Pushing method for direct current protection defect disposal strategy

Abstract

The invention relates to the technical field of electric power, and discloses a direct current protection defect handling strategy pushing method, which comprises the following steps of: acquiring direct current protection defect data and corresponding demand knowledge information and query statements in a current scene, and utilizing a preset direct current protection defect knowledge graph to push a plurality of knowledge handling strategies to the direct current protection defect data, determining a push knowledge set and a plurality of knowledge query results, obtaining the knowledge importance of each knowledge disposal strategy in a preset knowledge importance prediction model, and sorting the knowledge importance to determine knowledge push information of the DC protection defect disposal strategy; according to the method, multiple factors can be comprehensively considered by using the model, so that the importance of the knowledge in different scenes can be evaluated more accurately, and the knowledge pushing information with higher pertinence and practicability can be provided for workers, so that the workers can quickly obtain the most critical and valuable knowledge in the current scene, and the working efficiency is improved. And the efficiency of direct current protection defect disposal work and the scientificity of decision making are effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a method for pushing a direct current protection defect handling strategy. Background Art

[0002] During DC protection system operation, the intelligent generation and precise delivery of defect handling strategies are crucial for ensuring stable power system operation. Delivering DC protection defect handling strategies is a complex undertaking involving multiple steps and factors. Its core goal is to provide personnel with timely, accurate, and scenario-specific knowledge and information to assist them in efficiently handling DC protection defect issues.

[0003] Currently, existing push methods primarily rely on user interests and logs to push knowledge, but lack effective utilization of contextual information. DC protection defect resolution scenarios are complex and diverse, and the knowledge required and prioritized by personnel in different scenarios vary. Existing push methods fail to accurately assess the importance of knowledge resolution strategies based on their actual application value in different scenarios. As a result, the pushed knowledge may not be the most urgently needed or relevant content for personnel in the current scenario, hindering their effective acquisition and application of DC protection defect resolution strategy knowledge. Summary of the Invention

[0004] Based on this, it is necessary to propose a push method for DC protection defect handling strategies to address the above problems. That is, it can use the model to comprehensively consider multiple factors, so as to more accurately evaluate the importance of knowledge in different scenarios. This method can provide staff with more targeted and practical knowledge push information, enabling staff to quickly obtain the most critical and valuable knowledge in the current scenario, and effectively improve the efficiency of DC protection defect handling work and the scientific nature of decision-making.

[0005] To achieve the above-mentioned object, the present invention provides, in a first aspect, a method for pushing a DC protection defect handling strategy, the method comprising:

[0006] Acquire DC protection defect data in a current scenario, as well as required knowledge information and query statements corresponding to the DC protection defect data, wherein the DC protection defect data includes multiple knowledge processing strategies;

[0007] Determine a push knowledge set based on the required knowledge information and a preset DC protection defect knowledge graph, and use the query statement to query the preset DC protection defect knowledge graph to obtain multiple knowledge query results;

[0008] Input all knowledge disposal strategies into the preset knowledge importance prediction model to obtain the knowledge importance of each knowledge disposal strategy;

[0009] Using the knowledge importance of all knowledge disposal strategies, multiple knowledge query results are ranked to determine a target knowledge query result;

[0010] Determine knowledge push information of a DC protection defect handling strategy based on the target knowledge query result and the pushed knowledge set.

[0011] Optionally, the method further includes:

[0012] Acquiring historical DC protection defect data in historical scenarios, wherein the historical DC protection defect data includes a plurality of historical knowledge processing strategies;

[0013] Determine the knowledge importance of each historical knowledge disposal strategy;

[0014] The knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies are input into the initial BPNN model for training to obtain the preset knowledge importance prediction model.

[0015] Optionally, the knowledge importance of all historical knowledge handling strategies and all historical knowledge handling strategies are input into the initial BPNN model for training to obtain the preset knowledge importance prediction model, including:

[0016] Using SSA or ISSA, according to the initial BPNN model, determine an SSA-based BPNN model or an ISSA-based BPNN model;

[0017] The knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies are input into the SSA-based BPNN model or the ISSA-based BPNN model for training to obtain the preset knowledge importance prediction model.

[0018] Optionally, the using SSA or ISSA to determine an SSA-based BPNN model or an ISSA-based BPNN model according to the initial BPNN model includes:

[0019] Using the SSA or the ISSA, iteratively updating the position of each sparrow in the sparrow population until the global optimal fitness value of all iterations converges to a convergence threshold, or the number of iterations corresponding to the current iteration is equal to a preset maximum number of iterations, and obtaining the optimal model parameters corresponding to the global optimal fitness value of all sparrows in all iterations, wherein the position of each sparrow includes the model parameters;

[0020] According to the optimal model parameters corresponding to all iterations, the model parameters in the initial BPNN model are adjusted to obtain the SSA-based BPNN model or the ISSA-based BPNN model.

[0021] Optionally, under the condition of using the ISSA, the method further includes:

[0022] Determining the knowledge importance of each predicted knowledge disposal strategy obtained during the validation process of the ISSA-based BPNN model;

[0023] Determine a loss error value according to the knowledge importance of all prediction knowledge disposal strategies and the knowledge importance of all historical knowledge disposal strategies;

[0024] When the loss error value does not meet the preset error value, determining a target weight value according to the current number of iterations and the preset maximum number of iterations;

[0025] Adjusting the adaptive weight value of the ISSA according to the target weight value to obtain an adjusted ISSA;

[0026] The adjusted ISSA is used as the ISSA, and the ISSA is returned to be used to iteratively update the position of each sparrow in the sparrow population until the global optimal fitness value of all iterations converges to a convergence threshold, or the number of iterations corresponding to the current iteration is equal to the preset maximum number of iterations, and the optimal model parameters corresponding to the global optimal fitness value of all sparrows in all iterations are obtained, until the loss error value meets the preset error value, and the ISSA-based BPNN model corresponding to the loss error value meeting the preset error value is used as the final ISSA-based BPNN model.

[0027] Optionally, determining the target weight value according to the current number of iterations and the preset maximum number of iterations includes:

[0028] Using the formula Determining the target weight value;

[0029] Wherein, w is the target weight value, f randn is a random number that obeys the normal distribution, N is the preset maximum number of iterations, and n is the number of iterations at that time.

[0030] Optionally, determining the pushed knowledge set according to the demand knowledge information and a preset DC protection defect knowledge graph includes:

[0031] determining a required knowledge category according to the required knowledge information;

[0032] Determining multiple candidate knowledge sets according to the required knowledge category and the preset DC protection defect knowledge graph;

[0033] The pushed knowledge set is determined according to a plurality of candidate knowledge sets.

[0034] Optionally, determining the pushed knowledge set based on multiple candidate knowledge sets includes:

[0035] Determining the similarity between each candidate knowledge set and the required knowledge information;

[0036] The candidate knowledge set corresponding to the maximum similarity is used as the pushed knowledge set.

[0037] Optionally, the step of ranking the plurality of knowledge query results using the knowledge importance of all knowledge handling strategies to determine a target knowledge query result includes:

[0038] Using the knowledge importance of all knowledge disposal strategies, multiple knowledge query results are sorted in descending order to obtain a descending sort result;

[0039] The first preset knowledge query results in the descending sorting results are used as the target knowledge query results.

[0040] Optionally, determining the knowledge push information of the DC protection defect handling strategy according to the target knowledge query result and the pushed knowledge set includes:

[0041] The union of the target knowledge query result and the pushed knowledge set is used as the knowledge push information.

[0042] To achieve the above-mentioned object, the present invention provides, in a second aspect, a device for pushing a DC protection defect handling strategy, the device comprising:

[0043] an acquisition module, configured to acquire DC protection defect data in a current scenario, as well as required knowledge information and query statements corresponding to the DC protection defect data, wherein the DC protection defect data includes a plurality of knowledge handling strategies;

[0044] a determination and query module, configured to determine a push knowledge set based on the required knowledge information and a preset DC protection defect knowledge graph, and to use the query statement to query the preset DC protection defect knowledge graph to obtain a plurality of knowledge query results;

[0045] The model prediction module is used to input all knowledge disposal strategies into the preset knowledge importance prediction model to obtain the knowledge importance of each knowledge disposal strategy;

[0046] a ranking module, configured to use the knowledge importance of all knowledge disposal strategies to rank multiple knowledge query results to determine a target knowledge query result;

[0047] The determination module is used to determine the knowledge push information of the DC protection defect handling strategy according to the target knowledge query result and the pushed knowledge set.

[0048] To achieve the above-mentioned object, the present invention provides, in a third aspect, a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the method as described in any one of the first aspects.

[0049] To achieve the above-mentioned objectives, the present invention provides a computer device in a fourth aspect, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the method as described in any one of the first aspects.

[0050] The embodiment of the present invention has the following beneficial effects: the above method obtains DC protection defect data in the current scenario, as well as the required knowledge information and query statements corresponding to the DC protection defect data, wherein the DC protection defect data includes multiple knowledge handling strategies, and determines a push knowledge set based on the required knowledge information and a preset DC protection defect knowledge graph, and uses a query statement to query in the preset DC protection defect knowledge graph to obtain multiple knowledge query results, and then inputs all knowledge handling strategies into a preset knowledge importance prediction model to obtain the knowledge importance of each knowledge handling strategy, and then uses the knowledge importance of all knowledge handling strategies to sort the multiple knowledge query results to determine the target knowledge query result, and finally determines the knowledge push information of the DC protection defect handling strategy based on the target knowledge query result and the pushed knowledge set; that is, by inputting the knowledge handling strategy into the preset knowledge importance prediction model to determine the knowledge importance, the model can be used to comprehensively consider multiple factors, thereby more accurately evaluating the importance of knowledge in different scenarios. This method can provide staff with more targeted and practical knowledge push information, enabling staff to quickly obtain the most critical and valuable knowledge in the current scenario, effectively improving the efficiency of DC protection defect handling work and the scientific nature of decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] in:

[0053] Figure 1A schematic diagram of a method for pushing a DC protection defect handling strategy in an embodiment of the present application;

[0054] Figure 2 A schematic diagram of a push device for a DC protection defect handling strategy according to an embodiment of the present application;

[0055] Figure 3 1 is a diagram of the internal structure of a computer device in some embodiments. DETAILED DESCRIPTION

[0056] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0057] During DC protection system operation, the intelligent generation and precise delivery of defect handling strategies are crucial for ensuring stable power system operation. Delivering DC protection defect handling strategies is a complex undertaking involving multiple steps and factors. Its core goal is to provide personnel with timely, accurate, and scenario-specific knowledge and information to assist them in efficiently handling DC protection defect issues.

[0058] Currently, existing push methods primarily rely on user interests and logs to push knowledge, but lack effective utilization of contextual information. DC protection defect resolution scenarios are complex and diverse, and the knowledge required and prioritized by personnel in different scenarios vary. Existing push methods fail to accurately assess the importance of knowledge resolution strategies based on their actual application value in different scenarios. As a result, the pushed knowledge may not be the most urgently needed or relevant content for personnel in the current scenario, hindering their effective acquisition and application of DC protection defect resolution strategy knowledge.

[0059] In response to the above problems, this application proposes a method for pushing DC protection defect handling strategies, which can use a model to comprehensively consider multiple factors, so as to more accurately evaluate the importance of knowledge in different scenarios. This method can provide staff with more targeted and practical knowledge push information, enabling staff to quickly obtain the most critical and valuable knowledge in the current scenario, effectively improving the efficiency of DC protection defect handling and the scientific nature of decision-making. The specific implementation principles will be described in detail in the following embodiments.

[0060] In a first aspect, the present application provides a method for pushing a DC protection defect handling strategy.

[0061] See also Figure 1, is a schematic diagram of a method for pushing a DC protection defect handling strategy in an embodiment of the present application, the method comprising:

[0062] Step 110: obtaining DC protection defect data in the current scenario, as well as required knowledge information and query statements corresponding to the DC protection defect data, wherein the DC protection defect data includes multiple knowledge handling strategies.

[0063] Regarding the method of obtaining DC protection defect data, in some embodiments, after a fault corresponding to DC protection occurs in the power system, the relevant devices in the power system will generate some defect text data and electrical quantity data corresponding to the fault. Based on the defect text data and electrical quantity data, as well as the preset DC protection defect data elements, the DC protection defect data in the current scenario can be determined; wherein, the preset DC protection defect data elements can be obtained and pre-set by the operator based on a lot of experience, experiments or statistics, and of course, can also be set by the operator according to actual needs.

[0064] Regarding the information contained in the preset DC protection defect data element, in some embodiments, the preset DC protection defect data may include but is not limited to rule data and expert database, etc.; wherein the rule data includes data manuals, protection rules and operation rules, etc.

[0065] Regarding the required knowledge information and query statements, in some embodiments, the required knowledge information and query statements required by the user in the current scenario can be determined based on the DC protection defect data in the current scenario.

[0066] Regarding the type of query statement, in some embodiments, the query statement may be a Cypher language query statement.

[0067] Step 120: Determine a pushed knowledge set based on the required knowledge information and the preset DC protection defect knowledge graph, and use a query statement to query the preset DC protection defect knowledge graph to obtain multiple knowledge query results.

[0068] Among them, the preset DC protection defect knowledge graph can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics. Of course, it can also be set by the operator according to actual needs.

[0069] Regarding the method for determining the pushed knowledge set, in some embodiments, based on the required knowledge information, it can be extracted from the preset DC protection defect knowledge graph to obtain multiple alternative knowledge sets, and from the alternative knowledge sets, the alternative knowledge set closest to the required knowledge information is selected and used as the pushed knowledge set.

[0070] In this application, by constructing a knowledge graph in the field of DC protection defects and accurately matching the push knowledge set required in the current scenario, it can more accurately meet the knowledge needs of staff in different scenarios and significantly optimize the efficiency and accuracy of knowledge push.

[0071] Step 130: Input all knowledge handling strategies into a preset knowledge importance prediction model to obtain the knowledge importance of each knowledge handling strategy.

[0072] The preset knowledge importance prediction model here refers to a pre-trained model used to predict the knowledge importance corresponding to the output knowledge disposal strategy based on the input knowledge disposal strategy.

[0073] Step 140: Use the knowledge importance of all knowledge handling strategies to sort the multiple knowledge query results to determine the target knowledge query result.

[0074] It should be noted that since the query statement has a corresponding relationship with the DC protection defect data, and the DC protection defect data includes multiple knowledge handling strategies, the knowledge handling strategy and the knowledge query result also have a corresponding relationship. Therefore, in some embodiments, the knowledge importance of all knowledge handling strategies can be directly used to sort multiple knowledge query results to determine the target knowledge query result.

[0075] Furthermore, in some embodiments, if the knowledge handling strategy and the knowledge query result do not have a corresponding relationship, the corresponding relationship between the knowledge handling strategy and the knowledge query result can be determined based on the correlation or similarity between each knowledge handling strategy and each knowledge query result; wherein, the knowledge handling strategy and knowledge query result corresponding to the maximum correlation or similarity can be regarded as the knowledge handling strategy and knowledge query result having a corresponding relationship.

[0076] Regarding the method of determining the target knowledge query results, in some embodiments, the knowledge query results can be sorted from large to small according to the knowledge importance of the knowledge disposal strategy corresponding to each knowledge query result, and the sorted knowledge query results can be used as the target knowledge query results; in other embodiments, the preset number of knowledge query results can be used as the target knowledge query results among the sorted knowledge query results; wherein, the specific number of the preset number can be set by the operator according to actual needs, for example, the preset number can be set to 10, or the preset number can be set to 1, and this application does not limit it here.

[0077] In this application, by sorting the knowledge query results according to the knowledge importance of the knowledge handling strategy, the knowledge overload phenomenon can be effectively avoided and the effectiveness of knowledge transfer can be improved.

[0078] Step 150: Determine knowledge push information of the DC protection defect handling strategy based on the target knowledge query result and the pushed knowledge set.

[0079] Regarding the method of determining knowledge push information, in some embodiments, the target knowledge query result and the pushed knowledge set can be used as the knowledge push information of the DC protection defect handling strategy; in other embodiments, the union of the target knowledge query result and the pushed knowledge set can also be used as the knowledge push information of the DC protection defect handling strategy.

[0080] It is understandable that since there may be a small amount of duplicate data between the target knowledge query results and the pushed knowledge set, in order to avoid duplicate content in the pushed knowledge push information, it is necessary to delete a small amount of duplicate information, that is, to use the union of the target knowledge query results and the pushed knowledge set as the knowledge push information for the DC protection defect handling strategy.

[0081] In an embodiment of the present application, the importance of knowledge is determined by inputting the knowledge handling strategy into a preset knowledge importance prediction model. The model can be used to comprehensively consider multiple factors to more accurately evaluate the importance of knowledge in different scenarios. This method can provide staff with more targeted and practical knowledge push information, enabling staff to quickly obtain the most critical and valuable knowledge in the current scenario, effectively improving the efficiency of DC protection defect handling and the scientific nature of decision-making.

[0082] In addition to the advantages mentioned above, the push method of the DC protection defect handling strategy also has the following advantages: Improve fault processing speed: By comprehensively considering multiple factors to evaluate the importance of knowledge and sorting the knowledge query results accordingly, staff can quickly locate the key knowledge most relevant to the current scenario, thereby speeding up fault processing. The pushed knowledge push information has been screened and sorted, reducing the time for staff to screen useful knowledge from a large amount of information and improving work efficiency; Enhance scientific decision-making: The pushed knowledge push information is based on the DC protection defect data and preset knowledge graph in the current scenario, providing staff with data-based decision support, enhancing the scientific nature of decision-making, predicting the importance of knowledge through models, reducing the subjectivity and errors of human judgment, and improving the accuracy of decision-making; Optimize knowledge management: By constructing a DC protection defect field The knowledge graph realizes the systematic management and efficient utilization of knowledge, which helps to improve the overall level of knowledge management. With the accumulation of new fault cases and handling experience, the knowledge graph and the preset knowledge importance prediction model can be continuously updated and iterated to maintain the timeliness and accuracy of knowledge; improve system stability: by quickly pushing key knowledge, staff can quickly respond to and handle DC protection defects, reduce the impact of faults on the stability of the power system, and through learning and analyzing historical fault cases, the push method can predict and prevent potential DC protection defects and improve the overall stability of the power system; reduce operation and maintenance costs: by pushing accurate and practical knowledge push information, the staff's need for training on a large amount of unnecessary knowledge is reduced, and the operation and maintenance costs are reduced. By quickly handling DC protection defects, the equipment downtime due to faults is reduced, and the equipment utilization and economic benefits are improved.

[0083] In a feasible implementation, the method in the above embodiment also includes: obtaining historical DC protection defect data in historical scenarios, the historical DC protection defect data including multiple historical knowledge disposal strategies; determining the knowledge importance of each historical knowledge disposal strategy; inputting the knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies into the initial BPNN model for training to obtain a preset knowledge importance prediction model.

[0084] Among them, BPNN stands for Back Propagation Neural Network; the initial BPNN model refers to the original model that has not undergone any training.

[0085] Regarding the method of determining the knowledge importance of historical knowledge disposal strategies, in some embodiments, a match can be performed in a first preset mapping table between historical knowledge disposal strategies and knowledge importance according to each historical knowledge disposal strategy to determine the knowledge importance of each historical knowledge disposal strategy; in other embodiments, the knowledge node of each historical knowledge disposal strategy in the preset DC protection defect knowledge graph can also be determined, and then the knowledge importance of each historical knowledge disposal strategy can be determined based on the knowledge node of each historical knowledge disposal strategy; wherein, the first preset mapping table can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics, and of course, it can also be set by the operator according to actual needs.

[0086] In this application, the historical knowledge disposal strategy is introduced into the knowledge nodes in the preset DC protection defect knowledge graph to determine the knowledge importance of model training, so as to improve the accuracy and practicality of the preset knowledge importance prediction model.

[0087] In addition, by determining the knowledge nodes of each knowledge handling strategy in the preset DC protection defect knowledge graph and determining the knowledge importance based on the knowledge nodes, the structural relationship and semantic association of knowledge in the knowledge graph can be fully considered, so that the evaluation of knowledge importance is more in line with the actual business logic of DC protection defect handling, thereby providing staff with more accurate and more in line with the current scenario needs. Knowledge push information effectively improves the efficiency and accuracy of staff in acquiring relevant knowledge, and improves the quality and efficiency of DC protection defect handling work.

[0088] Furthermore, in some embodiments, a match can be made in a second preset mapping table between knowledge nodes and knowledge importance according to the knowledge nodes of each historical knowledge disposal strategy to determine the knowledge importance of each historical knowledge disposal strategy; wherein, the second preset mapping table can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics, and of course, can also be set by the operator according to actual needs.

[0089] Furthermore, in some embodiments, the retrieval sequence value and retrieval preference count corresponding to each historical knowledge disposal strategy in a preset historical scenario can be determined, and then the knowledge importance of each historical knowledge disposal strategy can be determined based on the retrieval sequence value and retrieval preference count corresponding to each historical knowledge disposal strategy in the preset historical scenario, as well as the knowledge nodes of each historical knowledge disposal strategy.

[0090] It should be noted that the retrieval order value corresponding to each historical knowledge disposal strategy in the preset historical scenarios refers to the retrieval order value of each historical knowledge disposal strategy corresponding to the user in the preset historical scenarios; the retrieval preference number corresponding to each historical knowledge disposal strategy in the preset historical scenarios refers to the retrieval preference number of each historical knowledge disposal strategy corresponding to the user in the preset historical scenarios.

[0091] Regarding the method of determining the number of retrieval preferences, in some embodiments, if there are M historical scenarios, then for the mth historical scenario, the number of retrieval preferences of the kth historical knowledge disposal strategy is equal to the sum of the number of retrievals of the kth historical knowledge disposal strategy in the 1st historical scenario to the mth historical scenario.

[0092] In this application, the importance of knowledge is determined by combining the retrieval order value and retrieval preference times in historical scenarios and knowledge nodes, thereby improving the comprehensiveness and accuracy of knowledge importance evaluation.

[0093] Furthermore, we can use the formula Determine the knowledge importance of each historical knowledge disposal strategy; where, is the knowledge importance of the k-th historical knowledge disposal strategy, M is the total number of historical scenarios in the preset historical scenarios, k a1 is the knowledge node of the k-th historical knowledge disposal strategy, is the retrieval order value corresponding to the k-th historical knowledge disposal strategy in the m-th historical scenario, is the number of retrieval preferences corresponding to the kth historical knowledge disposal strategy in the mth historical scenario, and K is the total number of historical knowledge disposal strategies.

[0094] In this application, by integrating multi-dimensional factors such as knowledge nodes, retrieval order values ​​and retrieval preference times, a formula is used to accurately calculate the knowledge importance of each historical knowledge disposal strategy, which significantly improves the accuracy and scientificity of knowledge importance assessment.

[0095] Regarding the method for determining the number of hidden neurons in the hidden layer of the initial BPNN model, in some embodiments, the formula Determine the number of hidden neurons; where hid is the number of hidden neurons, in is the number of input neurons in the input layer of the initial BPNN model, out is the number of output neurons in the output layer of the initial BPNN model, and α is any preset constant in the range [1,10].

[0096] Regarding the method of determining the number of input neurons and the number of output neurons, in some embodiments, the characteristic dimension of the historical knowledge handling strategy can be used as the number of input neurons, and the characteristic dimension of the knowledge importance of the historical knowledge handling strategy can be used as the number of output neurons.

[0097] Regarding the method for determining the preset constant, in other embodiments, when the number of input neurons is less than or equal to the preset feature dimension value, the optimal element can be selected from the preset discrete set {1, 3, 5, 8, 10} by cross-validation, and the optimal element can be used as the preset constant; when the number of input neurons is greater than the preset feature dimension value, the preset constant can be determined using the formula α = floor[log(in)]; wherein α is the preset constant, floor[] is the floor function, log() is the logarithm with the natural constant e as the base, or the logarithm with 2 as the base, and in is the number of input neurons.

[0098] Among them, the preset feature dimension value can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics. Of course, it can also be set by the operator according to actual needs. That is, in some embodiments, this application can preferably set the preset feature dimension value to 10.

[0099] In an embodiment of the present application, by using historical data to train the BPNN model to obtain a preset knowledge importance prediction model, it is possible to more accurately evaluate the importance of knowledge in different scenarios, provide staff with more targeted and practical knowledge push information, and further improve the efficiency of DC protection defect handling and the scientific nature of decision-making.

[0100] It is understandable that improving the accuracy of the model: by obtaining the historical DC protection defect data under historical scenarios and determining the knowledge importance of each historical knowledge disposal strategy, and using these data to train the initial BPNN model, a more accurate preset knowledge importance prediction model that can comprehensively consider multiple factors can be obtained. This model can learn the complex relationship between knowledge disposal strategy and knowledge importance based on historical data, so as to more accurately evaluate the importance of new knowledge in different scenarios; enhancing the pertinence of knowledge push: the preset knowledge importance prediction model can predict the knowledge importance of each knowledge disposal strategy based on the DC protection defect data in the current scenario, and sort the knowledge query results according to the knowledge importance, which can ensure the pushed knowledge The pushed information is more in line with the actual needs of the staff in the current scenario, and the pertinence of knowledge push is improved; the decision-making process is optimized: the pushed knowledge push information is based on the prediction results of the preset knowledge importance prediction model, providing staff with data-based decision support, which helps to reduce the subjectivity and errors of human judgment, improve the scientificity and accuracy of decision-making, and enable staff to make correct decisions more quickly; promote model iteration and optimization: with the accumulation of new fault cases and handling experience, historical DC protection defect data can be continuously updated and iterated, and then the preset knowledge importance prediction model can be continuously optimized, which helps to maintain the timeliness and accuracy of the model, and ensure that the knowledge push information can always reflect the latest DC protection defect handling strategies and knowledge.

[0101] In a feasible implementation method, the knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies in the above embodiment are input into the initial BPNN model for training to obtain a preset knowledge importance prediction model, including: using SSA or ISSA to determine the SSA-based BPNN model or the ISSA-based BPNN model according to the initial BPNN model; inputting the knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies into the SSA-based BPNN model or the ISSA-based BPNN model for training to obtain a preset knowledge importance prediction model.

[0102] Among them, SSA is Sparrow Search Algorithm, that is, sparrow search algorithm; ISSA is Improved Sparrow Search Algorithm, that is, improved sparrow search algorithm.

[0103] Regarding the method for determining the SSA-based BPNN model or the ISSA-based BPNN model, in some embodiments, SSA or ISSA can be used to adjust the model parameters of the initial BPNN model to determine the SSA-based BPNN model or the ISSA-based BPNN model.

[0104] In the embodiment of the present application, by introducing SSA or ISSA to optimize the BPNN model, the training effect and prediction accuracy of the preset knowledge importance prediction model are improved.

[0105] It is understandable that the optimization algorithm improves performance: using SSA (Sparrow Search Algorithm) or ISSA (Improved Sparrow Search Algorithm) to adjust the model parameters of the initial BPNN model can more effectively search for the optimal parameter combination. Compared with traditional methods, it can converge to the global optimal solution faster and avoid falling into the local optimal solution, thereby improving the training effect of the model; improving prediction accuracy: based on the optimized BPNN model (SSA-based BPNN model or ISSA-based BPNN model), the preset knowledge importance prediction model obtained by training can more accurately predict the importance of the knowledge disposal strategy. This is because the optimization algorithm helps the model better capture the complex patterns and relationships in the data, thereby improving the accuracy of the prediction; enhancing model adaptability: The BPNN model optimized by SSA or ISSA can better adapt to different data sets and scenarios, and can maintain good prediction performance even when the data distribution changes or there is noise, thereby enhancing the robustness and adaptability of the model; improving training efficiency: The application of the optimization algorithm can accelerate the model training process, reduce training time, and enable the model to be put into practical application more quickly, meeting the real-time and accuracy requirements of the power system.

[0106] In a feasible implementation, the use of SSA or ISSA in the above embodiment determines a BPNN model based on SSA or a BPNN model based on ISSA according to the initial BPNN model, including: using SSA or ISSA to iteratively update the position of each sparrow in the sparrow population until the global optimal fitness value of all iterations converges to a convergence threshold, or the number of iterations corresponding to the current iteration is equal to the preset maximum number of iterations, and the optimal model parameters corresponding to the global optimal fitness value of all sparrows in all iterations are obtained, wherein the position of each sparrow includes model parameters; according to the optimal model parameters corresponding to all iterations, the model parameters in the initial BPNN model are adjusted to obtain a BPNN model based on SSA or a BPNN model based on ISSA.

[0107] The convergence threshold and the preset maximum number of iterations can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics. Of course, they can also be set by the operator according to actual needs.

[0108] It should be noted that, when using SSA or ISSA to iteratively update the position of each sparrow in the sparrow population, a global fitness value is calculated for each iteration, and then the global fitness value calculated for each iteration is compared with a global optimal fitness value of all iterations, and the global optimal fitness value can be updated (if the global fitness value calculated for the tth iteration is less than the global optimal fitness value of all iterations from 1 to t-1, the global fitness value calculated for the tth iteration can be used as the global optimal fitness value of all iterations from 1 to t). a global optimal fitness value; if a global fitness value calculated in the t-th iteration is greater than or equal to a global optimal fitness value of all iterations from the 1st to the t-1th, then the global optimal fitness value of all iterations from the 1st to the t-1th can be used as a global optimal fitness value of all iterations from the 1st to the t-th), when the global optimal fitness values ​​of all iterations converge to the convergence threshold, or the number of iterations corresponding to the current iteration is equal to the preset maximum number of iterations, at this time, the optimal model parameters corresponding to the global optimal fitness value of all iterations are obtained.

[0109] It should be further explained that the initial position of each sparrow in the sparrow population can be set to a default initialization value; during the iterative process, the position of each sparrow in the next iteration can be calculated according to the position update formula in SSA or ISSA; and the global fitness value can be calculated according to the fitness formula in SSA or ISSA.

[0110] Regarding the adjustment method of the model parameters in the initial BPNN model, in some embodiments, the model parameters in the initial BPNN model can be directly adjusted to the optimal model parameters corresponding to all iterations to obtain a BPNN model based on SSA, or a BPNN model based on ISSA; of course, in other embodiments, the model parameters in the initial BPNN model can be adjusted to the sum or difference between the optimal model parameters corresponding to all iterations and the preset model error parameters to obtain a BPNN model based on SSA, or a BPNN model based on ISSA; wherein, the preset model error parameters can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics, and of course, can also be set by the operator according to actual needs.

[0111] Regarding the parameters included in the optimal model parameters, in some embodiments, the optimal model parameters include but are not limited to the weights and biases between the input layer and the hidden layer in the initial BPNN model, the weights and biases between the hidden layer and the output layer in the initial BPNN model, and the trigger threshold value of the neuron activation function in the initial BPNN model.

[0112] In the embodiment of the present application, the BPNN model parameters are optimized by SSA or ISSA, which improves the model training effect and prediction accuracy, and enhances the adaptability and robustness of the model.

[0113] It is understandable that improving the model training effect: using SSA or ISSA to iteratively update the position of each sparrow in the sparrow population can more effectively search for the optimal parameter combination of the BPNN model. Compared with the traditional method, this method can converge to the global optimal solution faster and avoid falling into the local optimal solution, thereby significantly improving the model training effect; improving prediction accuracy: based on the optimized BPNN model (SSA-based BPNN model or ISSA-based BPNN model), the preset knowledge importance prediction model obtained by training can more accurately predict the importance of knowledge disposal strategy. This is because the SSA and ISSA optimization algorithms help the model better capture the complex patterns and relationships in the data, thereby improving the accuracy of the prediction; enhancing model adaptability: the BPNN model optimized by SSA or ISSA can better It can adapt to different data sets and scenarios. Even when the data distribution changes or there is noise, the optimized model can maintain good prediction performance, which enhances the robustness and adaptability of the model; optimize model parameters: during the iterative update process, a global fitness value is calculated for each iteration to update the global optimal fitness value of all iterations, and the model parameters are adjusted according to this value. This dynamic adjustment mechanism ensures that the model parameters are always optimized in the optimal direction, further improving the performance of the model; provide flexible adjustment strategies: when adjusting the model parameters in the initial BPNN model, you can choose to directly adjust the parameters to the optimal model parameters, or adjust them to the sum or difference of the optimal model parameters and the preset model error parameters. This flexibility enables the model to be customized according to different needs, further enhancing the adaptability and practicality of the model.

[0114] In a feasible implementation, under the condition of using ISSA, the method in the above embodiment also includes: determining the knowledge importance of each predictive knowledge disposal strategy obtained by the ISSA-based BPNN model during the verification process; determining a loss error value based on the knowledge importance of all predictive knowledge disposal strategies and the knowledge importance of all historical knowledge disposal strategies; when the loss error value does not meet the preset error value, determining a target weight value based on the number of iterations and the preset maximum number of iterations; adjusting the adaptive weight value of the ISSA according to the target weight value to obtain an adjusted ISSA; using the adjusted ISSA as the ISSA, returning to execute the ISSA, iteratively updating the position of each sparrow in the sparrow population until the global optimal fitness value of all iterations converges to the convergence threshold, or the number of iterations corresponding to the current iteration is equal to the preset maximum number of iterations, and obtaining the optimal model parameters corresponding to the global optimal fitness value of all sparrows in all iterations, until the loss error value meets the preset error value, and using the ISSA-based BPNN model corresponding to the loss error value meeting the preset error value as the final ISSA-based BPNN model.

[0115] The preset error value may be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics. Of course, it may also be set by the operator based on actual needs.

[0116] Regarding the method of determining the loss error value, in some embodiments, the root mean square error value, the mean absolute error value, and the mean absolute percentage error value between the knowledge importance of all predictive knowledge disposal strategies and the knowledge importance of all historical knowledge disposal strategies can be determined, and the sum of the root mean square error value, the mean absolute error value, and the mean absolute percentage error value can be used as the loss error value.

[0117] Regarding the adjustment method of the adaptive weight value of ISSA, in some embodiments, the adaptive weight value of ISSA can be directly adjusted to the target weight value; of course, in other embodiments, the adaptive weight value of ISSA can be adjusted to the sum or difference between the target weight value and the preset weight error value; wherein, the preset weight error value can be obtained and pre-set by the operator based on a large amount of experience, experiments or statistics, and of course, can also be set by the operator according to actual needs.

[0118] In the embodiment of the present application, the BPNN model is optimized by dynamically adjusting the ISSA adaptive weight value, which significantly improves the model prediction accuracy and training effect, and enhances the robustness and adaptability of the model.

[0119] It can be understood that improving prediction accuracy: in the verification process, by calculating the loss error value between the knowledge importance of the predicted knowledge disposal strategy and the knowledge importance of the historical knowledge disposal strategy, the prediction performance of the model can be accurately evaluated. When the loss error value does not meet the preset error value, dynamically adjusting the adaptive weight value of ISSA will help the model to better capture the complex patterns and relationships in the data in subsequent iterations, thereby improving the accuracy of the prediction; enhancing model robustness: by dynamically adjusting the target weight value according to the number of iterations and the preset maximum number of iterations, and then adjusting the adaptive weight value of ISSA, the model can better cope with changes in data distribution at different stages during training, thereby enhancing the robustness of the model; optimizing the model Improved training effect: By continuously iteratively adjusting the adaptive weight values ​​of ISSA and re-executing the iterative update process of the position of each sparrow in the sparrow population, it can be ensured that the model parameters are always optimized in the optimal direction. This dynamic adjustment mechanism helps the model converge to the global optimal solution faster, avoiding falling into the local optimal solution, thereby significantly improving the training effect of the model; Improved model adaptability: By continuously adjusting the adaptive weight values ​​of ISSA until the loss error value meets the preset error value, the final ISSA-based BPNN model can better adapt to different data sets and scenarios. Even when the data distribution changes or there is noise, the optimized model can maintain good prediction performance, thereby enhancing the adaptability of the model.

[0120] In a feasible implementation, the above embodiment determines the target weight value according to the number of iterations and the preset maximum number of iterations, including:

[0121] Using the formula Determine the target weight value;

[0122] Among them, w is the target weight value, f randn is a random number that obeys the normal distribution, N is the preset maximum number of iterations, and n is the number of iterations at that time.

[0123] In the embodiment of the present application, the target weight value is dynamically determined by a formula, thereby improving the flexibility and effectiveness of the ISSA optimization BPNN model.

[0124] It can be understood that the flexibility of optimization is improved: the target weight value is dynamically determined according to the number of iterations and the preset maximum number of iterations using a formula, so that in the process of ISSA optimization of the BPNN model, the adjustment of the weight value is more flexible, and can be adaptively adjusted according to different iteration stages to better meet the needs of model training; the effectiveness of optimization is enhanced: random numbers that obey the normal distribution are introduced into the formula, which increases the randomness and diversity of the weight value adjustment, helps to avoid the model from falling into the local optimum, and improves the ability to search for the global optimal solution, thereby enhancing the effectiveness of the ISSA optimization BPNN model and enabling the model to more accurately predict the importance of the knowledge disposal strategy.

[0125] In a feasible implementation, step 120 in the above embodiment, determining the push knowledge set based on the demand knowledge information and the preset DC protection defect knowledge graph, includes: determining the demand knowledge category based on the demand knowledge information; determining multiple alternative knowledge sets based on the demand knowledge category and the preset DC protection defect knowledge graph; and determining the push knowledge set based on the multiple alternative knowledge sets.

[0126] Regarding the method of determining the pushed knowledge set, in some embodiments, one alternative knowledge set can be arbitrarily selected from multiple alternative knowledge sets as the pushed knowledge set; in other embodiments, the pushed knowledge set can also be determined based on the correlation or similarity, based on the correlation or similarity between each alternative knowledge set and the required knowledge information.

[0127] In the embodiment of the present application, by refining the required knowledge categories and screening the alternative knowledge sets based on the knowledge graph, and then determining the push knowledge sets, it is possible to more accurately match the knowledge needs of staff and improve the pertinence and effectiveness of knowledge push.

[0128] It is understandable that accurate matching of knowledge needs: determining the required knowledge category based on the required knowledge information can help to refine the specific needs of the staff into more specific categories, so as to more accurately understand their knowledge needs. Based on the required knowledge category and the preset DC protection defect knowledge graph, multiple alternative knowledge sets are determined to ensure that the alternative knowledge sets are closely related to the needs of the staff, thereby improving the accuracy of knowledge push; improving the pertinence of knowledge push: among multiple alternative knowledge sets, selecting the pushed knowledge set by correlation or similarity can ensure that the pushed knowledge is most consistent with the needs of the staff in the current scenario, thereby improving the pertinence of knowledge push. Compared with arbitrarily selecting alternative knowledge sets as pushed knowledge sets, based on correlation or similarity, The selection method is more scientific and reasonable, which can avoid pushing irrelevant or redundant knowledge; optimize the knowledge push process: by refining the required knowledge categories, screening alternative knowledge sets and determining the steps of pushing knowledge sets, a complete and systematic knowledge push process is formed. This process can ensure that each link of knowledge push is closely centered on the needs of staff, thereby improving the efficiency and effectiveness of knowledge push; enhance knowledge utilization efficiency: the accurately pushed knowledge can be understood and applied by staff more quickly, thereby improving the efficiency of DC protection defect handling work. At the same time, since the pushed knowledge is highly matched with the needs of staff, it can reduce the time for staff to screen useful knowledge from a large amount of information and improve knowledge utilization efficiency.

[0129] In a feasible implementation, the above embodiment determines the push knowledge set based on multiple candidate knowledge sets, including: determining the similarity between each candidate knowledge set and the required knowledge information; and using the candidate knowledge set corresponding to the maximum similarity as the push knowledge set.

[0130] In an embodiment of the present application, by calculating the similarity between the alternative knowledge set and the required knowledge information, and selecting the alternative knowledge set with the greatest similarity as the pushed knowledge set, the accuracy and pertinence of knowledge push can be significantly improved, ensuring that staff obtain the knowledge information that best meets their current needs.

[0131] It is understandable that the accuracy of push notifications is improved: by calculating similarity, the degree of match between each candidate knowledge set and the required knowledge information can be quantified, thereby selecting the candidate knowledge set that is closest to the required knowledge set as the pushed knowledge set. This method avoids the errors that may be caused by subjective judgment and random selection, and ensures the accuracy of the pushed knowledge set; the targeted nature of push notifications is enhanced: similarity calculation can accurately reflect the degree of association between the candidate knowledge set and the required knowledge information. Therefore, selecting the candidate knowledge set with the greatest similarity as the pushed knowledge set can ensure that the pushed knowledge is highly matched with the staff's needs in the current scenario. This targeted push method helps staff quickly obtain the required knowledge information and improve work efficiency; optimized knowledge utilization: accurately pushed knowledge sets can reduce the time staff spend sifting through large amounts of information for useful knowledge, allowing them to focus more on addressing DC protection defect issues. At the same time, because the pushed knowledge sets are highly matched with the staff's needs, they can improve knowledge utilization and avoid knowledge waste; improved user experience: when staff receive knowledge push notifications that are highly matched to their needs, they will feel more satisfied and recognized, thereby increasing their trust and reliance on the knowledge push system. This good user experience helps promote the continued use and promotion of the system.

[0132] In a feasible implementation, step 140 in the above embodiment uses the knowledge importance of all knowledge handling strategies to sort multiple knowledge query results to determine the target knowledge query result, including: using the knowledge importance of all knowledge handling strategies to sort multiple knowledge query results in descending order to obtain a descending sorting result; and using the first preset knowledge query results in the descending sorting result as the target knowledge query result.

[0133] Among them, the specific number of presets can be set by the operator according to actual needs. For example, the preset number can be set to 10, or the preset number can be set to 1. This application does not limit it here.

[0134] In an embodiment of the present application, by sorting the knowledge query results in descending order and selecting the first preset results as the target knowledge query results, it can ensure that the staff can obtain the most important and relevant knowledge information first, thereby significantly improving the effectiveness and practicality of knowledge push.

[0135] It is understandable that the effectiveness of knowledge push can be improved: by sorting in descending order, the knowledge query results that are most relevant and important to the current scenario can be placed in front, so that the staff can quickly locate the key information and avoid wasting time in a large amount of irrelevant or secondary information. Selecting the preset knowledge query results as the target knowledge query results can ensure that the amount of knowledge pushed is moderate, neither too much to cause information overload nor too little to meet the needs of the staff; enhancing the practicality of knowledge push: descending sorting and the selection method of the preset number make the pushed knowledge closer to the actual needs of the staff, improve the practicality and operability of the knowledge, and the staff can quickly understand the current situation based on the pushed target knowledge query results. Key issues and solutions in previous scenarios, so as to handle DC protection defects more efficiently; optimize knowledge utilization efficiency: by accurately pushing the most important knowledge information, the time invested by staff in screening and organizing knowledge can be reduced, and the efficiency of knowledge utilization can be improved. At the same time, because the pushed knowledge is highly matched with the needs of staff, it can be converted into actual actions and decisions more quickly, further improving work efficiency; improve user experience: when staff receive the target knowledge query results after descending sorting and preset number screening, they will feel more convenient and efficient, thereby improving their satisfaction and recognition of the knowledge push system. This good user experience helps promote the continuous use and optimization of the system, forming a virtuous circle.

[0136] In a feasible implementation, step 150 in the above embodiment, determining the knowledge push information of the DC protection defect handling strategy based on the target knowledge query result and the pushed knowledge set, includes: taking the union of the target knowledge query result and the pushed knowledge set as the knowledge push information.

[0137] In an embodiment of the present application, by taking the union of the target knowledge query result and the pushed knowledge set as the knowledge push information, the comprehensiveness and accuracy of the pushed content can be ensured, information omissions can be avoided, and the efficiency of staff in acquiring key knowledge can be improved.

[0138] It is understandable that ensuring the comprehensiveness of the pushed content: the target knowledge query results and the pushed knowledge set may each contain a portion of unique and important knowledge information. By taking the union, the two parts of information can be merged to ensure that the pushed knowledge content is more comprehensive and covers more possible staff needs; improving information accuracy: the union operation can avoid repeatedly pushing the same information while ensuring that all relevant and important knowledge is included, which helps to reduce information redundancy and improve the accuracy and pertinence of the pushed information; avoiding information omission: in the DC protection defect handling process, any omission of key information may lead to improper handling or delays. By taking the union, the risk of information omission can be minimized, ensuring that staff can obtain all necessary knowledge to efficiently handle defects; improving work efficiency: staff do not need to search and filter relevant knowledge in multiple information sources, because the union operation has integrated all important information together, which helps save time, improve work efficiency, and enable staff to make decisions and take actions faster; optimizing user experience: the comprehensiveness and accuracy of pushed information can improve staff satisfaction and trust in the knowledge push system. When staff can easily obtain all the key information they need, they will rely on and recognize the system more, thereby promoting the continuous use and optimization of the system.

[0139] In a second aspect, the present application provides a device for pushing a DC protection defect handling strategy.

[0140] See also Figure 2 , is a schematic diagram of a device for pushing a DC protection defect handling strategy according to an embodiment of the present application. The device 210 includes:

[0141] An acquisition module 211 is configured to acquire DC protection defect data in a current scenario, as well as required knowledge information and query statements corresponding to the DC protection defect data, wherein the DC protection defect data includes multiple knowledge handling strategies;

[0142] The determination and query module 212 is used to determine the pushed knowledge set based on the required knowledge information and the preset DC protection defect knowledge graph, and use a query statement to query the preset DC protection defect knowledge graph to obtain multiple knowledge query results;

[0143] The model prediction module 213 is used to input all knowledge handling strategies into a preset knowledge importance prediction model to obtain the knowledge importance of each knowledge handling strategy;

[0144] A ranking module 214 is configured to rank the plurality of knowledge query results using the knowledge importance of all knowledge handling strategies to determine a target knowledge query result;

[0145] The determination module 215 is configured to determine the knowledge push information of the DC protection defect handling strategy according to the target knowledge query result and the pushed knowledge set.

[0146] In the embodiment of the present application, the relevant contents of the acquisition module 211, the determination and query module 212, the model prediction module 213, the sorting module 214 and the determination module 215 can be referred to. Figure 1 The contents of the illustrated embodiments are not described in detail here.

[0147] It should be noted that the device 210 of the present application also includes some other modules. It can be understood that the method of the present application and the device 210 have a one-to-one correspondence. Therefore, the other modules of the device 210 of the present application are the contents corresponding to the method of the present application in the above-mentioned embodiment.

[0148] In an embodiment of the present application, the importance of knowledge is determined by inputting the knowledge handling strategy into a preset knowledge importance prediction model. The model can be used to comprehensively consider multiple factors to more accurately evaluate the importance of knowledge in different scenarios. This method can provide staff with more targeted and practical knowledge push information, enabling staff to quickly obtain the most critical and valuable knowledge in the current scenario, effectively improving the efficiency of DC protection defect handling and the scientific nature of decision-making.

[0149] In addition to the advantages mentioned above, the push device of the DC protection defect handling strategy also has the following advantages: Improve fault processing speed: By comprehensively considering multiple factors to evaluate the importance of knowledge and sorting the knowledge query results accordingly, staff can quickly locate the key knowledge most relevant to the current scenario, thereby speeding up fault processing. The pushed knowledge push information has been screened and sorted, reducing the time for staff to screen useful knowledge from a large amount of information and improving work efficiency; Enhance scientific decision-making: The pushed knowledge push information is based on the DC protection defect data and preset knowledge graph in the current scenario, providing staff with data-based decision support, enhancing the scientific nature of decision-making, predicting the importance of knowledge through models, reducing the subjectivity and errors of human judgment, and improving the accuracy of decision-making; Optimize knowledge management: By constructing a DC protection defect field The knowledge graph realizes the systematic management and efficient utilization of knowledge, which helps to improve the overall level of knowledge management. With the accumulation of new fault cases and handling experience, the knowledge graph and the preset knowledge importance prediction model can be continuously updated and iterated to maintain the timeliness and accuracy of knowledge; improve system stability: by quickly pushing key knowledge, staff can quickly respond to and handle DC protection defects, reduce the impact of faults on the stability of the power system, and through learning and analyzing historical fault cases, the push device can predict and prevent potential DC protection defects and improve the overall stability of the power system; reduce operation and maintenance costs: by pushing accurate and practical knowledge push information, the staff's need for training on a large amount of unnecessary knowledge is reduced, and the operation and maintenance costs are reduced. By quickly handling DC protection defects, the equipment downtime due to faults is reduced, and the equipment utilization and economic benefits are improved.

[0150] In a third aspect, the present application further provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes a method for pushing a DC protection defect handling strategy in the above method embodiment.

[0151] In a fourth aspect, the present application further provides a computer device including a memory and a processor, wherein the memory stores a computer program. When the computer program is executed by the processor, the processor executes a method for pushing a DC protection defect handling strategy in the above-mentioned method embodiment.

[0152] Figure 3 The internal structure diagram of the computer device in some embodiments is shown. The computer device can be a terminal, a server, or a gateway. Figure 3 As shown, the computer device includes a processor, a memory, and a network interface connected via a system bus.

[0153] The memory includes a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program. When the computer program is executed by the processor, the processor can implement the various steps in the above method embodiment. The internal memory may also store a computer program. When the computer program is executed by the processor, the processor can implement the various steps in the above method embodiment. It will be understood by those skilled in the art that Figure 3 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0154] Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing related hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods.

[0155] Among them, any reference to memory, storage, database or other media used in the various embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchl ink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0156] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0157] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person of ordinary skill in the art may make various modifications and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for pushing a DC protection defect handling strategy, characterized in that: The method comprises: Acquire DC protection defect data in a current scenario, as well as required knowledge information and query statements corresponding to the DC protection defect data, wherein the DC protection defect data includes multiple knowledge processing strategies; Determine a push knowledge set based on the required knowledge information and a preset DC protection defect knowledge graph, and use the query statement to query the preset DC protection defect knowledge graph to obtain multiple knowledge query results; Input all knowledge disposal strategies into the preset knowledge importance prediction model to obtain the knowledge importance of each knowledge disposal strategy; Using the knowledge importance of all knowledge disposal strategies, multiple knowledge query results are ranked to determine a target knowledge query result; Determine knowledge push information of a DC protection defect handling strategy based on the target knowledge query result and the pushed knowledge set.

2. The method for pushing a DC protection defect handling strategy according to claim 1, characterized in that: The method further comprises: Acquiring historical DC protection defect data in historical scenarios, wherein the historical DC protection defect data includes a plurality of historical knowledge processing strategies; Determine the knowledge importance of each historical knowledge disposal strategy; The knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies are input into the initial BPNN model for training to obtain the preset knowledge importance prediction model.

3. The method for pushing a DC protection defect handling strategy according to claim 2, characterized in that: The knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies are input into the initial BPNN model for training to obtain the preset knowledge importance prediction model, including: Using SSA or ISSA, according to the initial BPNN model, determine an SSA-based BPNN model or an ISSA-based BPNN model; The knowledge importance of all historical knowledge disposal strategies and all historical knowledge disposal strategies are input into the SSA-based BPNN model or the ISSA-based BPNN model for training to obtain the preset knowledge importance prediction model.

4. The method for pushing a DC protection defect handling strategy according to claim 3, characterized in that: The method of using SSA or ISSA to determine an SSA-based BPNN model or an ISSA-based BPNN model according to the initial BPNN model includes: Using the SSA or the ISSA, iteratively updating the position of each sparrow in the sparrow population until the global optimal fitness value of all iterations converges to a convergence threshold, or the number of iterations corresponding to the current iteration is equal to a preset maximum number of iterations, and obtaining the optimal model parameters corresponding to the global optimal fitness value of all sparrows in all iterations, wherein the position of each sparrow includes the model parameters; According to the optimal model parameters corresponding to all iterations, the model parameters in the initial BPNN model are adjusted to obtain the SSA-based BPNN model or the ISSA-based BPNN model.

5. The method for pushing a DC protection defect handling strategy according to claim 4, characterized in that: Under the condition of using the ISSA, the method further includes: Determining the knowledge importance of each predicted knowledge disposal strategy obtained during the validation process of the ISSA-based BPNN model; Determine a loss error value according to the knowledge importance of all prediction knowledge disposal strategies and the knowledge importance of all historical knowledge disposal strategies; When the loss error value does not meet the preset error value, determining a target weight value according to the current number of iterations and the preset maximum number of iterations; Adjusting the adaptive weight value of the ISSA according to the target weight value to obtain an adjusted ISSA; The adjusted ISSA is used as the ISSA, and the ISSA is returned to be used to iteratively update the position of each sparrow in the sparrow population until the global optimal fitness value of all iterations converges to a convergence threshold, or the number of iterations corresponding to the current iteration is equal to the preset maximum number of iterations, and the optimal model parameters corresponding to the global optimal fitness value of all sparrows in all iterations are obtained, until the loss error value meets the preset error value, and the ISSA-based BPNN model corresponding to the loss error value meeting the preset error value is used as the final ISSA-based BPNN model.

6. The method for pushing a DC protection defect handling strategy according to claim 5, characterized in that: The determining of the target weight value according to the current number of iterations and the preset maximum number of iterations includes: Using the formula determining the target weight value; Wherein, w is the target weight value, f randn is a random number that obeys the normal distribution, N is the preset maximum number of iterations, and n is the number of iterations at that time.

7. The method for pushing a DC protection defect handling strategy according to claim 1, characterized in that: The step of determining the pushed knowledge set based on the required knowledge information and the preset DC protection defect knowledge graph includes: determining a required knowledge category according to the required knowledge information; Determining multiple candidate knowledge sets according to the required knowledge category and the preset DC protection defect knowledge graph; The pushed knowledge set is determined according to a plurality of candidate knowledge sets.

8. The method for pushing a DC protection defect handling strategy according to claim 7, characterized in that: The step of determining the pushed knowledge set based on a plurality of candidate knowledge sets includes: Determining the similarity between each candidate knowledge set and the required knowledge information; The candidate knowledge set corresponding to the maximum similarity is used as the pushed knowledge set.

9. The method for pushing a DC protection defect handling strategy according to claim 1, characterized in that: The method of ranking the plurality of knowledge query results using the knowledge importance of all knowledge handling strategies to determine the target knowledge query result includes: Using the knowledge importance of all knowledge disposal strategies, multiple knowledge query results are sorted in descending order to obtain a descending sort result; The first preset knowledge query results in the descending sorting results are used as the target knowledge query results.

10. The method for pushing a DC protection defect handling strategy according to claim 1, characterized in that: The step of determining the knowledge push information of the DC protection defect handling strategy based on the target knowledge query result and the pushed knowledge set includes: The union of the target knowledge query result and the pushed knowledge set is used as the knowledge push information.

Images