A power distribution grid investment effectiveness method and system

Abstract

The application discloses a power distribution power supply grid investment effect method and system, which divides the power distribution network into power supply grids, and constructs a multi-dimensional quantitative evaluation index system. Through spatial matching and topological analysis, the investment project is associated to determine the expected improvement. The initial combination weight is calculated by using the entropy weight method and expert scoring, and the dynamic adjustment factor is generated by combining the development stage identifier to obtain the final dynamic weight and calculate the comprehensive score. Further, the unit investment benefit index is calculated, and the load data and equipment state data are collected to obtain the actual index improvement. When the deviation between the expected and actual improvement exceeds the first preset threshold, the dynamic adjustment factor is adjusted. A reinforcement learning model is constructed, and future investment planning direction data is generated based on historical data. The application realizes dynamic self-adaptation and feedback optimization of investment evaluation, and improves the intelligent level of investment decision and the efficiency of fund use.

Description

Technical Field

[0001] This invention belongs to the field of power system planning and management technology, and relates to a method and system for evaluating the investment effectiveness of power distribution grids. Background Technology

[0002] The power distribution network is a crucial component of the power system, directly serving a vast number of electricity users. With rapid economic and social development, users' demands for power supply reliability and power quality are increasing daily. To improve the overall level of the distribution network, power companies invest heavily in distribution network construction projects annually. These investments cover various aspects, including equipment upgrades, network structure optimization, and intelligent upgrades. Scientifically and rationally planning investment and accurately evaluating investment results are of great significance for improving the efficiency of capital utilization and promoting the high-quality development of the distribution network.

[0003] Existing power distribution network investment management largely adopts an extensive approach, with project initiation often relying on experience-based judgments or simple indicator summaries, lacking refined grid division of power supply areas. The weights of evaluation indicators across different dimensions are typically fixed, failing to reflect the actual needs of the power grid at different stages of development. For example, in the early stages of power grid construction, equipment level may be the primary concern. However, in the mature stage of the power grid, operational efficiency and green interaction may become more critical. A fixed weighting system is ill-suited to this dynamic change, leading to a discrepancy between investment direction and actual needs.

[0004] Furthermore, existing technologies suffer from a lag in post-investment evaluation, typically conducting a one-time evaluation only after the project has been in operation for a considerable period. This lack of real-time comparative analysis between expected and actual improvements means that when actual operational data deviates significantly from expected targets, subsequent investment strategies cannot be adjusted promptly. This absence of a feedback mechanism easily leads to the recurrence of inefficient investments and frequent waste of funds.

[0005] In terms of investment planning and decision-making, traditional methods rely heavily on human experts for scoring, which is highly subjective and difficult to process massive amounts of historical data. They fail to fully utilize big data and artificial intelligence technologies to uncover the patterns behind the data. There is still a lack of effective technical means to automatically generate future investment plans based on historical investment performance, making it difficult to achieve intelligent and precise investment decisions.

[0006] Existing evaluation systems have relatively simple functions, often focusing only on indicator calculation or data display. They lack a mechanism to link evaluation results with investment planning in a closed loop, and cannot form a complete technical chain from grid division, indicator evaluation, deviation analysis to strategy optimization, making it difficult to meet the current needs of lean management of power distribution networks.

[0007] Therefore, there is an urgent need for a method and system for evaluating the effectiveness of power distribution grid investment that can dynamically adjust weights, provide real-time feedback on deviations, and intelligently generate planning directions. Summary of the Invention

[0008] To address the problems existing in the background technology, this invention proposes a method and system for evaluating the investment efficiency of power distribution grids.

[0009] The first aspect of this application provides a method for evaluating the investment effectiveness of a power distribution grid, including:

[0010] The power distribution network is divided into several power supply grids, and a quantitative evaluation index system is constructed that includes six dimensions: safety and reliability, power supply quality, equipment level, operation efficiency, green interaction and service experience.

[0011] Establish an investment project database, associate each investment project with the corresponding power supply grid through spatial location matching and electrical topology analysis, and determine the expected improvement of each project in various indicators;

[0012] The comprehensive score of each power supply grid is calculated based on a dynamic weighting mechanism, and the unit investment benefit index is calculated based on the ratio of the change in comprehensive score before and after project implementation to the investment amount.

[0013] Data on power grid operation is collected according to a preset time period to calculate the actual improvement in indicators. The deviation between the expected improvement and the actual improvement is calculated. When the deviation exceeds a preset threshold, the dynamic adjustment factor for the next period is adjusted. Data on future investment planning direction is generated based on the historical unit investment benefit index.

[0014] Optionally, the construction of a quantitative evaluation index system comprising six dimensions specifically includes:

[0015] In terms of safety and reliability, power supply reliability rate and average fault repair time are selected as indicators.

[0016] In the power supply quality dimension, the voltage qualification rate and the average voltage deviation are selected as indicators.

[0017] In the dimension of equipment level, the equipment health index and the coverage rate of intelligent equipment are selected as indicators.

[0018] In terms of operational efficiency, the line load balance, the proportion of heavy-load distribution transformers, and the line loss rate are selected as indicators.

[0019] In the green interaction dimension, the penetration rate of distributed power sources, the coverage rate of electric vehicle charging facilities, and the proportion of adjustable load are selected as indicators.

[0020] In terms of service experience, the average number of power outages per user, the complaint rate, and the average time for business expansion and installation applications were selected as indicators.

[0021] Optionally, the calculation of the comprehensive score for each power supply grid based on the dynamic weighting mechanism specifically includes:

[0022] The objective weights of each dimension index are calculated using the entropy weight method, and the subjective weights of each dimension index are determined using the expert scoring method.

[0023] The initial combined weights are obtained by linearly combining the objective weights and subjective weights using a balance coefficient;

[0024] The dynamic adjustment factor is determined based on the development stage and policy orientation of the power grid. The initial combined weights are multiplied by the dynamic adjustment factor and normalized to obtain the final dynamic weights.

[0025] The comprehensive score of the power grid is obtained by multiplying the standardized values ​​of each dimension index with the corresponding final dynamic weights and then summing the results.

[0026] Optionally, the standardized value is calculated as follows:

[0027] For benefit-type indicators, the current indicator value is subtracted from the minimum indicator value and then divided by the difference between the maximum and minimum indicator values.

[0028] For cost-type indicators, the maximum indicator value minus the current indicator value is used, and then divided by the difference between the maximum and minimum indicator values.

[0029] The maximum and minimum index values ​​are respectively the maximum and minimum values ​​of the index in all power supply grids.

[0030] Optionally, the formula for calculating the unit investment benefit index is:

[0031] The unit investment benefit index is equal to the comprehensive score after project implementation minus the comprehensive score before project implementation, and then divided by the sum of the investment amounts of all projects invested in the power grid.

[0032] The overall score after project implementation refers to the overall score predicted after the plan is implemented or the overall score after actual operation.

[0033] Optionally, the dynamic adjustment of dimension weights based on the deviation results specifically includes:

[0034] When the actual improvement of a certain dimension indicator deviates from the expected improvement by more than a preset threshold, the dynamic adjustment factor of that dimension in the next cycle will be automatically reduced.

[0035] Alternatively, an expert review mechanism can be triggered to reset the subjective weights of that dimension.

[0036] Optionally, the proposed future investment planning direction includes:

[0037] The power grids are sorted according to the unit investment benefit index to identify inefficient and efficient grids;

[0038] For inefficient grids, generate a list of rectification measures; for efficient grids, summarize successful experiences and prioritize subsequent investment.

[0039] An investment decision agent is trained using a reinforcement learning algorithm with the goal of maximizing long-term cumulative investment benefits, and outputs the investment priority and recommended project types for each power grid.

[0040] A second aspect of this application provides a power distribution grid investment performance evaluation system, comprising:

[0041] The data acquisition module is used to collect power grid operation data, investment project data, and user service data.

[0042] The grid management module is used to perform power supply grid division, maintain grid attributes, and establish the association between investment projects and power supply grids;

[0043] The evaluation calculation module is used to perform standardization of indicators, dynamic weight calculation, comprehensive score calculation, and unit investment benefit index calculation.

[0044] The feedback optimization module is used to perform deviation analysis between the expected and actual improvement, dynamic correction of dimension weights, and generation of future investment planning directions.

[0045] The visualization module is used to display the distribution of investment benefits in the power grid, the trend of indicator changes, and investment planning recommendations.

[0046] Optionally, the evaluation calculation module integrates an entropy weight calculation unit, a subjective weight assignment unit, and a dynamic adjustment factor generation unit.

[0047] The entropy weight calculation unit is used to calculate objective weights based on the dispersion of the indicator data.

[0048] The subjective weight assignment unit is used to receive expert scoring data and calculate subjective weights;

[0049] The dynamic adjustment factor generation unit is used to output the dynamic adjustment factors of each power supply grid in each dimension according to a preset rule base or machine learning model.

[0050] Optionally, the feedback optimization module integrates a deviation analysis unit, a strategy correction unit, and a planning recommendation unit.

[0051] The deviation analysis unit is used to calculate the deviation value of the improvement of each dimension indicator and determine whether it exceeds the preset threshold.

[0052] The strategy correction unit is used to adjust the dynamic adjustment factor or generate a list of rectification measures based on the deviation analysis results.

[0053] The planning recommendation unit has a built-in reinforcement learning model, which is used to output a sequence of future investment priorities based on historical investment benefit data.

[0054] Compared with the prior art, the present invention has the following beneficial effects:

[0055] This invention provides a method and system for evaluating the effectiveness of power distribution grid investments. By dividing the power distribution network into several power supply grids, a quantitative evaluation index system with multiple dimensions is constructed. Through spatial location matching and electrical topology analysis, investment projects are accurately associated with corresponding power supply grids, achieving refined investment management and changing the previous extensive management model.

[0056] This invention introduces a dynamic weight adjustment mechanism. It uses the entropy weight method to calculate objective weights, combines this with expert scoring to determine subjective weights, and most importantly, generates a dynamic adjustment factor based on the development stage of the power grid. This factor reflects the actual demand differences of the power grid at different development stages, allowing the final dynamic weights to adapt adaptively. This solves the problem that fixed weights cannot adapt to the dynamic development of the power grid, improving the scientific rigor and accuracy of the evaluation results.

[0057] This invention establishes a real-time feedback optimization closed loop. Power grid operation data is collected according to a preset time period, the actual improvement in indicators is calculated, and this is compared with the expected improvement to obtain a deviation value. When the deviation value exceeds a first preset threshold, the dynamic adjustment factor for the next period is automatically adjusted. This mechanism can promptly detect deviations in investment execution, avoiding the recurrence of inefficient investments and significantly improving the efficiency of capital utilization.

[0058] This invention utilizes a reinforcement learning model to generate data on future investment planning directions, mapping historical unit investment benefit indices, load data, and equipment status data into state vectors. Investment project types and amounts are mapped into action commands, and an investment decision-making agent model is obtained through training. This model can uncover patterns behind massive amounts of historical data, outputting investment priority sequences and recommended project type data for each power grid. It overcomes the shortcomings of traditional manual decision-making, such as strong subjectivity and weak data processing capabilities, achieving intelligent and precise investment decision-making.

[0059] The system and method provided by this invention work in close coordination to achieve full-process automation from data acquisition, grid management, evaluation calculation to feedback optimization. The visualization module can intuitively present the distribution of grid benefits and indicator trends, providing managers with clear decision-making basis. The overall technical solution forms a complete closed-loop control chain, effectively promoting the high-quality development of the power distribution network. Attached Figure Description

[0060] Figure 1 This is a flowchart of a method for evaluating the investment efficiency of power distribution grids according to an embodiment of the present invention;

[0061] Figure 2 This is a schematic diagram of a power distribution grid investment performance evaluation system according to an embodiment of the present invention. Detailed Implementation

[0062] 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.

[0063] In one embodiment, such as Figure 1 As shown, a method for evaluating the investment efficiency of power distribution grids is provided, which is then applied to... Figure 1 Taking China as an example, the following specific steps will be used:

[0064] S10: Divide the distribution network into several power supply grids and construct a quantitative evaluation index system that includes six dimensions: safety and reliability, power supply quality, equipment level, operation efficiency, green interaction and service experience.

[0065] Specifically, geographic information system (GIS) data and electrical topology connection data of the power distribution network are acquired. Preliminary regional segmentation is performed based on administrative boundaries, natural road divisions, and substation power supply ranges. Combined with the electrical connections of medium-voltage lines, the electrical connections within each power supply grid are ensured to be tight, while the grids remain relatively independent. This ultimately forms several power supply grid units covering the entire region. Each power supply grid, as the basic management unit for investment effectiveness, has an independent coded identifier.

[0066] Based on the division of the power supply grid, a quantitative evaluation index system is constructed. This system comprises six dimensions. The first dimension is safety and reliability, which selects the power supply reliability rate and the mean time to repair (MTTR) for faults. The power supply reliability rate reflects the user's ability to obtain continuous power, while the MTR reflects the recovery speed after a grid fault. The second dimension is power quality, which selects the voltage qualification rate and the average voltage deviation. The voltage qualification rate measures the percentage of time the voltage is within the allowable range, while the average voltage deviation measures the degree to which the voltage deviates from the rated value.

[0067] The third dimension is the equipment level dimension, which selects the equipment health index and the intelligent equipment coverage index. The equipment health index comprehensively assesses the operating status of equipment such as transformers and lines, while the intelligent equipment coverage index counts the proportion of equipment with remote monitoring and automatic control functions. The fourth dimension is the operating efficiency dimension, which selects the line load balance index, the distribution transformer heavy load ratio index, and the line loss rate index. The line load balance index reflects the uniformity of load distribution on each line, the distribution transformer heavy load ratio index counts the proportion of distribution transformers under heavy load, and the line loss rate index measures the proportion of electrical energy lost during transmission.

[0068] The fifth dimension is the green interaction dimension, which selects distributed power generation penetration rate, electric vehicle charging facility coverage rate, and adjustable load ratio as indicators. The distributed power generation penetration rate reflects the proportion of clean energy sources such as photovoltaic and wind power connected within the grid. The electric vehicle charging facility coverage rate assesses the completeness of charging infrastructure, and the adjustable load ratio measures the ability of user-side loads to participate in grid interaction. The sixth dimension is the service experience dimension, which selects average number of power outages per user, complaint rate, and average time for business expansion application as indicators. The average number of power outages per user counts the frequency of power outages per unit time. The complaint rate reflects user satisfaction with power supply services, and the average time for business expansion application measures the efficiency of handling new electricity installation business.

[0069] In this embodiment, the present invention achieves comprehensive coverage of evaluation dimensions. The six dimensions encompass all aspects of information, from the physical state of the power grid to user perception. The quantitative evaluation index system provides a solid data foundation for subsequent weight calculations. The multi-dimensional index settings avoid the one-sidedness of evaluation from a single perspective, accurately identifying the shortcomings of different power supply grids and providing a scientific basis for the precise establishment of investment projects. Dynamic adjustment factors can be used to specifically weight weak links in different dimensions, significantly improving the objectivity and guiding significance of investment performance results.

[0070] S20: Establish an investment project database, associate each investment project with the corresponding power supply grid through spatial location matching and electrical topology analysis, and determine the expected improvement of each project on various indicators.

[0071] Specifically, information on all planned investment projects during the power distribution network planning period is collected. Project information includes project name, construction content, investment amount, implementation location coordinates, and expected commissioning time. This information is entered into a database to form an investment project database. Each project has a unique identifier in the database, and project types cover various categories such as line renovation, distribution transformer capacity expansion, automation upgrades, and new energy access.

[0072] Investment projects are associated with corresponding power supply grids through spatial location matching. Geographic coordinate data for each investment project is extracted. Geographic Information System (GIS) interfaces are used to obtain boundary vector data of the divided power supply grids. An algorithm determining whether a point lies within a polygon is used to determine the range of the power supply grid within which the project coordinates fall. If a project spans multiple grid boundaries, its affiliation is determined based on the area where the main equipment is located or the area with the largest investment proportion. Finally, each investment project is tagged with its assigned power supply grid.

[0073] Further verification and optimization of correlations are achieved through electrical topology analysis. Single-line diagram data and node connection relationships of the distribution network are read. The connection paths of equipment involved in the investment project within the electrical network are traced. The impact range of the project's operation on the electrical parameters of surrounding nodes is analyzed. When the project's spatial location is at the grid edge and its electrical connections mainly lead to adjacent grids, its assigned grid is adjusted based on the principle of shortest electrical distance. This ensures that the investment project and the power supply grid maintain a strong correlation both physically and electrically. This dual matching mechanism avoids assignment errors caused by inconsistencies between geographical boundaries and power supply ranges.

[0074] Determine the expected improvement of each dimension's indicators by the project. For each investment project, pre-determine its technological improvement model. For line renovation projects, calculate the reduction in line loss rate and the improvement in power supply reliability based on changes in conductor cross-section and length. For distribution transformer capacity expansion projects, calculate the decrease in the proportion of heavily loaded distribution transformers and the degree of improvement in voltage qualification rate based on capacity changes. For automation upgrade projects, calculate the reduction in mean time to repair (MTBL) based on terminal coverage. For green interactive projects, calculate the increase in distributed power penetration based on access capacity. Map the above calculation results to six evaluation dimensions to form a dataset of the expected improvement of each dimension's indicators for the project.

[0075] In this embodiment, the present invention achieves precise binding between investment projects and the power grid. The combination of spatial location matching and electrical topology analysis ensures the accuracy of the correlation. The quantitative calculation of the expected improvement provides a clear benchmark for subsequent investment benefit evaluation. This method can clearly define the specific contribution of each project to grid performance, avoiding problems such as unclear investment responsibility and ambiguous benefit attribution. It lays a solid foundation for subsequent calculations of the deviation between actual and expected improvement, significantly improving the precision and reliability of investment effectiveness evaluation.

[0076] S30: Calculate the comprehensive score of each power supply grid based on the dynamic weighting mechanism, and calculate the unit investment benefit index based on the ratio of the change in comprehensive score before and after project implementation to the investment amount.

[0077] Specifically, the raw data of quantitative indicators for each power supply grid across six dimensions are obtained. The raw data is then normalized to eliminate dimensional differences, resulting in standardized indicator values. Next, the initial combined weights are calculated. The entropy weight method is used to calculate objective weights based on the dispersion of the indicator data. Simultaneously, experts score the importance of each dimension, and the analytic hierarchy process (AHP) is used to determine subjective weights. Finally, the objective and subjective weights are linearly weighted and fused to generate the initial combined weights.

[0078] A dynamic adjustment factor is introduced to correct the initial combined weights. The development stage identifier for each power grid is identified. These stages include start-up, growth, maturity, and transformation. Different emphasis strategies are preset for different stages. For example, the focus is on safety and reliability during the start-up stage, and on operational efficiency during the maturity stage. The corresponding dynamic adjustment coefficient is obtained by looking up the development stage identifier in a table. This dynamic adjustment coefficient is applied to the initial combined weights to obtain the final dynamic weights. These weights adaptively reflect the actual needs of the power grid at different development stages. The final dynamic weights are weighted and summed with standardized index values ​​to calculate the comprehensive score for each power grid.

[0079] The specific steps for calculating the unit investment benefit index are as follows: First, collect the comprehensive score of each power supply grid before project implementation, recorded as the pre-implementation score. Second, after the project is put into operation and stabilizes, collect data again to calculate the post-implementation comprehensive score, recorded as the post-implementation score. Third, subtract the pre-implementation score from the post-implementation score to obtain the change in comprehensive score. This change reflects the overall performance improvement brought about by the investment project. Fourth, obtain the total investment amount of all related projects within the power supply grid. Fifth, divide the change in comprehensive score by the total investment amount to obtain the unit investment benefit index. This index characterizes the power grid performance improvement effect brought about by each unit of monetary investment.

[0080] In this embodiment, the present invention achieves dynamic adaptation of evaluation weights. This dynamic weighting mechanism overcomes the shortcomings of traditional fixed weights, which cannot adapt to changes in power grid development, making the evaluation results more closely reflect the actual operating state of the power grid. The introduction of the unit investment benefit index establishes a direct quantitative link between input and output. This index can intuitively reflect the efficiency of fund utilization. Managers can quickly identify high-efficiency projects and inefficient investment areas based on this index, providing clear data support for subsequent optimization of the investment structure, effectively avoiding blind investment and redundant construction, and improving the accuracy and scientific nature of distribution network project investment.

[0081] It is important to note that the calculation of the unit investment benefit index spans the entire lifecycle of distribution network investment management. The index is calculated using different comprehensive scoring data sources depending on the project's stage. Specifically, it is divided into two independent application scenarios: the planning and decision-making stage and the post-evaluation feedback stage.

[0082] During the planning and decision-making phase, before investment projects are formally implemented, the system cannot obtain actual operational data. Therefore, the system uses the predicted comprehensive score after project implementation as part of the numerator. The predicted comprehensive score is a theoretical value derived from historical data models and simulation techniques. The system reads the current comprehensive score of the power grid before project implementation, subtracts the current comprehensive score before implementation from the predicted comprehensive score after implementation, and obtains the predicted benefit increment. This predicted benefit increment is divided by the sum of the amounts of all investment projects invested in the power grid, and the result is the unit investment benefit index for the planning phase. This index is used to evaluate the economics of different schemes before project initiation. A higher value indicates a better expected unit capital input-output ratio for the scheme. The system prioritizes each candidate project based on this index, thereby guiding planners to select the optimal investment scheme.

[0083] In the post-evaluation feedback phase, the investment project has been completed and put into operation for the preset period. At this time, the system has collected real power grid operation data. The system uses the comprehensive score after actual operation as part of the calculation numerator. The comprehensive score after actual operation is an objective value calculated based on real monitoring indicators. The system reads the historical comprehensive score data archived before the project implementation. The comprehensive score after actual operation is subtracted from the comprehensive score archived before the project implementation to obtain the actual benefit increment. This actual benefit increment is divided by the sum of the investment amounts of all investment projects invested in the power grid. The result is the unit investment benefit index in the post-evaluation phase. This index is used to verify the authenticity of the investment effect after the project is put into operation. The higher the value, the better the actual unit capital input-output ratio of the project. The system uses this index to rate the effectiveness of completed projects, thereby verifying the accuracy of planning decisions and providing a basis for subsequent adjustments.

[0084] Through the above-described scenario definitions, this invention achieves closed-loop management of the entire investment process. The planning stage uses predicted values ​​to address the lack of quantitative basis for pre-decision decisions, while the post-evaluation stage uses actual values ​​to address the lack of objective standards for post-evaluation. Although the comprehensive scores in the two scenarios come from different sources, they both undergo the same standardized processing procedure, ensuring consistency in calculation logic and uniformity in evaluation criteria. Predicted values ​​and actual values ​​serve different management objectives. The former focuses on optimizing the solution, while the latter focuses on verifying effectiveness. Together, they constitute a complete unit investment benefit evaluation system, avoiding distortion of evaluation results due to data source confusion, and significantly improving the scientific nature of distribution network investment decisions and the controllability of investment effects. This provides solid technical support for power grid companies to achieve precise investment and lean management.

[0085] S40: Collect power grid operation data according to a preset time period to calculate the actual indicator improvement, calculate the deviation between the expected improvement and the actual indicator improvement, adjust the dynamic adjustment factor for the next period when the deviation exceeds a preset threshold, and generate future investment planning direction data based on the historical unit investment benefit index.

[0086] Specifically, this invention collects power grid operation data according to a preset time period to calculate the actual improvement in indicators. The preset time period can be set to monthly, quarterly, or annually. At the end of each period, the power grid operation data for that period is automatically extracted from the distribution automation system, electricity consumption information collection system, and marketing business system. The extracted data includes actual operation indicators in six dimensions, such as power supply reliability, voltage qualification rate, and line loss rate. The collected actual operation indicators are substituted into a quantitative evaluation indicator system to calculate the actual comprehensive score of each power supply grid at the end of the period. The actual comprehensive score at the end of the period is compared with the baseline comprehensive score before project implementation, and the difference is the actual improvement in indicators.

[0087] Calculate the deviation between the expected improvement and the actual improvement. Read the previously determined expected improvement data for each dimension of the project, and subtract the expected improvement from the actual improvement one by one to obtain the deviation value for each dimension. Further calculate the comprehensive deviation value, which can be the weighted sum or root mean square error of the deviation values ​​for each dimension. Set a first preset threshold, which is determined based on historical data statistical patterns and the allowable range of engineering errors. When the comprehensive deviation value exceeds the first preset threshold, a dynamic adjustment mechanism is triggered. Analyze the cause of the deviation. If it is due to changes in the power grid development stage that renders the original weights inapplicable, adjust the dynamic adjustment factor for the next cycle. Specifically, if the actual improvement is lower than expected, increase the dynamic adjustment coefficient for that grid weakness dimension and assign it a higher weight in the next cycle evaluation to guide subsequent investment towards weak links.

[0088] Based on historical unit investment efficiency indices, future investment planning direction data is generated, and a reinforcement learning model is constructed, using the unit investment efficiency indices of each historical period as reward signals input into the model. The indicator status, investment type, and investment amount of each power grid are used as the state space to train the agent to learn the mapping relationship between various investment strategies and final benefits under different states. After iterative training, the model can predict future return trends under different investment directions. Inputting the current state data of each power grid, the model outputs recommended future investment planning direction data, which includes suggested priority investment grid areas, recommended project types, and estimated investment scale rankings.

[0089] In this embodiment, the present invention achieves closed-loop feedback control for investment effectiveness. Periodic data collection ensures the timeliness of evaluation results, and the linkage mechanism between deviation monitoring and dynamic adjustment factors enables the evaluation system to have self-correcting capabilities. When project results do not meet expectations, the system can automatically adjust the evaluation orientation and correct investment deviations in a timely manner. The introduction of reinforcement learning models fully utilizes historical investment experience data, enabling the extraction of implicit investment patterns from massive amounts of historical data. The generated data on future investment planning directions is highly scientific and forward-looking, effectively guiding the precise planning of distribution network projects and avoiding inefficient and repetitive investments. This significantly improves the overall investment efficiency and operational level of the distribution network.

[0090] In one embodiment, step S30, namely calculating the comprehensive score of each power supply grid based on the dynamic weighting mechanism, further includes the following steps:

[0091] S31: Calculate the objective weights of each dimension indicator using the entropy weight method, and determine the subjective weights of each dimension indicator using the expert scoring method.

[0092] S32: The initial combined weights are obtained by linearly combining the objective weights and subjective weights using a balance coefficient;

[0093] S33: Determine the dynamic adjustment factor based on the development stage and policy orientation of the power grid, multiply the initial combined weights by the dynamic adjustment factor and normalize them to obtain the final dynamic weights;

[0094] S34: The comprehensive score of the power grid is obtained by multiplying the standardized values ​​of each dimension index with the corresponding final dynamic weights and summing the results.

[0095] Specifically, this invention utilizes the entropy weight method to calculate the objective weights of indicators across various dimensions. First, an original data matrix containing indicator values ​​for all power supply grids across six dimensions is constructed. This original data matrix is ​​then standardized, converting indicators of different dimensions into dimensionless relative numbers, and the proportion of each indicator in different power supply grids is calculated. Based on information entropy theory, the information entropy value of each indicator is calculated. The smaller the information entropy value, the greater the amount of information provided by the indicator and the higher its degree of variation. The difference coefficient of each indicator is calculated based on the information entropy value, and after normalization, the objective weights of each dimension's indicators are obtained. This process relies entirely on the inherent dispersion of the data, eliminating interference from human factors and ensuring the objectivity of the weights.

[0096] The subjective weights of each dimension's indicators were determined using an expert scoring method. A review panel comprised of experts in distribution network planning, operation and maintenance, and economic management was formed. Scoring criteria were developed, requiring experts to independently score the importance of six dimensions—safety and reliability, power quality, equipment level, operational efficiency, green interaction, and service experience—based on distribution network development strategies and practical operational experience. All expert scores were collected, outliers were removed, and the arithmetic mean was calculated. The average scores for each dimension were then normalized to obtain the subjective weights of each indicator. This process fully integrates the professional knowledge and experience of industry experts, reflecting management orientation and policy needs.

[0097] The initial combined weights are obtained by linearly combining objective and subjective weights using a balancing coefficient. A balancing coefficient between zero and one is set. This coefficient adjusts the contribution ratio of objective data-driven factors and subjective experience judgments in the final weights. The objective weights are multiplied by the balancing coefficient, and the subjective weights are multiplied by one and then subtracted from the balancing coefficient. The two products are then added together to obtain the initial combined weights for each dimension. The value of the balancing coefficient can be fine-tuned according to the specific needs of the evaluation period to achieve the optimal integration of objective and subjective information.

[0098] Dynamic adjustment factors are determined based on the development stage and policy orientation of the power supply grid. A power supply grid development stage database is established, dividing the grid into the initial stage, growth stage, maturity stage, and transformation stage. Basic adjustment strategies are preset for different development stages. For example, for grids in the initial stage, the weight of safety and reliability is increased. Policy orientation correction items are also introduced. When higher-level departments issue specific policies, such as green energy development policies or quality service improvement policies, the adjustment factors for the corresponding dimensions are increased. The stage-based basic strategies and policy orientation correction items are overlaid to generate the corresponding dynamic adjustment factors for each dimension.

[0099] The initial combined weights are multiplied by the dynamic adjustment factor and then normalized to obtain the final dynamic weights. The initial combined weights for each dimension obtained in the previous step are then multiplied by their corresponding dynamic adjustment factors. The sum of the products of all dimensions is calculated. The product of each dimension is divided by this sum to complete the normalization process. The result after this normalization is the final dynamic weight. The sum of the final dynamic weights for all dimensions equals one. This weighting system preserves the objective laws of the data, incorporates expert experience and judgment, and can respond in real time to changes in the power grid development stage and macroeconomic policy guidance.

[0100] The overall score of the power grid is obtained by multiplying the standardized values ​​of each dimension indicator by their corresponding final dynamic weights and then summing the results. The preprocessed and standardized values ​​of each dimension indicator are read. The standardized value of each dimension is multiplied by its corresponding final dynamic weight. The results of the six dimension multiplications are then summed. The sum is the overall score of the power grid. A higher overall score indicates a better overall performance level for the power grid.

[0101] In this embodiment, the present invention achieves a balance between the scientific nature and flexibility of weight determination. The combination of entropy weighting and expert scoring effectively overcomes the limitations of a single weighting method. The introduction of a balance coefficient makes the integration of subjective and objective weights smoother and more controllable. The application of dynamic adjustment factors enables the evaluation system to adapt to different stages of power grid development and changes in the external policy environment, ultimately ensuring that the dynamic weights accurately reflect the most important evaluation priorities at the current stage. The comprehensive score calculated based on this has extremely high reference value, accurately and comprehensively reflecting the operation and management level of the power supply grid, providing a reliable quantitative basis for the precise evaluation of investment projects, and significantly improving the scientific level of distribution network investment decisions.

[0102] In one embodiment, step S34, i.e., the calculation method of the standardized value, further includes the following steps:

[0103] S341: For benefit-type indicators, the current indicator value minus the minimum indicator value is divided by the difference between the maximum and minimum indicator values.

[0104] S342: For cost-type indicators, the maximum indicator value minus the current indicator value is used, and then divided by the difference between the maximum and minimum indicator values; where the maximum and minimum indicator values ​​are the maximum and minimum values ​​of the indicator in all power supply grids, respectively.

[0105] Specifically, this invention traverses all power supply grids, extracts all raw data under the same dimension index, and for each dimension index, selects the maximum and minimum index values. The maximum index value refers to the largest value of the index among all power supply grids in the current evaluation period, and the minimum index value refers to the smallest value of the index among all power supply grids in the current evaluation period. These two extreme values ​​constitute the boundary of the value range of the index data.

[0106] For benefit-type indicators, a specific normalization formula is used to calculate standardized values. Benefit-type indicators are those whose higher values ​​represent better grid performance, such as power supply reliability or voltage qualification rate. The calculation first obtains the current value of the indicator for the current power grid. The minimum indicator value, determined earlier, is subtracted from the current value to obtain the numerator. The minimum indicator value is then subtracted from the maximum indicator value to obtain the denominator. The numerator is divided by the denominator, and the quotient is the standardized value of the benefit-type indicator. This standardized value ranges from zero to one. If the current indicator value equals the minimum indicator value, the standardized value is 0. If the current indicator value equals the maximum indicator value, the standardized value is 1. This method ensures that the higher the value of the benefit-type indicator, the higher its standardized score, consistent with positive evaluation logic.

[0107] For cost-related indicators, a different normalization formula is used to calculate standardized values. Cost-related indicators, such as line loss rate or mean time to repair (MTBT), represent better grid performance with smaller values. The calculation involves obtaining the current value of this indicator for the current power grid. The numerator is obtained by subtracting the current value from the previously determined maximum value. The denominator is the difference between the maximum and minimum values. Dividing the numerator by the denominator yields the standardized value of the cost-related indicator. This standardized value also ranges from 0 to 1. If the current value equals the maximum value, the standardized value is zero. If the current value equals the minimum value, the standardized value is 1. This method achieves the reverse transformation of cost-related indicators, ensuring that smaller values ​​result in higher standardized scores and unifying the evaluation direction.

[0108] In this embodiment, the present invention effectively eliminates the dimensional differences between indicators of different dimensions. Power supply reliability is measured as a percentage, while line loss rate is also measured as a percentage but with the opposite meaning. Fault repair time is measured in minutes or hours. Directly weighting and summing these data with different units and properties leads to distorted evaluation results. The standardization method provided in this embodiment transforms all indicators into dimensionless pure numerical values, unifying the evaluation scale of all indicators and making indicators of different properties comparable. After processing, the values ​​of both benefit-type and cost-type indicators directly reflect the performance of the power grid; larger values ​​indicate better performance, and smaller values ​​indicate worse performance. This processing method ensures the accuracy and fairness of subsequent comprehensive score calculations, avoiding situations where excessively large values ​​or different units of individual indicators dominate the evaluation results. It lays a solid data foundation for constructing a scientific and reasonable evaluation system for distribution network investment effectiveness, significantly improving the reliability and persuasiveness of multi-indicator comprehensive evaluation results.

[0109] In one embodiment, step S40, which dynamically adjusts the dimension weights based on the deviation results, further includes the following steps: when the deviation between the actual improvement and the expected improvement of a certain dimension index exceeds a preset threshold, the dynamic adjustment factor of that dimension in the next period is automatically reduced; or an expert review mechanism is triggered to reset the subjective weight of that dimension.

[0110] Specifically, the present invention establishes a dynamic weight correction mechanism based on deviation feedback. This mechanism aims to address the disconnect between evaluation indicators and actual engineering results. The system monitors the difference between the actual and expected improvements in each dimension of the indicators in real time. First, it calculates the deviation value, which is equal to the absolute value of the actual improvement minus the expected improvement. The system then compares the calculated deviation value with a preset threshold. The preset threshold is a critical value pre-set based on the statistical distribution of historical data and the tolerance range of engineering management.

[0111] When the actual improvement of a certain dimension's indicator deviates from the expected improvement by more than a preset threshold, the system executes the first automatic correction strategy. This strategy automatically lowers the dynamic adjustment factor for that dimension in the next cycle. Specifically, the system identifies the specific dimension with excessive deviation. For example, if the actual improvement in the safety and reliability dimension is far lower than expected, it determines that the current weighting of that dimension may be excessive or distorted. The system then calls a preset attenuation coefficient and multiplies the current dynamic adjustment factor for that dimension by this coefficient. The attenuation coefficient is a positive number less than 1. After the multiplication, the dynamic adjustment factor for that dimension decreases. In the next evaluation cycle, due to the reduced dynamic adjustment factor, the proportion of that dimension in the final dynamic weight will decrease accordingly. This approach can quickly reduce excessive focus on that dimension in the evaluation system, prevent resources from continuing to be tilted towards inefficient areas, and guide investment towards other more promising dimensions.

[0112] Alternatively, when the deviation exceeds a preset threshold and meets specific triggering conditions, the system executes a second manual intervention strategy. This strategy triggers an expert review mechanism to reset the subjective weights of that dimension. Specific triggering conditions include deviations exceeding the limit for multiple consecutive periods, or deviations being extremely large, leading to severely abnormal evaluation results. The system automatically sends a review notification to the expert database. The notification includes historical data for that dimension, expected targets, actual completion status, and a deviation analysis report. An online or offline review meeting is organized with experts in relevant fields. Experts reassess the importance of that dimension based on new power grid operation trends and policy changes. Experts provide new weight recommendations using a scoring method. The system adopts the reset subjective weights from the experts, replacing the original subjective weight data. Subsequently, the system recalculates the initial combined weights based on the new subjective weights and updates the final dynamic weights.

[0113] In this embodiment, the present invention achieves adaptive optimization and closed-loop control of the evaluation system. The mechanism of automatically lowering the dynamic adjustment factor ensures the system's rapid response to short-term fluctuations, enabling timely correction of weight mismatch caused by sudden environmental changes without human intervention. The triggered expert review mechanism preserves the strategic judgment ability of human experts. It can handle complex and variable special cases that are difficult to quantify with simple algorithms. The two strategies complement each other, forming a management model that combines rigidity and flexibility. This effectively prevents overall investment decision-making errors caused by the distortion of a single indicator, ensures the accuracy and foresight of the unit investment benefit index calculation, significantly improves the precision and scientific nature of distribution network investment planning, and makes fund allocation more in line with the actual development needs of the power grid, avoiding ineffective investment and resource waste.

[0114] It is important to note that the dynamic adjustment factor is a continuous variable that evolves with the evaluation cycle. This variable has different numerical values ​​at different points in time. In the first evaluation cycle or at the beginning of a new policy implementation, the initial value of the dynamic adjustment factor is determined based on the development stage of the power grid and policy guidance. The development stage reflects the maturity of power grid construction. Policy guidance reflects the current management focus. At this time, the dynamic adjustment factor mainly reflects the macro-planning intent. The system uses this initial value to correct the initial combined weights to calculate the final dynamic weights for the current period.

[0115] Upon completion of the current evaluation and entry into the next cycle, the values ​​of the dynamic adjustment factors will be updated. This update process incorporates a deviation feedback mechanism. If the actual improvement of a certain dimension indicator in the previous cycle deviates from the expected improvement by more than a preset threshold, the system will correct the dynamic adjustment factor values ​​used in the previous cycle. Specifically, this means automatically lowering the dynamic adjustment factor value for that dimension, and using the corrected value as the input value for the dynamic adjustment factor in the next cycle. If the deviation does not exceed the preset threshold, the dynamic adjustment factor values ​​from the previous cycle will remain unchanged or be slightly adjusted.

[0116] Through the aforementioned mechanism, the dynamic adjustment factor achieves an organic integration of macro-level guidance and micro-level feedback. The initial value determined based on the development stage and policy orientation ensures that investment direction aligns with strategic planning. The automatic adjustment mechanism based on deviation feedback ensures the evaluation system possesses self-correcting capabilities; both operate on the same dynamic adjustment factor variable. The former establishes the variable's baseline starting point, while the latter determines the variable's evolution path. This interconnected relationship eliminates the possibility of ambiguous definitions, avoids misunderstandings of functional overlap, and ensures the closed-loop and consistency of the weight adjustment logic. This allows the final dynamic weight to respond to both long-term strategic needs and adapt to short-term operational fluctuations, significantly improving the robustness and adaptability of the distribution network investment performance evaluation system.

[0117] In one embodiment, step S40, namely generating suggestions for future investment planning directions, further includes the following steps:

[0118] S41: Sort the power supply grids according to the historical unit investment efficiency index and identify inefficient and efficient grids;

[0119] S42: Generate a list of corrective measures for inefficient grids, summarize successful experiences for efficient grids, and prioritize subsequent investment.

[0120] S43: Use reinforcement learning algorithms to train an investment decision agent with the goal of maximizing long-term cumulative investment benefits, and output the investment priority and recommended project types for each power grid.

[0121] Specifically, the process of generating future investment planning direction suggestions in this invention first sorts the power supply grids based on historical unit investment benefit indices. The system reads the unit investment benefit index calculated for all power supply grids in the previous evaluation period and uses this index as the sorting criterion, arranging all power supply grids in descending order. Based on the sorting results, the system automatically identifies high-efficiency grids at the top of the sequence and inefficient grids at the bottom. High-efficiency grids refer to power supply areas with a unit investment benefit index significantly higher than the average level. Inefficient grids refer to power supply areas with a unit investment benefit index significantly lower than the average level or failing to reach a preset passing score. This classification method intuitively reflects the current investment return level of each grid.

[0122] For identified inefficient grids, the system automatically generates a list of corrective measures. The system analyzes the shortcomings of inefficient grids across six dimensions of indicators. For example, if a grid's low efficiency is due to excessively high line loss rate, a loss reduction upgrade project will be included in the list. If poor efficiency is due to low power supply reliability, a grid structure optimization project will be included. The corrective measures list includes specific project types, estimated workload, and expected improvement targets. This list aims to guide maintenance personnel in addressing key issues hindering efficiency improvement. By implementing corrective measures, inefficient grids can be gradually transformed into medium- or high-efficiency grids, improving overall asset operation efficiency.

[0123] For identified high-efficiency power grids, the system summarizes successful experiences and prioritizes subsequent investment. The system extracts characteristic data of high-efficiency grids in terms of grid structure, equipment selection, and operating modes. This characteristic data is compared and analyzed with that of inefficient grids to extract replicable successful experience models. These models are then shared with planning departments as reference templates for other grid upgrades. Simultaneously, a priority strategy is implemented for high-efficiency grids in fund allocation. In the next investment cycle, expansion or upgrade projects are prioritized for high-efficiency grids. This is because high-efficiency grids typically possess favorable basic conditions and higher marginal returns, and continued investment can generate greater cumulative benefits. This differentiated strategy achieves optimal allocation of funds.

[0124] The system utilizes reinforcement learning algorithms to train an investment decision-making agent with the goal of maximizing long-term cumulative investment benefits. A reinforcement learning environment is constructed, defining the state of each power grid as the state space. The state space includes multi-dimensional features such as the current unit investment benefit index, the grid development stage, and load growth trends. The available investment project types and investment amounts are defined as the action space. The increment of the unit investment benefit index is defined as the reward signal. The objective function is set to maximize the long-term cumulative investment benefits over multiple future periods. Through multiple rounds of interaction and trial and error between the agent and the environment, the policy network parameters are continuously updated. During training, the agent learns which investment strategies to adopt under different states to obtain the maximum reward. After sufficient training, the investment decision-making agent possesses the ability to predict future return trends.

[0125] The trained investment decision-making agent outputs investment priorities and recommended project types for each power grid. The latest status data of all current power grids is input into the investment decision-making agent. The agent model evaluates each grid based on the learned optimal strategy. The output consists of two core parts. The first part is the investment priority ranking, which considers not only the current level of benefit but also future growth potential predictions. The second part is the recommended project type, providing specific optimal investment project suggestions for each grid, such as building new lines, replacing transformers, or installing smart terminals. This output constitutes the core content of data for future investment planning directions.

[0126] In this embodiment, the present invention achieves a leap from passive evaluation to proactive planning. Based on a ranking mechanism using a unit investment benefit index, the quality levels of power grid assets are clearly defined. A list of rectification measures is generated for inefficient grids, enabling precise problem identification and closed-loop management. Prioritizing investment in efficient grids leverages the demonstration and scale effects of high-quality assets. The introduction of reinforcement learning algorithms overcomes the limitations of traditional static planning methods. This method can handle nonlinear relationships and uncertainties in power grid development, aiming to maximize long-term cumulative investment benefits and avoiding short-sighted decision-making caused by short-term behavior. The output investment priorities and recommended project types are highly scientific and forward-looking, effectively guiding the precise implementation of distribution network projects, significantly improving the overall investment efficiency and sustainable development capabilities of the distribution network, and ensuring that every investment achieves maximum economic and social value.

[0127] In one embodiment, such as Figure 2 As shown, a power distribution grid investment performance evaluation system is provided. This system corresponds one-to-one with the power distribution grid investment performance evaluation method in the above embodiments. The power distribution grid investment performance evaluation system includes: a data acquisition module, a grid management module, an evaluation calculation module, and a visualization display module. Detailed descriptions of each functional module are as follows:

[0128] The data acquisition module is used to collect power grid operation data, investment project data, and user service data.

[0129] The grid management module is used to perform power supply grid division, maintain grid attributes, and establish the association between investment projects and power supply grids;

[0130] The evaluation calculation module is used to perform standardization of indicators, dynamic weight calculation, comprehensive score calculation, and unit investment benefit index calculation.

[0131] The feedback optimization module is used to perform deviation analysis between the expected and actual improvement, dynamic correction of dimension weights, and generation of future investment planning directions.

[0132] The visualization module is used to display the distribution of investment benefits in the power grid, the trend of indicator changes, and investment planning recommendations.

[0133] Furthermore, the evaluation calculation module integrates an entropy weight calculation unit, a subjective weight assignment unit, and a dynamic adjustment factor generation unit.

[0134] The entropy weight calculation unit is used to calculate objective weights based on the dispersion of the indicator data.

[0135] The subjective weight assignment unit is used to receive expert scoring data and calculate subjective weights;

[0136] The dynamic adjustment factor generation unit is used to output the dynamic adjustment factors of each power supply grid in each dimension according to a preset rule base or machine learning model.

[0137] Furthermore, the feedback optimization module integrates a deviation analysis unit, a strategy correction unit, and a planning recommendation unit.

[0138] The deviation analysis unit is used to calculate the deviation value of the improvement of each dimension indicator and determine whether it exceeds the preset threshold.

[0139] The strategy correction unit is used to adjust the dynamic adjustment factor or generate a list of rectification measures based on the deviation analysis results.

[0140] The planning recommendation unit has a built-in reinforcement learning model, which is used to output a sequence of future investment priorities based on historical investment benefit data.

[0141] Specific limitations regarding the power distribution grid investment performance evaluation system can be found in the limitations of the power distribution grid investment performance evaluation method described above, and will not be repeated here. Each module in the aforementioned power distribution grid investment performance evaluation system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0142] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0143] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for evaluating the investment effectiveness of a power distribution grid, characterized in that, include: The power distribution network is divided into several power supply grids, and a quantitative evaluation index system is constructed that includes six dimensions: safety and reliability, power supply quality, equipment level, operation efficiency, green interaction and service experience. Establish an investment project database, associate each investment project with the corresponding power supply grid through spatial location matching and electrical topology analysis, and determine the expected improvement of each project in various indicators; The comprehensive score of each power supply grid is calculated based on a dynamic weighting mechanism, and the unit investment benefit index is calculated based on the ratio of the change in comprehensive score before and after project implementation to the investment amount. Data on power grid operation is collected according to a preset time period to calculate the actual improvement in indicators. The deviation between the expected improvement and the actual improvement is calculated. When the deviation exceeds a preset threshold, the dynamic adjustment factor for the next period is adjusted. Data on future investment planning direction is generated based on the historical unit investment benefit index.

2. The method for evaluating the investment effectiveness of power distribution grids according to claim 1, characterized in that, The construction of the quantitative evaluation index system comprising six dimensions specifically includes: In terms of safety and reliability, power supply reliability rate and average fault repair time are selected as indicators. In the power supply quality dimension, the voltage qualification rate and the average voltage deviation are selected as indicators. In the dimension of equipment level, the equipment health index and the coverage rate of intelligent equipment are selected as indicators. In terms of operational efficiency, the line load balance, the proportion of heavy-load distribution transformers, and the line loss rate are selected as indicators. In the green interaction dimension, the penetration rate of distributed power sources, the coverage rate of electric vehicle charging facilities, and the proportion of adjustable load are selected as indicators. In terms of service experience, the average number of power outages per user, the complaint rate, and the average time for business expansion and installation applications were selected as indicators.

3. The method for evaluating the investment effectiveness of power distribution grids according to claim 1, characterized in that, The calculation of the comprehensive score for each power supply grid based on the dynamic weighting mechanism specifically includes: The objective weights of each dimension index are calculated using the entropy weight method, and the subjective weights of each dimension index are determined using the expert scoring method. The initial combined weights are obtained by linearly combining the objective weights and subjective weights using a balance coefficient; The dynamic adjustment factor is determined based on the development stage and policy orientation of the power grid. The initial combined weights are multiplied by the dynamic adjustment factor and normalized to obtain the final dynamic weights. The comprehensive score of the power grid is obtained by multiplying the standardized values ​​of each dimension index with the corresponding final dynamic weights and then summing the results.

4. The method for evaluating the investment effectiveness of power distribution grids according to claim 3, characterized in that, The standardized value is calculated as follows: For benefit-type indicators, the current indicator value is subtracted from the minimum indicator value and then divided by the difference between the maximum and minimum indicator values. For cost-related indicators, the maximum indicator value minus the current indicator value is used, and then divided by the difference between the maximum and minimum indicator values. The maximum and minimum indicator values ​​are the maximum and minimum values ​​of the indicator across all power grids, respectively.

5. The method for evaluating the investment effectiveness of power distribution grids according to claim 1, characterized in that, The formula for calculating the unit investment benefit index is as follows: The unit investment benefit index is equal to the comprehensive score after project implementation minus the comprehensive score before project implementation, and then divided by the sum of the investment amounts of all projects invested in the power grid. The overall score after project implementation refers to the overall score predicted after the plan is implemented or the overall score after actual operation.

6. The method for evaluating the investment effectiveness of power distribution grids according to claim 1, characterized in that, The adjustment factor for the next cycle includes, in detail: When the actual improvement of a certain dimension indicator deviates from the expected improvement by more than a preset threshold, the dynamic adjustment factor of that dimension in the next cycle will be automatically reduced. Alternatively, an expert review mechanism can be triggered to reset the subjective weights of that dimension.

7. The method for evaluating the investment effectiveness of power distribution grids according to claim 1, characterized in that, The data for generating future investment planning directions based on historical unit investment benefit indices includes: The power grid is sorted according to the historical unit investment efficiency index to identify inefficient and efficient grids; For inefficient grids, generate a list of rectification measures; for efficient grids, summarize successful experiences and prioritize subsequent investment. An investment decision agent is trained using a reinforcement learning algorithm with the goal of maximizing long-term cumulative investment benefits, and outputs the investment priority and recommended project types for each power grid.

8. A power distribution grid investment performance evaluation system, characterized in that, include: The data acquisition module is used to collect power grid operation data, investment project data, and user service data. The grid management module is used to perform power supply grid division, maintain grid attributes, and establish the association between investment projects and power supply grids; The evaluation calculation module is used to perform standardization of indicators, dynamic weight calculation, comprehensive score calculation, and unit investment benefit index calculation. The feedback optimization module is used to perform deviation analysis between the expected and actual improvement, dynamic correction of dimension weights, and generation of future investment planning directions. The visualization module is used to display the distribution of investment benefits in the power grid, the trend of indicator changes, and investment planning recommendations.

9. The power distribution grid investment performance evaluation system according to claim 8, characterized in that, The evaluation calculation module integrates an entropy weight method calculation unit, a subjective weight assignment unit, and a dynamic adjustment factor generation unit. The entropy weight calculation unit is used to calculate objective weights based on the dispersion of the indicator data. The subjective weight assignment unit is used to receive expert scoring data and calculate subjective weights; The dynamic adjustment factor generation unit is used to output the dynamic adjustment factors of each power supply grid in each dimension according to a preset rule base or machine learning model.

10. The power distribution grid investment performance evaluation system according to claim 8, characterized in that, The feedback optimization module integrates a deviation analysis unit, a strategy correction unit, and a planning recommendation unit. The deviation analysis unit is used to calculate the deviation value of the improvement of each dimension indicator and determine whether it exceeds the preset threshold. The strategy correction unit is used to adjust the dynamic adjustment factor or generate a list of rectification measures based on the deviation analysis results. The planning recommendation unit has a built-in reinforcement learning model, which is used to output a sequence of future investment priorities based on historical investment benefit data.

Images