

[0013] This invention also provides a power infrastructure project assessment model construction system for risk assessment, used to implement the power infrastructure project assessment model construction method for risk assessment as described above, including: The acquisition module is used to acquire historical risk data and real-time monitoring data of the power infrastructure projects to be evaluated; The preprocessing module is used to preprocess the historical risk data and real-time monitoring data to obtain a standardized dataset; An extraction module is configured to perform feature engineering on the standardized dataset, extract statistical features, time series features and correlation features, and perform screening to obtain key risk features; and a feature vector is constructed according to the key risk features; A construction module is configured to perform parameter initialization on a preset machine learning algorithm according to the dimension of the feature vector, and construct an initial risk assessment model; A training module is configured to train and optimize the initial risk assessment model by using the standardized dataset, and obtain a target risk assessment model; An output module is configured to perform risk assessment on real-time monitoring data of a power infrastructure project to be evaluated based on the target risk assessment model, and output a risk assessment result.

[0014] Compared with the prior art, the present application has the following advantages: (1) The present application constructs an initial risk assessment model according to a feature vector and a preset machine learning algorithm. The preset multiple machine learning algorithms provide flexibility for model construction. Different algorithms have different characteristics and application scenarios. By trying multiple algorithms, the most suitable algorithm can be selected to construct a model according to the specific characteristics and data characteristics of the power infrastructure project, so as to better fit the data and predict the risk. Moreover, the target risk assessment model can be used to perform real-time risk assessment on the current power infrastructure project, and output the risk level in time. This is crucial for the managers of the power infrastructure project, who can know the risk status of the project at any time during the implementation of the project, so as to take corresponding measures to respond to and control the risk in time, avoid the risk from further expanding and deteriorating, and ensure the smooth progress of the project.

[0015] (2) The present application extracts statistical features, time series features and correlation features at the same time, realizes deep fusion of multiple sources of information, and avoids information loss caused by a single feature dimension. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 A flowchart of a power infrastructure project evaluation model construction method for risk assessment provided in an embodiment of the present application is shown in the figure; Figure 2 A detailed flowchart of steps S1 and S2 provided in an embodiment of the present application is shown in the figure; Figure 3 A detailed flowchart of step S3 provided in an embodiment of the present application is shown in the figure; Figure 4This is a schematic diagram illustrating the process of obtaining key risk features provided in an embodiment of the present invention; Figure 5 This is a detailed flowchart illustrating step S4 provided in an embodiment of the present invention; Figure 6 This is a detailed flowchart of step S5 provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0018] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0019] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0020] Example 1 like Figure 1 As shown in the figure, this embodiment provides a method for constructing an assessment model for power infrastructure projects for risk assessment, including: S1. Obtain historical risk data and real-time monitoring data of the power infrastructure projects to be evaluated; S2. Preprocess historical risk data and real-time monitoring data to obtain a standardized dataset; S3. Perform feature engineering on the standardized dataset to extract statistical features, time series features, and correlation features, and then filter them to obtain key risk features; construct feature vectors based on the key risk features. S4. Initialize the parameters of the preset machine learning algorithm according to the dimension of the feature vector to build an initial risk assessment model; S5. Use standardized datasets to train and optimize the initial risk assessment model to obtain the target risk assessment model, which is used to assess the risk of the power infrastructure project to be assessed using real-time monitoring data, and outputs the risk assessment results.

[0021] The skilled in the art can understand that the present application comprehensively utilizes the experience information accumulated in the past and the real-time situation information occurring at present by acquiring the historical risk data and real-time monitoring data of the power infrastructure project, and the comprehensive integration enables the model to consider the project risk from multiple dimensions, avoids the evaluation deviation caused by relying on a single data source, and lays a solid foundation for accurate risk assessment; the acquired data is preprocessed to obtain a standardized data set, eliminating the differences in dimensions, formats, etc. that may exist between different data sources and different data types, and the standardized data is more convenient for subsequent analysis and modeling, improves the data processing efficiency, and also helps to improve the data processing ability of the model and the accuracy of the evaluation results; According to the feature vector and the preset machine learning algorithm, an initial risk assessment model is constructed, and the preset machine learning algorithm provides flexibility for model construction. Different algorithms have different characteristics and application scenarios. By trying multiple algorithms, the most suitable algorithm can be selected to construct the model according to the specific characteristics and data characteristics of the power infrastructure project, so as to better fit the data and predict the risk; the training data set is used to train and optimize the initial risk assessment model to obtain a target risk assessment model. During the training process, the model can continuously learn and adjust parameters to adapt to the changes of data and the dynamic characteristics of project risks. Through continuous optimization, the model can gradually improve the accuracy and reliability of the evaluation, and better cope with the complex and changeable risk situation in the actual project; based on the target risk assessment model, real-time risk assessment of the current power infrastructure project can be performed, and the risk level can be output in time. This is crucial for the managers of the power infrastructure project, which enables them to understand the risk situation of the project at any time during the implementation of the project, so as to take appropriate measures to respond to and control the risk in time, avoid the risk from further expanding and deteriorating, and ensure the smooth progress of the project.

[0022] Please refer to Figure 2 The historical risk data and real-time monitoring data are preprocessed to obtain a standardized data set, specifically including: S201, cleaning the historical risk data and real-time monitoring data to remove duplicate data, fill in missing values, and correct abnormal values; S202, normalizing the cleaned data, and the Min-Max normalization is used for normalization, and the formula is: Wherein, is the normalized data value, is the original data value, is the minimum value of all data of the feature, is the maximum value of all data of the feature; S203, time alignment is performed on the normalized data to ensure consistency of historical data and real-time monitoring data in the time dimension, and a standardized data set is obtained.

[0023] Please refer to Figure 3 Based on the standardized data set, feature engineering processing is performed, key risk features are extracted, and a feature vector is constructed, which specifically includes: S301, statistical features are extracted from the standardized data set, including mean, variance, maximum value, minimum value, and change rate; S302, time series features are extracted from the standardized data set, including trend features, periodic features, and autocorrelation features; S303, Pearson correlation coefficient is used to calculate the correlation features in the standardized data set, including the correlation between device status and construction progress, and the correlation between environmental parameters and risk events; The Pearson correlation coefficient is used to calculate the correlation features in the standardized data set, and the formula is: Wherein, is and is the covariance of and are the standard deviations of and , respectively, The value range of is [-1, 1], and the greater the absolute value, the stronger the correlation; S304, the statistical features, time series features, and correlation features are screened through a feature selection algorithm to obtain key risk features; S305, the key risk features are sorted according to the preset dimension to construct a feature vector.

[0024] Please refer to Figure 4 The statistical features, time series features, and correlation features are screened through a feature selection algorithm to obtain key risk features, which specifically includes: S401, calculate the Pearson correlation coefficient of each feature and the risk event, and select the features with an absolute value greater than a preset threshold; S402, calculate the importance score of each feature through the random forest algorithm, sort them from high to low, and select the top N features; S403, based on the L1 regularization logistic regression model, the features are sparsified, and the loss function formula is: Wherein, Loss function, is the original logistic regression loss term, is a regularization coefficient, is a model parameter L1 norm, i.e., the sum of the absolute values of each parameter, by adjusting force the parameters corresponding to unimportant features tend to zero, thereby achieving feature selection; S404, the intersection or union of the features screened by the Pearson correlation coefficient, the features screened by the random forest importance score, and the features reserved by the L1 regularization is processed to obtain key risk features.

[0025] Please refer to Figure 5 , an initial risk assessment model is constructed according to the feature vector and the preset machine learning algorithm, specifically including: S501, determine the input dimension of the feature vector, and match the input requirements of the preset machine learning algorithm; S502, select a preset machine learning algorithm, the preset machine learning algorithm including at least one of random forest, XGBoost, LightGBM, and neural network; S503, initialize the model parameters based on the dimension of the feature vector and the parameter space of the preset machine learning algorithm, and construct an initial risk assessment model.

[0026] Please refer to Figure 6 , the initial risk assessment model is trained and optimized using the training data set to obtain a target risk assessment model, specifically including: S601, divide the feature vector into a training data set and a validation data set, the ratio of the training data set to the validation data set being a preset value; S602, train the initial risk assessment model based on the training data set, and evaluate the model performance by the K-fold cross-validation method; The performance evaluation index of K-fold cross-validation adopts F1-Score, and its formula is: wherein, Precision is the precision, and Recall is the recall; S603, adjust the model hyperparameters according to the evaluation results of the validation data set, the hyperparameters including the depth of the tree, the learning rate, and the proportion of the sub-sample; S604, when the performance index of the model on the validation data set reaches a preset threshold, stop training, and obtain a target risk assessment model.

[0027] Based on the target risk assessment model, the current electric power infrastructure project is evaluated for risk, and a risk level is output, specifically including: Obtain real-time feature data of the current electric power infrastructure project, the real-time feature data including equipment status, construction progress, and environmental parameters; Convert real-time feature data into a format consistent with feature vectors and input it into the target risk assessment model; The risk score is output through the target risk assessment model. The risk score is the probability value of the project experiencing a risk event, which ranges from [0, 1]. According to the preset risk level classification rules, the risk score is mapped to the risk level, which includes low risk, medium risk and high risk. The contribution of each feature to the risk score is calculated based on the SHAP value, and the risk factor explanation is output. The risk factor explanation is used to indicate the key features that lead to the risk.

[0028] The contribution of each feature to the risk score is calculated based on the SHAP value, and the model output value can be expressed as: in, To explain the model, For the total number of features, The baseline value is denoted as , and the average predicted value is denoted as . For the first The SHAP value of each feature.

[0029] It also includes a multi-model fusion step: Construct at least two initial risk assessment models with different algorithm types, including tree models, neural network models, and probabilistic models; Each initial model was trained and optimized to obtain multiple target risk assessment models; The comprehensive risk score formula is obtained by fusing the outputs of multiple target risk assessment models using a weighted average method: in, For comprehensive risk scoring, For the first The output score of each model For the first The weights of each model, and satisfying Weight The score is determined based on the F1-Score of each model on the validation set.

[0030] Example 2 This embodiment provides a power infrastructure project assessment model construction system for risk assessment, used to implement a power infrastructure project assessment model construction method for risk assessment as described in Embodiment 1, including: The acquisition module is used to acquire historical risk data and real-time monitoring data of the power infrastructure projects to be evaluated; The preprocessing module is configured to preprocess historical risk data and real-time monitoring data to obtain a standardized data set. The extraction module is configured to perform feature engineering processing on the standardized data set, extract statistical features, time series features and correlation features, and perform screening to obtain key risk features; and construct a feature vector according to the key risk features. The construction module is configured to initialize parameters of a preset machine learning algorithm according to a dimension of the feature vector, and construct an initial risk assessment model. The training module is configured to train and optimize the initial risk assessment model by using the standardized data set to obtain a target risk assessment model. The output module is configured to perform risk assessment on real-time monitoring data of the power infrastructure project to be evaluated based on the target risk assessment model, and output a risk assessment result.

[0031] It should be noted that the specific content and beneficial effects of the device of the system can be referred to the above method embodiments, which will not be repeated here.

[0032] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code. The solutions in the embodiments of the present application can be implemented in various computer languages, such as object-oriented programming languages Java and interpreted scripting language JavaScript.

[0033] The present application is described with reference to flowcharts and / or block diagrams according to the method, device (system) and computer program product of the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 The device for implementing the functions specified in one flow or multiple flows and / or blocks Figure 1 The device for implementing the functions specified in one flow or multiple flows and / or blocks

[0034] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0035] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions that are executed on the computer or other programmable apparatus provide steps for implementing the Figure 1 function specified in the flow or flows and / or blocks Figure 1 of the block or blocks.

[0036] While the preferred embodiments of the application have been described, additional variations and modifications can be made to the preferred embodiments by those of skill in the art once they have the benefit of the present disclosure. Therefore, the appended claims are intended to encompass within their scope all possible variations and modifications of the preferred embodiments.

[0037] It is apparent that a person skilled in the art can make various changes and modifications to the application without departing from the spirit and scope thereof. Thus, if these modifications and variations of the application fall within the scope of the claims and their equivalents, it is intended to include them in the application.

Claims

1. A method for constructing an assessment model for power infrastructure projects for risk assessment, characterized in that, include: Obtain historical risk data and real-time monitoring data for the power infrastructure projects to be evaluated; The historical risk data and real-time monitoring data are preprocessed to obtain a standardized dataset; Feature engineering is performed on the standardized dataset to extract statistical features, time series features, and correlation features, and these features are then filtered to obtain key risk features; feature vectors are then constructed based on these key risk features. The parameters of the preset machine learning algorithm are initialized according to the dimension of the feature vector to construct an initial risk assessment model; The initial risk assessment model is trained and optimized using the standardized dataset to obtain the target risk assessment model, which is used to assess the risk of the power infrastructure project to be assessed based on real-time monitoring data, and outputs the risk assessment results.

2. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 1, characterized in that, The statistical characteristics include mean, variance, maximum value, minimum value, and rate of change; The time series features include trend features, periodic features, and autocorrelation features; The correlation features are extracted based on the calculation of the Pearson correlation coefficient, and the correlation features include the correlation between equipment status and construction progress, and the correlation between environmental parameters and risk events. The key risk features are sorted according to preset dimensions to construct the feature vector.

3. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 2, characterized in that, The formula for calculating the Pearson correlation coefficient is as follows: In the formula, The Pearson correlation coefficient is the coefficient between feature X and feature Y. Features and characteristics covariance, and Features and characteristics The standard deviation.

4. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 2, characterized in that, The screening process for the key risk characteristics includes: Calculate the Pearson correlation coefficients between each statistical feature, time series feature, and correlation feature and the risk event, and select features whose absolute Pearson correlation coefficient is greater than a preset correlation threshold as the first feature; The importance scores of each statistical feature, time series feature, and correlation feature are calculated using the random forest algorithm, and several features with higher importance scores are selected as the second feature. The statistical features, time series features, and correlation features are sparsified by using an L1 regularized logistic regression model to select multiple features as the third feature. The key risk features are obtained by intersecting or unifying the first, second, and third features. The expression for the loss function of the L1 regularized logistic regression model is as follows: In the formula, For loss function, This is the original logistic regression loss term. The regularization coefficient is . For model parameters The L1 norm is determined by adjusting the regularization coefficient. Make the parameters corresponding to unimportant features Approaching zero, thus achieving feature selection.

5. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 1, characterized in that, The process of constructing the initial risk assessment model includes: Based on the input dimension of the feature vector, determine the input requirements for the machine learning algorithm; Based on the input requirements of the machine learning algorithm, a machine learning algorithm is selected from the random forest model, XGBoost model, LightGBM model, and neural network model; Based on the input dimension of the feature vector and the parameter space of the selected machine learning algorithm, the model parameters of the machine learning algorithm are initialized to obtain the initial risk assessment model.

6. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 1, characterized in that, The initial risk assessment model is trained and optimized using the standardized dataset to obtain the target risk assessment model, including: The standardized dataset is divided into a training dataset and a validation dataset; The initial risk assessment model is trained based on the training dataset. The validation dataset is input into the trained initial risk assessment model for evaluation. The hyperparameters of the initial risk assessment model are adjusted based on the evaluation results until the preset training cutoff condition is reached, thus obtaining the target risk assessment model.

7. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 1, characterized in that, A risk assessment is conducted using real-time monitoring data of the power infrastructure projects to be evaluated, and the risk level is output, including: Based on the real-time monitoring data of the power infrastructure project to be evaluated, real-time characteristic data is obtained, including equipment status, construction progress and environmental parameters. The real-time feature data is converted into a feature vector format and then input into the target risk assessment model. The target risk assessment model outputs a risk score, which is the probability value of a risk event occurring in the project. According to the preset risk level classification rules, the risk score is mapped to different risk levels; The contribution of each feature to the risk score is calculated based on the SHAP value, and the risk factor explanation is output to indicate the key features that lead to the risk. The expression for calculating the SHAP value is: In the formula, To explain the model, For the total number of features, As the baseline value, For the first The SHAP value of each feature.

8. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 1, characterized in that, The method further includes: Construct at least two initial risk assessment models with different algorithm types to obtain multiple target risk assessment models; The outputs of each target risk assessment model are fused using a weighted average method to obtain a comprehensive risk score, which serves as the final risk assessment result. The formula for calculating the comprehensive risk score is as follows: In the formula, For comprehensive risk scoring, For the first The output score of the target risk assessment model For the first The weights of each target risk assessment model, and satisfying Weight The F1-Score of each target risk assessment model is determined based on the validation set.

9. The method for constructing an assessment model for power infrastructure projects for risk assessment according to claim 1, characterized in that, The preprocessing includes: Data cleaning is performed on historical risk data and real-time monitoring data. This data cleaning includes removing duplicate data, filling in missing values, and correcting outliers. The cleaned data is then normalized. The normalized data is time-aligned to obtain the standardized dataset.

10. A system for constructing an evaluation model for power infrastructure projects for risk assessment, used to implement the method for constructing an evaluation model for power infrastructure projects for risk assessment as described in any one of claims 1-9, characterized in that, include: The acquisition module is used to acquire historical risk data and real-time monitoring data of the power infrastructure projects to be evaluated; The preprocessing module is used to preprocess the historical risk data and real-time monitoring data to obtain a standardized dataset; The extraction module is used to perform feature engineering on the standardized dataset, extract statistical features, time series features, and correlation features, and filter them to obtain key risk features; and construct feature vectors based on the key risk features. The construction module is used to initialize the parameters of a preset machine learning algorithm based on the dimension of the feature vector, and construct an initial risk assessment model. The training module is used to train and optimize the initial risk assessment model using the standardized dataset to obtain the target risk assessment model. The output module is used to perform risk assessment on the real-time monitoring data of the power infrastructure project to be assessed based on the target risk assessment model, and output the risk assessment results.

Images