Fault plan determination method and device, computer equipment and storage medium

Abstract

The invention relates to a fault plan determination method and device, computer equipment and a storage medium. A terminal obtains fault phenomenon data of a power system; and inputting the fault phenomenon data into a plan matching model obtained by training the fault phenomenon and the fault plan in the historical time period for matching processing to obtain a target fault plan. Based on the complex semantic processing and natural language generation capability of the plan matching model, the accuracy and interpretation capability of the fault plan matching of the power automation system are improved, so that the automatic matching of the fault phenomenon and the fault plan of the power dispatching automation system can be realized, and the fault processing speed of dispatching personnel is improved.

Description

Technical Field

[0001] The present application relates to the field of electric power automation technology, and in particular to a method, device, computer equipment and storage medium for determining a fault plan. Background Art

[0002] With the large-scale access of new power systems and the expansion of the scale of dispatching automation systems, the operation and maintenance requirements of power automation systems are constantly increasing. In daily operation and maintenance monitoring, the power automation system needs to cover a large number of systems and equipment, and automation operation and maintenance personnel need to pay attention to a large amount of complex data in real time. When a power automation system fails, the dispatcher needs to find related problems from the huge emergency plan and select the most appropriate solution. This leads to slow fault handling and affects the stable operation of the power system. Summary of the invention

[0003] Based on this, it is necessary to provide a fault plan determination method, device, computer equipment and storage medium that can improve the stability of the power system in response to the above technical problems.

[0004] In a first aspect, the present application provides a method for determining a fault plan, characterized in that the method comprises:

[0005] Obtain fault phenomenon data of the power system;

[0006] The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

[0007] In one embodiment, the above-mentioned plan matching model includes a feature extraction layer and a classification layer; inputting the fault phenomenon data into the pre-built plan matching model for matching processing includes:

[0008] Input the fault phenomenon data into the feature extraction layer for feature extraction processing to obtain feature data;

[0009] The feature data is input into the classification layer for classification processing to obtain the target fault plan.

[0010] In one embodiment, the training process of the above-mentioned plan matching model includes:

[0011] Obtain fault phenomenon samples and fault contingency plan samples;

[0012] The fault phenomenon samples and the fault plan samples are respectively input into the pre-training model for plan prediction processing to obtain multiple prediction results;

[0013] If the matching accuracy determined according to the prediction result does not reach the accuracy threshold, the parameters in the pre-trained model are adjusted until the matching accuracy reaches the accuracy threshold, and the adjustment is completed to obtain the plan matching model.

[0014] In one embodiment, the above-mentioned adjustment of parameters in the pre-trained model includes:

[0015] Extracting and processing the original parameters of the feature extraction layer to obtain extraction parameters;

[0016] The extracted parameters are input into the fine-tuning model for low-rank fine-tuning to obtain updated parameters;

[0017] The updated parameters are combined with the original parameters to obtain the adjusted model parameters.

[0018] In one embodiment, the fine-tuning model includes a dimensionality reduction layer and a dimensionality increase layer, and the extracted parameters are input into the fine-tuning model for low-rank fine-tuning processing to obtain updated parameters, including:

[0019] The extracted parameters are input into the dimension reduction layer for dimension reduction processing to obtain a dimension reduction vector;

[0020] The obtained rank and dimension reduction vector are input into the dimension increasing layer for dimension increasing processing to obtain the updated parameters.

[0021] In one of the embodiments, the target fault plan includes at least one of preventive warning, closed management and regional control; the method further includes: sending the target fault plan to a terminal of a backend personnel.

[0022] In one of the embodiments, the method further includes: obtaining actual measures taken from the terminal, and retraining the plan matching model according to the actual measures taken.

[0023] In a second aspect, the present application further provides a fault plan determination device, the device comprising:

[0024] A data acquisition module, used to acquire fault phenomenon data of the power system;

[0025] The plan matching module inputs the fault phenomenon data into the pre-built plan matching model for matching processing to obtain the target fault plan; the plan matching model is obtained by training the fault phenomena and fault plans in the historical period.

[0026] In a third aspect, the present application further provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the following steps when executing the computer program:

[0027] Obtain fault phenomenon data of the power system;

[0028] The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

[0029] In a fourth aspect, the present application further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0030] Obtain fault phenomenon data of the power system;

[0031] The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

[0032] In a fifth aspect, the present application further provides a computer program product. The computer program product includes a computer program, and when the computer program is executed by a processor, the following steps are implemented:

[0033] Obtain fault phenomenon data of the power system;

[0034] The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

[0035] The above-mentioned fault plan determination method, device, computer equipment and storage medium, the terminal obtains the fault phenomenon data of the power system; the fault phenomenon data is input into the fault phenomenon of the historical period and the plan matching model obtained by the fault plan training for matching processing to obtain the target fault plan. This application is based on the complex semantic processing and natural language generation capabilities of the plan matching model to improve the accuracy and interpretation ability of the fault plan matching of the power automation system, thereby realizing the automatic matching of the fault phenomenon and the fault plan of the power dispatching automation system, and improving the speed of dispatching personnel in handling faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 An application environment diagram of a method for determining a fault plan in an embodiment;

[0037] Figure 2 A schematic diagram of a flow chart of a method for determining a fault plan in an embodiment;

[0038] Figure 3 A schematic diagram of a matching process in one embodiment;

[0039] Figure 4 A schematic diagram of a flow chart of a training process of a plan matching model in an embodiment;

[0040] Figure 5 A schematic diagram of a process for adjusting parameters in a pre-trained model in one embodiment;

[0041] Figure 6 is a schematic diagram of a process for obtaining updated parameters in one embodiment;

[0042] Figure 7 A structural block diagram of a device for determining a fault plan in an embodiment;

[0043] Figure 8 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0045] First of all, before specifically introducing the technical solution of the embodiments of the present application, the technical background based on the embodiments of the present application is introduced.

[0046] With the large-scale access of new power systems and the expansion of the scale of dispatching automation systems, the operation and maintenance requirements of power automation systems are constantly increasing. In daily operation and maintenance monitoring, the power automation system needs to cover a large number of systems and equipment, and automation operation and maintenance personnel need to pay attention to a large amount of complex data in real time. When a power automation system fails, the dispatcher needs to find related problems from the huge emergency plan and select the most appropriate solution. This leads to slow fault handling and affects the stable operation of the power system.

[0047] Based on this, the present application provides a method, apparatus, computer device and storage medium for determining a fault plan, aiming to solve the above-mentioned technical problems.

[0048] The fault plan determination method provided in the embodiment of the present application can be applied to Figure 1 In the application environment shown. Among them, the terminal 102 communicates with the server 104 through the network. The data storage system can store the data that the server 104 needs to process. The data storage system can be integrated on the server 104, or it can be placed on the cloud or other network servers. The terminal 102 obtains the fault phenomenon data of the power system; the fault phenomenon data is input into the fault phenomenon of the historical period and the plan matching model obtained by the fault plan training for matching processing to obtain the target fault plan. Among them, the terminal 102 can be but is not limited to various personal computers, laptops, smart phones, tablet computers, and Internet of Things devices. The server 104 can be implemented with an independent server or a server cluster consisting of multiple servers.

[0049] In an exemplary embodiment, Figure 2 As shown, the present application provides a method for determining a fault plan, which is applied to Figure 1 The terminal 102 in the example is used for explanation, including the following S201 to S202.

[0050] in:

[0051] S201, acquiring fault phenomenon data of the power system.

[0052] Among them, the fault phenomenon data includes data such as generator set failure, transmission line failure, substation failure, and bus failure.

[0053] In an embodiment of the present application, sensors or cameras may be provided in various devices of the power system, and the sensors or cameras periodically obtain various on-site operation data of various devices of the power system. After the above-mentioned fault occurs, these fault phenomenon data are sent to the terminal, and the terminal obtains data such as generator set fault, transmission line fault, substation fault, bus fault, etc. to perform fault analysis.

[0054] S202, inputting the fault phenomenon data into a pre-built plan matching model for matching processing to obtain a target fault plan.

[0055] Among them, the plan matching model is obtained by training the fault phenomena and fault plans in the historical period. The fault plan refers to the plan designed by the back-end personnel for different fault phenomena.

[0056] The plan matching model can adopt LLM (Large Language Model). The LLM model can handle a variety of natural language tasks, such as text classification, question and answer, and dialogue, and is an important path to artificial intelligence. From the technical characteristics, LLM is built on the Transformer architecture, in which multi-head attention layers are stacked in a very deep neural network. LLM greatly expands the model size, pre-training data, and total computation, so it can better understand natural language and generate high-quality text based on a given context. In addition, LLM also has emergent capabilities, including contextual learning, instruction following, and step-by-step reasoning, which enable LLM to be applied to solve various complex natural language processing tasks.

[0057] In an embodiment of the present application, based on the fault phenomenon data obtained in the above embodiment, the terminal inputs the fault phenomenon data into the plan matching model obtained by training the fault phenomenon and fault plan based on the historical period for matching processing to obtain the target fault plan. The specific process of matching processing can be data preprocessing: before performing data matching, we need to perform data preprocessing to ensure that the data format in the table is consistent, clean and correct. Then perform matching based on key fields: when performing data matching, matching is usually performed based on the key fields in the table. Key fields refer to those fields that uniquely identify a record, such as ID, name, etc. The following are some key field-based matching. The VLOOKUP function can be used to find matching data in one table based on one or more key fields, and return the results to another table. By setting the correct parameters and ranges, you can easily match data in two tables. Alternatively, SQL queries achieve data matching by writing query statements. Using SELECT statements and JOIN operations, you can select the same data from two tables and then save the results to a new table to achieve data matching.

[0058] The above fault plan determination method obtains the fault phenomenon data of the power system; inputs the fault phenomenon data into the fault phenomenon of the historical period and the plan matching model obtained by the fault plan training for matching processing to obtain the target fault plan. The embodiment of the present application improves the accuracy and interpretation ability of the fault plan matching of the power automation system based on the complex semantic processing and natural language generation capabilities of the plan matching model, thereby realizing the automatic matching of the fault phenomenon of the power dispatching automation system with the fault plan, and improving the speed of dispatching personnel in handling faults.

[0059] In an exemplary embodiment, based on the above embodiment, see Figure 3 The plan matching model in the embodiment of the present application includes a feature extraction layer and a classification layer; the embodiment of the present application involves a process of inputting fault phenomenon data into a pre-built plan matching model for matching processing, including the following S301 to S302. Among them:

[0060] S301, inputting the fault phenomenon data into the feature extraction layer for feature extraction processing to obtain feature data.

[0061] Among them, the feature extraction layer can use text feature extraction, image feature extraction, and other feature extraction, for example, using a clustering algorithm to cluster features together, selecting K optimal features, or using a recursive feature elimination algorithm to select features.

[0062] In an embodiment of the present application, based on the fault phenomenon data obtained in the above embodiment, the terminal performs feature extraction on the feature extraction layer of the fault phenomenon data plan matching model for different data types, such as text feature extraction, image feature extraction, and other feature extraction, for example, using a clustering algorithm to cluster features together, selecting K optimal features, or using a recursive feature elimination algorithm to select features, etc., to obtain feature data.

[0063] S302, input the feature data into the classification layer for classification processing to obtain a target fault plan.

[0064] Among them, the classification layer can use binary linear classifier, logistic regression, SVM support vector machine, KNN-K nearest neighbor algorithm.

[0065] In the embodiment of the present application, based on the feature data obtained in the above embodiment, the terminal inputs the feature data into the classification layer for classification processing to obtain a target fault plan. The specific process of classification processing can be to give a training set, treat a new instance, find the K points that are closest to it (distance measurement), and the new instance belongs to the class with the most of the K points (majority vote). Or a classification method based on Bayes' theorem and the conditional independence assumption of features requires that each sample is independent and identically distributed. The idea is: first learn the joint probability distribution of (x, y) based on the training data, and then find the y with the largest posterior probability based on the input x and the model.

[0066] In the embodiment of the present application, the characteristics of the data can be improved by extracting features from the fault phenomenon data so that the classification layer can classify the data and improve the accuracy of the classification.

[0067] In an exemplary embodiment, based on the above embodiment, see Figure 4 The present application relates to a training process of a plan matching model including the following S401 to S403. Among them:

[0068] S401, obtaining fault phenomenon samples and fault contingency plan samples.

[0069] Among them, the fault phenomenon samples include data samples of generator set faults, transmission line faults, substation faults, busbar faults, etc. The fault plan samples are plans designed by the back-end personnel corresponding to the fault phenomenon samples for different fault phenomena.

[0070] In the embodiment of the present application, sensors or camera devices can be set in various devices of the power system, and the sensors or camera devices regularly obtain various data of the on-site operation of various devices of the power system. After the above-mentioned fault occurs, these fault phenomenon samples are sent to the terminal, and the terminal obtains data such as generator set fault, transmission line fault, substation fault, bus fault, etc. to perform fault analysis. Fault plan samples can be obtained by input by background personnel.

[0071] S402, inputting the fault phenomenon samples and the fault plan samples into the pre-training model respectively to perform plan prediction processing to obtain multiple prediction results.

[0072] The pre-trained model refers to the model to be trained corresponding to the plan matching model.

[0073] In an embodiment of the present application, based on the fault phenomenon samples and fault plan samples obtained in the above-mentioned embodiment, the terminal inputs the fault phenomenon samples and fault plan samples into the pre-training model for plan prediction processing to obtain multiple prediction results. The specific process of plan prediction processing is as follows: the fault phenomenon data samples are input into the feature extraction layer for feature extraction processing to obtain feature data. The feature extraction layer can use text feature extraction, image feature extraction, and other feature extraction, for example, using a clustering algorithm to cluster features together, selecting K optimal features, or using a recursive feature elimination algorithm to select features. The feature data is input into the classification layer for classification processing to obtain multiple prediction results.

[0074] S403, if the matching accuracy determined according to the prediction result does not reach the accuracy threshold, the parameters in the pre-trained model are adjusted until the matching accuracy reaches the accuracy threshold, and the adjustment is completed to obtain the plan matching model.

[0075] In the embodiment of the present application, based on the prediction results obtained in the above embodiment, the terminal compares the prediction results with the fault plan samples, calculates the matching accuracy, and compares it with the accuracy threshold given in advance. If the matching accuracy determined according to the prediction results does not reach the accuracy threshold, the parameters in the pre-trained model are adjusted until the matching accuracy reaches the accuracy threshold and the adjustment is completed to obtain the plan matching model. The specific process of the adjustment process can be manually adjusted by the backend personnel, or automatically adjusted by the model.

[0076] In the embodiment of the present application, the model parameters are adjusted so that the model converges to the expected effect, thereby making the fault plan prediction result more accurate.

[0077] In an exemplary embodiment, based on the above embodiment, see Figure 5 The embodiment of the present application relates to a process of adjusting parameters in a pre-trained model, including:

[0078] S501, extracting the original parameters of the feature extraction layer to obtain extraction parameters.

[0079] In the embodiment of the present application, the terminal extracts the original parameters of the feature extraction layer to obtain the extracted parameters. Specifically, the given parameters of the model are traversed to obtain the extracted parameters.

[0080] S502, input the extracted parameters into the fine-tuning model for low-rank fine-tuning processing to obtain updated parameters.

[0081] Among them, the fine-tuning model can adopt the LoRA (Low-Rank Adaptation, fine-tuning model) model. The LoRA model indirectly trains some dense layers in the neural network by optimizing the rank decomposition matrix that changes in the dense layer during the adaptation process. In the process of fine-tuning the original parameters using the fine-tuning model, an adaptive optimizer is used to improve training efficiency and reduce resource usage.

[0082] S503: merge the updated parameters with the original parameters to obtain adjusted model parameters.

[0083] In an embodiment of the present application, based on the updated parameters obtained in the above embodiment, the terminal can replace the original parameters with the updated parameters, or merge the original parameters and the updated parameters to obtain adjusted model parameters.

[0084] In the embodiment of the present application, an adaptive optimizer is used to improve training efficiency and reduce resource usage by fine-tuning the original parameters using a fine-tuning model.

[0085] In an exemplary embodiment, based on the above embodiment, see Figure 6 The fine-tuning model of the embodiment of the present application includes a dimensionality reduction layer and a dimensionality increase layer. The embodiment of the present application involves a process of inputting the extracted parameters into the fine-tuning model for low-rank fine-tuning processing to obtain updated parameters, including the following S601 to S602. Among them:

[0086] S601, input the extracted parameters into the dimension reduction layer for dimension reduction processing to obtain a dimension reduction vector.

[0087] In the embodiment of the present application, based on the extraction parameters obtained in the above embodiment, the terminal can input the extraction parameters into the dimensionality reduction layer of the fine-tuning model for dimensionality reduction processing to obtain a dimensionality reduction vector. The dimensionality reduction processing can adopt a linear dimensionality reduction method, a distribution neighborhood embedding, a local linear embedding of an autoencoder, and the like.

[0088] S602, input the obtained rank and dimension reduction vector into the dimension increasing layer for dimension increasing processing to obtain update parameters.

[0089] The rank is calculated based on the vector matrix.

[0090] In an embodiment of the present application, the terminal inputs the acquired rank and dimension reduction vector into the dimension increase layer for dimension increase processing to obtain updated parameters. Specifically: the fine-tuning model adds a small number of trainable parameter models to the model, while the original model parameters remain frozen. The fine-tuning model decomposes the weight matrix into two smaller weight matrices, as shown below, to approximate fully supervised fine-tuning in a more parameter-efficient manner. The specific idea is: compared with all the weights in the weight matrix (W) of the fine-tuned pre-trained large language model, fine-tune two smaller matrices (A and B), which are approximate to the update of the original matrix. During reasoning, the results of the left and right parts can be added together, and the trained matrix product BA is added to the original weight matrix W as the new weight parameter to replace the weight matrix W of the original pre-trained language model.

[0091] In the embodiment of the present application, a fine-tuning model is adopted and fewer parameters are used, which significantly reduces the computational complexity and video memory usage during the model training process.

[0092] In an exemplary embodiment, based on the above embodiment, the method of the embodiment of the present application includes: sending the target fault plan to the terminal of the backend personnel, and also includes obtaining the actual measures taken from the terminal, and retraining the plan matching model according to the actual measures taken.

[0093] Among them, the target fault plan includes at least one of preventive warning, closed management and regional control.

[0094] In an embodiment of the present application, based on the target fault plans such as preventive warning, closed management and regional control obtained in the above embodiments, the target fault plan is sent to the terminal of the backend personnel so that the staff can actually take measures according to the target fault plan and retrain the plan matching model based on the actual measures taken.

[0095] In the embodiment of the present application, the fault phenomenon of the electric power dispatching automation system and the fault plan can be automatically matched by retraining the plan matching model, thereby improving the speed at which dispatching personnel handle faults.

[0096] In an exemplary embodiment, based on the above embodiment, the method of the present application further includes the following steps:

[0097] Step 1: Obtain fault phenomenon samples and fault plan samples; input the fault phenomenon samples and fault plan samples into the pre-training model for plan prediction processing to obtain multiple prediction results;

[0098] Step 2: If the matching accuracy determined according to the prediction result does not reach the accuracy threshold, the original parameters of the feature extraction layer are extracted to obtain the extracted parameters; the extracted parameters are input into the dimension reduction layer for dimension reduction to obtain the reduced dimension vector; the obtained rank and reduced dimension vector are input into the dimension increase layer for dimension increase to obtain the updated parameters, and the updated parameters are combined with the original parameters to obtain the adjusted model parameters, until the matching accuracy reaches the accuracy threshold and the adjustment is completed to obtain the plan matching model;

[0099] Step 3: Obtain fault phenomenon data of the power system;

[0100] Step 4: Input the fault phenomenon data into the feature extraction layer of the plan matching model for feature extraction processing to obtain feature data; input the feature data into the classification layer of the plan matching model for classification processing to obtain the target fault plan.

[0101] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0102] Based on the same inventive concept, the embodiment of the present application also provides a fault plan determination device for implementing the fault plan determination method involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more fault plan determination device embodiments provided below can refer to the limitations of the fault plan determination method above, and will not be repeated here.

[0103] In one embodiment, Figure 7 As shown, a fault plan determination device 700 is provided, comprising:

[0104] The data acquisition module 701 is used to acquire the fault phenomenon data of the power system;

[0105] The plan matching module 702 inputs the fault phenomenon data into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training the fault phenomena and fault plans in the historical period.

[0106] In one embodiment, the above-mentioned plan matching model includes a feature extraction layer and a classification layer; the above-mentioned plan matching module includes:

[0107] A feature extraction unit, used for inputting the fault phenomenon data into the feature extraction layer for feature extraction processing to obtain feature data;

[0108] The classification unit is used to input the feature data into the classification layer for classification processing to obtain the target fault plan.

[0109] In one embodiment, the above device further comprises:

[0110] A sample acquisition module is used to acquire fault phenomenon samples and fault contingency plan samples;

[0111] The plan prediction module is used to input the fault phenomenon samples and the fault plan samples into the pre-training model for plan prediction processing to obtain multiple prediction results;

[0112] The parameter adjustment module is used to adjust the parameters in the pre-trained model if the matching accuracy determined according to the prediction results does not reach the accuracy threshold, until the matching accuracy reaches the accuracy threshold and the adjustment is completed to obtain the plan matching model.

[0113] In one embodiment, the parameter adjustment module includes:

[0114] A parameter extraction unit, used for extracting the original parameters of the feature extraction layer to obtain extraction parameters;

[0115] A parameter updating unit, used for inputting the extracted parameters into the fine-tuning model for low-rank fine-tuning processing to obtain updated parameters;

[0116] The parameter merging unit is used to merge the updated parameters with the original parameters to obtain the adjusted model parameters.

[0117] In one embodiment, the fine-tuning model includes a dimensionality reduction layer and a dimensionality increase layer, and the parameter updating unit includes:

[0118] The parameter dimension reduction subunit is used to input the extracted parameters into the dimension reduction layer for dimension reduction processing to obtain a dimension reduction vector;

[0119] The parameter dimension-increasing subunit is used to input the acquired rank and dimension-reduced vector into the dimension-increasing layer for dimension-increasing processing to obtain updated parameters.

[0120] In one embodiment, the above-mentioned target fault plan includes at least one of preventive warning, closed management and regional control; the above-mentioned device also includes: a plan sending module, which is used to send the target fault plan to the terminal of the backend personnel.

[0121] In one embodiment, the above-mentioned device also includes: a model retraining module, which is used to obtain actual measures taken from the terminal and retrain the plan matching model according to the actual measures taken.

[0122] Each module in the above-mentioned fault plan determination device can be implemented in whole or in part by software, hardware and a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0123] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit and an input device. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface, the display unit and the input device are connected to the system bus via the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for determining a fault plan is implemented. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0124] Those skilled in the art will understand that Figure 8 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 those shown in the figure, or combine certain components, or have a different arrangement of components.

[0125] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0126] Obtain fault phenomenon data of the power system;

[0127] The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

[0128] In one embodiment, the above-mentioned plan matching model includes a feature extraction layer and a classification layer; inputting the fault phenomenon data into the pre-built plan matching model for matching processing includes:

[0129] Input the fault phenomenon data into the feature extraction layer for feature extraction processing to obtain feature data;

[0130] The feature data is input into the classification layer for classification processing to obtain the target fault plan.

[0131] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0132] Obtain fault phenomenon samples and fault contingency plan samples;

[0133] The fault phenomenon samples and the fault plan samples are respectively input into the pre-training model for plan prediction processing to obtain multiple prediction results;

[0134] If the matching accuracy determined according to the prediction result does not reach the accuracy threshold, the parameters in the pre-trained model are adjusted until the matching accuracy reaches the accuracy threshold, and the adjustment is completed to obtain the plan matching model.

[0135] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0136] Extracting and processing the original parameters of the feature extraction layer to obtain extraction parameters;

[0137] The extracted parameters are input into the fine-tuning model for low-rank fine-tuning to obtain updated parameters;

[0138] The updated parameters are combined with the original parameters to obtain the adjusted model parameters.

[0139] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0140] The extracted parameters are input into the dimension reduction layer for dimension reduction processing to obtain a dimension reduction vector;

[0141] The obtained rank and dimension reduction vector are input into the dimension increasing layer for dimension increasing processing to obtain the updated parameters.

[0142] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0143] Send the target fault plan to the backend personnel's terminal.

[0144] In one embodiment, when the processor executes the computer program, the following steps are also implemented:

[0145] The actual measures taken are obtained from the terminal, and the plan matching model is retrained based on the actual measures taken.

[0146] In one embodiment, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0147] Obtain fault phenomenon data of the power system;

[0148] The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

[0149] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0150] Input the fault phenomenon data into the feature extraction layer for feature extraction processing to obtain feature data;

[0151] The feature data is input into the classification layer for classification processing to obtain the target fault plan.

[0152] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0153] Obtain fault phenomenon samples and fault contingency plan samples;

[0154] The fault phenomenon samples and the fault plan samples are respectively input into the pre-training model for plan prediction processing to obtain multiple prediction results;

[0155] If the matching accuracy determined according to the prediction result does not reach the accuracy threshold, the parameters in the pre-trained model are adjusted until the matching accuracy reaches the accuracy threshold, and the adjustment is completed to obtain the plan matching model.

[0156] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0157] Extracting and processing the original parameters of the feature extraction layer to obtain extraction parameters;

[0158] The extracted parameters are input into the fine-tuning model for low-rank fine-tuning to obtain updated parameters;

[0159] The updated parameters are combined with the original parameters to obtain the adjusted model parameters.

[0160] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0161] The extracted parameters are input into the dimension reduction layer for dimension reduction processing to obtain a dimension reduction vector;

[0162] The obtained rank and dimension reduction vector are input into the dimension increasing layer for dimension increasing processing to obtain the updated parameters.

[0163] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0164] Send the target fault plan to the backend personnel's terminal.

[0165] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0166] The actual measures taken are obtained from the terminal, and the plan matching model is retrained based on the actual measures taken.

[0167] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0168] Obtain fault phenomenon data of the power system;

[0169] The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

[0170] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0171] Input the fault phenomenon data into the feature extraction layer for feature extraction processing to obtain feature data;

[0172] The feature data is input into the classification layer for classification processing to obtain the target fault plan.

[0173] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0174] Obtain fault phenomenon samples and fault contingency plan samples;

[0175] The fault phenomenon samples and the fault plan samples are respectively input into the pre-training model for plan prediction processing to obtain multiple prediction results;

[0176] If the matching accuracy determined according to the prediction result does not reach the accuracy threshold, the parameters in the pre-trained model are adjusted until the matching accuracy reaches the accuracy threshold, and the adjustment is completed to obtain the plan matching model.

[0177] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0178] Extracting and processing the original parameters of the feature extraction layer to obtain extraction parameters;

[0179] The extracted parameters are input into the fine-tuning model for low-rank fine-tuning to obtain updated parameters;

[0180] The updated parameters are combined with the original parameters to obtain the adjusted model parameters.

[0181] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0182] The extracted parameters are input into the dimension reduction layer for dimension reduction processing to obtain a dimension reduction vector;

[0183] The obtained rank and dimension reduction vector are input into the dimension increasing layer for dimension increasing processing to obtain the updated parameters.

[0184] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0185] Send the target fault plan to the backend personnel's terminal.

[0186] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0187] The actual measures taken are obtained from the terminal, and the plan matching model is retrained based on the actual measures taken.

[0188] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions.

[0189] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.

[0190] The technical features of the above embodiments may 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.

[0191] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for determining a fault plan, characterized in that: The method comprises: Obtain fault phenomenon data of the power system; The fault phenomenon data is input into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training fault phenomena and fault plans in historical periods.

2. The method according to claim 1, characterized in that The plan matching model includes a feature extraction layer and a classification layer; the inputting of the fault phenomenon data into the pre-built plan matching model for matching processing includes: Inputting the fault phenomenon data into the feature extraction layer for feature extraction processing to obtain feature data; The characteristic data is input into the classification layer for classification processing to obtain the target fault plan.

3. The method according to claim 2, characterized in that The training process of the plan matching model includes: Obtain fault phenomenon samples and fault contingency plan samples; Inputting the fault phenomenon samples and the fault plan samples into the pre-training model respectively for plan prediction processing to obtain multiple prediction results; If the matching accuracy determined according to the prediction result does not reach the accuracy threshold, the parameters in the pre-trained model are adjusted until the matching accuracy reaches the accuracy threshold, and the adjustment is terminated to obtain the plan matching model.

4. The method according to claim 3, characterized in that The adjusting process of the parameters in the pre-trained model includes: Extracting the original parameters of the feature extraction layer to obtain extraction parameters; Inputting the extracted parameters into a fine-tuning model for low-rank fine-tuning processing to obtain updated parameters; The updated parameters and the original parameters are combined to obtain adjusted model parameters.

5. The method according to claim 4, characterized in that The fine-tuning model includes a dimensionality reduction layer and a dimensionality increase layer, and the extraction parameters are input into the fine-tuning model for low-rank fine-tuning to obtain updated parameters, including: Inputting the extracted parameters into a dimensionality reduction layer for dimensionality reduction processing to obtain a dimensionality reduction vector; The obtained rank and the dimension reduction vector are input into the dimension increasing layer for dimension increasing processing to obtain the update parameter.

6. The method according to claim 5, characterized in that The target fault plan includes at least one of preventive warning, closed management and regional control; the method also includes: sending the target fault plan to a terminal of a backstage personnel.

7. The method according to any one of claims 1 to 6, characterized in that: The method further includes: acquiring actual measures taken from the terminal, and retraining the plan matching model according to the actual measures taken.

8. A device for determining a fault plan, characterized in that: The device comprises: A data acquisition module, used to acquire fault phenomenon data of the power system; The plan matching module inputs the fault phenomenon data into a pre-built plan matching model for matching processing to obtain a target fault plan; the plan matching model is obtained by training the fault phenomena and fault plans in the historical period.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images