Electric power engineering technology economic analysis system based on big data and big model

Abstract

The invention provides an electric power engineering technology economic analysis system based on big data and a big model. The system comprises a data acquisition integration end, a model analysis prediction end, a decision support end and a user interaction end. The data acquisition integrated end is used for acquiring various data related to the electric power engineering; the model analysis and prediction end is used for modeling, predicting and analyzing technical and economic characteristics of an electric power engineering project through various models; the decision support end is used for generating an intelligent decision suggestion according to an analysis result of the model analysis prediction end; the user interaction end is used for completing interaction between a user and the system; through combination of big data and a big model, accurate modeling and intelligent decision support of technical and economic characteristics of an electric power engineering project are realized, and efficiency and precision of management and decision are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power engineering data processing systems, and in particular to a power engineering technical economic analysis system based on big data and big models. Background Art

[0002] In modern power engineering projects, as project scale and complexity increase, the technical and economic characteristics of the projects become more diverse and complex. Power engineering construction requires comprehensive consideration of technical and economic analysis, including project cost input, resource allocation, construction efficiency, and return on investment. Traditional decision-making and evaluation methods often cannot meet the multi-dimensional needs of complex projects and lack the ability to deeply explore and comprehensively analyze historical and real-time data. In addition, how to transform huge amounts of historical project data into effective knowledge that can be used to guide current and future projects and improve the efficiency and quality of project construction and management is a problem that needs to be solved urgently.

[0003] After consulting relevant publicly available technical solutions, the technology with publication number CN109858751A proposes a project review system, which consists of a user layer, an application layer, a platform layer, and a data layer, including a work task module, a project review module, and a system management module. The work task module mainly displays automatic push message reminders of the project review work status within the system; the project review module realizes online review and evaluation of scientific decision-making, investment control, and benchmarking management at all stages of power engineering project construction; the system management module is responsible for system process design, monitoring, and historical data management. The system solves problems such as unclear work processes, unclear responsibilities, missing information, incomplete content, and insufficient management benchmarking data in the review process of power engineering project construction, realizes standardization, normalization, and information management of project review work, improves work quality and efficiency, and promotes the improvement of the refined management level of engineering project construction. However, this solution mainly focuses on the standardization and management optimization of the project review process, and is relatively insufficient in technical and economic forecasting and optimization, lacking in-depth support for multi-dimensional data analysis and dynamic optimization capabilities of power engineering projects. Summary of the Invention

[0004] The purpose of this invention is to address the current deficiencies and propose a power engineering technical and economic analysis system based on big data and big models.

[0005] The present invention adopts the following technical solutions:

[0006] A technical and economic analysis system for electric power engineering based on big data and big models is characterized in that the system includes a data acquisition and integration terminal, a model analysis and prediction terminal, a decision support terminal and a user interaction terminal; the data acquisition and integration terminal is used to collect various types of data related to electric power engineering; the model analysis and prediction terminal is used to model and predict the technical and economic characteristics of electric power engineering projects through various models; the decision support terminal is used to generate intelligent decision-making suggestions based on the analysis results of the model analysis and prediction terminal; and the user interaction terminal is used to complete the interaction between the user and the system.

[0007] The data acquisition integration terminal includes a data acquisition module, a historical data integration module and a data processing and storage module; the data acquisition module is used to collect various types of data related to real-time power engineering; the historical data integration module is used to obtain and integrate historical power engineering project data; the data processing and storage module is used to pre-process various types of data related to real-time power engineering and historical power engineering project data, and store them according to project type classification after pre-processing.

[0008] Furthermore, the model analysis and prediction end includes a feature processing module, a multi-model module and a prediction input module; the feature processing module is used to perform feature processing on various data related to real-time power engineering and historical power engineering project data to generate standardized features that can be recognized and utilized by subsequent prediction models; the multi-model module includes multiple different types of prediction models related to power engineering, each of which is used in a specific analysis field and provides accurate prediction results; the multiple different types of prediction models jointly complete the comprehensive analysis and prediction of the technical and economic related fields of an entire power project; the prediction input module is used to input various types of data related to real-time power engineering into each prediction model in the multi-model module after processing in the feature processing module, and coordinate the data flow and input management between the prediction models to ensure that each prediction model can receive the feature data it requires for efficient and accurate prediction analysis.

[0009] Furthermore, each prediction model in the multi-model module is constructed based on historical power engineering project data; after the feature processing module performs feature processing on the historical power engineering project data, the processed feature data is divided into a training set and a validation set to effectively train and evaluate the prediction model, thereby constructing a prediction model with good prediction ability; the training set is used to optimize the model parameters and learn the model structure in the prediction model, and the validation set is used to evaluate the performance of the prediction model on unseen data to ensure the generalization ability and prediction accuracy of the model.

[0010] Furthermore, the training set contains a large amount of sample data extracted from historical power engineering project data and processed by a feature processing module; each sample data represents all the features contained in a historical power engineering project; during the training process of the prediction model, the training set generates multiple groups of variant samples of the original sample data by perturbation processing each sample data, thereby simulating different possible situations, and using the original sample data and the variant samples to jointly train the prediction model; and the variant samples share the same output label with the original sample data.

[0011] Furthermore, the specific execution process of the perturbation processing for the sample data is as follows:

[0012] S11: For each feature in a certain sample data, the disturbance factor is calculated in combination with the historical data characteristics of the feature;

[0013]

[0014] Among them, ∈ i is the disturbance factor of the i-th feature, γ is the disturbance adjustment coefficient, which is used to adjust the baseline intensity of data disturbance; is the local volatility measure of the i-th feature; is the global volatility measure of the i-th feature; k is the sensitivity adjustment coefficient, which is used to adjust the sensitivity of the disturbance factor.

[0015] Among them, for satisfy:

[0016]

[0017] in, It represents the set of characteristic values of the i-th feature in all similar sample data within the time window of the current sample data in the historical power engineering project data, where the similar sample data are sample data with the same power engineering project type; T is the time window, which is pre-set according to the periodic characteristics of the power engineering project or the data timeliness requirements; std() is the standard deviation calculation function.

[0018] Among them, for satisfy:

[0019]

[0020] in, Represents the set of eigenvalues of the i-th feature in all similar samples in historical power engineering project data.

[0021] S12: For each feature in the sample data, determine its perturbation range according to the perturbation factor; for the i-th feature of a certain sample data, its perturbation range is defined as (xi -∈ i ,x i +∈ i ); where x i is the characteristic value of the i-th feature in the sample data.

[0022] S13: For each feature in the sample data, a value is randomly generated within its perturbation range. When all features are generated, a new sample is formed. The new sample is a variant sample of the original sample data.

[0023] S14: Repeat step S13 multiple times to obtain multiple variant samples of the sample data.

[0024] S15: Execute steps S11 to S14 for all sample data in the training set, thereby generating variant samples of all sample data in the training set, thereby improving the generalization ability of model training.

[0025] Furthermore, the decision support end includes a result receiving module, a difference analysis module, an optimization module and a decision output module; the result receiving module is used to receive the analysis and prediction results from the model analysis and prediction end; the difference analysis module is used to receive the ideal optimization target set by the user, and identify the user's ideal optimization target and the analysis and prediction results provided by the analysis and prediction end for difference analysis; the optimization module is used to adjust and optimize various types of data related to real-time power engineering in combination with the difference analysis results; the decision output module is used to output the optimization results of the optimization module and display them to the user in a visual manner.

[0026] The beneficial effects achieved by the present invention are:

[0027] By integrating data acquisition, predictive modeling and intelligent decision-making functions, the present invention achieves a comprehensive analysis and accurate prediction of the technical and economic characteristics of power engineering projects. By perturbing sample data during the prediction model training process, the generalization ability and prediction accuracy of the prediction model are improved, and a flexible disturbance processing mechanism is provided, so that the model can adapt to complex engineering scenarios and changes. At the same time, through intelligent optimization and decision support functions, more valuable and refined guidance is provided for the management and decision-making of power engineering projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] The present invention can be further understood from the following description in conjunction with the accompanying drawings. The components in the figures are not necessarily drawn to scale, but rather the emphasis is placed on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0029] Figure 1 Schematic diagram of the internal framework of the system of the present invention.

[0030] Figure 2 Schematic diagram of the disturbance processing workflow of sample data according to the present invention.

[0031] Figure 3 Schematic diagram of the workflow of the optimization module of the present invention. DETAILED DESCRIPTION

[0032] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with its embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention; for those skilled in the art, after reviewing the following detailed description, other systems, methods and / or features of the present embodiment will become apparent; it is intended that all such additional systems, methods, features and advantages are included in this specification; included within the scope of the present invention and protected by the appended claims; additional features of the disclosed embodiments are described in the following detailed description, and these features will be apparent from the following detailed description.

[0033] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if there are terms such as "upper", "lower", "left", "right", etc. indicating directions or positional relationships, they are based on the directions or positional relationships shown in the drawings. This is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or component referred to must have a specific direction, be constructed and operate in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0034] Example 1:

[0035] like Figure 1 As shown, this embodiment provides a power engineering technical and economic analysis system based on big data and big models, and the system includes a data acquisition and integration terminal, a model analysis and prediction terminal, a decision support terminal and a user interaction terminal; the data acquisition and integration terminal is used to collect various types of data related to power engineering; the model analysis and prediction terminal is used to model and predict the technical and economic characteristics of power engineering projects through various models; the decision support terminal is used to generate intelligent decision-making suggestions based on the analysis results of the model analysis and prediction terminal; the user interaction terminal is used to complete the interaction between the user and the system.

[0036] The data acquisition integration terminal includes a data acquisition module, a historical data integration module and a data processing and storage module; the data acquisition module is used to collect various types of data related to real-time power engineering; the historical data integration module is used to obtain and integrate historical power engineering project data; the data processing and storage module is used to pre-process various types of data related to real-time power engineering and historical power engineering project data, and store them according to project type classification after pre-processing.

[0037] Furthermore, the model analysis and prediction end includes a feature processing module, a multi-model module and a prediction input module; the feature processing module is used to perform feature processing on various data related to real-time power engineering and historical power engineering project data to generate standardized features that can be recognized and utilized by subsequent prediction models; the multi-model module includes multiple different types of prediction models related to power engineering, each of which is used in a specific analysis field and provides accurate prediction results; the multiple different types of prediction models jointly complete the comprehensive analysis and prediction of the technical and economic related fields of an entire power project; the prediction input module is used to input various types of data related to real-time power engineering into each prediction model in the multi-model module after processing in the feature processing module, and coordinate the data flow and input management between the prediction models to ensure that each prediction model can receive the feature data it requires for efficient and accurate prediction analysis.

[0038] Furthermore, each prediction model in the multi-model module is constructed based on historical power engineering project data; after the feature processing module performs feature processing on the historical power engineering project data, the processed feature data is divided into a training set and a validation set to effectively train and evaluate the prediction model, thereby constructing a prediction model with good prediction ability; the training set is used to optimize the model parameters and learn the model structure in the prediction model, and the validation set is used to evaluate the performance of the prediction model on unseen data to ensure the generalization ability and prediction accuracy of the model.

[0039] Furthermore, the training set contains a large amount of sample data extracted from historical power engineering project data and processed by a feature processing module; each sample data represents all the features contained in a historical power engineering project; during the training process of the prediction model, the training set generates multiple groups of variant samples of the original sample data by perturbation processing each sample data, thereby simulating different possible situations, and using the original sample data and the variant samples to jointly train the prediction model; and the variant samples share the same output label with the original sample data.

[0040] Further, such as Figure 2 As shown, the specific execution process of the perturbation processing for the sample data is as follows:

[0041] S11: For each feature in a certain sample data, the disturbance factor is calculated in combination with the historical data characteristics of the feature;

[0042]

[0043] Among them, ∈ i is the disturbance factor of the i-th feature, γ is the disturbance adjustment coefficient, which is used to adjust the baseline intensity of data disturbance; is the local volatility measure of the i-th feature; is the global volatility measure of the i-th feature; k is the sensitivity adjustment coefficient, which is used to adjust the sensitivity of the disturbance factor.

[0044] Among them, for satisfy:

[0045]

[0046] in, It represents the set of characteristic values of the i-th feature in all similar sample data within the time window of the current sample data in the historical power engineering project data, where the similar sample data are sample data with the same power engineering project type; T is the time window, which is pre-set according to the periodic characteristics of the power engineering project or the data timeliness requirements; std() is the standard deviation calculation function.

[0047] Among them, for satisfy:

[0048]

[0049] in, Represents the set of eigenvalues of the i-th feature in all similar samples in the historical power engineering project data;

[0050] S12: For each feature in the sample data, determine its perturbation range according to the perturbation factor; for the i-th feature of a certain sample data, its perturbation range is defined as (x i -∈ i ,x i +∈ i ); where x i is the characteristic value of the i-th feature in the sample data.

[0051] S13: For each feature in the sample data, a value is randomly generated within its perturbation range. When all features are generated, a new sample is formed. The new sample is a variant sample of the original sample data.

[0052] S14: Repeat step S13 multiple times to obtain multiple variant samples of the sample data.

[0053] S15: Execute steps S11 to S14 for all sample data in the training set, thereby generating variant samples of all sample data in the training set, thereby improving the generalization ability of model training.

[0054] Furthermore, in this embodiment, the partial implementation code for the perturbation processing of the sample data is as follows:

[0055]

[0056]

[0057]

[0058]

[0059] Where: calculate_perturbation_factor is the function that calculates the perturbation factor.

[0060] perturb_samples is a function that perturbs sample data to generate variant samples.

[0061] The code examples assume that the historical and sample data have project_type features for classification and calculation of local volatility.

[0062] This solution constructs multiple prediction models by combining historical power engineering project data, thereby achieving a comprehensive analysis and prediction of the technical and economic characteristics of power engineering projects; by generating variant samples through feature processing and perturbation of sample data, the diversity of training data is effectively increased, thereby improving the generalization ability and prediction accuracy of the model; at the same time, when performing perturbation processing, the perturbation factor is dynamically set according to the statistical characteristics of the historical data of the features in the sample data and the statistical characteristics of the data within the time limit, thereby ensuring the rationality and adaptability of the perturbation and avoiding the prediction model from being too sensitive or having excessive deviations; ultimately, the prediction accuracy and stability of the prediction model in practical applications are enhanced, providing more reliable information support for subsequent decision-making.

[0063] Example 2:

[0064] This embodiment should be understood to include at least all the features of any of the aforementioned embodiments and be further improved thereon;

[0065] This embodiment provides a technical and economic analysis system for electric power engineering projects based on big data and big models, and the system includes a data acquisition and integration terminal, a model analysis and prediction terminal, a decision support terminal and a user interaction terminal; the data acquisition and integration terminal is used to collect various types of data related to electric power engineering projects; the model analysis and prediction terminal is used to model and predict the technical and economic characteristics of electric power engineering projects through various models; the decision support terminal is used to generate intelligent decision-making recommendations based on the analysis results of the model analysis and prediction terminal; and the user interaction terminal is used to complete the interaction between the user and the system.

[0066] Furthermore, the decision support end includes a result receiving module, a difference analysis module, an optimization module and a decision output module; the result receiving module is used to receive the analysis and prediction results from the model analysis and prediction end; the difference analysis module is used to receive the ideal optimization target set by the user, and identify the user's ideal optimization target and the analysis and prediction results provided by the analysis and prediction end for difference analysis; the optimization module is used to adjust and optimize various types of data related to real-time power engineering in combination with the difference analysis results; the decision output module is used to output the optimization results of the optimization module and display them to the user in a visual manner.

[0067] Furthermore, the specific workflow of the optimization module is as follows:

[0068] S21: Receive the difference analysis results of the difference analysis module and use them as the target optimization function of the subsequent steps; the expression is:

[0069]

[0070] Among them, F represents the target optimization function value, which specifically represents the quantitative difference between the user's ideal optimization target and the model analysis prediction end; d j is the quantitative difference between the ideal optimization target of the jth user and the analysis and prediction result output by the model analysis and prediction end, c j is the weight of the jth user's ideal optimization goal, and m is the total number of user's ideal optimization goals.

[0071] S22: According to the characteristics of real-time power engineering related data, the valid range of each input feature is set, and an initial population is generated within this range to ensure that each input feature value meets the set constraints; each individual in the initial population represents a possible solution, which contains a set of specific input feature combinations as the starting point of the optimization process; the input features are specifically standardized features input to the prediction model in the multi-model module to complete the prediction.

[0072] S23: Iteratively perform the selection, crossover, and mutation operations of the genetic algorithm on the initial population to gradually optimize the individuals in the population. The specific implementation process is as follows:

[0073] Selection operation: Based on the fitness value of each individual in the current population, select the individuals with better performance from the current population; the fitness value calculation process is as follows:

[0074] S231: Input each individual in the current population into the multi-model module for prediction to obtain analysis and prediction results;

[0075] S232: Use the analysis and prediction results in step S231 to calculate the target optimization function value of each individual, and calculate its fitness value based on the target optimization function value; in this scheme, the inverse of the target optimization function value can be used as the fitness value.

[0076] Crossover operation: After the selection operation is completed, the input features of the selected individuals are cross-combined to generate new offspring individuals. The specific crossover operation process is as follows:

[0077] S233: Calculate the crossover selection probability for each input feature:

[0078]

[0079] Among them, P k,1 is the cross selection probability of the kth input feature under the current number of iterations; α is the proportional adjustment factor used to control the benchmark strength of cross selection; std(Y k ) is the standard deviation of the k-th input feature in the current population, Y k is the set of all feature values of the k-th input feature in the current population; ∈ is a very small positive number that is not 0.

[0080] S234: Crossover and combination of input features in the selected individuals are performed according to the calculated crossover selection probability to generate new offspring individuals.

[0081] Mutation operation: After the crossover operation is completed, the mutation operation is performed on the generated offspring individuals, that is, the input feature values of certain input features are randomly changed; the specific implementation process is as follows:

[0082] S235: Calculate the mutation rate of each individual:

[0083]

[0084] Among them, P2(t) is the mutation rate of each input feature in the individual at the current iteration number t; P start is the mutation rate of the initial population, set in the initial stage; β is the mutation rate adjustment parameter, used to control the decay rate of the mutation rate; T max The preset maximum number of iterations.

[0085] S236: Randomly change the input features of the offspring individuals generated after the crossover operation is completed according to the calculated mutation rate, and ensure that the input feature values after the random change are within a preset valid range.

[0086] S24: Integrate the offspring individuals after the crossover and mutation operations with the current population, select individuals with better performance according to the fitness value to form a new generation population, and repeat the selection, crossover and mutation operation steps.

[0087] S25: When the number of iterations reaches the preset maximum number of iterations, the iteration process is stopped, and the individual with the highest fitness value in the current population is selected as the adjustment optimization result and transmitted to the decision output module; the adjustment optimization result is the optimal input feature combination.

[0088] Furthermore, the functions that can be realized by the user interaction terminal include but are not limited to:

[0089] User input and demand collection: Users can input their optimization goals, constraints, and weight preferences through the interactive interface; the system will guide the model analysis and prediction end and the decision support end to perform corresponding analysis and optimization based on the user's input.

[0090] Result feedback and display: Users can use the interactive interface to display model prediction results and adjustment optimization results in an intuitive and visual way, making it easier to make decisions.

[0091] Adjustment and interactive feedback: Users can adjust and provide feedback on their past input content based on the adjustment and optimization results, and trigger the system's optimization process again.

[0092] This solution uses the target optimization function to clarify the difference between the user's ideal goal and the actual prediction result, which serves as the direction and driving force of optimization. It uses a genetic algorithm to iteratively optimize the population and dynamically adjust the combination of different features in cross selection to improve the diversity and search ability of the population. During the mutation operation, the dynamic mutation rate is adjusted to enable the population to have stronger exploration capabilities in the early stage and gradually converge in the later stage to avoid falling into the local optimum too early. The sorting process helps to generate efficient, stable optimization results that meet the user's goals.

[0093] The contents disclosed above are only preferred feasible embodiments of the present invention and do not limit the scope of protection of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention description and drawings are included in the scope of protection of the present invention. In addition, the elements therein can be updated as technology develops.

Claims

1. A power engineering technical economic analysis system based on big data and big models, characterized by: The system includes a data acquisition and integration terminal, a model analysis and prediction terminal, a decision support terminal, and a user interaction terminal; the data acquisition and integration terminal is used to collect various types of data related to power engineering projects; the model analysis and prediction terminal is used to model and predict the technical and economic characteristics of power engineering projects through various models; and the decision support terminal is used to generate intelligent decision suggestions based on the analysis results of the model analysis and prediction terminal. The user interaction terminal is used to complete the interaction between the user and the system; The data acquisition integration terminal includes a data acquisition module, a historical data integration module and a data processing and storage module; the data acquisition module is used to collect various types of data related to real-time power engineering; the historical data integration module is used to obtain and integrate historical power engineering project data; the data processing and storage module is used to pre-process various types of data related to real-time power engineering and historical power engineering project data, and store them according to project type classification after pre-processing.

2. The power engineering technical economic analysis system based on big data and big models according to claim 1 is characterized in that: The model analysis and prediction end includes a feature processing module, a multi-model module and a prediction input module; the feature processing module is used to perform feature processing on various data related to real-time power engineering and historical power engineering project data to generate standardized features that can be recognized and utilized by subsequent prediction models; the multi-model module includes multiple different types of prediction models related to power engineering, each of which is used in a specific analysis field and provides accurate prediction results; the multiple different types of prediction models jointly complete the comprehensive analysis and prediction of the technical and economic related fields of an entire power project; the prediction input module is used to input various types of data related to real-time power engineering into each prediction model in the multi-model module after processing in the feature processing module, and coordinate the data flow and input management between the prediction models to ensure that each prediction model can receive the feature data it requires for efficient and accurate prediction analysis.

3. The power engineering technical economic analysis system based on big data and big models according to claim 2 is characterized in that: Each prediction model in the multi-model module is constructed based on historical power engineering project data; After the feature processing module performs feature processing on the historical power engineering project data, the processed feature data is divided into a training set and a validation set to effectively train and evaluate the prediction model, thereby constructing a prediction model with good prediction capabilities; the training set is used to optimize the model parameters and learn the model structure in the prediction model, and the validation set is used to evaluate the performance of the prediction model on unseen data to ensure the generalization ability and prediction accuracy of the model.

4. The power engineering technical economic analysis system based on big data and big models according to claim 3 is characterized in that: The training set contains a large amount of sample data extracted from historical power engineering project data and processed by a feature processing module; each sample data represents all the features contained in a historical power engineering project; during the training process of the prediction model, the training set generates multiple groups of variant samples of the original sample data by perturbing each sample data, thereby simulating different possible situations, and using the original sample data and the variant samples to jointly train the prediction model; and the variant samples share the same output label with the original sample data.

5. The power engineering technical economic analysis system based on big data and big models according to claim 4 is characterized in that: The specific execution process of the disturbance processing for the sample data is as follows: S11: For each feature in a certain sample data, the disturbance factor is calculated in combination with the historical data characteristics of the feature; Among them, ∈ i is the disturbance factor of the i-th feature, γ is the disturbance adjustment coefficient, which is used to adjust the baseline intensity of data disturbance; is the local volatility measure of the i-th feature; is the global volatility measure of the i-th feature; k is the sensitivity adjustment coefficient, which is used to adjust the sensitivity of the disturbance factor; Among them, for satisfy: in, represents the set of feature values of the i-th feature in all similar sample data within the time window of the current sample data in the historical power engineering project data, where the similar sample data are sample data of the same power engineering project type; T is the time window, which is pre-set based on the periodic characteristics of the power engineering project or the timeliness requirements of the data; std() is the standard deviation calculation function; Among them, for satisfy: in, Represents the set of eigenvalues of the i-th feature in all similar samples in the historical power engineering project data; S12: For each feature in the sample data, determine its perturbation range according to the perturbation factor; for the i-th feature of a certain sample data, its perturbation range is defined as (x i -∈ i ,x i +∈ i ); where x i is the characteristic value of the i-th feature in the sample data; S13: For each feature in the sample data, a value is randomly generated within its perturbation range. When all features are generated, a new sample is formed. The new sample is a variant sample of the original sample data. S14: Repeat step S13 multiple times to obtain multiple variant samples of the sample data; S15: Execute steps S11 to S14 for all sample data in the training set, thereby generating variant samples of all sample data in the training set, thereby improving the generalization ability of model training.

6. The electric power engineering technical economic analysis system based on big data and big models according to claim 5 is characterized in that: The decision support end includes a result receiving module, a difference analysis module, an optimization module and a decision output module; the result receiving module is used to receive the analysis and prediction results from the model analysis and prediction end; the difference analysis module is used to receive the ideal optimization target set by the user, and identify the user's ideal optimization target and the analysis and prediction results provided by the analysis and prediction end for difference analysis; the optimization module is used to adjust and optimize various types of data related to real-time power engineering in combination with the difference analysis results; the decision output module is used to output the optimization results of the optimization module and display them to the user in a visual manner.

Images