Electric power survey and design integration method and system for intelligent power grid construction

Abstract

The invention belongs to the technical field of electric power survey design, and discloses an electric power survey design integration method and system for intelligent power grid construction, and the method comprises the steps: extracting preset engineering features of an existing survey result in a current engineering region, and forming a sample set in combination with current engineering planning condition features; the sample set is input into a pre-trained graph neural network, the engineering application probability and the instability risk score of each survey result are output, nodes of the graph neural network correspond to single survey results, and edges participate in design relation construction based on spatial adjacency, engineering association or joint; and generating a design consistency constraint for the survey result meeting the preset threshold, and finally checking the conformity of the design scheme and the constraint. The method realizes substantive integration of survey and design, improves the design efficiency and quality, and reduces the rework risk.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric power survey design, in particular to an electric power survey design integration method and system for smart grid construction. BACKGROUND

[0002] In the construction and expansion project of smart grid, survey design integration has become the mainstream implementation mode in the industry. The core is to integrate existing survey results and design process to improve engineering efficiency and quality. However, the existing implementation mode still faces many difficult-to-solve technical bottlenecks, which seriously affects the project advancement effect.

[0003] The existing survey results in the same engineering area are complex, covering different survey stages, engineering purposes and time periods, and have significant heterogeneity in coordinate reference, accuracy level, survey depth, etc., resulting in great difficulty in integrating the results. At the same time, the existing technology highly depends on the personal experience of designers for the usability judgment of survey results, lacks a unified and verifiable judgment standard, and is prone to over-trust or over-rejection of results, thereby causing design hazards or repeated survey problems.

[0004] The gradual clarification of planning conditions makes the conflict between survey and design present a lagging exposure feature. The design scheme based on preliminary survey results in the early stage often exposes uncoordinated contradictions after the planning details are clarified. In addition, the design stage lacks an engineering-level constraint mechanism for the use of historical survey results, so even if the risks of the results are identified, it is difficult to convert them into specific design constraints, ultimately leading to problems such as local discontinuity and parameter instability in path design and station design, causing engineering rework.

[0005] The existing method cannot effectively deal with the deep problems of difficult judgment of the credibility of existing survey results, difficult prediction of planning conflicts, and difficult conversion of design constraints, and has been difficult to meet the high-quality development needs of smart grid construction. SUMMARY

[0006] To at least partially solve the above technical problems, the present application provides an electric power survey design integration method and system for smart grid construction.

[0007] The present application provides an electric power survey design integration method for smart grid construction, comprising the following method steps:

[0008] Based on the existing survey results in the current smart grid construction engineering area, extract a preset engineering feature for each of the existing survey results, combine the extracted preset engineering features with the planning condition features of the current smart grid construction project to form a current engineering sample set;

[0009] inputting the current engineering sample set into a pre-trained deep learning model, the deep learning model outputting an engineering applicability probability and an instability risk score of each of the existing survey results; wherein the pre-trained deep learning model is a graph neural network, nodes of the graph neural network corresponding to the existing survey results, a feature vector of a node containing the engineering characteristics and planning condition characteristics corresponding to the existing survey results, edges of the graph neural network being determined according to an association relationship between two of the existing survey results, the association relationship including a spatial adjacency relationship, an engineering association relationship, and a common design participation relationship;

[0010] generating a design consistency constraint for the existing survey results with an engineering applicability probability meeting a preset threshold condition according to the engineering applicability probability and the instability risk score output by the deep learning model;

[0011] checking a design scheme against the design consistency constraint to ensure that the design scheme meets the design consistency constraint.

[0012] In one possible embodiment, the method further includes monitoring changes in planning conditions of the current engineering, updating planning condition characteristics in the current engineering sample set when the planning conditions are updated in stages, and inputting the updated current engineering sample set into the pre-trained deep learning model to obtain updated engineering applicability probability and instability risk score; calculating a combined instability risk score of a candidate design scheme according to the updated engineering applicability probability and the instability risk score of multiple existing survey results used to support the candidate design scheme, comparing the combined instability risk score with a preset engineering tolerance interval, and determining whether the conflict escalates and the conflict type according to the comparison result.

[0013] In one possible embodiment, determining whether the conflict escalates and the conflict type specifically includes: if the combined instability risk score exceeds an upper limit of the engineering tolerance interval, or a change amount of the combined instability risk score compared to a last calculation exceeds a preset change threshold, determining that the conflict escalates.

[0014] In the case of conflict escalation, if replacing or supplementing survey results with instability risk scores meeting the requirements can reduce the combined instability risk score to within the engineering tolerance interval, the conflict is determined to be a reconcilable conflict; if adjusting design parameters of the candidate design scheme can reduce the combined instability risk score to within the engineering tolerance interval, the conflict is determined to be a condition limitation conflict; otherwise, the conflict is determined to be an irreconcilable conflict.

[0015] In a possible embodiment, corresponding subsequent operations are performed according to the determined conflict type: if the conflict is determined to be a reconcilable conflict, prompt information is output to replace or supplement new survey results; if the conflict is determined to be a conditional restriction conflict, a new design consistency constraint is generated according to the updated instability risk score, and the checking level is increased for a checking process related to the candidate design scheme; if the conflict is determined to be an irreconcilable conflict, warning information is output and it is suggested to restart the survey process or make significant scheme changes.

[0016] In a possible embodiment, the candidate design scheme is a candidate line path scheme or a candidate substation site scheme, and the plurality of existing survey results supporting the same candidate design scheme cover or adjoin a geographical space range related to the candidate design scheme.

[0017] In a possible embodiment, the preset engineering features include a time difference between a survey time of the existing survey result and a current project start time, a similarity between an original project type of the existing survey result and a current project type, a matching degree between a survey accuracy level of the existing survey result and a minimum accuracy level required by a current design stage, and an actual use result feature of the existing survey result in a historical project.

[0018] In a possible embodiment, the planning condition features include a planning explicitness feature and a constraint density feature, the planning explicitness feature is determined according to completeness of a planning document and specificity of planning indicators, and the constraint density feature is determined according to a ratio of a number of constraint clauses in a current planning condition to an average number of constraint clauses.

[0019] In a possible embodiment, generating the design consistency constraint includes determining a constraint strength coefficient according to the instability risk score, the constraint strength coefficient being positively correlated with the instability risk score, and generating a geometric constraint, a parameter constraint or a process constraint for the design scheme based on the constraint strength coefficient.

[0020] Another aspect of the present application provides an electric power survey and design integrated system for smart grid construction, comprising:

[0021] A sample set construction module is configured to extract preset engineering features from each of the existing survey results based on the existing survey results in a current smart grid construction project area, and combine the extracted preset engineering features with planning condition features of the current smart grid construction project to form a current project sample set.

[0022] a deep learning analysis module configured to input the current engineering sample set into a pre-trained deep learning model, the pre-trained deep learning model outputting an engineering applicability probability and a risk of instability score for each of the existing survey results; wherein the pre-trained deep learning model is a graph neural network, nodes of the graph neural network corresponding to individual ones of the existing survey results, a feature vector of a node including the corresponding engineering features and planning condition features, edges of the graph neural network being determined according to an association relationship between two of the existing survey results, the association relationship including a spatial adjacency relationship, an engineering association relationship, and a common design participation relationship;

[0023] a constraint generation module configured to generate a design consistency constraint for an existing survey result whose engineering applicability probability meets a preset threshold condition according to the engineering applicability probability and the risk of instability score output by the deep learning model;

[0024] a scheme checking module configured to check a design scheme against the design consistency constraint to ensure that the design scheme meets the design consistency constraint.

[0025] In another aspect, the present application provides an electronic device, which includes at least one processor, and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the power survey and design integration method for smart grid construction as described above.

[0026] In another aspect, the present application provides a storage medium having stored thereon computer program instructions executable by a processor to implement the power survey and design integration method for smart grid construction as described above.

[0027] In another aspect, the present application provides a computer program product comprising a computer program, which, when executed by a processor, implements the power survey and design integration method for smart grid construction as described above.

[0028] This application achieves substantial integration of power survey and design in smart grid construction through processes such as current engineering feature extraction and sample construction, deep learning model training, automatic judgment of survey result credibility, automatic generation of design constraints, constraint verification of design schemes, and determination of planning conflict evolution. The method, with graph neural networks at its core, automatically determines the engineering applicability probability and instability risk score of survey results, avoiding the subjectivity and limitations of manual experience-based judgment. It transforms model output into design consistency constraints, establishing a close link between survey results and the design process, solving the problem of lacking engineering-level constraint mechanisms. By dynamically monitoring changes in planning conditions and inferring conflict evolution, it achieves early identification and classification of planning conflicts, avoiding engineering rework caused by delayed conflict exposure. This significantly reduces design rework and engineering risks caused by misuse of survey results and untimely handling of planning conflicts, improving the design efficiency and quality of smart grid construction projects, and is particularly suitable for complex scenarios such as smart grid expansion and existing project renovation. Attached Figure Description

[0029] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0030] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0031] Figure 1 This is a schematic diagram of an integrated power survey and design method for smart grid construction, provided as an embodiment of this application.

[0032] Figure 2 This is a schematic diagram of the graph neural network construction and training process.

[0033] Figure 3 This is a schematic diagram of the neural network model structure in an embodiment of this application.

[0034] Figure 4 This is a schematic diagram of the design constraint generation process provided for an embodiment of this application.

[0035] Figure 5 This is a schematic diagram of the planning conflict identification and classification process provided in an embodiment of this application.

[0036] Figure 6 This is a schematic diagram of the structure of an integrated power survey and design system for smart grid construction provided in an embodiment of this application.

[0037] Figure 7 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0039] It should be noted that all user information (including but not limited to user device information, user personal information, object information corresponding to device usage data, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, device usage data, etc.) involved in all embodiments of this application are information and data authorized by the user or fully authorized by all parties.

[0040] This method is applicable to smart grid construction and expansion projects, and is particularly suitable for engineering scenarios involving substation site selection and transmission line route planning. It can handle situations where there are existing survey results across multiple stages, purposes, and cycles within the same project area, such as smart grid expansion projects and existing engineering renovation projects. Implementing this method typically requires meeting the following conditions: having readily available existing survey data, including key information such as survey time, accuracy level, and historical usage records; having relevant planning conditions data for the current project, from which features such as planning clarification and constraints can be extracted; having sufficient historical project case data, including design result feedback information; and having computing resources to support deep learning model training and inference, meeting the operational requirements of graph neural networks.

[0041] The following detailed description, in conjunction with specific embodiments, illustrates the implementation process of the integrated power survey and design method for smart grid construction described in this application. It should be noted that these embodiments are merely for explaining this application and not for limiting its scope of protection. Any conventional adjustments or substitutions made by those skilled in the art to the steps without departing from the concept of this application should be included within the scope of protection of this application.

[0042] like Figure 1 As shown in the figure, this application discloses a schematic diagram of an integrated power survey and design method for smart grid construction, including the following method steps:

[0043] S1. Based on the existing survey results within the current smart grid construction project area, extract preset engineering features for each of the existing survey results, and combine the extracted preset engineering features with the planning condition features of the current smart grid construction project to form the current project sample set.

[0044] S2. Input the current engineering sample set into a pre-trained deep learning model. The deep learning model outputs the engineering applicability probability and instability risk score for each existing survey result. The pre-trained deep learning model is a graph neural network. The nodes of the graph neural network correspond to a single existing survey result. The feature vector of the node contains the corresponding engineering features and planning condition features. The edges of the graph neural network are determined according to the association between two corresponding existing survey results. The association includes spatial adjacency, engineering association, and joint design participation.

[0045] S3. Based on the engineering applicability probability and instability risk score output by the deep learning model, design consistency constraints are generated for existing survey results whose engineering applicability probability meets the preset threshold condition.

[0046] S4. Verify the design scheme against the design consistency constraints to ensure that the design scheme meets the design consistency constraints.

[0047] In some embodiments, step S1 provides basic data support for subsequent deep learning model inference. The core of this step is to transform the scattered existing survey results and planning conditions into standardized samples. The specific implementation process is as follows:

[0048] Collect existing survey results within the current smart grid construction project area. Comprehensively review all survey results generated at different stages (preliminary survey, detailed survey, resurvey), for different project purposes (new construction, expansion / relocation), and at different time periods within the area, ensuring no omissions. Preprocess the collected existing survey results, including data format standardization, invalid data removal, and marking missing key information. For example, convert survey data of different formats to a pre-defined standard format, and separately mark results lacking survey accuracy information, excluding them from subsequent feature extraction processes.

[0049] For each pre-processed existing survey result, pre-defined engineering features are extracted. Specifically, the extracted engineering features may include four items: the first is the time difference between the survey time of the existing survey result and the start time of the current project, thus obtaining the completion time of the existing survey result. Compared with the start time of the current project Through calculation Once this characteristic is obtained, the time unit can be standardized to month or year. For example, if the current project starts in January of a certain year, and an existing survey result was completed in March of a previous year, then the calculation will be based on the difference between the corresponding months. The second item is the similarity between the original project type and the current project type based on existing survey results. First, the project types are divided into preset categories, such as new substation projects, substation renovation and expansion projects, and new transmission line projects. A corresponding feature vector is constructed for each category. Then, the feature vector of the original project type is calculated using a cosine similarity algorithm. Feature vector of the current project type similarity The calculation formula is: ,in Representing vectors and dot product, and Representing vectors respectively and The model, The value range is [0,1];

[0050] The third item is the matching degree between the accuracy level of existing survey results and the minimum accuracy level required in the current design stage. A sequence of survey accuracy levels is preset, such as dividing them into four levels from low to high accuracy. The minimum accuracy requirement level for each design stage of the current project, such as the preliminary design stage and the construction drawing design stage, is also clearly defined. Based on the comparison results between the accuracy level of existing survey results and the minimum requirement level for the current design stage, the matching degree is calculated using a matching degree function. If the accuracy level of existing survey results is higher than or equal to the minimum required level, The value is [0.6, 1.0], otherwise it is [0, 0.6); the fourth item is the actual use result characteristics of existing survey results in historical projects. By querying historical project archives, the application records of the survey results in similar or related projects in the past are obtained. If the use did not lead to rework or adjustment, it is marked as the first characteristic value. If it led to rework or adjustment, it is marked as the second characteristic value. If there is no historical use record, it is marked as a neutral characteristic value.

[0051] Then, the planning condition characteristics of the current project are extracted. These characteristics include the degree of planning clarity and the density of constraints. The degree of planning clarity is determined based on the completeness of the planning documents and the specificity of the planning indicators. If only the general planning direction is clear without specific indicator constraints, the value of this characteristic is low. If core indicators such as planning red lines, land use, and plot ratio are clear, the value is high. If the indicators are completely clear and there is no room for adjustment, the value is the highest, ranging from [0,1]. The constraint density characteristic is determined based on the ratio of the number of constraint clauses in the current planning conditions to the average number of constraint clauses. The calculation formula is as follows: ,in The number of constraint clauses in the current planning conditions. This represents the average number of constraint clauses in planning conditions for similar projects. This is a weighting coefficient, with a value range of [0.8, 1.2], which can be adjusted according to the specific characteristics of the project area. The value range is [0,1].

[0052] The engineering features of each existing survey result are combined with the planning conditions features of the current project to form a single sample in the current project sample set. All samples are then aggregated to form the current project sample set. All samples in the current project sample set are normalized, mapping each feature value to the [0,1] interval to eliminate the influence of dimensional differences on subsequent model inference.

[0053] In some embodiments, step S2 addresses the problem in the prior art that relying on human experience or rule bases cannot accurately determine the credibility of survey results. By constructing and training a graph neural network, it enables the network to automatically determine the engineering applicability probability and instability risk score of the survey results.

[0054] Please see Figure 2 , Figure 2 This is a schematic diagram of the graph neural network construction and training process. In S201, historical engineering case data is collected. A wide range of historical engineering cases in the field of smart grid construction and expansion are collected, ensuring that the cases cover different project types, regions, and planning conditions. Each historical engineering case must include existing survey data, corresponding planning condition data, and historical design result feedback data. The historical design result feedback data includes information on whether route adjustments, site changes, rework, etc., occurred, ensuring the completeness and diversity of the data to provide sufficient and effective data support for model training.

[0055] In S202, a training sample set is constructed. For each historical engineering case, existing survey results are used to extract four pre-defined engineering features using the same method as for current engineering features: the time difference between the survey time and the start time of the corresponding historical engineering project, the similarity between the original engineering type and the corresponding historical engineering type, the matching degree between the survey accuracy level and the minimum accuracy level required in the design stage of the corresponding historical engineering project, and the actual usage results in earlier historical engineering projects. Simultaneously, historical planning condition features corresponding to each historical engineering case are extracted, including the clarity of historical planning and the density of historical constraints, using the same extraction method as for current engineering planning condition features. The extracted engineering features of each historical engineering project's existing survey results are combined with the corresponding historical planning condition features and historical design results feedback to form a single training sample. All training samples are then aggregated to form the training sample set. The training sample set is divided into a training subset, a validation subset, and a test subset according to a pre-defined ratio, for example, 70% for the training subset, 20% for the validation subset, and 10% for the test subset. The training subset is used for model parameter training, the validation subset for model hyperparameter adjustment, and the test subset for model performance evaluation.

[0056] In S203, the specific structure of the deep learning model is determined. The deep learning model adopts a graph neural network. The nodes of the graph neural network correspond to a single existing survey result. The feature vector of each node contains the corresponding engineering features and planning condition features. The dimension of the feature vector is the sum of the number of engineering features and the number of planning condition features. The edges of the graph neural network are determined according to the relationship between the corresponding two existing survey results. The relationship includes spatial adjacency, engineering relationship, and joint participation in design. Spatial adjacency means that the survey areas of the two survey results overlap or are adjacent. Engineering relationship means that the two survey results correspond to the same historical project or serve the same type of engineering purpose. Joint participation in design means that the two survey results have jointly participated in the same design process.

[0057] The edges of a graph neural network have corresponding weights. These weights are calculated by weighting and summing the weights of spatial adjacency, engineering association, and joint participation relationships. The formula is as follows: ,in For the first The node and the first The total weight of the edges between nodes The weight of spatial adjacency relationship. Weights for project-related relationships. To jointly participate in relationship weighting, , , For the weighting coefficients, satisfying This can be optimized and adjusted by validating subset data.

[0058] Please see Figure 3 , Figure 3 This is a schematic diagram of the graph neural network model structure in an embodiment of this application. The network layer structure of the graph neural network includes an input layer, a graph convolutional layer, an activation layer, a pooling layer, and an output layer. The input layer receives node feature vectors and edge weights, converts them into tensor forms that the model can process, and the input dimension is the dimension of the node feature vectors. The graph convolutional layer uses the GCN convolution operator to aggregate and update the node features, and the calculation formula is as follows: ,in For the first The node feature matrix of the layer For the first The node feature matrix of the layer To add the adjacency matrix after adding self-loop edges, This is the original adjacency matrix. It is the identity matrix. for The degree matrix, For the first The weight matrix of the layer, For the first The layer's bias vector, The activation function is ReLU. The number of graph convolutional layers can be configured from 2 to 4. For example, setting 3 graph convolutional layers, the first layer has an output dimension of 64, the second layer has an output dimension of 32, and the third layer has an output dimension of 16. The activation layer uses the ReLU function as the activation function, and the calculation formula is... This is used to introduce non-linear features and enhance the expressive power of the model. The pooling layer adopts topological pooling, selecting the top-K nodes to retain based on the importance scores of the node features. The value of K can be set to 50%-80% of the total number of nodes. The output layer adopts a fully connected layer with an output dimension of 2, corresponding to the engineering applicability probability and instability risk score, respectively. The output results are normalized by the Sigmoid function, and the values ​​range from [0,1].

[0059] In S204, model training and optimization are performed. This involves adjusting the weight matrix of the graph neural network. Bias vector Initialize the model using the Xavier initialization method to ensure consistent variance across all layers. Set the training batch size, such as 32 or 64, adjusting based on available computing resources; set the number of training epochs, such as 50-100. Calculate the model's prediction error using the joint loss function, calculated as follows: ,in This is a binary cross-entropy loss function used to calculate the prediction error of the engineering applicability probability. The mean squared error loss function is used to calculate the prediction error of the instability risk score. The balancing coefficient has a value range of [0.4, 0.6]. Binary cross-entropy loss function. The calculation formula is ,in This represents the number of samples in the training batch. For the first The true label of the engineering applicability probability of each sample. For the first Predicted engineering applicability probability values ​​for each sample; mean squared error loss function. The calculation formula is ,in For the first The true label of the instability risk score for each sample. For the first Predicted instability risk score for each sample.

[0060] The Adam optimizer is used to update the model parameters, with a learning rate set between 0.001 and 0.01. A learning rate decay strategy is employed, for example, the learning rate decreases to 0.9 every 10 epochs. Training subset samples are batch-wise input into the graph neural network. Forward propagation calculates the model output, and the prediction error is calculated using the loss function. Backpropagation updates the model parameters. In each training epoch, the model performance is evaluated using a validation subset, calculating the loss and accuracy on the validation subset. If the loss on the validation subset does not decrease for five consecutive epochs, training is stopped. After training, the model performance is evaluated using a test subset. Evaluation metrics include the AUC (Aspect-Understanding Value) for engineering applicability and the MAE (Maintenance-Based Estimation Value) for instability risk. The AUC is required to be no less than 0.85, and the MAE no greater than 0.1. If these metrics are not met, the model structure is adjusted or the training parameters are optimized, and training is repeated until the model performance meets the requirements, resulting in a pre-trained deep learning model.

[0061] In some embodiments, a pre-trained graph neural network is used to automatically determine the credibility of existing survey results, solving the problems of inconsistent judgment standards and low accuracy caused by the reliance on human experience in traditional methods. The specific implementation process is as follows:

[0062] The completed current project sample set is input into the pre-trained deep learning model. During the input process, ensure that the sample format in the current project sample set is consistent with the sample format during model training. The dimension of the feature vectors and the order of each feature should also be consistent with the training stage to avoid abnormal model inference results due to inconsistent formats.

[0063] The pre-trained deep learning model processes the input current engineering sample set, aggregates and updates node features through graph convolutional layers, and fully captures the influence of the correlation between existing survey results on their credibility. After processing by activation and pooling layers, the output layer outputs the engineering applicability probability and instability risk score of each existing survey result. The engineering applicability probability reflects the degree to which the survey result is applicable to the current engineering design, with a value range of [0,1]. A higher value indicates stronger applicability. The instability risk score reflects the risk that the survey result, when used in the current engineering design, will lead to local discontinuities in the design, parameter instability, or rework in the later stages. The value range is [0,1], with a higher value indicating greater risk.

[0064] After the model outputs the results, the validity of the output results is verified to ensure that the output engineering applicability probability and instability risk score are both within the range of [0,1]. If there are outliers outside this range, the current engineering sample set is re-inputted for inference. If outliers occur multiple times, the construction of the current engineering sample set is checked for problems, or the pre-trained deep learning model is fine-tuned until all output results are valid values.

[0065] In some embodiments, step S4 directly transforms the confidence results output by the model into design constraints, solving the problem in the prior art of lacking an engineering-level constraint mechanism for the use of historical survey results, and realizing a substantial connection between surveying and design. (See also...) Figure 4 , Figure 4 This is a schematic diagram of the design constraint generation process provided for an embodiment of this application. The specific implementation process is as follows:

[0066] In S401, a preset threshold for the applicability probability of the project is set. The preset threshold ranges from [0.5, 0.7] and can be adjusted according to factors such as the type, importance, and design stage of the current project. For example, the preset threshold for the preliminary design stage can be configured to 0.5, the preset threshold for the construction drawing design stage can be configured to 0.65, the preset threshold for important transmission line projects can be configured to 0.6, and the preset threshold for general distribution network projects can be configured to 0.55.

[0067] In S402, existing survey results that meet the criteria are screened. The engineering applicability probability of each existing survey result is compared with a preset threshold. If the engineering applicability probability is greater than or equal to the preset threshold, the existing survey result meets the preset threshold condition and is allowed to enter the subsequent design process; if the engineering applicability probability is less than the preset threshold, the existing survey result does not meet the preset threshold condition and is prohibited from entering the design process to avoid design risks caused by results with low applicability.

[0068] In S403, design consistency constraints are generated for existing survey results that meet the conditions. The generation process includes two steps: the first step is to determine the constraint strength coefficient based on the instability risk score. The constraint strength coefficient is positively correlated with the instability risk score, and the calculation formula is as follows: ,in The constraint strength coefficient, Assess the risk of instability. and Let be the coefficient, satisfying when hour, As a low-strength benchmark value, when hour, For high-strength benchmark values, such as configuration , ,but , The value range is [0.2, 0.9];

[0069] The second step involves generating geometric constraints, parametric constraints, or process constraints for the design scheme based on the constraint strength coefficients. Geometric constraints define the spatial scope of the design results. For example, the geometric constraint for route design is that the route must be located centered on the path indicated by the survey results, with a width of... Within the strip-shaped area, The allowable path deviation is calculated using the following formula: , This represents the maximum permissible deviation value under unconstrained conditions, preset according to engineering standards; parameter constraints specify the range of values ​​for design parameters, such as conductor cross-section and tower height for transmission lines, which can be determined based on the parameters indicated by the survey results. Determine the fluctuation range. The higher the value, the smaller the fluctuation range; process constraints specify the verification process and validation requirements during the design process. The higher the value, the more checks are performed and the stricter the verification standards. For example, low The corresponding design scheme only requires one round of internal verification. The corresponding design scheme requires two rounds of internal verification and one round of external review. The corresponding design scheme requires 3 rounds of internal verification, 2 rounds of external review, and 1 round of on-site verification.

[0070] Finally, the generated design consistency constraints are organized and stored according to a preset format to form a constraint file, which can be easily accessed during subsequent design scheme constraint verification. The constraint file must clearly specify the existing survey results identifier, constraint type, constraint parameters, and scope of application for each constraint to ensure that designers can clearly understand and apply each constraint.

[0071] In some embodiments, step S4 ensures that the design scheme conforms to the generated design consistency constraints, avoiding rework due to conflicts between the design scheme and the constraints. The specific implementation process is as follows:

[0072] In the specific design phase, after completing the preliminary design scheme, the designers input the relevant data of the design scheme into the verification system. The design scheme includes candidate line route schemes or candidate substation site schemes, and the relevant data of the design scheme includes route coordinates, site parameters, design parameter values, spatial range information, etc.

[0073] The verification system calls the stored constraint file and extracts the design consistency constraints related to the design scheme. For candidate line route schemes, it extracts the corresponding geometric constraints, parametric constraints, and process constraints; for candidate substation site schemes, it extracts the corresponding geometric constraints and parametric constraints.

[0074] The verification system employs a combination of automated and manual verification to check the design scheme against the extracted design consistency constraints. During automated verification, the system verifies whether the design scheme meets the requirements one by one according to the constraint type. For example, for geometric constraints, the system calculates the overlap between the spatial range of the design scheme and the spatial range specified by the constraint; if the overlap reaches a preset ratio, the geometric constraint is deemed satisfied. For parametric constraints, the system checks whether the values ​​of the design parameters are within the range specified by the constraint; if all parameters are within the range, the parametric constraint is deemed satisfied. For process constraints, the system verifies whether the design process has completed the verification process and verification steps specified by the constraint; if all are completed, the process constraint is deemed satisfied. If the design scheme meets all relevant constraints, the automated verification passes; if there are items that do not meet the constraints, the system marks the specific problems and provides feedback to the designer.

[0075] For design schemes that pass automated verification, based on constraint strength coefficients Determine the depth of manual review. At lower levels, manual review only requires sampling to check the satisfaction of key constraints; When the constraint level is high, manual review requires a comprehensive examination of the satisfaction of all constraints to ensure no constraint violations occur. If issues are found during manual review, they should be promptly reported to the designers for adjustments; if no issues are found during manual review, the design scheme passes the verification.

[0076] If the design scheme fails the automated verification or manual review, the designer adjusts the design scheme based on the feedback, and then resubmits it to the verification system for verification until the design scheme fully meets all relevant design consistency constraints.

[0077] In some embodiments, to address the problem that existing technologies cannot predict conflict transitions caused by the progression of planning conditions, planning conflicts can be identified and classified in advance by monitoring changes in planning conditions and making dynamic inferences. Please refer to [link to relevant documentation]. Figure 5 , Figure 5 This is a schematic diagram illustrating the planning conflict identification and classification process provided in an embodiment of this application. The specific implementation process is as follows:

[0078] In S501, a planning condition update monitoring mechanism will be established to track changes in the planning conditions of the current project in real time. Monitoring will include updates to planning documents, adjustments to planning indicators, additions or modifications to constraint clauses, and improvements in planning clarity. The monitoring frequency can be set according to the project progress and the planning update cycle, for example, monitoring once a day, or monitoring promptly after the planning department issues relevant notices.

[0079] In S502, when a phased update of the planning conditions is detected, the planning condition features in the current engineering sample set are updated. Following the same method as the initial planning condition feature extraction, the planning clarification features and constraint density features are re-extracted and replaced with the original planning condition features in the current engineering sample set, forming an updated current engineering sample set. During the update process, it is ensured that the newly extracted planning condition features accurately reflect the updated planning situation, avoiding the impact of inaccurate feature extraction on subsequent model inference results.

[0080] In S503, the updated engineering applicability probability and instability risk score are obtained. The updated current engineering sample set is then re-input into the pre-trained deep learning model. The model processes the data according to the same procedure as the initial inference, outputting the updated engineering applicability probability and instability risk score for each existing survey result.

[0081] In S504, multiple existing survey results are identified to support the same candidate design scheme; the survey areas of these multiple existing survey results must cover or be adjacent to the geographical space involved in the candidate design scheme. Based on the updated engineering applicability probability and instability risk score of these existing survey results, the combined instability risk score of the candidate design scheme is calculated. The combined instability risk score is calculated using a weighted summation method, and the calculation formula is as follows: ,in To score the portfolio's instability risk, For the first An updated instability risk score based on existing survey results. For the first The weight of existing survey results in the portfolio is determined based on their updated engineering applicability probability. Confirmed, the calculation formula is as follows: , The total number of existing survey results used to support the candidate design scheme is used to ensure that the sum of the weights is 1.

[0082] In S505, obtain the preset engineering tolerance range and the preset change threshold; the engineering tolerance range is... , The minimum tolerable risk score, The maximum tolerable risk score can be adjusted based on factors such as project type and importance, for example, for important transmission line projects. It can be configured to 0.3, which is generally used in power distribution network projects. It can be configured to 0.45; the preset change threshold is the maximum allowable change range of the combined instability risk score, for example, it can be configured to 0.1.

[0083] In S506, the calculated combined instability risk score is compared with the engineering tolerance range, and the change in the combined instability risk score compared to the previous calculation is calculated. If the combined instability risk score exceeds the upper limit of the engineering tolerance range, or the change exceeds the preset change threshold, the conflict is determined to have escalated; if the combined instability risk score is within the engineering tolerance range and the change does not exceed the preset change threshold, the conflict is determined not to have escalated, and the current candidate design scheme can continue to be used.

[0084] In S507, in cases of escalating conflict, the conflict type is further determined. If replacing or supplementing survey results that meet the instability risk score requirements can reduce the combined instability risk score to within the engineering tolerance range, and the replacement or supplementation cost is within an acceptable range, it is determined to be a reconcilable conflict. If adjusting the design parameters of the candidate design scheme, such as the route alignment and the construction scale of the substation, can reduce the combined instability risk score to within the engineering tolerance range, and the adjusted design scheme can still meet the engineering requirements, it is determined to be a conditional conflict. If replacing all survey results with high instability risk scores, supplementing with new survey results, or significantly adjusting the design parameters still cannot reduce the combined instability risk score to within the engineering tolerance range, or the adjustment cost is too high and does not meet the engineering construction requirements, it is determined to be an unreconcilable conflict.

[0085] In S508, the corresponding follow-up operations are performed based on the determined conflict type. If the conflict is determined to be resolvable, a prompt message is output, clearly informing the designer of the type, quantity, and related requirements of the survey results that need to be replaced or supplemented. After replacing or supplementing the survey results according to the prompt message, the designer recalculates the combined instability risk score until the combined instability risk score meets the requirements. If the conflict is determined to be a condition constraint conflict, the automatic design constraint generation step is re-executed. New design consistency constraints are generated based on the updated instability risk score. At the same time, the verification level of the design scheme constraint verification process related to the candidate design scheme is increased, the number of verifications is increased, or the verification standard is raised. After the designer adjusts the design scheme according to the new design consistency constraints, the verification is performed again. If the conflict is determined to be incompatible, a warning message is output, clearly informing the designer that the conflict cannot be resolved through conventional methods, and suggesting restarting the survey process or making major scheme changes. The designer restarts the relevant work according to the suggestions.

[0086] This method achieves substantial integration of power survey and design in smart grid construction through processes such as current engineering feature extraction and sample construction, deep learning model training, automatic judgment of survey result credibility, automatic generation of design constraints, constraint verification of design schemes, and determination of planning conflict evolution. Using a graph neural network as its core, the method automatically determines the engineering applicability probability and instability risk score of survey results, avoiding the subjectivity and limitations of manual experience-based judgment. It transforms model output into design consistency constraints, establishing a close link between survey results and the design process, solving the problem of lacking engineering-level constraint mechanisms. By dynamically monitoring changes in planning conditions and inferring conflict evolution, it achieves early identification and classification of planning conflicts, avoiding engineering rework caused by delayed conflict exposure. This significantly reduces design rework and engineering risks caused by misuse of survey results and untimely handling of planning conflicts, improving the design efficiency and quality of smart grid construction projects, and is particularly suitable for complex scenarios such as smart grid expansion and existing project renovation.

[0087] It should be noted that although the operations of the method of this application are described in a specific order in the accompanying drawings, this does not require or imply that these operations must be performed in that specific order, or that all the operations shown must be performed to achieve the desired result. On the contrary, the steps depicted in the flowchart can be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into one step, and / or one step may be broken down into multiple steps.

[0088] Please see Figure 6 , Figure 6 This is a schematic diagram of a power survey and design integrated system for smart grid construction provided in an embodiment of this application. As shown in the figure, the system includes:

[0089] The sample set construction module 601 is used to extract preset engineering features from each existing survey result within the current smart grid construction project area, and combine the extracted preset engineering features with the planning condition features of the current smart grid construction project to form the current project sample set.

[0090] The deep learning analysis module 602 is used to input the current engineering sample set into a pre-trained deep learning model. The deep learning model outputs the engineering applicability probability and instability risk score for each existing survey result. The pre-trained deep learning model is a graph neural network. The nodes of the graph neural network correspond to a single existing survey result. The feature vector of the node contains the corresponding engineering features and planning condition features. The edges of the graph neural network are determined according to the association between two corresponding existing survey results. The association includes spatial adjacency, engineering association, and joint design participation.

[0091] The constraint generation module 603 is used to generate design consistency constraints for existing survey results whose engineering applicability probability meets the preset threshold condition based on the engineering applicability probability and instability risk score output by the deep learning model.

[0092] The scheme verification module 604 is used to verify the design scheme against the design consistency constraints to ensure that the design scheme meets the design consistency constraints.

[0093] Those skilled in the art will clearly understand that the technical solutions of the embodiments of this application can be implemented by means of software and / or hardware. In this specification, "unit" and "module" refer to software and / or hardware that can independently complete or cooperate with other components to complete a specific function, wherein the hardware may be, for example, a field-programmable gate array (FPGA), an integrated circuit (IC), etc.

[0094] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0095] Please see Figure 7 It shows a schematic diagram of the structure of an electronic device according to an embodiment of this application, which can be used to implement... Figure 1 The method in the illustrated embodiment. (As shown) Figure 7 As shown, the electronic device may include:

[0096] The system includes at least one processor 701, at least one network interface 704, a user interface 703, a memory 705, and at least one communication bus 702. The communication bus 702 is used to enable connection and communication between the components. The user interface 703 may include buttons, and optionally include a standard wired or wireless interface. The network interface 704 may include, but is not limited to, a Bluetooth module, an NFC module, a Wi-Fi module, etc.

[0097] The processor 701 may include one or more processing cores and connect to various parts within the electronic device 700 via various interfaces and lines. It implements various functions and data processing of the device 700 by running or executing instructions, programs, code sets, or instruction sets stored in the memory 705, and by accessing data in the memory 705. Optionally, the processor 701 may be implemented using at least one hardware form of DSP, FPGA, or PLA. The processor 701 may also integrate one or more combinations of CPU, GPU, and modem.

[0098] The memory 705 may include random access memory (RAM) or read-only memory (ROM). Optionally, the memory 705 includes a non-transitory computer-readable medium for storing instructions, programs, code, code sets, or instruction sets. The memory 705 may be divided into a program storage area and a data storage area, wherein the program storage area can be used to store instructions for implementing an operating system and instructions for implementing the foregoing method embodiments; the data storage area can be used to store data related to the relevant method embodiments. The memory 705 may also be at least one storage device located remotely from the processor 701. Figure 7 As shown, the memory 705, which serves as a computer storage medium, may contain an operating system, a network communication module, a user interface module, and program instructions.

[0099] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by processor 701, the functions defined in the methods of this application are performed.

[0100] Another embodiment of this application provides a storage medium storing computer program instructions thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application.

[0101] In the above embodiments, the descriptions of each embodiment have different focuses. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. The above descriptions are merely preferred embodiments of this application and explanations of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by specific combinations of the above technical features, but should also cover other technical solutions formed by arbitrary combinations of the above technical features or their equivalent features without departing from the inventive concept.

Claims

1. An integrated method for power survey and design for smart grid construction, characterized in that, include: Based on the existing survey results within the current smart grid construction project area, pre-defined engineering features are extracted from each of the existing survey results. The extracted pre-defined engineering features are then combined with the planning condition features of the current smart grid construction project to form the current project sample set. The current engineering sample set is input into a pre-trained deep learning model, which outputs the engineering applicability probability and instability risk score for each existing survey result. The pre-trained deep learning model is a graph neural network, where each node corresponds to a single existing survey result. The feature vector of each node contains the corresponding engineering features and planning condition features. The edges of the graph neural network are determined based on the association between two corresponding existing survey results, including spatial adjacency, engineering association, and joint design participation. Based on the engineering applicability probability and instability risk score output by the deep learning model, design consistency constraints are generated for existing survey results whose engineering applicability probability meets the preset threshold condition. The design scheme is checked against the design consistency constraints to ensure that the design scheme meets the design consistency constraints.

2. The integrated power survey and design method for smart grid construction according to claim 1, characterized in that, It also includes monitoring changes in the planning conditions of the current project, updating the planning condition features in the current project sample set when the planning conditions are updated in stages, and re-inputting the updated current project sample set into the pre-trained deep learning model to obtain the updated project applicability probability and instability risk score. Based on the updated engineering applicability probability and instability risk score of multiple existing survey results used to support the same candidate design scheme, the combined instability risk score of the candidate design scheme is calculated. The combined instability risk score is compared with the preset engineering tolerance range, and the conflict escalation and conflict type are determined based on the comparison results.

3. The integrated power survey and design method for smart grid construction according to claim 2, characterized in that: Determining whether a conflict has escalated and the specific type of conflict includes: If the combined instability risk score exceeds the upper limit of the engineering tolerance range, or if the change in the combined instability risk score compared to the previous calculation exceeds a preset change threshold, then the conflict is determined to have escalated. In the event of escalating conflict, if replacing or supplementing the survey results that meet the requirements for instability risk score can reduce the combined instability risk score to within the engineering tolerance range, it is determined to be a reconcilable conflict; if adjusting the design parameters of the candidate design scheme can reduce the combined instability risk score to within the engineering tolerance range, it is determined to be a conditional conflict; otherwise, it is determined to be an unreconcilable conflict.

4. The integrated power survey and design method for smart grid construction according to claim 3, characterized in that, Perform the corresponding follow-up operations based on the determined conflict type: If the conflict is determined to be reconcilable, a prompt message will be output to replace or supplement the new survey results; If the condition constraint conflict is determined, a new design consistency constraint is generated based on the updated instability risk score, and the verification level is increased for the verification process related to the candidate design scheme. If the conflict is determined to be irreconcilable, a warning message will be issued and it will be recommended to restart the survey process or make major changes to the plan.

5. The integrated power survey and design method for smart grid construction according to claim 2, characterized in that, The candidate design scheme is a candidate route scheme or a candidate substation site scheme. The multiple existing survey results used to support the same candidate design scheme cover or are adjacent to the geographical space involved in the candidate design scheme.

6. The integrated power survey and design method for smart grid construction according to claim 1, characterized in that, The preset engineering characteristics include: the time difference between the survey time of the existing survey results and the start time of the current project; the similarity between the original project type of the existing survey results and the current project type; the matching degree between the survey accuracy level of the existing survey results and the minimum accuracy level required in the current design stage; and the actual usage results of the existing survey results in historical projects.

7. The integrated power survey and design method for smart grid construction according to claim 1, characterized in that, The planning conditions characteristics include planning clarity characteristics and constraint density characteristics. The planning clarity characteristics are determined based on the completeness of the planning documents and the specificity of the planning indicators. The constraint density characteristics are determined based on the ratio of the number of constraint clauses in the current planning conditions to the average number of constraint clauses.

8. The integrated power survey and design method for smart grid construction according to claim 1, characterized in that, The generation of design consistency constraints includes: determining constraint strength coefficients based on the instability risk score, wherein the constraint strength coefficients are positively correlated with the instability risk score; and generating geometric constraints, parametric constraints, or process constraints for the design scheme based on the constraint strength coefficients.

9. An integrated power survey and design system for smart grid construction, characterized in that, include: The sample set construction module is used to extract preset engineering features from each existing survey result within the current smart grid construction project area, and combine the extracted preset engineering features with the planning condition features of the current smart grid construction project to form the current project sample set. The deep learning analysis module is used to input the current engineering sample set into a pre-trained deep learning model. The deep learning model outputs the engineering applicability probability and instability risk score for each existing survey result. The pre-trained deep learning model is a graph neural network. The nodes of the graph neural network correspond to a single existing survey result. The feature vector of the node contains the corresponding engineering features and planning condition features. The edges of the graph neural network are determined according to the association between two corresponding existing survey results. The association includes spatial adjacency, engineering association, and joint design participation. The constraint generation module is used to generate design consistency constraints for existing survey results whose engineering applicability probability meets the preset threshold condition based on the engineering applicability probability and instability risk score output by the deep learning model. The scheme verification module is used to verify the design scheme against the design consistency constraints to ensure that the design scheme meets the design consistency constraints.

10. An electronic device, characterized in that, include: At least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 8.

Images