Operation information determination method and device, equipment, storage medium and program product

Abstract

The invention relates to an operation information determination method and device, equipment, a storage medium and a program product. The method comprises the following steps: receiving question information, wherein the question information is used for constructing a to-be-operated risk scene in the field of power enterprises; inputting the question information into a preset risk decision model to determine target operation information in the current risk scene according to the risk decision model, the preset risk decision model being obtained based on historical operation data training of the power enterprise; and sending the target operation information to an operator, wherein the target operation information is used for guiding the operator to perform operation processing on the power equipment in the risk scene. By adopting the method, the accuracy of determining the operation information can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a method, device, equipment, storage medium and program product for determining operation information. Background Art

[0002] Many safety accidents in the power industry have caused people's concerns, especially personal injury and damage to power equipment caused by improper operation. In order to greatly reduce or even eliminate the occurrence of these situations, it is necessary to identify the risks and then measure and analyze the risks, and determine the corresponding operation information to guide operators to perform their operations.

[0003] At present, the existing daily operations in the power industry mainly rely on the theoretical knowledge and experience of operators to determine the operation information under risk scenarios.

[0004] However, when determining the operation information in this way, it is highly dependent on the operator's theoretical knowledge and experience, contains subjective factors, and is not very accurate. Summary of the invention

[0005] Based on this, it is necessary to provide an operation information determination method, device, equipment, storage medium and program product that can improve the accuracy of determining operation information in response to the above technical problems.

[0006] In a first aspect, the present application provides a method for determining operation information, comprising:

[0007] Receive question information, which is used to construct risk scenarios for operations to be performed in the field of power enterprises; input the question information into a preset risk decision model to determine target operation information under the current risk scenario based on the risk decision model, and the preset risk decision model is trained based on the historical operation data of the power enterprise; send the target operation information to the operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario.

[0008] In one embodiment, the training process of the risk decision model includes:

[0009] Obtain the historical operation data of the power company stored in the data warehouse, which includes the historical operation data of power equipment, the historical operation data of business personnel and the historical safety accident data; obtain the basic model and pre-train the basic model based on the historical operation data; adjust the pre-trained basic model to obtain the risk decision model.

[0010] In one embodiment, obtaining a basic model and pre-training the basic model based on historical job data includes:

[0011] Perform data preprocessing on historical operation data; determine feature data from the preprocessed historical operation data according to a preset feature extraction method, the feature data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents; pre-train the basic model based on the feature data.

[0012] In one embodiment, the pre-trained basic model is adjusted to obtain a risk decision model, including:

[0013] Based on the preset LoRA model, the low-rank parameter matrix of the pre-trained basic model is adjusted; the pre-trained basic model is updated according to the trained low-rank parameter matrix to adjust the pre-trained basic model.

[0014] In one embodiment, the training process of the risk decision model includes supervised learning and unsupervised learning. Supervised learning is used to learn from the annotated historical operation data and establish a prediction sub-model; unsupervised learning is used to learn from the unlabeled historical operation data and determine the inherent laws and characteristics of the historical operation data.

[0015] In one embodiment, the weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; the operation content related to historical safety accidents includes operation type, operation duration and operation area; the operator information related to historical safety accidents includes the operator's qualifications, age, gender and health condition.

[0016] In a second aspect, the present application further provides an operation information determination device, comprising:

[0017] A receiving module, used for receiving question information, and the question information is used to construct a risk scenario to be operated in the field of power enterprises;

[0018] A determination module is used to input the question information into a preset risk decision model to determine the target operation information under the current risk scenario according to the risk decision model, wherein the preset risk decision model is trained based on the historical operation data of the power enterprise;

[0019] The sending module is used to send target operation information to the operator, and the target operation information is used to guide the operator to operate the power equipment under risk scenarios.

[0020] In one of the embodiments, the operation information determination device also includes a training module, which is used to obtain historical operation data of the power enterprise stored in the data warehouse, the historical operation data including historical operation data of power equipment, historical operation data of business personnel and historical safety accident data; obtain a basic model, and pre-train the basic model based on the historical operation data; adjust the pre-trained basic model to obtain a risk decision model.

[0021] In one of the embodiments, the training module is specifically used to perform data preprocessing on historical operation data; determine feature data from the preprocessed historical operation data according to a preset feature extraction method, the feature data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents; and pre-train the basic model based on the feature data.

[0022] In one of the embodiments, the training module is specifically used to adjust the low-rank parameter matrix of the pre-trained basic model based on the preset LoRA model; and update the pre-trained basic model according to the trained low-rank parameter matrix to adjust the pre-trained basic model.

[0023] In one embodiment, the training process of the risk decision model includes supervised learning and unsupervised learning. Supervised learning is used to learn from the annotated historical operation data and establish a prediction sub-model; unsupervised learning is used to learn from the unlabeled historical operation data and determine the inherent laws and characteristics of the historical operation data.

[0024] In one embodiment, the weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; the operation content related to historical safety accidents includes operation type, operation duration and operation area; the operator information related to historical safety accidents includes the operator's qualifications, age, gender and health condition.

[0025] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, any of the methods described in the first aspect is implemented.

[0026] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements any of the methods described in the first aspect above.

[0027] In a fifth aspect, the present application further provides a computer program product, including a computer program, which, when executed by a processor, implements any of the methods described in the first aspect above.

[0028] The above-mentioned operation information determination method, device, equipment, storage medium and program product receive question information, which is used to construct a risk scenario to be operated in the field of the power enterprise, and then input the question information into a preset risk decision model to determine the target operation information in the current risk scenario according to the risk decision model. The preset risk decision model is trained based on the historical operation data of the power enterprise. Finally, the target operation information is sent to the operator, and the target operation information is used to guide the operator to operate the power equipment in the risk scenario. Before performing the power operation in the risk scenario, the question information is input into the preset risk decision model. Since the risk decision model is trained based on the historical operation data of the power enterprise, it can accurately understand the question information and identify the risk items in the current operation scenario according to the question information. Then, the target operation information is accurately determined based on the historical operation process data, and the target operator can operate according to the target operation information. In this way, compared with the target operation information determined by the experience and professional knowledge of the operator, the target operation information determined based on the risk decision model is more accurate, thereby providing the operator with professional knowledge and safety guidance to effectively control or reduce risks when operating in risk scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0030] FIG1 is a schematic flow chart of a method for determining operation information in one embodiment;

[0031] FIG2 is a schematic diagram of a process flow of a risk decision model training process in one embodiment;

[0032] FIG3 is a schematic flow chart of the steps of pre-training a basic model in one embodiment;

[0033] FIG4 is a schematic flow chart of the steps of adjusting the pre-trained basic model in one embodiment;

[0034] FIG5 is a schematic diagram of the architecture of a LoRA model in one embodiment;

[0035] FIG6 is a schematic flow chart of a method for determining operation information in another embodiment;

[0036] FIG7 is a structural block diagram of an operation information determination device in one embodiment;

[0037] FIG8 is a structural block diagram of an operation information determination device in another embodiment;

[0038] FIG. 9 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0039] 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.

[0040] Many safety accidents in the power industry have caused people's concerns, especially personal injury and damage to power equipment caused by improper operation. In order to greatly reduce or even eliminate the occurrence of these situations, it is necessary to identify the risks and then measure and analyze the risks, and determine the corresponding operation information to guide operators to perform their operations.

[0041] At present, the existing daily operations in the power industry mainly rely on the theoretical knowledge and experience of operators to determine the operation information under risk scenarios.

[0042] However, when determining the operation information in this way, it is highly dependent on the operator's theoretical knowledge and experience, contains subjective factors, and is not very accurate.

[0043] In view of this, the present application provides an operation information determination method that can improve the accuracy when determining operation information. The operation information determination method provided in the embodiment of the present application, its execution subject can be an operation information determination device, and the operation information determination device can be implemented by software, hardware, or a combination of software and hardware. It can be embedded in or independent of the processor in the computer device in the form of hardware, or it can be stored in the memory of the computer device in the form of software. The computer device can be a server, and the server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides cloud computing services. The computer device can also be a terminal or a robot, and the embodiment of the present application does not limit this. In the following method embodiments, the execution subject is a computer device as an example for explanation.

[0044] In an exemplary embodiment, as shown in FIG1 , a method for determining operation information is provided, comprising the following steps 101 to 103. Among them:

[0045] S101, receiving question information.

[0046] Among them, the question information is used to construct the risk scenarios to be carried out in the field of power enterprises.

[0047] Optionally, the question information may be input by the operator through the input device of the terminal. The question information may include text format, image format or audio format, etc. Correspondingly, the input device of the terminal may include a keyboard, a touch screen, an image acquisition device, a sound acquisition device or a voice device engine, etc. The embodiment of the present application does not limit this and can be set according to actual needs.

[0048] Optionally, risk scenarios to be performed in the power enterprise field may include the installation, commissioning, maintenance or repair scenarios of power equipment, or scenarios such as performing operations when the power equipment is energized. The power equipment may include power equipment in power plants, power equipment such as transformers, circuit breakers and disconnect switches in substations, and power equipment involving urban and rural distribution lines.

[0049] S102, inputting the question information into a preset risk decision model to determine target operation information under the current risk scenario according to the risk decision model.

[0050] Among them, the preset risk decision model is trained based on the historical operation data of the power company.

[0051] In one possible implementation, the risk decision model can be a multimodal model that can process data in multiple different modalities at the same time. That is, the risk decision model can process question information input in text format, question information input in image format, and question information input in audio format to determine the target operation information under the current risk scenario.

[0052] In another possible implementation, the risk decision model may include multiple sub-models, each sub-model corresponding to data of a different modality, so that the corresponding sub-model is determined according to the format of the input question information, and then the sub-model determines the target operation information under the current risk scenario based on the question information.

[0053] Optionally, when the executor of the embodiment of the present application is a terminal or a robot, the risk decision model can be trained on the terminal or robot based on the historical operation data of the power enterprise, or it can be trained on the server based on the historical operation data of the power enterprise and integrated on the terminal or robot.

[0054] Optionally, the risk decision model can be a model that is built based on natural language processing technology and visual image processing technology and is capable of automatically answering question information. It can learn how to understand question information and generate appropriate answers from a large amount of historical operating data of power companies and / or from a large amount of professional knowledge of power companies, and use the generated answers as target operation information under the current risk scenario.

[0055] Optionally, when the risk decision model determines the target operation information under the current risk scenario based on the question information, it can include understanding the question information, information retrieval, context understanding, generating operation information and continuous learning, wherein understanding the question information can include identifying the subject, intention and related keywords of the question information, and identifying the risk items in the current operation scenario; information retrieval can be after understanding the question information, retrieving information from the knowledge base or related data of the risk decision model, and the knowledge base and related data can be composed of the historical operation data and professional knowledge of the power company; context understanding can be by combining previous question information to better understand and answer the current question information; generating operation information can be converting the retrieved information into a concise and accurate answer, which can include combining, summarizing or re-expressing the retrieved information so that the answer matches the question information; continuous learning can be that as the number of question information received increases, the risk decision model can continuously improve its problem understanding ability and operation information generation ability, and continuously optimize.

[0056] Optionally, the target operation information may include target operation procedures and target operation precautions, etc.

[0057] S103, sending the target operation information to the operator.

[0058] Among them, the target operation information is used to guide operators to operate power equipment under risk scenarios.

[0059] Optionally, the target operation information may be sent to an operator in the format of voice, text or image, and the operator may operate the power equipment according to the target operation information.

[0060] Exemplarily, the risk decision model can be connected through a visual operation interface and a background connection interface, question information can be received through the visual operation interface, and after the target operation information is determined, it can be displayed to the operator through the visual operation interface. The visual operation interface can be implemented by using the front-end VUE framework technology.

[0061] The above-mentioned operation information determination method receives question information, which is used to construct the risk scenario to be operated in the field of the power enterprise, and then inputs the question information into a preset risk decision model to determine the target operation information under the current risk scenario according to the risk decision model. The preset risk decision model is trained based on the historical operation data of the power enterprise. Finally, the target operation information is sent to the operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario. Before performing the power operation in the risk scenario, the question information is input into the preset risk decision model. Since the risk decision model is trained based on the historical operation data of the power enterprise, it can accurately understand the question information and identify the risk items in the current operation scenario according to the question information. Then, the target operation information is accurately determined based on the historical operation process data, and the target operator can operate according to the target operation information. In this way, compared with the target operation information determined by the operator's experience and professional knowledge, the target operation information determined based on the risk decision model is more accurate, thereby providing the operator with professional knowledge and safety guidance to effectively control or reduce risks when operating in risk scenarios.

[0062] In an exemplary embodiment, as shown in FIG2 , optionally, the training process of the risk decision model includes the following steps 201 to 203 . Among them:

[0063] S201, obtaining the historical operation data of the power enterprise stored in the data warehouse.

[0064] Among them, historical operation data includes historical operation data of power equipment, historical operation data of business personnel and historical safety accident data.

[0065] Optionally, the data warehouse may also store information such as the power company's professional knowledge and standard operating procedures.

[0066] Optionally, the data warehouse may be a hive data warehouse, a cloud data warehouse, or other types of data warehouses, etc. The embodiment of the present application does not limit the type of the data warehouse.

[0067] Optionally, the hive data warehouse can be built using a distributed cluster environment, consisting of a monitoring module, a load module, a data storage module, and an API module. It can support millions of concurrent connections and meet the different data collection needs of dedicated transformers, distribution transformers, all metering points of power plants and stations, and low-voltage users. In other words, the historical operating data of power companies can be collected and saved through the hive data warehouse.

[0068] S202, obtaining a basic model, and pre-training the basic model based on historical operation data.

[0069] Optionally, the basic model can be a new model based on a CNN convolutional neural network for processing question information, or it can be a model that has been trained on a large-scale general corpus and has extensive language understanding capabilities, for example, a basic model in a Python machine learning library.

[0070] Optionally, after obtaining the basic model, the basic model can be trained based on the historical operating data of the power company, and the basic model can also be pre-trained based on the professional knowledge and standard operating procedures of the power company stored in the data warehouse.

[0071] S203, adjusting the pre-trained basic model to obtain a risk decision model.

[0072] Optionally, by adjusting the pre-trained basic model, a risk decision model that can better understand and process the language and terminology in the power enterprise domain can be obtained. In this way, when training the risk decision model, instead of learning a specific task from scratch, the knowledge points learned on the pre-trained basic model are used as a starting point to learn and solve new tasks. In this way, the amount of training data required for a specific task can be reduced, while improving the performance of the risk decision model.

[0073] Optionally, the pre-trained base model is adjusted by freezing most of the parameters in the pre-trained base model and fine-tuning a small number or additional parameters.

[0074] In one possible implementation, the adjustment of the pre-trained base model can be achieved by adding additional parameters. For example, the pre-trained base model is adjusted through an adapter adjustment method, and a small neural network module called an "adapter" is inserted between each layer or selected layers of the base model. During the adjustment process, the model parameters of the pre-trained base model remain unchanged, and the parameters of the adapter are trained to obtain a risk decision model.

[0075] In another possible implementation, the adjustment of the pre-trained base model can be achieved by introducing re-parameterization, for example, adjusting the pre-trained base model through a low-rank adapter method (Low-Rank Adapation, LoRA) or a quantized low-rank adapter method (Quantized Low-Rank Adapation, QLoRA).

[0076] In another possible implementation, the adjustment of the pre-trained base model can be achieved by adding a trainable, task-specific prefix before the pre-trained base model input sequence. Similarly, during the adjustment process, the model parameters of the pre-trained base model remain unchanged, but the prefix is ​​added to the input to adjust the pre-trained base model. The prefix can be fixed, that is, a manually designed static prompt, or it can be trainable, that is, a dynamic prompt learned by the model during the adjustment process.

[0077] Optionally, the training process of the risk decision model includes supervised learning and unsupervised learning. Supervised learning is used to learn from the annotated historical operation data and establish a prediction sub-model; unsupervised learning is used to learn from the unlabeled historical operation data and determine the inherent laws and characteristics of the historical operation data.

[0078] Optionally, unsupervised learning can also be used in the pre-training process of the basic model, receiving training on large-scale unlabeled historical job data, so that the basic model can learn the meaning of various words, sentence construction rules, and basic information, context, internal laws and characteristics of historical job data.

[0079] Optionally, the amount of data can be increased to deepen the training of the risk decision-making model, improve the accuracy and generalization ability of the risk decision-making model, and continuously adjust the model parameters while increasing the training data. The model parameters may include the learning rate, number of iterations, and number of hidden layer nodes, so as to improve the accuracy and stability of the risk decision-making model.

[0080] In an exemplary embodiment, as shown in FIG3 , optionally, a basic model is obtained and pre-trained based on historical operation data, including the following steps 301 to 303 . Among them:

[0081] S301, performing data preprocessing on historical operation data.

[0082] Optionally, data preprocessing of historical operation data may include data cleaning and data organization, removing erroneous data, duplicate data and incomplete data, and normalizing and standardizing the data.

[0083] S302: determining feature data from the preprocessed historical operation data according to a preset feature extraction method.

[0084] Among them, characteristic data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents.

[0085] Optionally, a preset feature extraction method can determine all relevant feature data that may cause personal safety from the preprocessed historical operation data. The preset feature extraction method can be a feature extraction standard determined by business experts of the power company, and data that meets the feature extraction standard is found from the preprocessed historical operation data. Then, these data that meet the feature extraction standard are used as feature data. It can also be implemented based on a feature selection method. For example, the preset feature extraction method can be a maximum mutual information coefficient method, a support vector machine recursive feature elimination method, or a maximum correlation minimum redundancy algorithm.

[0086] Optionally, the preprocessed historical operation data may be processed based on a preset feature extraction method to obtain the features most relevant to the risk decision model.

[0087] Optionally, weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; operation content related to historical safety accidents includes operation type, operation duration and operation area; operator information related to historical safety accidents includes operator qualifications, age, gender and health condition.

[0088] For example, weather conditions related to historical safety accidents, such as heavy rain, lightning, strong winds, etc., may affect the operation of the power grid and the maintenance of power equipment, thereby increasing personal safety risks; the work content related to historical safety accidents includes such as high-altitude work, contact with high-voltage equipment, operation of dangerous chemicals, etc., and the working time, such as work intensity level, work schedule, etc., high-intensity work may cause operator fatigue and lack of concentration, thereby increasing the probability of accidents; the qualifications of operators related to historical safety accidents include the operator's academic qualifications, work experience, skill certificates, etc. The qualifications and skill levels of operators have an important impact on personal safety. Age includes the operator's date of birth, length of service, etc. Young operators may lack experience and skills, while older operators may have problems such as decreased physical function, which will increase personal safety risks. The operator's gender and health status also have an impact on the risk. Among them, men and obese people may be at higher risk in certain high-intensity jobs or those requiring physical coordination. The operator's health status may affect their work ability and judgment, thereby increasing the risk of accidents. The operator's health status includes characteristic data such as the operator's physical examination report and illness status; the determined characteristic data can be saved in the characteristic library of the data warehouse system.

[0089] S303, pre-training the basic model according to the feature data.

[0090] Optionally, the feature data is used as a training sample set, and a basic model is pre-trained based on the feature data through unsupervised learning.

[0091] In an exemplary embodiment, as shown in FIG4 , optionally, the pre-trained basic model is adjusted to obtain a risk decision model, including the following steps 401 to 402. Among them:

[0092] S401, adjusting the low-rank parameter matrix of the pre-trained basic model based on the preset LoRA model.

[0093] S402: Update the pre-trained basic model according to the adjusted low-rank parameter matrix to adjust the pre-trained basic model.

[0094] Optionally, the low-rank parameters of the pre-trained base model can be the weight matrices of the multi-head self-attention and feed-forward neural network parts of the base model, or the weight matrices in the fully connected layers of the base model that contain matrix multiplication.

[0095] Optionally, the LoRA model can simulate the change of parameters through low-rank decomposition, thereby achieving model-based indirect training with extremely small parameters.

[0096] Optionally, as shown in Figure 5, after determining the weight matrix that needs to be fine-tuned, two low-rank matrices A and B with smaller dimensions can be introduced. Assuming that the size of the original weight matrix W0 (Pretrained Weights) of the pre-trained basic model is d×d, the sizes of matrices A and B may be d×r and r×d, where r is much smaller than d; a new matrix is ​​generated by multiplying the two low-rank matrices A and B by BA. The rank of this new matrix BA is r, which is much smaller than the rank of the original weight matrix. During the adjustment process, the weight parameters of matrix A can be initialized by a Gaussian function, and the weight parameters of matrix B can be initialized to a zero matrix; the generated low-rank matrix BA is superimposed on the original weight matrix to obtain a new weight matrix, that is, W=W0+BA, and the original weight matrix is ​​fine-tuned by continuously adjusting the weight parameters of matrices A and B.

[0097] Optionally, the adjusted risk decision model is tested and reviewed by industry experts based on the test results. Meanwhile, the risk decision model is continuously optimized to continuously improve the accuracy of the determined target operation information.

[0098] The above-mentioned preset LoRA model adjusts the low-rank parameter matrix of the pre-trained basic model, and updates the pre-trained basic model according to the adjusted low-rank parameter matrix to adjust the pre-trained basic model. In this way, the adjustment of the model is achieved by introducing a small, low-rank matrix in the decisive layer of the pre-trained basic model. There is no need to make major modifications to the entire model structure. Without increasing the computational burden, the model can be effectively adjusted so that it can better understand and process the language and terminology in the field of power enterprises, thereby improving the accuracy of determining target operation information.

[0099] As an optional implementation, as shown in FIG6 , the operation information determination method provided in the embodiment of the present application may include the following specific steps:

[0100] S601, obtaining the historical operation data of the power enterprise stored in the data warehouse.

[0101] Among them, historical operation data includes historical operation data of power equipment, historical operation data of business personnel and historical safety accident data.

[0102] S602: preprocess the historical operation data.

[0103] S603: determining feature data from the preprocessed historical operation data according to a preset feature extraction method.

[0104] Among them, characteristic data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents.

[0105] S604, obtaining a basic model, and pre-training the basic model according to the feature data.

[0106] S605, adjusting the low-rank parameter matrix of the pre-trained basic model based on the preset LoRA model.

[0107] S606, updating the pre-trained basic model according to the trained low-rank parameter matrix to adjust the pre-trained basic model to obtain a risk decision model.

[0108] S607, receiving question information.

[0109] Among them, the question information is used to construct the risk scenarios to be carried out in the field of power enterprises.

[0110] S608: Input the question information into a preset risk decision model to determine target operation information under the current risk scenario according to the risk decision model.

[0111] Among them, the preset risk decision model is trained based on the historical operation data of the power company.

[0112] S609, sending the target operation information to the operator.

[0113] Among them, the target operation information is used to guide operators to operate power equipment under risk scenarios.

[0114] 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.

[0115] Based on the same inventive concept, the embodiment of the present application also provides an operation information determination device for implementing the operation information 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 the one or more operation information determination device embodiments provided below can refer to the limitations of the operation information determination method above, and will not be repeated here.

[0116] In an exemplary embodiment, as shown in FIG. 7 , an operation information determination device 700 is provided, comprising: a receiving module 701, a determining module 702 and a sending module 703, wherein:

[0117] The receiving module 701 is used to receive question information, and the question information is used to construct a risk scenario to be operated in the field of the power enterprise.

[0118] The determination module 702 is used to input the question information into a preset risk decision model to determine the target operation information under the current risk scenario according to the risk decision model. The preset risk decision model is trained based on the historical operation data of the power enterprise.

[0119] The sending module 703 is used to send the target operation information to the operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario.

[0120] In an exemplary embodiment, as shown in FIG8 , the operation information determination device 700 also includes a training module 704, which is used to obtain historical operation data of the power enterprise stored in the data warehouse, the historical operation data including historical operation data of power equipment, historical operation data of business personnel and historical safety accident data; obtain a basic model, and pre-train the basic model based on the historical operation data; and adjust the pre-trained basic model to obtain a risk decision model.

[0121] In an exemplary embodiment, the training module 704 is specifically used to perform data preprocessing on historical operation data; determine feature data from the preprocessed historical operation data according to a preset feature extraction method, and the feature data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents; and pre-train the basic model according to the feature data.

[0122] In an exemplary embodiment, the training module 704 is specifically used to adjust the low-rank parameter matrix of the pre-trained basic model based on the preset LoRA model; update the pre-trained basic model according to the trained low-rank parameter matrix to adjust the pre-trained basic model.

[0123] In an exemplary embodiment, the training process of the risk decision model includes supervised learning and unsupervised learning. Supervised learning is used to learn from the historical operation data that has been labeled to establish a prediction sub-model; unsupervised learning is used to learn from the historical operation data that has not been labeled to determine the inherent laws and characteristics of the historical operation data.

[0124] In an exemplary embodiment, weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; operation content related to historical safety accidents includes operation type, operation duration and operation area; operator information related to historical safety accidents includes operator qualifications, age, gender and health condition.

[0125] Each module in the above-mentioned operation information determination device can be implemented in whole or in part by software, hardware, or 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.

[0126] In an exemplary embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be shown in FIG9. The computer device includes a processor, a memory, an input / output interface, and a communication interface. 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, a near field communication (NFC) or other technologies. When the computer program is executed by the processor, a method for determining operation information is implemented.

[0127] Those skilled in the art will understand that the structure shown in FIG. 9 is merely a block diagram of a partial 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 arrangement of components.

[0128] In an exemplary 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:

[0129] Receive question information, which is used to construct risk scenarios for operations to be performed in the field of power enterprises; input the question information into a preset risk decision model to determine target operation information under the current risk scenario based on the risk decision model, and the preset risk decision model is trained based on the historical operation data of the power enterprise; send the target operation information to the operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario.

[0130] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0131] Obtain the historical operation data of the power company stored in the data warehouse, which includes the historical operation data of power equipment, the historical operation data of business personnel and the historical safety accident data; obtain the basic model and pre-train the basic model based on the historical operation data; adjust the pre-trained basic model to obtain the risk decision model.

[0132] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0133] Perform data preprocessing on historical operation data; determine feature data from the preprocessed historical operation data according to a preset feature extraction method, the feature data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents; pre-train the basic model based on the feature data.

[0134] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0135] Based on the preset LoRA model, the low-rank parameter matrix of the pre-trained basic model is adjusted; the pre-trained basic model is updated according to the trained low-rank parameter matrix to adjust the pre-trained basic model.

[0136] In one embodiment, the training process of the risk decision model includes supervised learning and unsupervised learning. Supervised learning is used to learn from the annotated historical operation data and establish a prediction sub-model; unsupervised learning is used to learn from the unlabeled historical operation data and determine the inherent laws and characteristics of the historical operation data.

[0137] In one embodiment, weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; operation content related to historical safety accidents includes operation type, operation duration and operation area; operator information related to historical safety accidents includes operator qualifications, age, gender and health condition.

[0138] 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:

[0139] Receive question information, which is used to construct risk scenarios for operations to be performed in the field of power enterprises; input the question information into a preset risk decision model to determine target operation information under the current risk scenario based on the risk decision model, and the preset risk decision model is trained based on the historical operation data of the power enterprise; send the target operation information to the operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario.

[0140] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0141] Obtain the historical operation data of the power company stored in the data warehouse, which includes the historical operation data of power equipment, the historical operation data of business personnel and the historical safety accident data; obtain the basic model and pre-train the basic model based on the historical operation data; adjust the pre-trained basic model to obtain the risk decision model.

[0142] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0143] Perform data preprocessing on historical operation data; determine feature data from the preprocessed historical operation data according to a preset feature extraction method, the feature data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents; pre-train the basic model based on the feature data.

[0144] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0145] Based on the preset LoRA model, the low-rank parameter matrix of the pre-trained basic model is adjusted; the pre-trained basic model is updated according to the trained low-rank parameter matrix to adjust the pre-trained basic model.

[0146] In one embodiment, the training process of the risk decision model includes supervised learning and unsupervised learning. Supervised learning is used to learn from the annotated historical operation data and establish a prediction sub-model; unsupervised learning is used to learn from the unlabeled historical operation data and determine the inherent laws and characteristics of the historical operation data.

[0147] In one embodiment, weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; operation content related to historical safety accidents includes operation type, operation duration and operation area; operator information related to historical safety accidents includes operator qualifications, age, gender and health condition.

[0148] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0149] Receive question information, which is used to construct risk scenarios for operations to be performed in the field of power enterprises; input the question information into a preset risk decision model to determine target operation information under the current risk scenario based on the risk decision model, and the preset risk decision model is trained based on the historical operation data of the power enterprise; send the target operation information to the operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario.

[0150] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0151] Obtain the historical operation data of the power company stored in the data warehouse, which includes the historical operation data of power equipment, the historical operation data of business personnel and the historical safety accident data; obtain the basic model and pre-train the basic model based on the historical operation data; adjust the pre-trained basic model to obtain the risk decision model.

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

[0153] Perform data preprocessing on historical operation data; determine feature data from the preprocessed historical operation data according to a preset feature extraction method, the feature data is used to describe weather conditions, operation content, operation intensity and operator information related to historical safety accidents; pre-train the basic model based on the feature data.

[0154] In one embodiment, when the computer program is executed by a processor, the following steps are also implemented:

[0155] Based on the preset LoRA model, the low-rank parameter matrix of the pre-trained basic model is adjusted; the pre-trained basic model is updated according to the trained low-rank parameter matrix to adjust the pre-trained basic model.

[0156] In one embodiment, the training process of the risk decision model includes supervised learning and unsupervised learning. Supervised learning is used to learn from the annotated historical operation data and establish a prediction sub-model; unsupervised learning is used to learn from the unlabeled historical operation data and determine the inherent laws and characteristics of the historical operation data.

[0157] In one embodiment, weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; operation content related to historical safety accidents includes operation type, operation duration and operation area; operator information related to historical safety accidents includes operator qualifications, age, gender and health condition.

[0158] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method 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 memory 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), magnetic 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, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0159] 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 application.

[0160] 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 operation information, characterized in that: The method comprises: Receiving question information, wherein the question information is used to construct a risk scenario to be operated in the field of the power enterprise; Inputting the question information into a preset risk decision model to determine target operation information under the current risk scenario according to the risk decision model, wherein the preset risk decision model is trained based on the historical operation data of the power enterprise; The target operation information is sent to an operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario.

2. The method according to claim 1, characterized in that The training process of the risk decision model includes: Acquire the historical operation data of the power enterprise stored in the data warehouse, wherein the historical operation data includes the historical operation data of power equipment, the historical operation data of business personnel and the historical safety accident data; Acquire a basic model, and pre-train the basic model based on the historical operation data; The pre-trained basic model is adjusted to obtain the risk decision model.

3. The method according to claim 2, characterized in that The obtaining of the basic model and pre-training the basic model based on the historical operation data includes: Performing data preprocessing on the historical operation data; Determining feature data from the preprocessed historical operation data according to a preset feature extraction method, wherein the feature data is used to describe weather conditions, operation content, operation intensity, and operator information related to historical safety accidents; The basic model is pre-trained according to the feature data.

4. The method according to claim 2, characterized in that: The step of adjusting the pre-trained basic model to obtain the risk decision model includes: Adjusting the low-rank parameter matrix of the pre-trained basic model based on a preset LoRA model; The pre-trained basic model is updated according to the trained low-rank parameter matrix to adjust the pre-trained basic model.

5. The method according to claim 2, characterized in that: The training process of the risk decision model includes supervised learning and unsupervised learning. The supervised learning is used to learn from the annotated historical operation data to establish a prediction sub-model; the unsupervised learning is used to learn from the unlabeled historical operation data to determine the inherent laws and characteristics of the historical operation data.

6. The method according to claim 3, characterized in that The weather conditions related to historical safety accidents include weather type, weather duration, wind speed and rainfall; The operation contents related to the historical safety accidents include the operation type, operation duration and operation area; The operator information related to historical safety accidents includes the operator's qualifications, age, gender and health condition.

7. An operation information determination device, characterized in that: The device comprises: A receiving module, used for receiving question information, wherein the question information is used for constructing a risk scenario to be operated in the field of the power enterprise; A determination module, used for inputting the question information into a preset risk decision model to determine target operation information under the current risk scenario according to the risk decision model, wherein the preset risk decision model is trained based on the historical operation data of the power enterprise; The sending module is used to send the target operation information to the operator, and the target operation information is used to guide the operator to operate the power equipment under the risk scenario.

8. 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 6 are implemented.

9. 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 6 are implemented.

10. A computer program product, comprising a computer program, 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 6 are implemented.

Images