An adaptive optimization method and system for power distribution network planning

Abstract

The application relates to the technical field of power distribution networks, and discloses a self-adaptive optimization method and system for power distribution network planning. The system comprises a data acquisition and processing module, a prediction analysis module, a scheme generation module, an evaluation and decision module, an execution and feedback module and a self-adaptive optimization module. When the optimization of the power distribution network planning scheme is carried out, a power distribution network multi-source data set is constructed and standardized processing is carried out, and a multi-objective optimization function is set for different planning scenarios, so that the pertinence of power distribution network planning schemes of different scales is ensured, a space-time distribution prediction algorithm is used for load and new energy output prediction, the uncertainty caused by power grid operation data fluctuation can be effectively coped with, the unity of the forward-looking and scientific nature of the planning scheme is ensured, planning decision deviation is further reduced, and a preference learning algorithm is introduced in the scheme evaluation process, so that the finally recommended planning scheme meets the technical and economic requirements and the actual management requirements.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power distribution network, in particular to a self-adaptive optimization method and system for power distribution network planning. BACKGROUND

[0002] The power distribution network refers to the power network that accepts electric energy from the power transmission network and regional power plants, and distributes it to various users through power distribution facilities and step-by-step distribution according to voltage. It is composed of overhead lines, cables, towers, distribution transformers, disconnectors, reactive power compensators and some auxiliary facilities, and plays an important role in distributing electric energy in the power network.

[0003] At present, in the process of power distribution network planning, multiple influencing factors are involved. When the traditional evaluation model is used for optimization of the power distribution network planning scheme, it cannot respond to the changes in the power grid operation state in real time. When there is load fluctuation and abnormal new energy output, the planning scheme will deviate greatly from the actual demand, and the applicability of the planning scheme cannot be guaranteed.

[0004] Therefore, the present application provides a self-adaptive optimization method and system for power distribution network planning to solve the above problems. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a self-adaptive optimization method and system for power distribution network planning, which solves the problem of large deviation between the planning scheme and the actual demand in the background technology, and cannot guarantee the applicability of the planning scheme.

[0006] To achieve the above purpose, the present application provides the following technical scheme: A self-adaptive optimization method and system for power distribution network planning, the method comprising the following steps:

[0007] S1, collect multi-source heterogeneous data of the power distribution network, including historical load data, real-time operation data, equipment state data, geographic information data and distributed energy access data, and generate a multi-source data set of the power distribution network;

[0008] S2, data cleaning, normalization and feature extraction processing are performed on the multi-source data set of the power distribution network, and a power distribution network planning feature data set is constructed;

[0009] S3, based on the power distribution network planning feature data set, a time-space distribution prediction algorithm is used for load prediction and new energy output prediction, and a power distribution network planning prediction data is generated;

[0010] S4, according to the power distribution network planning prediction data, combining with the planning constraint condition, a plurality of power distribution network planning alternative schemes are generated through a multi-objective optimization algorithm;

[0011] S5, comprehensively evaluate the multiple power distribution network planning alternative schemes by using the adaptive evaluation model, and generate scheme evaluation result data;

[0012] S6, based on the scheme evaluation result data, identify the decision preference by using the preference learning algorithm, and output the recommended scheme through the interactive decision interface;

[0013] S7, execute the power distribution network planning scheme deployment according to the recommended scheme, and monitor the operation effect in real time to generate feedback data;

[0014] S8, dynamically update the adaptive evaluation model and the preference learning algorithm based on the feedback data, and realize the continuous optimization of the power distribution network planning scheme.

[0015] Preferably, the collection of power distribution network multi-source heterogeneous data in step S1 comprises:

[0016] S11, collect power distribution network multi-source heterogeneous data through smart meters, SCADA systems, Internet of Things sensors and GIS platforms;

[0017] S12, perform time alignment and space correlation processing on the collected data to generate a power distribution network multi-source data set with unified space-time reference.

[0018] Preferably, the construction of power distribution network planning feature data set in step S2 comprises:

[0019] S21, use wavelet transform and principal component analysis to filter noise and reduce dimensionality of features of the power distribution network multi-source data set;

[0020] S22, extract load time sequence features, network topology features and environmental influence factor features to construct a power distribution network planning feature data set.

[0021] Preferably, the generation of power distribution network planning prediction data in step S3 comprises:

[0022] S31, perform short-term and medium-long-term load prediction based on long short-term memory network LSTM and space-time graph convolution network STGCN;

[0023] S32, consider weather factors and equipment aging effects, and use a probabilistic prediction method to generate a new energy output prediction interval.

[0024] Preferably, the generation of multiple power distribution network planning alternative schemes in step S4 comprises:

[0025] S41, set economic, reliability, environmental protection and scalability multi-objective functions;

[0026] S42, use multi-objective particle swarm optimization algorithm MOPSO to solve the Pareto optimal solution set, and generate multiple power distribution network planning alternative schemes.

[0027] Preferably, the generation scheme evaluation result data in step S5 includes:

[0028] S51, construct a comprehensive evaluation index system including technical indicators, economic indicators and risk indicators;

[0029] S52, calculate the index weight based on entropy weight method and CRITIC method, and sort the schemes by TOPSIS method to generate scheme evaluation result data.

[0030] Preferably, the output recommended scheme in step S6 includes:

[0031] S61, learn the decision preference from historical decision data using a preference learning algorithm based on deep reinforcement learning;

[0032] S62, display the scheme evaluation result through a human-computer interaction interface and receive the decision maker's feedback to dynamically adjust the recommended scheme.

[0033] Preferably, the continuous optimization of the power distribution network planning scheme in step S8 includes:

[0034] S81, real-time collection of power distribution network operation data, calculation of the deviation between the actual effect and the expected effect of the scheme;

[0035] S82, update the evaluation model weight and the preference learning model parameter through an online learning mechanism to realize model self-adaptive optimization.

[0036] Preferably, the system includes:

[0037] A data acquisition and processing module acquires power distribution network heterogeneous data through a multi-source data access unit, and generates a power distribution network multi-source data set using a data cleaning and fusion unit;

[0038] A prediction analysis module receives the power distribution network multi-source data set, and generates power distribution network planning prediction data through a load prediction unit and a new energy prediction unit;

[0039] A scheme generation module generates multiple power distribution network planning alternative schemes through a multi-objective optimization algorithm unit based on the prediction data;

[0040] An evaluation and decision module receives the alternative schemes, generates scheme evaluation results using an adaptive evaluation unit, and outputs recommended schemes through a preference learning unit and an interactive decision unit;

[0041] An execution and feedback module executes and deploys according to the recommended scheme, collects feedback data through a running monitoring unit, and transmits the feedback data to a model updating unit;

[0042] An adaptive optimization module receives the feedback data, dynamically adjusts the evaluation model and the preference learning algorithm through a model updating unit, and realizes continuous optimization of the system.

[0043] Preferably, the evaluation and decision module further comprises:

[0044] A visual interaction unit displays the scheme evaluation results and the recommended scheme, and provides a man-machine interaction interface for the decision maker to input preference feedback;

[0045] A real-time correction unit dynamically adjusts the recommended scheme according to the feedback data to ensure that the scheme meets the actual decision-making requirements.

[0046] Compared with the prior art, the present application provides an adaptive optimization method and system for power distribution network planning, which has the following beneficial effects:

[0047] 1. In the present application, when the power distribution network planning scheme is optimized, a multi-source data set of the power distribution network is constructed and standardized, and a multi-objective optimization function is set for different planning scenarios to ensure the pertinence of the generated power distribution network planning scheme of different scales. At the same time, a time and space distribution prediction algorithm is used for load and new energy output prediction, which can effectively cope with the uncertainty caused by fluctuations in power grid operation data, ensure the unity of forward-looking and scientific planning scheme, and further reduce planning decision bias.

[0048] 2. In the present application, when the power distribution network planning scheme is comprehensively evaluated, a comprehensive evaluation system including technical indicators, economic indicators and risk indicators is established, and an adaptive evaluation model is used to dynamically calculate the indicator weights, so that the system can objectively quantify the comprehensive benefits of each alternative scheme. In the scheme evaluation process, a preference learning algorithm is introduced to real-time integrate the subjective preference of the decision maker and the objective evaluation results, so that the finally recommended planning scheme meets the technical and economic requirements and the actual management requirements, and ensures the implementability of the planning scheme.

[0049] 3. In the present application, during the implementation of the power distribution network planning scheme, a real-time feedback mechanism is established to continuously collect operation data, and an online learning technology is used to dynamically update the evaluation model and the optimization algorithm, so that the system can continuously optimize the planning scheme according to the actual operation state of the power grid, realize the whole life cycle management of the planning scheme, effectively improve the adaptability and long-term operation benefit of the power distribution network planning, and further improve the fine level and dynamic optimization ability of the power distribution network planning management. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figure 1 The flowchart of the adaptive optimization method for power distribution network planning of the present application;

[0051] Figure 2 The framework diagram of the adaptive optimization system for power distribution network planning of the present application. Detailed Implementation

[0052] 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. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Please see Figure 1 - Figure 2 This is an adaptive optimization method and system for distribution network planning. The method includes the following steps:

[0054] S1. Collect multi-source heterogeneous data of the distribution network, including historical load data, real-time operation data, equipment status data, geographic information data and distributed energy access data, and generate a multi-source data set of the distribution network;

[0055] S2. Perform data cleaning, normalization, and feature extraction processing on the multi-source data set of the distribution network to construct a distribution network planning feature dataset;

[0056] Data cleaning: refers to the process of identifying and processing errors, missing values, outliers, and inconsistencies in raw data to ensure data quality;

[0057] Normalization: also known as standardization, is the process of scaling data of different scales and dimensions proportionally to make them fall into a specific range, thereby eliminating the influence of dimensions and accelerating model convergence. Commonly used methods include Min-Max normalization and Z-score standardization.

[0058] formula:

[0059] Min-Max normalization:

[0060] ;

[0061] in, Here, X represents the normalized value, and X represents the original data value. , These are the minimum and maximum values ​​of this feature in the sample;

[0062] Z-score standardization:

[0063] ;

[0064] in, The standardized value. The mean of the sample. The standard deviation of the sample;

[0065] Feature extraction: the process of constructing new features from raw data that are more informative, relevant, and suitable for modeling;

[0066] S3, based on the distribution network planning feature dataset, using the spatio-temporal distribution prediction algorithm to predict the load and new energy output, and generate the distribution network planning prediction data;

[0067] Spatio-temporal distribution prediction: refers to the prediction method considering both time and space dimensions, load and new energy output change with time and are affected by geographical location and network topology factors;

[0068] S4, according to the distribution network planning prediction data, combined with the planning constraints, generate multiple distribution network planning alternative schemes through multi-objective optimization algorithm;

[0069] Multi-objective optimization function: multiple conflicting objectives need to be optimized simultaneously in planning, economic efficiency, reliability, environmental protection, and scalability are mentioned in the claims;

[0070] Example of dimensionless normalization of objective function:

[0071] Economic efficiency objective ;

[0072] Total investment cost of the scheme, The maximum cost among all alternative schemes;

[0073] Reliability objective ;

[0074] SAIDI is the system average interruption time index;

[0075] S5, use adaptive evaluation model to comprehensively evaluate multiple distribution network planning alternative schemes, and generate scheme evaluation result data;

[0076] S6, based on the scheme evaluation result data, use preference learning algorithm to identify decision preference, and output recommended scheme through interactive decision interface;

[0077] S7, according to the recommended scheme, execute the distribution network planning scheme deployment, and monitor the operation effect in real time, and generate feedback data;

[0078] S8, based on the feedback data, dynamically update the adaptive evaluation model and the preference learning algorithm, and realize the continuous optimization of the distribution network planning scheme;

[0079] The collection of multi-source heterogeneous data of distribution network in step S1 includes:

[0080] S11, collect multi-source heterogeneous data of distribution network through smart meters, SCADA systems, Internet of Things sensors and GIS platforms;

[0081] S12, time alignment and space correlation processing are performed on the collected data to generate a power distribution network multi-source data set with unified space-time reference;

[0082] The construction of the power distribution network planning feature data set in step S2 includes:

[0083] S21, wavelet transform and principal component analysis are used to filter noise and reduce the dimensionality of the power distribution network multi-source data set;

[0084] S22, load time sequence features, network topology features and environmental influence factor features are extracted to construct a power distribution network planning feature data set;

[0085] The generation of the power distribution network planning prediction data in step S3 includes:

[0086] S31, short-term and medium-long-term load prediction is performed based on long short-term memory network LSTM and space-time graph convolution network STGCN;

[0087] LSTM: a special recurrent neural network that solves the gradient vanishing and explosion problem in long sequence training by introducing a "gate" mechanism, which is very suitable for time series prediction;

[0088] Simplified diagram of core calculation of LSTM unit:

[0089] For a time step t:

[0090] Forget gate: ;

[0091] Input gate: ;

[0092] Candidate cell state: ;

[0093] Cell state update: ;

[0094] Output gate: ;

[0095] Hidden state output: ;

[0096] where, is the Sigmoid activation function, is the input at time t, is the hidden state at time t-1, is the cell state at time t-1, , is the parameter matrix and bias vector to be trained, and * is element-wise multiplication;

[0097] STGCN: a deep learning architecture that combines graph convolutional networks and processing time series models, which abstracts the power distribution network as a graph, captures spatial features with graph convolution, and captures temporal features with time convolution and recurrent units, which is very suitable for spatiotemporal data prediction of power grids with topological structure;

[0098] S32, considering the weather factors and the influence of equipment aging, a probability prediction method is used to generate a new energy output prediction interval;

[0099] The generation of multiple power distribution network planning alternatives in step S4 includes:

[0100] S41, set multi-objective functions of economy, reliability, environmental protection and scalability;

[0101] S42, use multi-objective particle swarm optimization algorithm MOPSO to solve the Pareto optimal solution set, and generate multiple power distribution network planning alternatives;

[0102] Pareto optimal solution set: in multi-objective optimization, there is no unique solution that is best in all objectives, and the Pareto optimal solution refers to a solution that cannot be modified without degrading any other objective, and the collection of all these solutions constitutes the Pareto frontier.

[0103] The general form of the multi-objective optimization problem is:

[0104] ;

[0105] ;

[0106] ;

[0107] ;

[0108] Wherein, is a vector containing m objective functions, is a decision variable vector, is the mth objective function, , is an inequality and equality constraint, is the feasible region of the decision variable, T is the transpose symbol, which means that the objective function is a column vector, j is the index symbol, which means the jth inequality constraint, which represents various physical restrictions and safety boundaries that must be met in power distribution network planning, k is the index symbol, which means the kth equality constraint, which represents the physical laws that the power distribution network must strictly comply with;

[0109] MOPSO: is the extension of particle swarm optimization algorithm in the field of multi-objective, it maintains an external archive to store non-dominated solutions, and based on this to guide the flight and search of particle swarm, finally output a set of approximate Pareto optimal solution set;

[0110] The generating scheme evaluation result data in step S5 includes:

[0111] S51, construct a comprehensive evaluation index system including technical indicators, economic indicators and risk indicators;

[0112] S52, calculate the index weight based on entropy weight method and CRITIC method, and sort the schemes by TOPSIS method, generate scheme evaluation result data;

[0113] Entropy weight method: an objective weighting method, according to the variation degree of each index value to determine the weight, the greater the difference of data sequence of a certain index, the more information it contains, the greater the weight.

[0114] CRITIC method: another objective weighting method, which not only considers the contrast intensity of the index, but also considers the conflict between the indexes, the greater the contrast intensity and the stronger the conflict with other indexes, the greater the weight.

[0115] TOPSIS method: a multi-attribute decision-making method, which constructs a "positive ideal solution" and a "negative ideal solution", and sorts the schemes by calculating the relative distance between the two ideal solutions, the closer to the positive ideal solution, the farther from the negative ideal solution, the better the scheme;

[0116] The core steps of TOPSIS method:

[0117] Construct a normalized decision matrix: normalize the original data matrix;

[0118] Construct a weighted normalized matrix: multiply each column of the normalized matrix by its corresponding weight;

[0119] Determine the positive ideal solution:

[0120] Positive ideal solution , wherein ;

[0121] Negative ideal solution , wherein ;

[0122] Calculate the distance: the distance of each scheme to the positive ideal solution ;

[0123] The distance of each scheme to the negative ideal solution ;

[0124] Calculate the relative closeness:

[0125] ;

[0126] The larger, the closer the scheme to the ideal solution, the higher the ranking;

[0127] The output recommended scheme in step S6 includes:

[0128] S61, learn the decision preference from the historical decision data by using the preference learning algorithm based on deep reinforcement learning;

[0129] Preference learning: learn the implicit preference of the decision maker for each target and attribute from his past choices and feedback, rather than directly specifying fixed weights.

[0130] Deep reinforcement learning DRL: combines the perception ability of deep learning and the decision-making ability of reinforcement learning. In this context, a framework is constructed:

[0131] Agent: recommendation system;

[0132] State: current set of alternative schemes and their evaluation index values;

[0133] Action: recommend a certain scheme and adjust the weights to generate a new recommended ranking;

[0134] Reward: the decision maker finally adopts the recommended scheme, and rejects the recommended scheme and makes other choices;

[0135] Through continuous trial and reward, the DRL model learns the recommended strategy that best meets the decision maker's preference;

[0136] S62, display the scheme evaluation results through the human-computer interaction interface, receive the decision maker's feedback, and dynamically adjust the recommended scheme;

[0137] The continuous optimization of the power distribution network planning scheme in step S8 includes:

[0138] S81, real-time collection of power distribution network operation data, calculation of the deviation between the actual effect and the expected effect of the scheme;

[0139] S82, update the evaluation model weights and preference learning model parameters through online learning mechanism to realize model self-adaptive optimization;

[0140] Online learning: contrary to batch learning, the model receives new data in the form of data stream and small batches one by one, and immediately updates the model parameters incrementally, so that the model can quickly adapt to changes in data distribution;

[0141] The system includes:

[0142] The data acquisition and processing module collects heterogeneous data of the power distribution network through a multi-source data access unit, and generates a power distribution network multi-source data set using a data cleaning and fusion unit.

[0143] The prediction analysis module receives the power distribution network multi-source data set, and generates power distribution network planning prediction data through a load prediction unit and a new energy prediction unit.

[0144] The scheme generation module generates multiple power distribution network planning alternative schemes based on the prediction data through a multi-objective optimization algorithm unit.

[0145] The evaluation and decision module receives the alternative schemes, generates scheme evaluation results using an adaptive evaluation unit, and outputs a recommended scheme through a preference learning unit and an interactive decision unit.

[0146] The execution and feedback module executes the deployment according to the recommended scheme, collects feedback data through a running monitoring unit, and transmits the feedback data to the model updating unit.

[0147] The adaptive optimization module receives the feedback data, dynamically adjusts the evaluation model and the preference learning algorithm through the model updating unit, and realizes continuous optimization of the system.

[0148] The evaluation and decision module further includes:

[0149] The visual interactive unit displays the scheme evaluation results and the recommended scheme, and provides a human-computer interaction interface for the decision maker to input preference feedback.

[0150] The real-time correction unit dynamically adjusts the recommended scheme according to the feedback data to ensure that the scheme meets the actual decision-making needs.

[0151] The adaptive optimization method and system for power distribution network planning run as follows:

[0152] Step 1: Multi-source data fusion and feature engineering

[0153] The starting point of the system is to collect a wide range of multi-source heterogeneous data of the power distribution network, which includes historical and real-time load data, device status, geographic information, and new energy access data obtained from smart meters, SCADA systems, and sensors. These raw data are strictly cleaned, time-aligned, and spatially correlated to form a unified spatio-temporal reference data set. Subsequently, wavelet transform and principal component analysis techniques are used for noise filtering and feature dimensionality reduction to extract key load time series features, network topology features, and environmental influence features, and to construct a high-quality planning feature data set, laying a solid data foundation for subsequent analysis.

[0154] Step 2: Spatio-temporal prediction and uncertainty quantification

[0155] The system uses the feature data set constructed in the previous step to perform load prediction and new energy output prediction using advanced spatio-temporal distribution prediction algorithms. This step not only generates point predictions, but more importantly, it considers weather and equipment aging factors and outputs prediction intervals through probabilistic prediction methods, quantifying the uncertainty of future grid states and ensuring the forward-looking nature and adaptability to fluctuations of the planning scheme.

[0156] Step 3: Multi-objective optimization to generate Pareto solutions

[0157] Based on the predicted data, the system combines the multiple objectives of economic efficiency, reliability, environmental protection, and scalability of distribution network planning with physical constraints and uses the multi-objective particle swarm optimization (MOPSO) algorithm for solution. The core is to find and generate a set of Pareto optimal solutions, i.e., a series of alternative planning schemes that achieve the best balance among multiple objectives, rather than a single solution, providing a rich selection space for decision-makers.

[0158] Step 4: Adaptive evaluation and preference learning decision-making

[0159] The system constructs a comprehensive evaluation index system and uses an adaptive model to dynamically calculate index weights for objective quantitative evaluation and ranking of alternative schemes. At the same time, the system uses a preference learning algorithm based on deep reinforcement learning to mine and learn the implicit preferences of decision-makers from historical decision-making data. Through the human-computer interaction interface, the evaluation results and recommended schemes are presented to the decision-makers, and feedback can be received to dynamically adjust the output, ensuring that the final recommended scheme is both scientific and objective and meets actual management needs.

[0160] Step 5: Execution feedback and model self-evolution

[0161] After the recommended scheme is deployed and executed, the system does not terminate operation. It collects actual operation effect data of the planning scheme through real-time monitoring and compares it with the expected effect to generate feedback deviation data. These valuable field data are input into the system, and the weights of the evaluation model and the parameters of the preference learning algorithm are continuously and dynamically updated through online learning mechanisms, forming a "evaluation-decision-execution-feedback-optimization" closed loop. This enables the entire system and planning scheme to continuously evolve with the operation of the power grid, achieving continuous adaptive optimization of the entire life cycle of distribution network planning.

[0162] In summary, the core principle of the system is to build a complete adaptive cycle of "data-driven prediction, multi-objective optimization generation, human-machine collaborative decision-making, and closed-loop feedback optimization," ensuring that the planning scheme from formulation to implementation and updating has forward-looking, scientific, and high adaptability.

[0163] It is to be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting; it is not intended to exclude myriad other embodiments of the present application that other inventors can develop based on the same general inventive concepts embodied by the described embodiments. That is, although the present application is described in terms of particular embodiments and implementations, it is to be understood that the terminology used is for the purpose of descriptive clarity and that it should be taken in a descriptive sense and not a limiting sense.

[0164] While the embodiments of the application have been shown and described herein, it is to be understood that the application is not limited to these embodiments. Rather, many modifications, changes, substitutions, and alterations can be made to the embodiments of the application without departing from the spirit and scope of the application as defined by the appended claims and their equivalents.

Claims

1. An adaptive optimization method for distribution network planning, characterized in that, The method includes the following steps: S1. Collect multi-source heterogeneous data of the distribution network, including historical load data, real-time operation data, equipment status data, geographic information data and distributed energy access data, and generate a multi-source data set of the distribution network; S2. Perform data cleaning, normalization, and feature extraction processing on the multi-source data set of the distribution network to construct a distribution network planning feature dataset; S3. Based on the aforementioned power distribution network planning feature dataset, load forecasting and new energy output forecasting are performed using a spatiotemporal distribution prediction algorithm to generate power distribution network planning prediction data. S4. Based on the power distribution network planning and forecasting data and combined with the planning constraints, generate multiple alternative power distribution network planning schemes through a multi-objective optimization algorithm. S5. Use an adaptive evaluation model to comprehensively evaluate the multiple alternative power distribution network planning schemes and generate scheme evaluation result data; S6. Based on the evaluation results of the above schemes, a preference learning algorithm is used to identify decision preferences, and a recommended scheme is output through an interactive decision interface; S7. Implement the distribution network planning scheme according to the recommended scheme, monitor the operation effect in real time, and generate feedback data; S8. Based on the feedback data, dynamically update the adaptive evaluation model and preference learning algorithm to achieve continuous optimization of the power distribution network planning scheme.

2. The adaptive optimization method for distribution network planning according to claim 1, characterized in that, The collection of multi-source heterogeneous data from the distribution network in step S1 includes: S11. Collect multi-source heterogeneous data of the power distribution network through smart meters, SCADA systems, IoT sensors and GIS platforms; S12. Perform time alignment and spatial correlation processing on the collected data to generate a multi-source data set of the distribution network with a unified spatiotemporal reference.

3. The adaptive optimization method for distribution network planning according to claim 1, characterized in that, The construction of the power distribution network planning feature dataset in step S2 includes: S21. Wavelet transform and principal component analysis are used to perform noise filtering and feature dimensionality reduction on the multi-source data set of the power distribution network. S22. Extract load time sequence characteristics, network topology characteristics, and environmental influencing factor characteristics to construct a distribution network planning feature dataset.

4. The adaptive optimization method for distribution network planning according to claim 1, characterized in that: The generation of distribution network planning and forecasting data in step S3 includes: S31. Short-term and medium-to-long-term load forecasting based on Long Short-Term Memory Network (LSTM) and Spatiotemporal Graph Convolutional Network (STGCN); S32. Considering the impact of weather factors and equipment aging, a probabilistic prediction method is used to generate a new energy output prediction range.

5. The adaptive optimization method for distribution network planning according to claim 1, characterized in that, The generation of multiple distribution network planning alternatives in step S4 includes: S41. Define a multi-objective function that considers economy, reliability, environmental friendliness, and scalability. S42. The multi-objective particle swarm optimization algorithm (MOPSO) is used to solve the Pareto optimal solution set and generate multiple alternative schemes for distribution network planning.

6. The adaptive optimization method for distribution network planning according to claim 1, characterized in that, The generated scheme evaluation result data in step S5 includes: S51. Construct a comprehensive evaluation indicator system that includes technical indicators, economic indicators, and risk indicators; S52. Calculate the index weights based on the entropy weight method and the CRITIC method, and rank the schemes using the TOPSIS method to generate scheme evaluation result data.

7. The adaptive optimization method for distribution network planning according to claim 1, characterized in that, The output recommendation scheme in step S6 includes: S61. A preference learning algorithm based on deep reinforcement learning is used to learn decision preferences from historical decision data; S62. Display the evaluation results of the solution through a human-computer interaction interface, receive feedback from decision-makers, and dynamically adjust the recommended solution.

8. The adaptive optimization method for distribution network planning according to claim 1, characterized in that, The continuous optimization of the distribution network planning scheme in step S8 includes: S81. Collect distribution network operation data in real time and calculate the deviation between the actual effect and the expected effect of the scheme; S82. The evaluation model weights and preference learning model parameters are updated through an online learning mechanism to achieve adaptive optimization of the model.

9. An adaptive optimization system for distribution network planning, implementing the adaptive optimization method for distribution network planning as described in any one of claims 1-8, characterized in that, The system includes: The data acquisition and processing module collects heterogeneous data from the distribution network through the multi-source data access unit and generates a multi-source data set of the distribution network using the data cleaning and fusion unit. The predictive analysis module receives the multi-source data set of the distribution network and generates distribution network planning and prediction data through the load prediction unit and the new energy prediction unit. The scheme generation module generates multiple alternative schemes for power distribution network planning based on the predicted data through a multi-objective optimization algorithm unit. The evaluation and decision-making module receives the alternative solutions, generates solution evaluation results using the adaptive evaluation unit, and outputs recommended solutions through the preference learning unit and the interactive decision-making unit. The execution and feedback module performs deployment according to the recommended scheme, collects feedback data through the operation monitoring unit, and transmits it to the model update unit; The adaptive optimization module receives the feedback data and dynamically adjusts the evaluation model and preference learning algorithm through the model update unit to achieve continuous system optimization.

10. An adaptive optimization system for distribution network planning according to claim 9, characterized in that, The assessment and decision-making module also includes: The visual interactive unit displays the evaluation results and recommended solutions, and provides a human-computer interaction interface for decision-makers to input their preferences and feedback. The real-time correction unit dynamically adjusts the recommended solution based on feedback data to ensure that the solution meets the actual decision-making needs.

Images