Power grid dispatching service interface system, method and equipment and storage medium
By introducing interface modules with multi-level cache and fault tolerance mechanisms, combined with fault diagnosis modules and scheduling suggestions modules, the problem of inefficiency of traditional power grid scheduling systems is solved, efficient and intelligent power grid scheduling is achieved, and the real-time and stability of the system is improved.
Patent Information
- Application Number
- CN202510514957.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-23
- Publication Date
- 2025-07-11
AI Technical Summary
Traditional power grid scheduling systems rely on dispatcher experience and static rule bases, resulting in inefficiency and cannot meet the needs of modern power systems for efficient and accurate scheduling.
The interface module of a multi-level caching mechanism and fault tolerance mechanism is introduced, and the fault diagnosis module (based on statistical analysis, Bayesian reasoning, knowledge graph and large model) is used to identify faults. It generates efficient scheduling strategies through the scheduling suggestion module, and uses knowledge graph and large model for logical reasoning and complex pattern recognition.
It improves the intelligence level and work efficiency of the power grid scheduling system, reduces manual interference, and realizes the functions of real-time query, fault diagnosis and scheduling suggestions, ensuring the efficient and stable operation of the system in complex environments.
Smart Images
Figure CN120297684A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the technical field of power grid dispatching, and in particular to a power grid dispatching service interface system, method, device, and storage medium. Background Art
[0002] With the continuous expansion of the scale and the increasing complexity of the power system, power grid dispatching faces unprecedented challenges. Traditional dispatching systems, i.e., traditional power grid dispatching systems, usually rely on dispatchers' experience and static rule bases, and have certain limitations in providing dispatching suggestions and real-time queries, unable to meet the requirements of modern power systems for efficient and accurate dispatching. Summary of the Invention
[0003] This application provides a power grid dispatching service interface system, method, device, and storage medium, which are used to improve the technical problem that the existing dispatching system relies on dispatchers' experience and static rule bases to obtain dispatching strategies, with a lot of manual intervention and low efficiency.
[0004] In view of this, the first aspect of this application provides a power grid dispatching service interface system, including:
[0005] An interface module for establishing a connection with the power grid dispatching system; among them, a multi-level caching mechanism and a fault tolerance mechanism are introduced in the interface module;
[0006] A query module for obtaining a query result from the power grid dispatching system with the goal of minimizing the query cost according to the query conditions input by the user; the query result includes the current operating state of the power grid
[0007] A fault diagnosis module for fault identification of the current operating state of the power grid based on a fault identification model of statistical analysis and Bayesian inference to generate a first fault identification result; or, fault identification of the current operating state of the power grid based on a knowledge graph to generate a second fault identification result; or, fault identification of the current operating state of the power grid based on a large model to generate a third fault identification result;
[0008] A dispatching suggestion module for effect evaluation of each dispatching strategy according to the current operating state of the power grid and the fault identification result output by the fault diagnosis module, and selecting the dispatching strategy with the highest effect evaluation value to generate a first dispatching strategy; or, reasoning based on the second fault identification result and the current operating state of the power grid through a knowledge graph to generate a second dispatching strategy; or, predicting a dispatching strategy based on the third fault identification result and the current operating state of the power grid through a large model to generate a third dispatching strategy; or, predicting a dispatching strategy based on the second fault identification result, the third fault identification result, and the current operating state of the power grid by combining a knowledge graph and a large model to generate a fourth dispatching strategy.
[0009] Optionally, when the scheduling recommendation module is used to predict the scheduling strategy according to the second fault identification result, the third fault identification result and the current operating state of the power grid by combining the knowledge graph and the large model, and generate the fourth scheduling strategy, it specifically includes:
[0010] Based on the second fault identification result and the current operating state of the power grid, reasoning is carried out based on the knowledge graph to generate the second scheduling strategy;
[0011] Based on the third fault identification result and the current operating state of the power grid, predict the scheduling strategy based on the large model to generate the third scheduling strategy;
[0012] Perform weighted summation on the second scheduling strategy and the third scheduling strategy to obtain the fourth scheduling strategy.
[0013] Optionally, the effect evaluation formula is:
[0014]
[0015] Among them, is the utility value of the scheduling strategy s, is the benefit of the scheduling strategy s on the index k, is the cost of the scheduling strategy s on the cost item l; and are the weights of the benefit and the cost respectively.
[0016] Optionally, the system further includes:
[0017] A knowledge graph construction module, which is used to perform entity recognition and entity relationship recognition according to the target data of the power grid scheduling system, and construct a knowledge graph based on the recognition results. The target data includes scheduling data, equipment data, and power grid operation state data.
[0018] Optionally, the system further includes:
[0019] An update module, which is used to update the weights of the entities in the knowledge graph according to the new data when new data is input into the power grid scheduling system. The weight update formula is:
[0020]
[0021] Among them, is the weight of the entity after update, is the weight of the entity before update, is the weight of the entity calculated from the new data, is the update coefficient.
[0022] Optionally, the objective function corresponding to minimizing the query cost is:
[0023]
[0024] Among them, is the query condition, N is the number of subtasks involved in the query, is the subtask weight, is the subtask execution time.
[0025] Optionally, the system further includes: a repair suggestion module, configured to generate repair suggestions with the goal of minimizing the repair cost according to the fault identification result, and the objective function corresponding to the minimum repair cost is:
[0026]
[0027] Among them, is the optimal repair strategy, S is the set of all feasible repair strategies, is the cost of the repair strategy s on the target j, is the weight coefficient of the target j.
[0028] Optionally, the query module is further configured to calculate the relevance between the query result and the query condition, and sort the query results based on the relevance and the calculation cost of obtaining the query results from the power grid dispatching system. The sorting formula is:
[0029]
[0030] Among them, represents the k-th item in the query result, represents the relevance between the query result and the query condition, is the calculation cost of obtaining the query results from the power grid dispatching system, and are adjustment weights, satisfying .
[0031] The second aspect of this application provides a power grid dispatching service interface method, which is applied to any one of the power grid dispatching service interface systems in the first aspect. The method includes:
[0032] Establish a connection with the power grid dispatching system through the interface module; among them, a multi-level caching mechanism and a fault tolerance mechanism are introduced in the interface module;
[0033] Obtain query results from the power grid dispatching system with the goal of minimizing the query cost according to the query conditions input by the user through the query module; the query results include the current operating state of the power grid;
[0034] The fault diagnosis module uses a fault identification model based on statistical analysis and Bayesian inference to identify faults in the current operating state of the power grid, generating a first fault identification result; alternatively, uses a knowledge graph to identify faults in the current operating state of the power grid, generating a second fault identification result; alternatively, uses a large model to identify faults in the current operating state of the power grid, generating a third fault identification result;
[0035] The scheduling recommendation module evaluates the effects of each scheduling strategy according to the current operating state of the power grid and the fault identification result output by the fault diagnosis module, and selects the scheduling strategy with the highest effect evaluation value to generate a first scheduling strategy; alternatively, uses the knowledge graph to reason according to the second fault identification result and the current operating state of the power grid, generating a second scheduling strategy; alternatively, uses the large model to predict the scheduling strategy according to the third fault identification result and the current operating state of the power grid, generating a third scheduling strategy; alternatively, combines the knowledge graph and the large model to predict the scheduling strategy according to the second fault identification result, the third fault identification result and the current operating state of the power grid, generating a fourth scheduling strategy.
[0036] The third aspect of this application provides an electronic device, which includes a processor and a memory;
[0037] The memory is used to store program code and transmit the program code to the processor;
[0038] The processor is used to execute the power grid scheduling service interface method described in the second aspect according to the instructions in the program code.
[0039] The fourth aspect of this application provides a computer-readable storage medium, which is used to store program code, and when the program code is executed by a processor, it implements the power grid scheduling service interface method described in the second aspect.
[0040] From the above technical solutions, it can be seen that this application has the following advantages:
[0041] Through the power grid scheduling service interface system provided by this application, functions such as real-time query, fault diagnosis, and scheduling recommendation of the scheduling system can be realized, which helps to improve the intelligent level and working efficiency of the power grid scheduling system and reduces manual intervention;
[0042] Furthermore, by introducing the knowledge graph and the large model, the system is given powerful logical reasoning and complex pattern recognition capabilities. The knowledge graph uses structured data and logical relationships to quickly perform preliminary fault diagnosis and scheduling strategy recommendations; the large model extracts complex features and patterns from a large amount of operating data through deep learning technology to provide more accurate fault prediction and scheduling decisions; this multi-level and multi-dimensional intelligent processing enables the power grid scheduling system to operate efficiently and stably in a more complex operating environment. Brief Description of the Drawings
[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.
[0044] Figure 1 A structural schematic diagram of a power grid dispatching service interface system provided by an embodiment of the present application. Detailed Embodiments
[0045] In order to enable those skilled in the art to better understand the solution of the present application, the following will clearly and completely describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, rather than all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the protection scope of the present application.
[0046] For ease of understanding, please refer to Figure 1 , an embodiment of the present application provides a power grid dispatching service interface system, including:
[0047] An interface module for establishing a connection with the power grid dispatching system; among them, a multi-level caching mechanism and a fault tolerance mechanism are introduced in the interface module;
[0048] A query module for obtaining a query result from the power grid dispatching system with the goal of minimizing the query cost according to the query conditions input by the user; the query result includes the current operating state of the power grid;
[0049] A fault diagnosis module for performing fault identification on the current operating state of the power grid based on a fault identification model of statistical analysis and Bayesian inference to generate a first fault identification result; or, performing fault identification on the current operating state of the power grid based on a knowledge graph to generate a second fault identification result; or, performing fault identification on the current operating state of the power grid based on a large model to generate a third fault identification result;
[0050] The scheduling recommendation module is used to evaluate the effects of each scheduling strategy based on the current operating state of the power grid and the fault identification results output by the fault diagnosis module, and select the scheduling strategy with the highest effect evaluation value to generate the first scheduling strategy; or, perform reasoning through the knowledge graph based on the second fault identification result and the current operating state of the power grid to generate the second scheduling strategy; or, predict the scheduling strategy through the large model based on the third fault identification result and the current operating state of the power grid to generate the third scheduling strategy; or, combine the knowledge graph and the large model to predict the scheduling strategy based on the second fault identification result, the third fault identification result and the current operating state of the power grid to generate the fourth scheduling strategy.
[0051] To seamlessly integrate the interface module into the existing power grid scheduling system, this application adopts standardized data structures and communication protocols. The input and output data formats use standardized JSON or XML formats, and the data fields include device status, historical records, scheduling rules, etc. For each field, its possible value range and default value are defined to ensure data integrity and consistency. The data standardization of the interface is the key to ensuring smooth information exchange between the internal system and external systems.
[0052] The query module is one of the core functions of the interface system of this application, and is used for users to quickly obtain the required information during the scheduling process. The query results need to be returned within a short time and be able to be flexibly adjusted according to different query conditions. The query process involves the retrieval and processing of large-scale data. To improve the query efficiency, this application introduces a query optimization strategy in the query module. The query module obtains the query results from the power grid scheduling system with the goal of minimizing the query cost to improve the query efficiency. The objective function corresponding to minimizing the query cost is:
[0053]
[0054] where is the query condition, N is the number of subtasks involved in the query, is the subtask weight, is the subtask execution time.
[0055] To ensure the relevance of the query results, the query module is also used to calculate the relevance between the query results and the query conditions, and sort the query results based on the relevance and the calculation cost of obtaining the query results from the power grid scheduling system. The sorting formula is:
[0056]
[0057] where represents the k-th item in the query results, represents the relevance between the query results and the query conditions, The computational costs for obtaining query results from the power grid dispatching system, including execution time, resource consumption, etc., can be calculated by the system; and is for adjusting the weights to meet . Among them, the relevance calculation depends on the query conditions and query results. The Euclidean distance can be used to calculate the relevance of data, and the TF-IDF method can be used to estimate the relevance of text. The query conditions refer to the requirements input by the user, including the equipment status, historical dispatching data, operating status, etc. of the power grid. The query results include the operating status of the current power grid, fault records, the working status of equipment, etc.
[0058] To ensure the efficient operation of the interface module under large-scale concurrent requests, this application introduces performance optimization and fault tolerance mechanisms in the interface design.
[0059] To reduce the latency problem caused by frequent queries, this application designs a multi-level caching mechanism in the interface module. The selection of the cache is based on the access frequency and timeliness of the data. The cache hit rate formula is as follows:
[0060]
[0061] In the formula, Cache Hits is the number of cache hits, and Total Requests is the total number of accesses.
[0062] When processing large-scale query requests, the interface system decomposes the query tasks into several subtasks for parallel execution. The efficiency of parallel processing is evaluated by the following method:
[0063]
[0064] Among them, is the serial execution time, is the parallel execution time.
[0065] To ensure the stable operation of the interface system under various abnormal conditions, this application designs a fault tolerance mechanism in the interface module, including a retry policy and a failover mechanism. The design of the retry policy is based on the exponential backoff algorithm, and its waiting time is as follows:
[0066]
[0067] Among them, is the waiting time for the nth retry, is the base waiting time.
[0068] The fault diagnosis module is used to analyze the operating status of the power grid in real time (such as real-time operating parameters of voltage, current, power, frequency, etc. at each node of the power grid) and identify possible faults.
[0069] In one embodiment, the fault diagnosis module may use a fault identification model based on statistical analysis and Bayesian inference to identify faults and obtain a first fault identification result. The fault identification model is as follows:
[0070]
[0071] where is the probability of fault F occurring under the observed power grid state E, is the probability of observing state when fault occurs, is the prior probability of the fault, is the marginal probability of observing state .
[0072] In another embodiment, the fault diagnosis module identifies faults in the current operating state of the power grid based on a knowledge graph and generates a second fault identification result. The fault diagnosis module uses the knowledge graph for preliminary reasoning. The knowledge graph can perform logical reasoning in structured data by defining entities, relationships, and rules. Specifically, the knowledge graph can infer possible fault patterns based on the current operating state of the power grid, historical data, and rules defined by experts.
[0073] Furthermore, the system in this application further includes a knowledge graph construction module, which is used to perform entity recognition, entity relationship recognition, and attribute recognition based on the target data of the power grid dispatching system, and construct a knowledge graph based on the recognition results. The target data includes dispatching data, equipment data, and power grid operating state data.
[0074] The knowledge graph construction module constructs a power grid dispatching knowledge graph based on rich data sources of the power grid dispatching system (such as historical dispatching data, equipment archives, fault records, dispatching rules, etc.). The knowledge graph construction module quantifies the knowledge graph using the following steps:
[0075] (1) Entity weight calculation: The weight of each entity in the knowledge graph is determined by its frequency and importance , and the calculation formula is:
[0076]
[0077] where is a constant to prevent the denominator from being zero, usually taking a very small positive value.
[0078] (2) Relationship strength calculation: For each pair of entities and in the knowledge graph, its relationship strength The calculation formula is:
[0079]
[0080] Among them, represents the shortest path distance between entities and in the knowledge graph. The relationship strength is used to measure the connection strength between two entities, reflecting the logic between them. The greater the strength, the closer the association between the two entities; it helps to infer the power grid state, identify fault patterns, and generate dispatching suggestions through the relationships between entities in the graph; when analyzing the power grid state, entities with greater relationship strength are given priority, enhancing the prediction ability of the knowledge graph.
[0081] Furthermore, the system also includes a weight update module, which is used to update the weights of entities in the knowledge graph according to new data when new data is input into the power grid dispatching system. The weight update formula is:
[0082]
[0083] Among them, is the updated weight of entity , is the weight of entity before update, is the weight of entity calculated from the new data, is the update coefficient, .
[0084] In another embodiment, the fault diagnosis module performs fault identification on the current operating state of the power grid based on a large model, generating a third fault identification result. The large model can provide more complex pattern recognition and prediction functions by learning the big data of the power grid operating state.
[0085] Furthermore, the system also includes a large model construction module, which is used to train a convolutional neural network according to the target data of the power grid dispatching system to obtain the large model.
[0086] The training of the large model depends on a large amount of power grid historical data, including operation records, fault logs, operating state data, etc. To ensure the accuracy and effectiveness of model training, the large model construction module first needs to preprocess the original data, including:
[0087] (1) Data cleaning: The original data may contain missing values, outliers, or noise data, so data cleaning is required. Data cleaning methods include deleting missing values, filling in missing data, and removing obvious outliers.
[0088] (2)Feature Processing: Feature engineering is an important step in improving the performance of large models. Features can be the direct use of raw data or generated by transforming, combining, or aggregating raw data. For example, for fault log data, features related to faults can be extracted through time series analysis.
[0089] (3)Data Normalization: Data for different features may have different dimensions and ranges. To enable effective learning by large models, it is usually necessary to normalize the data.
[0090] The architecture design of large models is the core of their performance. To capture the complex relationships and patterns in power grid operation, this application preferably adopts the architecture of a multi-layer neural network (DNN), which consists of several layers of linear transformations and non-linear activation functions, specifically including:
[0091] Input Layer: The input layer receives the preprocessed feature data, and each input node represents a feature. Assuming the dimension of the input data is d, the input layer will have d nodes;
[0092] Hidden Layer: The hidden layer is a key component of the model and is used to learn the complex patterns in the data. Each node in the hidden layer is connected to the nodes in the previous layer through a weighted sum activation function;
[0093] Output Layer: The output layer generates the final prediction result according to the specific tasks of power grid scheduling. For classification tasks, the output layer usually uses the softmax function for multi-classification, and the formula is:
[0094]
[0095] where, is the predicted probability of class i, is the activation value of the i-th node in the output layer, and C is the total number of classes.
[0096] The loss function is used to measure the difference between the model's prediction result and the true result and is the core indicator for model optimization. In the training process of this large model, the cross-entropy loss function is used, and the loss function is defined as the cross-entropy loss:
[0097]
[0098] where, is the true label of sample , is the probability predicted by the model, is the total number of samples. The cross-entropy loss function optimizes the model parameters by measuring the uncertainty between the model's predicted probability and the true label.
[0099] To make the prediction results of the large model more accurate, the model is optimized by adjusting the model parameters to minimize the value of the loss function. This application uses the gradient descent algorithm for model optimization. The basic idea of gradient descent is to calculate the gradient of the loss function with respect to the model parameters and adjust the parameters in the opposite direction of the gradient to gradually reduce the value of the loss function. The parameter update formula is as follows:
[0100]
[0101] where, is the model parameter at the t-th iteration, is the learning rate, is the gradient of the loss function L with respect to the parameter.
[0102] To prevent the model from overfitting, this application adds a regularization term to the loss function. The introduction of the regularization term can effectively constrain the magnitude of the model parameters and prevent the model from performing well on the training data but poorly on the test data. The formula for the regularization term is as follows:
[0103]
[0104] where, is the regularization coefficient, and m is the total number of model parameters.
[0105] Furthermore, the system also includes a repair suggestion module. The repair suggestion module generates repair suggestions based on the fault identification results with the goal of minimizing the repair cost. The objective function corresponding to the minimum repair cost is:
[0106]
[0107] where, is the optimal repair strategy, S is the set of all feasible repair strategies, is the cost of repair strategy s on objective j, is the weight coefficient of objective j. The repair suggestion module analyzes the fault identification results to determine the possible fault types and impact scopes; selects a repair plan with the lowest cost (such as the least impact on grid equipment, the shortest fault recovery time, the lowest repair cost, etc.) according to the operating state of the power grid to restore the normal operation of the power grid; and reasonably arranges the repair steps according to the urgency of the power grid and the importance of the equipment; when determining the repair strategy, priority is given to those repair plans with less impact on other parts of the power grid.
[0108] The scheduling suggestion module is responsible for generating optimized scheduling operation suggestions during the operation of the power grid according to the current operating state of the power grid and the fault identification results.
[0109] In one embodiment, the scheduling recommendation module is specifically configured to evaluate the effects of each scheduling strategy based on the current operating state of the power grid and the fault identification result, and select the scheduling strategy with the highest effect evaluation value to generate the first scheduling strategy; wherein, the effect evaluation formula is:
[0110]
[0111] Wherein, is the utility value of the scheduling strategy s, is the benefit of the scheduling strategy s on the index k, is the cost of the scheduling strategy s on the cost item l; and are the weights of the benefit and the cost respectively.
[0112] The index is used to evaluate different performance criteria for the effect of the scheduling strategy, including power grid load stability, system operation efficiency, fault recovery time, power loss, etc.; the cost item is various costs involved in implementing the scheduling strategy. It includes power supply cost, equipment maintenance cost, fault repair cost, time cost of system operation, implicit costs such as system instability caused by scheduling operations, etc.
[0113] The benefit calculation depends on the index and is calculated according to different indexes. Benefit(s, load stability) = amplitude of load fluctuation before scheduling - amplitude of load fluctuation after scheduling; Benefit(s, system efficiency) = energy loss before scheduling - energy loss after scheduling; Benefit(s, fault recovery time) = recovery time before scheduling - recovery time after scheduling.
[0114] The cost depends on the cost item: Cost(s, power supply cost) = power required by the scheduling strategy × unit price of power purchase; Cost(s, equipment maintenance cost) = number of equipment maintenance times × cost per maintenance; Cost(s, fault repair cost) = repair time required × cost per unit time + cost of replacing equipment; Cost(s, time cost of system operation) = additional operation time × cost per unit time consumption; Cost(s, system instability cost) = number of faults caused by instability × cost per fault.
[0115] In another embodiment, based on the second fault identification result and the current operating state of the power grid, reasoning is performed based on the knowledge graph to generate the second scheduling strategy.
[0116] In yet another embodiment, based on the third fault identification result and the current operating state of the power grid, scheduling strategy prediction is performed based on the large model to generate the third scheduling strategy.
[0117] In another embodiment, based on the second fault identification result and the current operating state of the power grid, reasoning is performed based on the knowledge graph to generate a second scheduling strategy; based on the third fault identification result and the current operating state of the power grid, scheduling strategy prediction is performed based on the large model to generate a third scheduling strategy; the second scheduling strategy and the third scheduling strategy are weighted and summed to obtain a fourth scheduling strategy.
[0118] In this embodiment, the system first performs preliminary reasoning using the knowledge graph. The knowledge graph can infer possible fault patterns and scheduling strategies based on the current state of the power grid, historical data, and rules defined by experts. Secondly, the system inputs the preliminary reasoning results of the knowledge graph into the large model for further in-depth analysis. By learning the big data of the power grid operating state, the large model can provide more complex pattern recognition and prediction functions. The in-depth analysis process of the large model includes steps such as feature extraction, pattern recognition, and result generation. To make full use of the respective advantages of the knowledge graph and the large model, the system adopts a joint reasoning strategy to fuse the reasoning results of the two to generate the final decision-making suggestions. This fusion process is carried out by weighted summation, which can flexibly adjust the proportion of the two methods in the final decision. Among them, the result of joint reasoning is calculated using the following method:
[0119]
[0120] Among them, R is the final scheduling decision result, R KG is the scheduling strategy inferred by the knowledge graph, R LM is the scheduling strategy predicted by the large model, is the fusion weight coefficient, . This weight coefficient can be dynamically adjusted according to the specific requirements of the system to optimize the decision-making effect in different scenarios.
[0121] By constructing the knowledge graph of power grid equipment, the system realizes real-time monitoring and logical reasoning of equipment status; combined with the trained large model, the system can accurately predict the possibility of faults and generate the optimal scheduling suggestions in actual operation; when a certain equipment anomaly is detected, the system quickly identifies the fault type through the fusion strategy and proposes corresponding repair solutions, ultimately greatly reducing the impact of faults and ensuring the stable operation of the power grid.
[0122] To enable the system to maintain efficient operation in a changing power grid environment, the update module in the system of this application is specifically used to update the knowledge graph and the large model according to user feedback and real-time data. This adaptive adjustment mechanism enables the system to continuously optimize and adapt to new power grid operating states and external environment changes. After receiving user feedback, the system will adaptively adjust the knowledge graph and the large model through the following formula:
[0123]
[0124] Among them, is the weight after the update of the knowledge graph or the large model, is the current weight of the knowledge graph or the large model, is the weight change caused by the feedback, is the feedback update coefficient, .
[0125] When new data is input into the power grid dispatching system, the update process of the network parameters (weights) of the large model is similar to the process of updating the weights of the knowledge graph according to the new data, which will not be elaborated here.
[0126] This application endows the system with powerful logical reasoning and complex pattern recognition capabilities by introducing a knowledge graph and a large model; the knowledge graph can quickly perform preliminary fault diagnosis and dispatching strategy suggestions by using structured data and logical relationships; the large model extracts complex features and patterns from a large amount of operation data through deep learning technology to provide more accurate fault prediction and dispatching decisions; this multi-level and multi-dimensional intelligent processing enables the power grid dispatching system to operate efficiently and stably in a more complex operating environment;
[0127] Through the design and implementation of the interface, the functions of the knowledge graph and the large model can be efficiently utilized. The interface design adopts a standardized data format and communication protocol, ensuring the compatibility and scalability of the system. Users can implement functions such as real-time query, fault diagnosis, and dispatching suggestions of the dispatching system through the interface, which greatly improves the usability and practicality of the system. At the same time, the efficiency and stability of the interface enable the system to maintain good performance under large-scale concurrent requests, meeting the strict requirements of the power grid dispatching system for real-time and reliability.
[0128] Through joint reasoning, the system can comprehensively consider structured logical information and unstructured complex data, thereby generating more accurate and comprehensive decision-making suggestions. This fusion strategy not only improves the application effect of a single technology but also ensures that the system can continuously optimize its performance with the change of the power grid operating environment through an adaptive adjustment mechanism.
[0129] This application embodiment also provides a power grid dispatching service interface method, which is applied to the power grid dispatching service interface system in the foregoing system embodiment. The method includes:
[0130] S1. Establish a connection with the power grid dispatching system through the interface module; among them, a multi-level caching mechanism and a fault tolerance mechanism are introduced in the interface module;
[0131] S2. The query module obtains query results from the power grid dispatching system according to the query conditions input by the user with the goal of minimizing the query cost; the query results include the current operating state of the power grid.
[0132] S3. The fault diagnosis module performs fault identification on the current operating state of the power grid based on a fault identification model of statistical analysis and Bayesian inference to generate a first fault identification result; or, performs fault identification on the current operating state of the power grid based on a knowledge graph to generate a second fault identification result; or, performs fault identification on the current operating state of the power grid based on a large model to generate a third fault identification result.
[0133] S4. The dispatching recommendation module evaluates the effects of each dispatching strategy according to the current operating state of the power grid and the fault identification result output by the fault diagnosis module, and selects the dispatching strategy with the highest effect evaluation value to generate a first dispatching strategy; or, the knowledge graph performs reasoning according to the second fault identification result and the current operating state of the power grid to generate a second dispatching strategy; or, the large model predicts the dispatching strategy according to the third fault identification result and the current operating state of the power grid to generate a third dispatching strategy; or, combines the knowledge graph and the large model to predict the dispatching strategy according to the second fault identification result, the third fault identification result and the current operating state of the power grid to generate a fourth dispatching strategy.
[0134] Through the power grid dispatching service interface system provided by this application, functions such as real-time query, fault diagnosis, and dispatching recommendation of the dispatching system can be realized, which helps to improve the intelligent level and working efficiency of the power grid dispatching system and reduces manual interference.
[0135] Furthermore, by introducing the knowledge graph and the large model, the system is given powerful logical reasoning and complex pattern recognition capabilities. The knowledge graph uses structured data and logical relationships to quickly perform preliminary fault diagnosis and dispatching strategy suggestions; the large model extracts complex features and patterns from massive operation data through deep learning technology to provide more accurate fault prediction and dispatching decisions; this multi-level and multi-dimensional intelligent processing enables the power grid dispatching system to maintain efficient and stable operation in a more complex operating environment.
[0136] The embodiment of this application also provides an electronic device, which includes a processor and a memory.
[0137] The memory is used to store program code and transmit the program code to the processor.
[0138] The processor is used to execute the power grid dispatching service interface method in the foregoing method embodiment according to the instructions in the program code.
[0139] The embodiment of the present application also provides a computer-readable storage medium, which is used to store program codes. When the program codes are executed by a processor, the power grid dispatching service interface method in the foregoing method embodiment is implemented.
[0140] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific processes of the methods described above can refer to the corresponding processes in the foregoing system embodiments, and will not be elaborated herein.
[0141] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not have to be used to describe a specific order or sequence. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device that includes a series of steps or units does not have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.
[0142] As mentioned above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims
1. A power grid dispatching service interface system, characterized in that, Including: An interface module for establishing a connection with the power grid dispatching system; among them, a multi-level caching mechanism and a fault tolerance mechanism are introduced in the interface module; A query module for obtaining a query result from the power grid dispatching system with the goal of minimizing the query cost according to the query conditions input by the user; the query result includes the current operating state of the power grid; A fault diagnosis module for performing fault identification on the current operating state of the power grid based on a fault identification model of statistical analysis and Bayesian inference to generate a first fault identification result; or, performing fault identification on the current operating state of the power grid based on a knowledge graph to generate a second fault identification result; or, performing fault identification on the current operating state of the power grid based on a large model to generate a third fault identification result; A scheduling recommendation module for evaluating the effects of each scheduling strategy according to the current operating state of the power grid and the fault identification result output by the fault diagnosis module, and selecting the scheduling strategy with the highest effect evaluation value to generate a first scheduling strategy; or, reasoning through the knowledge graph based on the second fault identification result and the current operating state of the power grid to generate a second scheduling strategy; or, predicting the scheduling strategy through the large model based on the third fault identification result and the current operating state of the power grid to generate a third scheduling strategy; or, combining the knowledge graph and the large model to predict the scheduling strategy based on the second fault identification result, the third fault identification result and the current operating state of the power grid to generate a fourth scheduling strategy.
2. The power grid dispatching service interface system according to claim 1, wherein When the scheduling recommendation module is used to combine the knowledge graph and the large model to predict the scheduling strategy based on the second fault identification result, the third fault identification result and the current operating state of the power grid to generate a fourth scheduling strategy, it specifically includes: Reasoning through the knowledge graph based on the second fault identification result and the current operating state of the power grid to generate a second scheduling strategy; Predicting the scheduling strategy through the large model based on the third fault identification result and the current operating state of the power grid to generate a third scheduling strategy; Performing a weighted sum on the second scheduling strategy and the third scheduling strategy to obtain a fourth scheduling strategy.
3. The power grid dispatching service interface system according to claim 1, characterized in that The effect evaluation formula is: Among them, is the utility value of the scheduling policy s, is the revenue of the scheduling policy s on the metric k, is the cost of the scheduling policy s on the cost item l; and are the weights of the revenue and the cost respectively.
4. The power grid dispatching service interface system according to claim 1, wherein The system further includes: A knowledge graph construction module for performing entity recognition and entity relationship recognition on the target data of the power grid dispatching system, and constructing a knowledge graph based on the recognition results, where the target data includes scheduling data, equipment data and power grid operating state data.
5. The power grid dispatching service interface system according to claim 1, characterized in that The system further includes: An update module for updating the weights of entities in the knowledge graph according to the new data when new data is input into the power grid dispatching system, and the weight update formula is: Among them, is the entity The updated weight, is the entity The weight before update, is the weight of the entity calculated from the new data and is the update coefficient.
6. The power grid dispatching service interface system according to claim 1, characterized in that, The objective function corresponding to minimizing the query cost is: Among them, is the query condition, N is the number of subtasks involved in the query, is the subtask weight, is the subtask execution time.
7. The power grid dispatching service interface system according to claim 1, characterized in that The system further includes: a repair recommendation module for generating repair recommendations with the goal of minimizing the repair cost according to the fault identification result, and the objective function corresponding to the minimum repair cost is: Among them, is the optimal repair strategy, S is the set of all feasible repair strategies, is the cost of the repair strategy s on the target j, is the weight coefficient of the target j.
8. The power grid dispatching service interface system according to claim 1, characterized in that The query module is further used to calculate the relevance between the query result and the query conditions, and sort the query results based on the relevance and the calculation cost of obtaining the query result from the power grid dispatching system, and the sorting formula is: Among them, represents the k-th item in the query result, represents the relevance between the query result and the query condition, is the calculation cost of obtaining the query result from the power grid dispatching system, and are the adjusted weights, satisfying .
9. A power grid dispatching service interface method, characterized in that, Applied to the power grid dispatching service interface system described in any one of claims 1-7, the method includes: Establishing a connection with the power grid dispatching system through the interface module; among them, a multi-level caching mechanism and a fault tolerance mechanism are introduced in the interface module; The query module obtains query results from the power grid dispatching system according to the query conditions input by the user with the goal of minimizing the query cost; the query results include the current operating status of the power grid. The fault diagnosis module performs fault identification on the current operating status of the power grid through a fault identification model based on statistical analysis and Bayesian inference to generate a first fault identification result; or, performs fault identification on the current operating status of the power grid based on a knowledge graph to generate a second fault identification result; or, performs fault identification on the current operating status of the power grid based on a large model to generate a third fault identification result. The dispatching recommendation module evaluates the effects of each dispatching strategy according to the current operating status of the power grid and the fault identification results output by the fault diagnosis module, and selects the dispatching strategy with the highest effect evaluation value to generate a first dispatching strategy; or, the knowledge graph performs reasoning according to the second fault identification result and the current operating status of the power grid to generate a second dispatching strategy; or, the large model predicts the dispatching strategy according to the third fault identification result and the current operating status of the power grid to generate a third dispatching strategy; or, combines the knowledge graph and the large model to predict the dispatching strategy according to the second fault identification result, the third fault identification result and the current operating status of the power grid to generate a fourth dispatching strategy.
10. An electronic device, characterized in that, The device includes a processor and a memory. The memory is used to store program code and transmit the program code to the processor. The processor is used to execute the power grid dispatching service interface method described in claim 9 according to the instructions in the program code.
11. A computer-readable storage medium, characterized in that, The computer-readable storage medium is used to store program code, and when the program code is executed by the processor, it implements the power grid dispatching service interface method described in claim 9.