A power distribution network project execution process evaluation method, system, device and medium

Abstract

This invention provides a method, system, equipment, and medium for evaluating the execution progress of distribution network projects, relating to the field of distribution network management technology. The method includes: acquiring operational data of the distribution network for the current time period and node information for each execution node; the operational data includes source-load data for conventional equipment and renewable energy equipment; based on this data, constructing a multi-dimensional vector and calculating the initial progress score for each execution node; extracting features from the source-load data to generate a source-load trend-fluctuation hierarchical representation, and determining the source-load coupling scenario category and fluctuation weight through dual-channel feature extraction; combining the scenario category and node information to determine the dynamic sensitivity coefficient and dynamic correction coefficient for each node; performing asymmetric correction on the initial progress score based on the dynamic correction coefficient to obtain the final progress score, and integrating the scores of all nodes to obtain the execution progress evaluation result for the corresponding distribution network project. This invention improves the accuracy and reliability of distribution network project execution progress evaluation.

Description

Technical Field

[0001] This invention relates to the field of power distribution network management technology, and more specifically, to a method, system, equipment, and medium for evaluating the progress of power distribution network projects. Background Technology

[0002] The core of evaluating the execution progress of a distribution network project is to verify whether each execution node in the distribution network has achieved the goal of effectively implementing the corresponding equipment functions. Since source load data can directly reflect the actual operating status after equipment is connected, it is usually necessary to convert source load data into progress indicators. Existing distribution network project execution progress evaluations mostly adopt static logic that matches the process with fixed prediction thresholds, comparing the predicted source load data with preset thresholds to determine the completion status of the execution nodes.

[0003] In related technologies, with the large-scale integration of new energy equipment such as distributed photovoltaics and electric vehicle charging piles into the distribution network, the source-load characteristics of the distribution network exhibit strong volatility and uncertainty. Therefore, when source-load fluctuations cause a large deviation between actual data and predicted values, if a unified standard is used to process the process conversion of all execution nodes in the distribution network, source-load fluctuations may be mistakenly identified as execution nodes failing to meet standards, leading to significant deviations in process evaluation and consequently distorted process evaluation results. Summary of the Invention

[0004] The problem addressed by this invention is how to improve the accuracy of the evaluation of the execution process of power distribution network projects that include new energy equipment.

[0005] To address the aforementioned issues, this invention provides a method, system, device, and medium for evaluating the progress of power distribution network projects.

[0006] In a first aspect, the present invention provides a method for evaluating the execution progress of a power distribution network project, comprising:

[0007] Obtain the current time period operation data of the distribution network and the node information of each execution node of the distribution network in the current time period. The operation data includes the conventional source-load data of conventional equipment and the new energy source-load data of new energy equipment in the distribution network.

[0008] Based on the conventional source load data and the new energy source load data, a multi-dimensional vector of the operation data is determined, and based on the multi-dimensional vector, an initial process score for each execution node is obtained.

[0009] Feature extraction is performed on the conventional source-load data and the new energy source-load data to generate a hierarchical representation of the source-load trend and fluctuation of the distribution network in the current time period;

[0010] Dual-channel feature extraction is performed on the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period, as well as the conventional fluctuation weight and new energy fluctuation weight corresponding to the conventional equipment and the new energy equipment, respectively.

[0011] Based on the source-load coupling scenario category, and combined with the equipment type in the node information and the preset completion standard of the corresponding project of the power distribution network, the dynamic sensitivity coefficient of each execution node is determined;

[0012] Based on the dynamic sensitivity coefficient, and in combination with the conventional fluctuation weight and the new energy fluctuation weight, the dynamic correction coefficient of each execution node is determined;

[0013] The initial process score is asymmetrically corrected according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node. The execution process evaluation result of the corresponding project of the distribution network is obtained by combining the final process scores of all execution nodes.

[0014] Optionally, determining a multi-dimensional vector of the operating data based on the conventional source-load data and the new energy source-load data, and obtaining an initial process score for each execution node based on the multi-dimensional vector, includes:

[0015] According to the preset window size, the conventional source load data and the new energy source load data are extracted by sliding to obtain multiple window features;

[0016] Based on the window features, the multi-dimensional vector corresponding to each execution node is obtained;

[0017] The multi-dimensional vector and the device type in the node information are coupled with the preset completion standard input of the project to form a causal convolutional network to obtain the state curve corresponding to the execution node.

[0018] The initial process score is obtained by performing residual integration and Sigmoid mapping based on the state curve.

[0019] Optionally, the step of extracting features from the conventional source-load data and the new energy source-load data to generate a hierarchical representation of the source-load trend and fluctuation of the distribution network in the current time period includes:

[0020] The conventional source load data and the new energy source load data are subjected to adaptive denoising processing by variational mode decomposition algorithm to obtain the effective sequence of conventional source load and the effective sequence of new energy source load.

[0021] Based on a preset multi-scale decomposition strategy, ensemble empirical mode decomposition is performed on the effective sequence of conventional source load and the effective sequence of new energy source load respectively to obtain the trend component and multi-order fluctuation component corresponding to the effective sequence of conventional source load and the effective sequence of new energy source load respectively.

[0022] The trend components and multi-order fluctuation components of the conventional source-load data are dimensionally aligned with the trend components and multi-order fluctuation components of the new energy source-load data to construct a four-dimensional hierarchical structure including a conventional trend layer, a conventional fluctuation layer, a new energy trend layer, and a new energy fluctuation layer, thereby generating a hierarchical representation of the source-load trend-fluctuation of the distribution network in the current time period.

[0023] Optionally, the dual-channel feature extraction of the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period and the conventional fluctuation weights and new energy fluctuation weights corresponding to the conventional equipment and the new energy equipment, respectively, includes:

[0024] Input the conventional trend layer and the conventional fluctuation layer in the source load trend-fluctuation hierarchical representation into the conventional source load feature extraction channel to obtain the amplitude change features of the conventional trend layer and the fluctuation features of the conventional fluctuation layer.

[0025] The new energy trend layer and the new energy fluctuation layer are input into the new energy source load feature extraction channel to obtain the steady-state features of the new energy trend layer and the abrupt change features of the new energy fluctuation layer;

[0026] The amplitude variation feature and the fluctuation feature are normalized and fused to obtain a conventional source-load comprehensive feature vector;

[0027] By combining feature concatenation and principal component analysis, the steady-state features and the abrupt change features are simplified in dimensionality to obtain a comprehensive feature vector of new energy source and load.

[0028] Based on the conventional source-load integrated feature vector and the new energy source-load integrated feature vector, the source-load coupling scenario category is obtained by prediction.

[0029] The coupling correlation degree and the cooperative fluctuation coefficient of the conventional source-load integrated feature vector and the new energy source-load integrated feature vector are obtained, and the conventional fluctuation weight and the new energy fluctuation weight are determined by the entropy method based on the coupling correlation degree and the cooperative fluctuation coefficient.

[0030] Optionally, determining the dynamic sensitivity coefficient of each execution node based on the source-load coupling scenario category, combined with the device type in the node information and the preset completion standard of the corresponding project in the distribution network, includes:

[0031] Based on the source-load coupling scenario category, a sensitivity benchmark matrix is ​​generated;

[0032] Based on the device type and preset completion standard in the node information of each execution node, the baseline sensitivity coefficient of each execution node is obtained by matching in the sensitivity baseline matrix;

[0033] Based on the operating years, construction complexity level, and historical progress deviation rate of the associated equipment of the execution node, the additional impact characteristics of the execution node are obtained;

[0034] Based on the additional impact characteristics, the sensitivity correction factor for each execution node is obtained;

[0035] The dynamic sensitivity coefficient of each execution node is determined based on the baseline sensitivity coefficient and the sensitivity correction factor of the execution node.

[0036] Optionally, determining the dynamic correction coefficient for each execution node based on the dynamic sensitivity coefficient, combined with the conventional fluctuation weight and the new energy fluctuation weight, includes:

[0037] Based on the device type in the node information of each execution node, a target fluctuation weight is determined for the execution node, wherein the target fluctuation weight is the conventional fluctuation weight or the new energy fluctuation weight;

[0038] Based on the target fluctuation weight and combined with the preset cross-device type fluctuation correlation coefficient, the comprehensive fluctuation weight of the execution node is obtained.

[0039] The initial correction coefficient of each execution node is obtained by weighting the dynamic sensitivity coefficient and the comprehensive fluctuation weight of each execution node.

[0040] Based on the source-load coupling scenario category, a scenario adaptation correction coefficient is determined, and combined with the initial correction coefficient, the dynamic correction coefficient for each execution node is obtained.

[0041] Optionally, the step of asymmetrically correcting the initial process score according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node, and then combining the final process scores of all execution nodes to obtain the execution process evaluation result of the corresponding project of the distribution network, includes:

[0042] Based on the dynamic correction coefficients of the execution node, determine the asymmetric correction rule of the execution node;

[0043] Based on the initial process score of each execution node and in conjunction with the asymmetric correction rule, a target correction formula corresponding to the execution node is generated.

[0044] The intermediate process score of the execution node is obtained through the target correction formula;

[0045] Based on the preset valid range of the distribution network execution process score, the intermediate process score is subjected to boundary constraint processing to obtain the final process score of each execution node;

[0046] A depth weight is generated according to the preset topology depth of each execution node;

[0047] The final process score is weighted and summed with the corresponding depth weight to obtain the weighted total process score.

[0048] The weighted process total score is mapped to a percentage range to obtain the execution process evaluation result of the distribution network in the current time period.

[0049] Secondly, the present invention provides a power distribution network project execution progress evaluation system, comprising:

[0050] The data acquisition unit is used to acquire the operating data of the distribution network in the current time period and the node information of each execution node of the distribution network in the current time period. The operating data includes the conventional source-load data of conventional equipment and the new energy source-load data of new energy equipment in the distribution network.

[0051] The initial analysis unit is used to determine the multi-dimensional vector of the running data based on the conventional source load data and the new energy source load data, and to obtain the initial process score of each execution node based on the multi-dimensional vector.

[0052] The characterization generation module unit is used to extract features from the conventional source load data and the new energy source load data to generate a hierarchical characterization of the source load trend and fluctuation of the distribution network in the current time period.

[0053] The feature extraction unit is used to perform dual-channel feature extraction on the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period and the conventional fluctuation weight and new energy fluctuation weight corresponding to the conventional equipment and the new energy equipment, respectively.

[0054] The first coefficient determination unit is used to determine the dynamic sensitivity coefficient of each execution node based on the source-load coupling scenario category, combined with the equipment type in the node information and the preset completion standard of the corresponding project of the power distribution network;

[0055] The second coefficient determination unit is used to determine the dynamic correction coefficient of each execution node based on the dynamic sensitivity coefficient, combined with the conventional fluctuation weight and the new energy fluctuation weight.

[0056] The comprehensive evaluation unit is used to perform asymmetric correction on the initial process score according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node, and to combine the final process scores of all execution nodes to obtain the execution process evaluation result of the corresponding project of the distribution network.

[0057] Thirdly, the electronic device of the present invention includes: a processor and a memory, the memory being used to store a computer program;

[0058] When the computer program is loaded by the processor, it causes the processor to execute the aforementioned method for evaluating the progress of power distribution network projects.

[0059] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the above-described method for evaluating the progress of a power distribution network project.

[0060] The present invention relates to a method, system, equipment, and medium for evaluating the execution progress of distribution network projects. First, it integrates the operational data of the distribution network into a multi-dimensional vector, overcoming the limitations of traditional methods that rely solely on single source-load data (such as power). This multi-dimensional data fusion approach comprehensively reflects the actual operating status of different devices in the distribution network, providing more accurate foundational data for calculating the initial progress score and significantly improving the comprehensiveness and accuracy of the evaluation. Next, feature extraction is performed on the source-load data to generate a layered representation of source-load trends and fluctuations, decomposing the source-load data into trend and fluctuation components. This process accurately identifies fluctuations and trends in the source-load data, avoiding misjudging normal source-load fluctuations as non-compliance of execution nodes. For example, for short-term fluctuations in photovoltaic power generation (such as those caused by weather changes), the layered representation clearly identifies these as normal source-load fluctuations rather than problems with execution nodes, thereby effectively reducing evaluation bias and improving the reliability of the evaluation results.

[0061] Furthermore, this invention not only considers the fluctuation characteristics of the equipment itself, but also analyzes the coupling relationships between different devices through dual-channel feature extraction. In the power distribution network, the operating states of different devices affect each other; for example, the connection of electric vehicle charging piles may affect the voltage stability of the power grid. Through dual-channel feature extraction and source-load coupling scenario classification, the evaluation process can dynamically adapt to different scenarios and adjust the evaluation logic according to different scenarios, thereby more accurately reflecting the actual operating state of the execution nodes and further improving the adaptability and accuracy of the evaluation. Based on the source-load coupling scenario category, combined with the equipment type and preset completion standard in the node information, the dynamic sensitivity coefficient of each execution node is determined, and the dynamic correction coefficient is determined based on the dynamic sensitivity coefficient and the fluctuation weight. This dynamic correction mechanism can make personalized corrections according to the specific situation of each execution node (such as equipment type and scenario), avoiding the one-size-fits-all problem of traditional static evaluation methods. For example, for a node sensitive to the fluctuation of new energy source-load, its new energy fluctuation weight is large, and the dynamic correction coefficient will be adjusted according to this weight, thereby more reasonably correcting its process score. Through dynamic correction, the process score of each node can more accurately reflect its actual execution status, thereby improving the accuracy and reliability of the overall evaluation results. Finally, by comprehensively considering the final progress scores of all execution nodes, the execution status of each node is taken into account, avoiding excessive impact of abnormal situations at individual nodes on the overall evaluation results. This comprehensive evaluation process ensures the reliability of the overall evaluation results, avoids misjudgments caused by local deviations, and provides more accurate data support for distribution network management and decision-making.

[0062] In summary, this invention comprehensively improves the accuracy and reliability of power distribution network project execution progress assessment through innovative technologies such as multi-dimensional data fusion, accurate source-load fluctuation identification, equipment coupling relationship analysis, dynamic correction mechanism, and comprehensive evaluation, providing strong technical support for the efficient management and optimization of power distribution networks. Attached Figure Description

[0063] Figure 1 This is a flowchart illustrating the power distribution network project execution progress evaluation method according to an embodiment of the present invention.

[0064] Figure 2 This is a schematic diagram of the structure of the power distribution network project execution process evaluation system according to an embodiment of the present invention. Detailed Implementation

[0065] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present invention. It should be understood that the accompanying drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0066] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0067] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to"; the term "based on" means "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; and the term "optionally" means "optional embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first," "second," etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0068] It should be noted that the terms "one" and "more" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0069] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, data stored, data displayed, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties. The collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation portals are provided for users to choose to authorize or refuse.

[0070] Combination Figure 1 As shown, this embodiment of the invention provides a method for evaluating the execution progress of a power distribution network project, including:

[0071] The system acquires the current time period's operational data of the distribution network and the node information of each execution node of the distribution network during the current time period. The operational data includes the conventional source-load data of conventional equipment and the new energy source-load data of new energy equipment in the distribution network.

[0072] Specifically, the real-time monitoring system of the distribution network collects conventional source-load data (including parameters such as power, current, and voltage) from all conventional equipment (such as transformers and lines) in the distribution network during the current time period, and renewable energy source-load data (such as power generation and charging power) from renewable energy equipment (such as distributed photovoltaic systems and electric vehicle charging piles). Simultaneously, it obtains node information for each execution node during the current time period from the distribution network management system, including equipment type, equipment parameters, and the preset completion standards for the corresponding node. This data can be transmitted via a communication network to the evaluation system for storage and processing.

[0073] Based on the conventional source-load data and the new energy source-load data, a multi-dimensional vector of the operating data is determined, and based on the multi-dimensional vector, an initial process score for each execution node is obtained.

[0074] Specifically, the collected conventional source-load data and new energy source-load data are normalized to eliminate the influence of different data units and magnitudes. Then, based on a preset feature extraction algorithm (such as Principal Component Analysis (PCA), these data are integrated into a multi-dimensional vector. Each dimension represents a key operating parameter, such as power, current, and voltage. Based on the multi-dimensional vector, a preset scoring model (such as a linear regression model or a machine learning model) is used to calculate the initial process score for each execution node. The scoring model comprehensively evaluates whether the operating status of the equipment meets the preset process objectives based on the feature values ​​of the multi-dimensional vector, thereby obtaining the initial process score.

[0075] Feature extraction is performed on the conventional source-load data and the new energy source-load data to generate a hierarchical representation of the source-load trend and fluctuation of the distribution network in the current time period.

[0076] Specifically, time series analysis methods (such as wavelet transform or sliding window averaging) are used to extract features from conventional and new energy source-load data. The data is decomposed into trend and fluctuation components. The trend component is extracted using methods such as moving average or multinomial fitting to reflect the long-term trend of equipment operation; the fluctuation component is obtained by calculating the difference between the data and the trend component, reflecting the short-term fluctuations in equipment operation. The resulting source-load trend-fluctuation hierarchical representation clearly shows the operating trend and fluctuation characteristics of each device in the current time period.

[0077] Dual-channel feature extraction is performed on the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period, as well as the conventional fluctuation weight and new energy fluctuation weight corresponding to the conventional equipment and the new energy equipment, respectively.

[0078] Specifically, a dual-channel feature extraction process is performed on the generated source-load trend-fluctuation hierarchical representation. One channel extracts features from conventional equipment, and the other channel extracts features from renewable energy equipment. Using clustering analysis or classification algorithms (such as K-means clustering or decision tree classification), the operating state of the distribution network is divided into different source-load coupling scenario categories. Each scenario category corresponds to a specific equipment operating mode, such as a high photovoltaic power generation and low load demand scenario or a high charging pile utilization and voltage fluctuation scenario. Simultaneously, based on each scenario category, statistical analysis methods (such as analysis of variance) are used to calculate the conventional fluctuation weights and renewable energy fluctuation weights for conventional equipment and renewable energy equipment, respectively, reflecting the degree of fluctuation impact of different equipment in the current scenario.

[0079] Based on the source-load coupling scenario category, and combined with the equipment type in the node information and the preset completion standard of the corresponding project of the power distribution network, the dynamic sensitivity coefficient of each execution node is determined.

[0080] Specifically, based on the determined source-load coupling scenario category, and combined with the device type and preset completion standards of each execution node, the dynamic sensitivity coefficient is calculated by querying a preset sensitivity coefficient table or using a machine learning model (such as a support vector machine or neural network). The sensitivity coefficient table is pre-established based on historical data and expert experience, containing sensitivity values ​​for different device types in different scenarios. The machine learning model automatically calculates the dynamic sensitivity coefficient for each node by learning the relationship between device operating status and sensitivity in historical data. The dynamic sensitivity coefficient reflects the sensitivity of each execution node to source-load fluctuations, providing a basis for subsequent dynamic correction.

[0081] Based on the dynamic sensitivity coefficient, and in combination with the conventional fluctuation weight and the new energy fluctuation weight, the dynamic correction coefficient of each execution node is determined.

[0082] Specifically, based on the dynamic sensitivity coefficient of each execution node, combined with the conventional fluctuation weight and the new energy fluctuation weight, a dynamic correction coefficient is calculated using mathematical formulas or optimization algorithms. For example, a weighted sum can be used, multiplying the sensitivity coefficient by the fluctuation weight and summing the results to obtain the dynamic correction coefficient. The dynamic correction coefficient is used to adjust the initial process score to better reflect actual operating conditions. For nodes sensitive to new energy fluctuations, their new energy fluctuation weight is larger, and the dynamic correction coefficient will be increased accordingly, thus more reasonably correcting their process score.

[0083] The initial process score is asymmetrically corrected according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node. The execution process evaluation result of the corresponding project of the distribution network is obtained by combining the final process scores of all execution nodes.

[0084] Specifically, the initial process score is asymmetrically corrected based on the dynamic correction coefficient of each execution node. Asymmetric correction means using different correction strategies for deviations in different directions (such as process lead or lag). For example, for nodes lagging behind, the correction magnitude can be larger to more significantly reflect their failure to meet the target; while for nodes leading ahead, the correction magnitude can be relatively smaller. The corrected process score is the final process score for each execution node. Finally, the final process scores of all execution nodes are weighted and averaged or comprehensively analyzed to obtain the overall execution process evaluation result of the distribution network. The evaluation results can be expressed as a percentage, grade, or other form, providing an intuitive reference for the management and decision-making of the distribution network.

[0085] The present invention relates to a method, system, equipment, and medium for evaluating the execution progress of distribution network projects. First, it integrates the operational data of the distribution network into a multi-dimensional vector, overcoming the limitations of traditional methods that rely solely on single source-load data (such as power). This multi-dimensional data fusion approach comprehensively reflects the actual operating status of different devices in the distribution network, providing more accurate foundational data for calculating the initial progress score and significantly improving the comprehensiveness and accuracy of the evaluation. Next, feature extraction is performed on the source-load data to generate a layered representation of source-load trends and fluctuations, decomposing the source-load data into trend and fluctuation components. This process accurately identifies fluctuations and trends in the source-load data, avoiding misjudging normal source-load fluctuations as non-compliance of execution nodes. For example, for short-term fluctuations in photovoltaic power generation (such as those caused by weather changes), the layered representation clearly identifies these as normal source-load fluctuations rather than problems with execution nodes, thereby effectively reducing evaluation bias and improving the reliability of the evaluation results.

[0086] Furthermore, this invention not only considers the fluctuation characteristics of the equipment itself, but also analyzes the coupling relationships between different devices through dual-channel feature extraction. In the power distribution network, the operating states of different devices affect each other; for example, the connection of electric vehicle charging piles may affect the voltage stability of the power grid. Through dual-channel feature extraction and source-load coupling scenario classification, the evaluation process can dynamically adapt to different scenarios and adjust the evaluation logic according to different scenarios, thereby more accurately reflecting the actual operating state of the execution nodes and further improving the adaptability and accuracy of the evaluation. Based on the source-load coupling scenario category, combined with the equipment type and preset completion standard in the node information, the dynamic sensitivity coefficient of each execution node is determined, and the dynamic correction coefficient is determined based on the dynamic sensitivity coefficient and the fluctuation weight. This dynamic correction mechanism can make personalized corrections according to the specific situation of each execution node (such as equipment type and scenario), avoiding the one-size-fits-all problem of traditional static evaluation methods. For example, for a node sensitive to the fluctuation of new energy source-load, its new energy fluctuation weight is large, and the dynamic correction coefficient will be adjusted according to this weight, thereby more reasonably correcting its process score. Through dynamic correction, the process score of each node can more accurately reflect its actual execution status, thereby improving the accuracy and reliability of the overall evaluation results. Finally, by comprehensively considering the final progress scores of all execution nodes, the execution status of each node is taken into account, avoiding excessive impact of abnormal situations at individual nodes on the overall evaluation results. This comprehensive evaluation process ensures the reliability of the overall evaluation results, avoids misjudgments caused by local deviations, and provides more accurate data support for distribution network management and decision-making.

[0087] In summary, this invention comprehensively improves the accuracy and reliability of power distribution network project execution progress assessment through innovative technologies such as multi-dimensional data fusion, accurate source-load fluctuation identification, equipment coupling relationship analysis, dynamic correction mechanism, and comprehensive evaluation, providing strong technical support for the efficient management and optimization of power distribution networks.

[0088] Optionally, determining a multi-dimensional vector of the operating data based on the conventional source-load data and the new energy source-load data, and obtaining an initial process score for each execution node based on the multi-dimensional vector, includes:

[0089] According to the preset window size, the conventional source load data and the new energy source load data are extracted by sliding to obtain multiple window features;

[0090] Based on the window features, the multi-dimensional vector corresponding to each execution node is obtained;

[0091] The multi-dimensional vector and the device type in the node information are coupled with the preset completion standard input of the project to form a causal convolutional network to obtain the state curve corresponding to the execution node.

[0092] The initial process score is obtained by performing residual integration and Sigmoid mapping based on the state curve.

[0093] Specifically, first, a window size is set, determined based on the data's temporal resolution and evaluation requirements. For example, if the data is recorded in minutes, the window size can be set to 15 minutes. Then, sliding extraction is performed from the time series of conventional source-load data and new energy source-load data according to the set window size. That is, starting from the beginning of the data series, each time unit is moved forward (e.g., 1 minute), and a data segment of length equal to the window size is extracted, forming a feature window. In this way, multiple feature windows can be obtained, each containing source-load data over a period of time, reflecting the operating status of the equipment in different time periods. For each feature window, key features are extracted. These features include average power, power standard deviation, peak current, and voltage fluctuation range. These features are used as dimensions to construct a vector. For example, a feature window may correspond to a 5-dimensional vector, where the first dimension is average power, the second dimension is the power standard deviation, the third dimension is the peak current, the fourth dimension is the voltage fluctuation range, and the fifth dimension is the equipment's operating time. In this way, each execution node has a corresponding multi-dimensional vector for different time periods (i.e., different feature windows), and these vectors can comprehensively reflect the node's operating state within that time period. A Coupled Causal Convolutional Network (CCCN) is constructed. This network contains multiple convolutional layers to extract spatiotemporal features from the multi-dimensional vector. Simultaneously, the device type and the preset completion standard corresponding to the project in the node information are used as additional input features, which are transformed into vectors of the same dimension as the multi-dimensional vector through an embedding layer, and then concatenated with the multi-dimensional vector. Causal convolution is used in the convolutional layers to ensure that the network can only use information from the current and previous time steps when predicting the current state, avoiding information leakage. Through multiple convolutional operations, the network can learn the state change patterns of execution nodes in different time periods, ultimately outputting a state curve for each execution node. The state curve is a time series reflecting the changes in the node's operating state within the evaluation time period. Residual integration is then performed on the state curve. Residual integration refers to calculating the difference between the value of the state curve at each time point and the preset completion standard, and then integrating these differences. The integral result reflects the cumulative deviation of a node from the completion standard throughout the entire evaluation period. Next, the integral result is mapped using the Sigmoid function, which maps the integral result to a normalized process score between 0 and 1. This score represents the progress of the node in the current time period; a score close to 1 indicates good progress, while a score close to 0 indicates lagging progress. In this way, an initial process score is obtained for each execution node, providing a basis for subsequent dynamic correction and comprehensive evaluation.

[0094] In this embodiment of the invention, by using a sliding window to extract source payload data features, local information within different time periods can be captured, enabling refined analysis of node operating status. By combining multi-dimensional vectors with device type and preset standards as input to a coupled causal convolutional network, state curves are generated, achieving dynamic adaptation to different devices and scenarios in the evaluation. At the same time, causal convolution ensures the logic and accuracy of the evaluation. By obtaining a normalized initial process score through residual integrals and sigmoid mapping, not only is the deviation of the node from the preset standard quantified, but the interpretation of the evaluation results is also simplified, providing a unified benchmark for subsequent correction and comprehensive evaluation.

[0095] Optionally, the step of extracting features from the conventional source-load data and the new energy source-load data to generate a hierarchical representation of the source-load trend and fluctuation of the distribution network in the current time period includes:

[0096] The conventional source load data and the new energy source load data are subjected to adaptive denoising processing by variational mode decomposition algorithm to obtain the effective sequence of conventional source load and the effective sequence of new energy source load.

[0097] Based on a preset multi-scale decomposition strategy, ensemble empirical mode decomposition is performed on the effective sequence of conventional source load and the effective sequence of new energy source load respectively to obtain the trend component and multi-order fluctuation component corresponding to the effective sequence of conventional source load and the effective sequence of new energy source load respectively.

[0098] The trend components and multi-order fluctuation components of the conventional source-load data are dimensionally aligned with the trend components and multi-order fluctuation components of the new energy source-load data to construct a four-dimensional hierarchical structure including a conventional trend layer, a conventional fluctuation layer, a new energy trend layer, and a new energy fluctuation layer, thereby generating a hierarchical representation of the source-load trend-fluctuation of the distribution network in the current time period.

[0099] Specifically, firstly, Variational Mode Decomposition (VMD) is applied to both conventional and new energy source-load data to decompose the complex signal into multiple modal components, each representing a specific frequency range. During denoising, VMD optimizes a variational problem to decompose the signal into multiple modal components, adaptively adjusting based on the center frequency and bandwidth of each component. By setting appropriate parameters (such as the number of modal components and the noise threshold), the noise component in the signal is separated, resulting in effective sequences of conventional and new energy source-load data. These effective sequences, freed from noise interference, retain the main features of the signal, providing a clearer data foundation for subsequent feature extraction. Then, Ensemble Empirical Mode Decomposition (EEMD) is applied to both the denoised effective sequences of conventional and new energy source-load data. EEMD is an improved Empirical Mode Decomposition (EMD) method that effectively reduces mode aliasing and improves the stability and accuracy of decomposition by adding white noise to the original signal and averaging the results after multiple decompositions. The pre-defined multi-scale decomposition strategy determines the parameter settings of EEMD, such as the amplitude of white noise and the number of decompositions. EEMD decomposes each effective sequence into trend components and multi-order fluctuation components. The trend component reflects the long-term trend of the signal, while the multi-order fluctuation components capture the short-term fluctuations of the signal at different time scales. This multi-scale decomposition can more comprehensively reveal the dynamic characteristics of source-load data, providing detailed feature information for subsequent hierarchical characterization. Dimensional alignment is performed on the trend and multi-order fluctuation components of conventional source-load data and new energy source-load data. That is, trend and fluctuation components from different sources are aligned to the same time axis to ensure temporal consistency. Then, these components are combined into a four-dimensional hierarchical structure, including a conventional trend layer, a conventional fluctuation layer, a new energy trend layer, and a new energy fluctuation layer. Each layer stores the corresponding type of trend or fluctuation information. This hierarchical structure clearly displays the source-load trends and fluctuations of the distribution network in the current time period, providing structured data support for subsequent analysis and evaluation. This hierarchical characterization can effectively distinguish different types of source-load data and their dynamic characteristics, facilitating further feature analysis and evaluation logic design.

[0100] In this embodiment of the invention, a variational mode decomposition algorithm is used to adaptively denoise conventional source-load data and new energy source-load data, effectively separating the noise component in the signal and obtaining clear effective sequences of conventional and new energy source-load data, providing a high-quality data foundation for subsequent analysis. Furthermore, based on a preset multi-scale decomposition strategy, an ensemble empirical mode decomposition algorithm is used to decompose these effective sequences, obtaining trend components and multi-order fluctuation components. This multi-scale decomposition can comprehensively reveal the dynamic characteristics of source-load data, effectively reducing mode aliasing and improving the stability and accuracy of the decomposition. Finally, by dimensionally aligning the trend and fluctuation components, a four-dimensional hierarchical structure is constructed, including a conventional trend layer, a conventional fluctuation layer, a new energy trend layer, and a new energy fluctuation layer, generating a source-load trend-fluctuation hierarchical representation. This hierarchical representation clearly shows the source-load trend and fluctuation of the distribution network in the current time period, providing structured data support for subsequent evaluation and analysis, significantly improving the accuracy and reliability of the evaluation, effectively distinguishing different types of source-load data and their dynamic characteristics, facilitating further feature analysis and evaluation logic design.

[0101] Optionally, the dual-channel feature extraction of the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period and the conventional fluctuation weights and new energy fluctuation weights corresponding to the conventional equipment and the new energy equipment, respectively, includes:

[0102] Input the conventional trend layer and the conventional fluctuation layer in the source load trend-fluctuation hierarchical representation into the conventional source load feature extraction channel to obtain the amplitude change features of the conventional trend layer and the fluctuation features of the conventional fluctuation layer.

[0103] The new energy trend layer and the new energy fluctuation layer are input into the new energy source load feature extraction channel to obtain the steady-state features of the new energy trend layer and the abrupt change features of the new energy fluctuation layer;

[0104] The amplitude variation feature and the fluctuation feature are normalized and fused to obtain a conventional source-load comprehensive feature vector;

[0105] By combining feature concatenation and principal component analysis, the steady-state features and the abrupt change features are simplified in dimensionality to obtain a comprehensive feature vector of new energy source and load.

[0106] Based on the conventional source-load integrated feature vector and the new energy source-load integrated feature vector, the source-load coupling scenario category is obtained by prediction.

[0107] The coupling correlation degree and the cooperative fluctuation coefficient of the conventional source-load integrated feature vector and the new energy source-load integrated feature vector are obtained, and the conventional fluctuation weight and the new energy fluctuation weight are determined by the entropy method based on the coupling correlation degree and the cooperative fluctuation coefficient.

[0108] Specifically, the conventional trend layer and conventional fluctuation layer in the source-load trend-fluctuation hierarchical representation are input into the conventional source-load feature extraction channel. The amplitude variation characteristics of the conventional trend layer are extracted using time-domain analysis methods (such as calculating maximum, minimum, and mean values) and frequency-domain analysis methods (such as Fourier transform to calculate frequency distribution). For the conventional fluctuation layer, statistical analysis methods (such as calculating standard deviation and variance) and wavelet transform (used to capture local fluctuation features) are employed to extract fluctuation features. These features reflect the operational status changes of the conventional source-load data within the current time period. The new energy trend layer and new energy fluctuation layer are input into the new energy source-load feature extraction channel. For the new energy trend layer, smoothing processing (such as moving average) and stability analysis (such as calculating fluctuation coefficient) are used to extract steady-state features, reflecting the stability of new energy equipment in long-term operation. For the new energy fluctuation layer, abrupt change features are extracted using abrupt change detection algorithms (such as Canny edge detection or abrupt change detection module in wavelet transform) to capture abrupt changes in the short-term operation of new energy equipment, such as rapid fluctuations in photovoltaic power generation or instantaneous power changes in electric vehicle charging piles. The extracted amplitude variation and fluctuation features are normalized to eliminate the influence of different feature dimensions and magnitudes. Normalization methods can include min-max normalization or Z-score normalization. Then, the normalized features are merged using feature fusion methods (such as weighted summation or feature concatenation) to obtain a comprehensive feature vector representing the overall operational characteristics of conventional source-load data within the current time period. The extracted steady-state and abrupt change features are concatenated to form a high-dimensional feature vector. Principal component analysis (PCA) is then used to reduce the dimensionality of this high-dimensional feature vector, extracting the principal components to obtain the comprehensive feature vector of new energy source-load data. PCA effectively reduces feature dimensionality by calculating the covariance matrix of the feature vector and extracting its principal components, while retaining key information and improving the efficiency and accuracy of subsequent processing. The comprehensive feature vectors of conventional and new energy source-load data are then input into a pre-trained classification model (such as Support Vector Machine (SVM) or a deep learning classification network). Based on the learned mapping relationship between features and scene categories, the classification model predicts the source-load coupling scene category within the current time period. For example, the model can identify scenario categories such as high photovoltaic power generation and low load demand, or high charging pile utilization and voltage fluctuations, providing a scenario basis for subsequent dynamic correction. The coupling correlation degree (e.g., calculated through correlation coefficient) and the cooperative fluctuation coefficient (e.g., calculated through covariance) between the comprehensive feature vectors of conventional and renewable energy sources are calculated. The coupling correlation degree reflects the correlation between the two vectors, while the cooperative fluctuation coefficient reflects their synergy in fluctuation. Then, these coefficients are comprehensively analyzed using the entropy method to determine the weights for conventional and renewable energy fluctuations. The entropy method assesses the importance of each feature in decision-making by calculating its entropy value, thereby rationally allocating weights and ensuring the scientific and reasonable nature of the weight allocation.

[0109] In a preferred embodiment of the present invention, a Support Vector Machine (SVM) is used as a classification model to determine the source-load coupling scenario category of the distribution network in the current time period. Specifically, the model structure is based on the traditional SVM framework, employing linear or nonlinear kernel functions (such as radial basis functions, RBF) to handle the complex relationship between feature vectors and scenario categories. During the training phase, historical data containing both conventional and new energy source-load integrated feature vectors are collected, and the corresponding source-load coupling scenario categories are labeled to construct a training dataset. The SVM model is trained using an optimization algorithm (such as Sequential Minimum Optimization, SMO). The model learns the mapping relationship between feature vectors and scenario categories and finds the optimal separating hyperplane to maximize the margin between different categories. In the application phase, the integrated feature vector of the current time period is input into the trained SVM model. The model quickly and accurately determines the source-load coupling scenario category to which the current feature vector belongs by calculating the distance between the input feature vector and the separating hyperplane. For example, if the input feature vector is determined to belong to a high charging pile utilization and voltage fluctuation scenario, this result will serve as the basis for subsequent dynamic correction, thereby achieving accurate evaluation of the execution process under different scenarios and significantly improving the adaptability and accuracy of the evaluation.

[0110] In this embodiment of the invention, dual-channel feature extraction and comprehensive analysis significantly improve the accuracy and reliability of power distribution network source-load coupling scenario identification and weight determination. Specifically, feature extraction is performed on the trend and fluctuation layers of conventional source loads and new energy source loads respectively, obtaining amplitude change, fluctuation, steady state, and abrupt change features. These features comprehensively reflect the operating status of different devices. Furthermore, through normalization processing and feature fusion, comprehensive feature vectors of conventional source loads and new energy source loads are generated, eliminating the dimensional differences between different features and enhancing the comparability of features. Principal component analysis (PCA) is used to reduce the dimensionality of new energy features, simplifying the feature space and improving computational efficiency. Based on these comprehensive feature vectors, prediction is performed to accurately identify the source-load coupling scenario category in the current time period, providing a scientific basis for subsequent dynamic correction. Furthermore, by combining the entropy method with coupling correlation and cooperative fluctuation coefficient, the weights of conventional fluctuations and new energy fluctuations are dynamically determined. This data-driven weight allocation method fully considers the actual impact of different devices in the current scenario, avoids the irrationality of subjectively setting weights, and makes the weight allocation more scientific and objective, thereby further improving the accuracy and adaptability of the performance evaluation of power distribution network projects.

[0111] Optionally, determining the dynamic sensitivity coefficient of each execution node based on the source-load coupling scenario category, combined with the device type in the node information and the preset completion standard of the corresponding project in the distribution network, includes:

[0112] Based on the source-load coupling scenario category, a sensitivity benchmark matrix is ​​generated;

[0113] Based on the device type and preset completion standard in the node information of each execution node, the baseline sensitivity coefficient of each execution node is obtained by matching in the sensitivity baseline matrix;

[0114] Based on the operating years, construction complexity level, and historical progress deviation rate of the associated equipment of the execution node, the additional impact characteristics of the execution node are obtained;

[0115] Based on the additional impact characteristics, the sensitivity correction factor for each execution node is obtained;

[0116] The dynamic sensitivity coefficient of each execution node is determined based on the baseline sensitivity coefficient and the sensitivity correction factor of the execution node.

[0117] Specifically, based on the determined source-load coupling scenario categories, a sensitivity benchmark matrix is ​​pre-constructed. This matrix is ​​a multi-dimensional table where each row represents a source-load coupling scenario category, and each column represents the benchmark sensitivity coefficient for different equipment types (such as transformers, distributed photovoltaics, and electric vehicle charging piles) within that scenario. The benchmark sensitivity coefficients are pre-set based on historical data and expert experience, reflecting the sensitivity of different equipment types to source-load fluctuations in a specific scenario. For example, in a scenario of high photovoltaic power generation and low load demand, the benchmark sensitivity coefficient for distributed photovoltaic equipment may be higher, while that for transformers may be lower. For each execution node, the corresponding benchmark sensitivity coefficient is searched in the sensitivity benchmark matrix based on the equipment type and preset completion standard in its node information. The equipment type in the node information directly determines the column searched in the matrix, and the preset completion standard affects the specific search position; for example, nodes with high completion standards may use a higher sensitivity coefficient. In this way, a benchmark sensitivity coefficient is matched for each execution node, serving as the basis for subsequent dynamic sensitivity coefficient calculations.

[0118] Information such as the service life, construction complexity level, and historical progress deviation rate of the associated equipment at each execution node is collected. Service life can be calculated from the equipment's installation time; the construction complexity level is assessed based on the equipment's installation and maintenance records, typically categorized as low, medium, or high; and the historical progress deviation rate is calculated by analyzing the node's progress data over a past period. This information collectively constitutes the additional impact characteristics, reflecting the actual operating status and historical performance of the equipment, providing a basis for adjusting the sensitivity coefficient. Based on the additional impact characteristics, a sensitivity correction factor for each execution node is calculated using pre-defined correction rules or models. Correction rules can be based on empirical formulas or machine learning models. For example, equipment with a long service life may be given a higher correction factor to reflect the impact of aging on sensitivity; equipment with a high construction complexity level may be given a lower correction factor to reflect the impact of maintenance difficulty on sensitivity; and nodes with a high historical progress deviation rate may be given a higher correction factor to reflect the impact of process instability on sensitivity. These rules or models transform the additional impact characteristics into specific correction factors. The dynamic sensitivity coefficient is calculated by combining the baseline sensitivity coefficient of each execution node with a sensitivity correction factor. The combination method can be simple multiplication or weighted summation, depending on the design of the correction factor. For example, the dynamic sensitivity coefficient can be expressed as the baseline sensitivity coefficient multiplied by the correction factor, or the baseline sensitivity coefficient plus a weighted value of the correction factor. The final dynamic sensitivity coefficient comprehensively reflects the sensitivity of equipment type, operating status, and historical performance to source load fluctuations, providing a personalized adjustment basis for subsequent dynamic correction.

[0119] In this embodiment of the invention, a sensitivity benchmark matrix is ​​generated to provide benchmark sensitivity coefficients for equipment types in different scenarios, ensuring the basic scientific nature and adaptability of the assessment. Secondly, by matching the benchmark sensitivity coefficients with node information, the assessment's targeting is further refined, making the sensitivity assessment of each node more closely aligned with its actual equipment type and completion standards. Furthermore, by considering additional influencing factors such as the operating years of associated equipment, construction complexity level, and historical progress deviation rate, and calculating sensitivity correction factors accordingly, the actual operating conditions and historical performance of the equipment are fully considered, making the adjustment of sensitivity coefficients more reasonable and comprehensive. Finally, the benchmark sensitivity coefficients are combined with the correction factors to determine the dynamic sensitivity coefficients, achieving personalized sensitivity assessment for each execution node. This method of calculating dynamic sensitivity coefficients not only improves the accuracy of the assessment but also enhances the adaptability and reliability of the assessment results, providing strong technical support for the refined management and optimization decision-making of power distribution networks.

[0120] Optionally, determining the dynamic correction coefficient for each execution node based on the dynamic sensitivity coefficient, combined with the conventional fluctuation weight and the new energy fluctuation weight, includes:

[0121] Based on the device type in the node information of each execution node, a target fluctuation weight is determined for the execution node, wherein the target fluctuation weight is the conventional fluctuation weight or the new energy fluctuation weight;

[0122] Based on the target fluctuation weight and combined with the preset cross-device type fluctuation correlation coefficient, the comprehensive fluctuation weight of the execution node is obtained.

[0123] The initial correction coefficient of each execution node is obtained by weighting the dynamic sensitivity coefficient and the comprehensive fluctuation weight of each execution node.

[0124] Based on the source-load coupling scenario category, a scenario adaptation correction coefficient is determined, and combined with the initial correction coefficient, the dynamic correction coefficient for each execution node is obtained.

[0125] Specifically, based on the equipment type in the node information of each execution node, it is determined whether the node mainly involves conventional equipment or new energy equipment. For example, if a node mainly contains transformers and transmission lines, its equipment type is conventional equipment; if a node mainly contains distributed photovoltaics and electric vehicle charging piles, its equipment type is new energy equipment. Based on this determination, a corresponding target fluctuation weight is matched for each execution node, i.e., conventional fluctuation weight or new energy fluctuation weight. This process ensures that the subsequent calculation of correction coefficients can be accurately adjusted for the actual equipment type of the node. Combined with a preset cross-equipment type fluctuation correlation coefficient, the target fluctuation weight is adjusted to obtain the comprehensive fluctuation weight. The cross-equipment type fluctuation correlation coefficient is a preset coefficient used to reflect the degree of fluctuation correlation between different equipment types. For example, if the fluctuation correlation between conventional equipment and new energy equipment is strong, the coefficient value is high; otherwise, it is low. By multiplying the target fluctuation weight by the cross-equipment type fluctuation correlation coefficient, the comprehensive fluctuation weight can be obtained. This weight comprehensively considers the fluctuation characteristics of the equipment type within the node and its correlation with other equipment types. A weighted calculation is performed based on the dynamic sensitivity coefficient and the comprehensive fluctuation weight of each execution node. Weighted calculations can be performed using simple multiplication or weighted summation. For example, the initial correction coefficient can be expressed as the product of the dynamic sensitivity coefficient and the comprehensive fluctuation weight, or a weighted sum of the two. This process combines the node's sensitivity and fluctuation characteristics to obtain an initial correction coefficient that reflects the node's comprehensive correction requirements. The initial correction coefficient provides the basic value for subsequent dynamic correction. Based on the source-load coupling scenario category, a scenario adaptation correction coefficient is determined. This coefficient is pre-set according to the current source-load coupling scenario category and is used to reflect the adjustment requirements of the correction coefficient under different scenarios. For example, in a scenario with high photovoltaic power generation and low load demand, a higher scenario adaptation correction coefficient may be needed to reflect the impact of new energy equipment fluctuations on the correction coefficient under this scenario. Then, the scenario adaptation correction coefficient is combined with the initial correction coefficient to obtain the final dynamic correction coefficient. The combination method can be multiplication or weighted summation, depending on the design of the correction coefficient. The final dynamic correction coefficient comprehensively reflects the node's sensitivity, fluctuation characteristics, and correction requirements under the current scenario, providing a precise adjustment basis for the dynamic correction of the execution process score.

[0126] In this embodiment of the invention, by determining the target fluctuation weight, the correction process is ensured to be precisely adjusted for the actual equipment type of each execution node, avoiding a one-size-fits-all approach. Secondly, by combining the cross-equipment type fluctuation correlation coefficient to calculate the comprehensive fluctuation weight, the fluctuation correlation between different equipment types is considered, making the correction more comprehensive and scientific. Next, the initial correction coefficient is obtained through a weighted calculation of the dynamic sensitivity coefficient and the comprehensive fluctuation weight, further combining the node's sensitivity characteristics and fluctuation characteristics, providing a more accurate basis for correction. Finally, the scenario adaptation correction coefficient is determined according to the source-load coupling scenario category, and combined with the initial correction coefficient to obtain the dynamic correction coefficient. This embodiment fully considers the impact of the current scenario on the correction requirements, enabling the correction coefficient to dynamically adapt to different operating states. This dynamic correction mechanism can effectively address the problems of diverse equipment types and complex operating states in the distribution network, significantly improving the accuracy and reliability of the execution process evaluation.

[0127] Optionally, the step of asymmetrically correcting the initial process score according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node, and then combining the final process scores of all execution nodes to obtain the execution process evaluation result of the corresponding project of the distribution network, includes:

[0128] Based on the dynamic correction coefficients of the execution node, determine the asymmetric correction rule of the execution node;

[0129] Based on the initial process score of each execution node and in conjunction with the asymmetric correction rule, a target correction formula corresponding to the execution node is generated.

[0130] The intermediate process score of the execution node is obtained through the target correction formula;

[0131] Based on the preset valid range of the distribution network execution process score, the intermediate process score is subjected to boundary constraint processing to obtain the final process score of each execution node;

[0132] A depth weight is generated according to the preset topology depth of each execution node;

[0133] The final process score is weighted and summed with the corresponding depth weight to obtain the weighted total process score.

[0134] The weighted process total score is mapped to a percentage range to obtain the execution process evaluation result of the distribution network in the current time period.

[0135] Specifically, based on the dynamic correction coefficient of each execution node, corresponding asymmetric correction rules are formulated. Asymmetric correction rules refer to adjusting the initial process score in different directions and to different degrees according to the node's dynamic correction coefficient. For example, if the dynamic correction coefficient is high, it indicates that the node is more sensitive to source load fluctuations and requires a larger adjustment to the process score; if the dynamic correction coefficient is low, the adjustment is smaller. Asymmetric correction rules can be implemented using piecewise functions or nonlinear functions to ensure more flexible and precise adjustments to different process scores. Combining the initial process score of each execution node and the asymmetric correction rules, a corresponding target correction formula is generated.

[0136] The target correction formula can be a mathematical expression, such as: Intermediate process score = Initial process score × (1 + Dynamic correction coefficient × f(Initial process score)); where f(Initial process score) is a function that adjusts the correction magnitude based on the initial process score, and can be linear or non-linear. This formula can dynamically adjust the correction magnitude based on the specific value of the initial process score to generate the intermediate process score for each execution node.

[0137] The intermediate process score for each execution node is calculated using a target correction formula. Substituting the initial process score into the target correction formula yields the intermediate process score after asymmetric correction. This intermediate process score reflects the node's progress after considering dynamic correction, providing a basis for subsequent boundary constraint processing. Boundary constraint processing is applied to the intermediate process score based on a preset valid range (e.g., 0 to 100) for the distribution network execution process score. If the intermediate process score exceeds the valid range, it is adjusted to the boundary value. For example, if the intermediate process score is less than 0, it is adjusted to 0; if it is greater than 100, it is adjusted to 100. Boundary constraint processing ensures that the final process score is within a reasonable range, avoiding unreasonable scores caused by correction. A depth weight is generated based on the preset topology depth of each execution node. Topology depth reflects the node's position and importance in the distribution network topology. For example, nodes closer to power sources or critical loads may have a higher topology depth and are therefore assigned a higher depth weight.

[0138] Depth weights can be calculated using preset weight allocation rules, such as: Weighted Depth = Topology Depth / Sum of Topology Depths of All Nodes. This weight allocation method ensures that the weight of different nodes in the comprehensive evaluation matches their importance in the distribution network. The final process score of each executing node is weighted and summed with its corresponding depth weight to obtain the weighted total process score. Specifically: Weighted Total Process Score = Sum of the products of each node's final process score and its corresponding node depth weight.

[0139] This method comprehensively considers the process score of each node and its importance in the distribution network to obtain a weighted total score reflecting the overall execution process of the distribution network. The weighted total process score is then mapped to a percentage range (0 to 100) to obtain the performance evaluation result of the distribution network in the current time period. This mapping can be implemented using linear or nonlinear functions, for example: Performance evaluation result = (Weighted total process score - Minimum possible total score) / (Maximum possible total score - Minimum possible total score) × 100. This mapping method ensures that the evaluation results have a unified measurement standard, making them easy for management and decision-makers to understand and use.

[0140] In this embodiment of the invention, an asymmetric correction rule is determined based on dynamic correction coefficients, and a target correction formula is generated to flexibly adjust the initial process score. This asymmetric correction can accurately correct the dynamic characteristics of different nodes, avoiding the irrationality that may arise from traditional symmetric correction. Next, boundary constraint processing ensures that the intermediate process score is within a reasonable range, avoiding outliers caused by correction and enhancing the rationality of the evaluation results. Furthermore, depth weights are generated based on the topology depth of the nodes and weighted summation is performed, enabling the evaluation results to comprehensively reflect the importance of different nodes in the distribution network, improving the comprehensiveness and scientific rigor of the evaluation results. Finally, the weighted process total score is mapped to a percentage range to obtain a unified execution process evaluation result, facilitating intuitive understanding and use by management and decision-makers. This series of technical means works together to achieve a refined evaluation of the distribution network execution process, providing strong technical support for efficient management and optimization decision-making in the distribution network.

[0141] Combination Figure 2 As shown, the present invention provides a power distribution network project execution progress evaluation system, comprising:

[0142] The data acquisition unit is used to acquire the operating data of the distribution network in the current time period and the node information of each execution node of the distribution network in the current time period. The operating data includes the conventional source-load data of conventional equipment and the new energy source-load data of new energy equipment in the distribution network.

[0143] The initial analysis unit is used to determine the multi-dimensional vector of the running data based on the conventional source load data and the new energy source load data, and to obtain the initial process score of each execution node based on the multi-dimensional vector.

[0144] The characterization generation module unit is used to extract features from the conventional source load data and the new energy source load data to generate a hierarchical characterization of the source load trend and fluctuation of the distribution network in the current time period.

[0145] The feature extraction unit is used to perform dual-channel feature extraction on the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period and the conventional fluctuation weight and new energy fluctuation weight corresponding to the conventional equipment and the new energy equipment, respectively.

[0146] The first coefficient determination unit is used to determine the dynamic sensitivity coefficient of each execution node based on the source-load coupling scenario category, combined with the equipment type in the node information and the preset completion standard of the corresponding project of the power distribution network;

[0147] The second coefficient determination unit is used to determine the dynamic correction coefficient of each execution node based on the dynamic sensitivity coefficient, combined with the conventional fluctuation weight and the new energy fluctuation weight.

[0148] The comprehensive evaluation unit is used to perform asymmetric correction on the initial process score according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node, and to combine the final process scores of all execution nodes to obtain the execution process evaluation result of the corresponding project of the distribution network.

[0149] The advantages of the power distribution network project execution progress evaluation system of the present invention compared with the prior art are the same as those of the above-mentioned power distribution network project execution progress evaluation method compared with the prior art, and will not be repeated here.

[0150] The electronic device of the present invention includes: a processor and a memory, wherein the memory is used to store a computer program;

[0151] When the computer program is loaded by the processor, it causes the processor to execute the aforementioned method for evaluating the progress of power distribution network projects.

[0152] The electronic device of the present invention has the same advantages over the prior art as the aforementioned method for evaluating the progress of power distribution network projects, and will not be repeated here.

[0153] The present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method for evaluating the progress of a power distribution network project.

[0154] The computer-readable storage medium of the present invention has the same advantages over the prior art as the aforementioned method for evaluating the progress of power distribution network projects, and will not be repeated here.

[0155] While the present invention has been disclosed above, its scope of protection is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and all such changes and modifications will fall within the scope of protection of the present invention.

Claims

1. A method for evaluating the progress of a power distribution network project, characterized in that, include: Obtain the current time period operation data of the distribution network and the node information of each execution node of the distribution network in the current time period. The operation data includes the conventional source-load data of conventional equipment and the new energy source-load data of new energy equipment in the distribution network. Based on the conventional source load data and the new energy source load data, a multi-dimensional vector of the operation data is determined, and based on the multi-dimensional vector, an initial process score for each execution node is obtained. Feature extraction is performed on the conventional source-load data and the new energy source-load data to generate a hierarchical representation of the source-load trend and fluctuation of the distribution network in the current time period; Dual-channel feature extraction is performed on the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period, as well as the conventional fluctuation weight and new energy fluctuation weight corresponding to the conventional equipment and the new energy equipment, respectively. Based on the source-load coupling scenario category, and combined with the equipment type in the node information and the preset completion standard of the corresponding project of the power distribution network, the dynamic sensitivity coefficient of each execution node is determined; Based on the dynamic sensitivity coefficient, and in combination with the conventional fluctuation weight and the new energy fluctuation weight, the dynamic correction coefficient of each execution node is determined; The initial process score is asymmetrically corrected according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node. The execution process evaluation result of the corresponding project of the distribution network is obtained by combining the final process scores of all execution nodes.

2. The method for evaluating the progress of power distribution network projects according to claim 1, characterized in that, The step of determining a multi-dimensional vector of the operational data based on the conventional source-load data and the new energy source-load data, and obtaining an initial process score for each execution node based on the multi-dimensional vector, includes: According to the preset window size, the conventional source load data and the new energy source load data are extracted by sliding to obtain multiple window features; Based on the window features, the multi-dimensional vector corresponding to each execution node is obtained; The multi-dimensional vector and the device type in the node information are coupled with the preset completion standard of the project input to a causal convolutional network to obtain the state curve corresponding to the execution node. The coupled causal convolutional network contains multiple convolutional layers and embedding layers. The device type and the preset completion standard in the node information are used as additional input features. The embedding layer converts the additional input features into a vector with the same dimension as the multi-dimensional vector, and then concatenates them with the multi-dimensional vector. Through multiple convolutional operations of the convolutional layers, the state change pattern of the execution node in different time periods is obtained, and finally the state curve corresponding to each execution node is output. The initial process score is obtained by performing residual integration and Sigmoid mapping based on the state curve.

3. The method for evaluating the progress of power distribution network projects according to claim 1, characterized in that, The step of extracting features from the conventional source-load data and the new energy source-load data to generate a hierarchical representation of the source-load trend and fluctuation of the distribution network in the current time period includes: The conventional source load data and the new energy source load data are subjected to adaptive denoising processing by variational mode decomposition algorithm to obtain the effective sequence of conventional source load and the effective sequence of new energy source load. Based on a preset multi-scale decomposition strategy, ensemble empirical mode decomposition is performed on the effective sequence of conventional source load and the effective sequence of new energy source load respectively to obtain the trend component and multi-order fluctuation component corresponding to the effective sequence of conventional source load and the effective sequence of new energy source load respectively. The trend components and multi-order fluctuation components of the conventional source-load data are dimensionally aligned with the trend components and multi-order fluctuation components of the new energy source-load data to construct a four-dimensional hierarchical structure including a conventional trend layer, a conventional fluctuation layer, a new energy trend layer, and a new energy fluctuation layer, thereby generating a hierarchical representation of the source-load trend-fluctuation of the distribution network in the current time period.

4. The method for evaluating the progress of power distribution network projects according to claim 3, characterized in that, The dual-channel feature extraction of the source-load trend-fluctuation hierarchical representation determines the source-load coupling scenario category of the distribution network in the current time period, as well as the conventional fluctuation weights and new energy fluctuation weights corresponding to the conventional equipment and the new energy equipment, respectively, including: Input the conventional trend layer and the conventional fluctuation layer in the source load trend-fluctuation hierarchical representation into the conventional source load feature extraction channel to obtain the amplitude change features of the conventional trend layer and the fluctuation features of the conventional fluctuation layer. The new energy trend layer and the new energy fluctuation layer are input into the new energy source load feature extraction channel to obtain the steady-state features of the new energy trend layer and the abrupt change features of the new energy fluctuation layer; The amplitude variation feature and the fluctuation feature are normalized and fused to obtain a conventional source-load comprehensive feature vector; By combining feature concatenation and principal component analysis, the steady-state features and the abrupt change features are simplified in dimensionality to obtain a comprehensive feature vector of new energy source and load. Based on the conventional source-load integrated feature vector and the new energy source-load integrated feature vector, the source-load coupling scenario category is obtained by prediction. The coupling correlation degree and the cooperative fluctuation coefficient of the conventional source-load integrated feature vector and the new energy source-load integrated feature vector are obtained, and the conventional fluctuation weight and the new energy fluctuation weight are determined by the entropy method based on the coupling correlation degree and the cooperative fluctuation coefficient.

5. The method for evaluating the progress of power distribution network projects according to claim 1, characterized in that, The dynamic sensitivity coefficient of each execution node is determined based on the source-load coupling scenario category, combined with the equipment type in the node information and the preset completion standard of the corresponding project in the distribution network, including: Based on the source-load coupling scenario category, a sensitivity benchmark matrix is ​​generated; Based on the device type and preset completion standard in the node information of each execution node, the baseline sensitivity coefficient of each execution node is obtained by matching in the sensitivity baseline matrix; Based on the operating years, construction complexity level, and historical progress deviation rate of the associated equipment of the execution node, the additional impact characteristics of the execution node are obtained; Based on the additional impact characteristics, the sensitivity correction factor for each execution node is obtained; The dynamic sensitivity coefficient of each execution node is determined based on the baseline sensitivity coefficient and the sensitivity correction factor of the execution node.

6. The method for evaluating the progress of power distribution network projects according to claim 1, characterized in that, The step of determining the dynamic correction coefficient for each execution node based on the dynamic sensitivity coefficient, combined with the conventional fluctuation weight and the new energy fluctuation weight, includes: Based on the device type in the node information of each execution node, a target fluctuation weight is determined for the execution node, wherein the target fluctuation weight is the conventional fluctuation weight or the new energy fluctuation weight; Based on the target fluctuation weight and combined with the preset cross-device type fluctuation correlation coefficient, the comprehensive fluctuation weight of the execution node is obtained. The initial correction coefficient of each execution node is obtained by weighting the dynamic sensitivity coefficient and the comprehensive fluctuation weight of each execution node. Based on the source-load coupling scenario category, a scenario adaptation correction coefficient is determined, and combined with the initial correction coefficient, the dynamic correction coefficient for each execution node is obtained.

7. The method for evaluating the progress of power distribution network projects according to claim 1, characterized in that, The initial process score is asymmetrically corrected according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node. The final process scores of all execution nodes are then combined to obtain the execution process evaluation result of the corresponding project in the distribution network, including: Based on the dynamic correction coefficients of the execution node, determine the asymmetric correction rule of the execution node; Based on the initial process score of each execution node and in conjunction with the asymmetric correction rule, a target correction formula corresponding to the execution node is generated. The intermediate process score of the execution node is obtained through the target correction formula; Based on the preset valid range of the distribution network execution process score, the intermediate process score is subjected to boundary constraint processing to obtain the final process score of each execution node; A depth weight is generated according to the preset topology depth of each execution node; The final process score is weighted and summed with the corresponding depth weight to obtain the weighted total process score. The weighted process total score is mapped to a percentage range to obtain the execution process evaluation result of the distribution network in the current time period.

8. A system for evaluating the progress of a power distribution network project, characterized in that, include: The data acquisition unit is used to acquire the operating data of the distribution network in the current time period and the node information of each execution node of the distribution network in the current time period. The operating data includes the conventional source-load data of conventional equipment and the new energy source-load data of new energy equipment in the distribution network. The initial analysis unit is used to determine the multi-dimensional vector of the running data based on the conventional source load data and the new energy source load data, and to obtain the initial process score of each execution node based on the multi-dimensional vector. The characterization generation module unit is used to extract features from the conventional source load data and the new energy source load data to generate a hierarchical characterization of the source load trend and fluctuation of the distribution network in the current time period. The feature extraction unit is used to perform dual-channel feature extraction on the source-load trend-fluctuation hierarchical representation to determine the source-load coupling scenario category of the distribution network in the current time period and the conventional fluctuation weight and new energy fluctuation weight corresponding to the conventional equipment and the new energy equipment, respectively. The first coefficient determination unit is used to determine the dynamic sensitivity coefficient of each execution node based on the source-load coupling scenario category, combined with the equipment type in the node information and the preset completion standard of the corresponding project of the power distribution network; The second coefficient determination unit is used to determine the dynamic correction coefficient of each execution node based on the dynamic sensitivity coefficient, combined with the conventional fluctuation weight and the new energy fluctuation weight. The comprehensive evaluation unit is used to perform asymmetric correction on the initial process score according to the dynamic correction coefficient of the execution node to obtain the final process score of each execution node, and to combine the final process scores of all execution nodes to obtain the execution process evaluation result of the corresponding project of the distribution network.

9. An electronic device, characterized in that, include: Processor and memory, the memory being used to store computer programs; When the computer program is loaded by the processor, it causes the processor to execute the power distribution network project execution process evaluation method as described in any one of claims 1-7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for evaluating the progress of power distribution network projects as described in any one of claims 1-7.

Images