Nuclear power experience feedback model training method and device and electronic equipment

Abstract

The invention is suitable for the technical field of nuclear power, and provides a nuclear power experience feedback model training method and device and electronic equipment, and the method comprises the steps: obtaining an experience feedback report of a nuclear power plant; extracting corresponding content from the experience feedback report according to a preset event element to obtain the content of the event element; acquiring a work instruction corresponding to the experience feedback report; according to the event elements, the content of the event elements, the experience feedback report and a work instruction corresponding to the experience feedback report, training a nuclear power experience feedback model to be trained to obtain a trained nuclear power experience feedback model, the trained nuclear power experience feedback model is used for outputting corresponding experience feedback according to the received work instruction. Through the mode, the comprehensiveness of the output experience feedback is improved, and the efficiency of the output experience feedback is also improved.

Description

Technical Field

[0001] The present application belongs to the field of nuclear power technology, and in particular relates to a training method for a nuclear power experience feedback model, an experience feedback acquisition method, a training device for a nuclear power experience feedback model, an experience feedback acquisition device, an electronic device, a computer-readable storage medium, and a computer program product. Background Art

[0002] A nuclear power plant, also known as a nuclear power plant, is a facility that uses a nuclear reactor to convert nuclear energy into electrical energy. Since the nuclear reaction process produces radioactive waste, it is extremely important to use the experience feedback of nuclear power plants to maintain nuclear power plants and improve their safety. Among them, the experience feedback of nuclear power plants refers to the root cause analysis (RCA), solutions and experience summary of important production management issues of nuclear power plants throughout their life cycle.

[0003] The experience feedback of nuclear power plants is the most important part of nuclear power plant operation. It is also an important part of ensuring nuclear safety, improving operational efficiency, avoiding recurrence of incidents, saving operating costs, and promoting the sustainable development of new quality productivity. IAEA research points out that during the operating cycle of more than 50 years, nuclear power plants need to maintain the required skills and capabilities through knowledge management. The importance of collecting key and important knowledge through operating experience cases of operating units is particularly significant, and the primary challenge of retaining and utilizing this important knowledge is how to scientifically organize and plan and quickly and conveniently obtain this nuclear power experience and pass it on to current staff in a timely manner.

[0004] Nuclear power plants at home and abroad attach great importance to experience feedback, and a large number of experience feedback reports are accumulated during the decades of operation. However, the application efficiency of these valuable experience feedback reports in the work process of nuclear power plants is still not high. Specifically, the staff often consults the relevant experience feedback reports in history after the incident occurs, and refers to these experience feedback reports to deal with related production problems. When the experience feedback reports are used in this way, the utilization rate of the experience feedback reports is low. Summary of the invention

[0005] The embodiments of the present application provide a training method, device and electronic equipment for a nuclear power experience feedback model, which can solve the problem of low utilization rate of experience feedback reports in existing methods.

[0006] In a first aspect, an embodiment of the present application provides a training method for a nuclear power experience feedback model, comprising:

[0007] Obtain experience feedback reports from nuclear power plants;

[0008] Extracting corresponding content from the experience feedback report according to preset event elements to obtain the content of the event elements;

[0009] Obtaining work instructions corresponding to the experience feedback report;

[0010] The nuclear power experience feedback model to be trained is trained according to the event elements, the content of the event elements, the experience feedback report and the work instructions corresponding to the experience feedback report to obtain a trained nuclear power experience feedback model, wherein the trained nuclear power experience feedback model is used to output corresponding experience feedback according to the received work instructions.

[0011] In an embodiment of the present application, the nuclear power experience feedback model to be trained is trained according to the event elements, the content of the event elements, the experience feedback report, and the work instructions corresponding to the experience feedback report to obtain a trained nuclear power experience feedback model, and the trained nuclear power experience feedback model is used to output the corresponding experience feedback according to the received work instructions. Since the nuclear power experience feedback model can output the corresponding experience feedback according to the input work instructions, that is, there is no need to perform manual search, thus improving the efficiency of obtaining experience feedback. In addition, since the experience feedback report, the preset event elements, the content of the event elements, and the work instructions are used in the training process, the trained nuclear power experience feedback model can perform a more comprehensive experience feedback search based on the experience feedback report, or perform a faster experience feedback search based on the content of the event elements, thereby improving the comprehensiveness of the output experience feedback, and improving the efficiency of the output experience feedback, thereby improving the utilization rate of the experience feedback report.

[0012] In a second aspect, an embodiment of the present application provides a method for obtaining experience feedback, including:

[0013] Obtain work orders to be evaluated;

[0014] The work instruction to be evaluated is used as the input of the trained nuclear power experience feedback model as described in the first aspect to obtain the experience feedback output by the trained nuclear power experience feedback model.

[0015] In a third aspect, an embodiment of the present application provides a training device for a nuclear power experience feedback model, comprising:

[0016] An experience feedback report acquisition module is used to obtain experience feedback reports from nuclear power plants;

[0017] An event element extraction module, used to extract corresponding content from the experience feedback report according to preset event elements to obtain the content of the event elements;

[0018] A work instruction acquisition module, used to acquire a work instruction corresponding to the experience feedback report;

[0019] The nuclear power experience feedback model training module is used to train the nuclear power experience feedback model to be trained according to the event elements, the content of the event elements, the experience feedback report and the work instructions corresponding to the experience feedback report to obtain the trained nuclear power experience feedback model, wherein the trained nuclear power experience feedback model is used to output corresponding experience feedback according to the received work instructions.

[0020] In a fourth aspect, an embodiment of the present application provides an experience feedback acquisition device, characterized in that it includes:

[0021] A work instruction acquisition module, used to acquire the work instructions to be evaluated;

[0022] The experience feedback output module is used to use the work instruction to be evaluated as the input of the trained nuclear power experience feedback model as described in the first aspect, and obtain the experience feedback output by the trained nuclear power experience feedback model.

[0023] In a fifth aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method described in the first aspect is implemented, or the method described in the second aspect is executed.

[0024] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and wherein when the computer program is executed by a processor, the computer program implements the method described in the first aspect, or executes the method described in the second aspect.

[0025] In a seventh aspect, an embodiment of the present application provides a computer program product, which, when executed on an electronic device, enables the electronic device to execute the method described in the first aspect above, or enables the electronic device to execute the method described in the second aspect above.

[0026] It can be understood that the beneficial effects of the second to seventh aspects mentioned above can be found in the relevant description of the first aspect mentioned above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the embodiments or the description of the prior art are briefly introduced below.

[0028] Figure 1 It is a flowchart of a training method for a nuclear power experience feedback model provided by an embodiment of the present application;

[0029] Figure 2It is a logical relationship diagram of event elements provided by an embodiment of the present application;

[0030] Figure 3 It is a flowchart of a method for obtaining experience feedback provided in one embodiment of the present application;

[0031] Figure 4 It is a structural schematic diagram of a training device for a nuclear power experience feedback model provided in one embodiment of the present application;

[0032] Figure 5 is a structural schematic diagram of an experience feedback acquisition device provided by another embodiment of the present application;

[0033] Figure 6 It is a structural schematic diagram of an electronic device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0034] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present application. However, it should be clear to those skilled in the art that the present application may also be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present application.

[0035] It should be understood that when used in the present specification and the appended claims, the term "comprising" indicates the presence of described features, wholes, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or combinations thereof.

[0036] It should also be understood that the term “and / or” used in the specification and appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0037] References to "one embodiment" or "some embodiments" etc. described in the specification of this application mean that one or more embodiments of the present application include specific features, structures or characteristics described in conjunction with the embodiment. Therefore, the phrases "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. appearing in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways.

[0038] During the operation of nuclear power plants, a large number of experience feedback reports are usually accumulated. Taking a nuclear power base as an example, after more than 30 years of production and operation, thousands of experience feedback reports have been accumulated. Each experience feedback report records the whole process of the incident and the valuable handling experience in detail. These experience feedback reports are rare knowledge assets to guide the safe production and operation of nuclear power plants.

[0039] At present, nuclear power plant staff often consult historical experience feedback reports only after an event (here, an event refers to matters that need to be reported according to international, domestic and nuclear power plant-related standards, regulations and work procedures) occurs. Since there are many historical experience feedback reports, it takes a long time to find relevant experience feedback reports. In addition, since the event has already occurred, that is, the consequences of the event have already occurred, the relevant experience feedback reports found can provide low utilization value, resulting in a low utilization rate of the experience feedback reports.

[0040] In order to improve the utilization rate of the experience feedback report, the embodiment of the present application provides a training method for a nuclear power experience feedback model. In the method, the nuclear power experience feedback model is trained using the experience feedback report, preset event elements, the content of the event elements extracted from the experience feedback report according to the preset event elements, and the work instructions corresponding to the experience feedback report.

[0041] The training method of the nuclear power experience feedback model provided in the embodiment of the present application is described below with reference to the accompanying drawings.

[0042] Figure 1 A flow chart of a training method for a nuclear power experience feedback model provided in an embodiment of the present application is shown. The method can be applied to electronic equipment in a nuclear power plant, and is described in detail as follows:

[0043] S11, obtain experience feedback reports from nuclear power plants.

[0044] The experience feedback report here can be in text form or in table form, which is not limited here.

[0045] In the embodiment of the present application, experience feedback reports of different nuclear power plants in different periods can be obtained to obtain as comprehensive experience feedback reports as possible.

[0046] Optionally, to improve the diversity of the acquired experience feedback reports, the above S11 includes:

[0047] A plurality of the above-mentioned experience feedback reports of the nuclear power plant are obtained, wherein among the plurality of the above-mentioned experience feedback reports, at least two of the above-mentioned experience feedback reports have different contents of event elements corresponding to the above-mentioned preset event elements.

[0048] The event element may be the occurrence process of the event, or may be the cause of the event or a corrective action, etc., which is not limited here. For example, the event element corresponding to the occurrence process of the event may be "occurrence process".

[0049] In an embodiment of the present application, when the contents of the event elements corresponding to the same event element in two experience feedback reports are different, it indicates that the two experience feedback reports are reports reflecting different events. For example, when the event element is the cause of the event, since the event causes corresponding to the two events are different, the two events are different events, that is, the two experience feedback reports corresponding to the two events are different experience feedback reports. Since among the multiple experience feedback reports obtained, there are at least two experience feedback reports with different contents of the event elements corresponding to the same event element, the obtained experience feedback reports can cover more types of events, which is beneficial to improve the accuracy of the nuclear power experience feedback model obtained by subsequent training.

[0050] Optionally, in order to improve the accuracy of the acquired experience feedback report, after acquiring the experience feedback report of the nuclear power plant, each acquired experience feedback report is preprocessed, and the preprocessing includes at least one of the following: filtering duplicate experience feedback reports, filtering incomplete experience feedback reports, and filtering erroneous experience feedback reports.

[0051] The duplicate experience feedback reports here refer to experience feedback reports with exactly the same event elements. For example, when the event elements are the cause of the event, the process of the event and the corrective action, if the descriptions of the cause of the event, the process of the event and the corrective action in two experience feedback reports are the same, then the two experience feedback reports are determined to be duplicate experience feedback reports.

[0052] The incomplete experience feedback report here refers to an experience feedback report that is missing information, such as an experience feedback report that does not include all event elements.

[0053] The erroneous experience feedback report here refers to an experience feedback report that has been proven to be erroneous. For example, if a corrective action recorded in an experience feedback report is judged to be erroneous after subsequent analysis, then the experience feedback report will be judged as an erroneous experience feedback report. Optionally, in order to improve the accuracy of determining erroneous experience feedback reports, each experience feedback report can be analyzed to find out the contradictory experience feedback reports, and then analyze whether there are erroneous experience feedback reports in the contradictory experience feedback reports. Since there is a contradiction between two experience feedback reports, it usually indicates that one or both of the experience feedback reports are erroneous. Therefore, determining the erroneous experience feedback report from the contradictory experience feedback reports can improve the accuracy of the determination results.

[0054] S12, extracting corresponding content from the above-mentioned experience feedback report according to preset event elements to obtain the content of the above-mentioned event elements.

[0055] Among them, the preset event elements can be set according to the actual situation. For example, the number of preset event elements can be 1, such as "occurrence process"; for another example, the number of preset event elements can be greater than 1, such as "equipment", "work type", "occurrence process", "consequences", "cause of event", "corrective action", etc. Here, "equipment" refers to the equipment where the event occurs, and "work type" refers to the scene of the event, such as the scene of operation by the operating personnel, the scene of valve maintenance, etc. Further, the above-mentioned "cause of event" can be further subdivided into "contributing factors", "root causes" and "direct causes", and correspondingly, the above-mentioned "corrective actions" can be further subdivided into corrective actions for "contributing factors", "root causes" and "direct causes".

[0056] In order to increase the probability of extracting more detailed event elements from the experience feedback report, multiple event elements can be set, and at least two event elements in each event element are logically related, and the contents corresponding to different event elements may also be cross-related. Assume that the preset multiple event elements include: power plant, unit, system, component type, component failure mode, failure principle, unit status, when a certain part is working, work type, a certain profession, event, human failure, management failure, equipment failure, event cause factors, affected systems, other units, contributing causes, root causes, direct causes, component A, failure phenomenon, failure mode, event description, event consequences, etc. If a straight line (solid or dotted line) is used to connect two event elements to represent the logical relationship between the two event elements, then when the content of the event elements is extracted from the nuclear power experience report according to these event elements, it is equivalent to following Figure 2 The connection relationship shown extracts the content of the event element.

[0057] exist Figure 2 In the diagram, the power plant (i.e., nuclear power plant) is connected to the unit, indicating that the unit belongs to the power plant; the unit is connected to the system, indicating that the system is the system applied by the unit. Similarly, the connection relationship between equipment, component type, component failure mode, failure principle, etc. is similar to the connection relationship between the power plant and the unit, and will not be repeated here.

[0058] Since nuclear power experience reports usually correspond to events at nuclear power plants, Figure 2 The "event" in the nuclear power experience report is equivalent to the nuclear power experience report. Before extracting the content of the event element, it is necessary to first determine the various event elements of the nuclear power experience report. The following is a method flow for determining the various event elements of the nuclear power experience report. Specifically, determine the unit where the event occurred, the unit status of the unit, the work performed by the unit (i.e. Figure 2 After the crew where the incident occurred is determined, if the crew has a standard work instruction set, the work type and specialty specified in the standard work instruction set are determined, and the work type and specialty of the crew when performing a certain work are determined.

[0059] Determine whether the event that occurred when the crew was engaged in a certain work was a human failure, management failure, or equipment failure, and determine the cause of the event, the system affected by the event, and the consequences of the event. In addition, determine other crews that were fed back by the event. Among them, the determined cause factors of the event include contributing factors, root causes, and direct causes. After determining the cause factors of the event, determine the corrective actions corresponding to the contributing factors, root causes, and direct causes.

[0060] In the case where the event is characterized as equipment failure, determine the component where the failure point is located (i.e. Figure 2 ), determine the component type to which component A belongs, the failure phenomenon displayed when component A fails, the failure mode to which the failure phenomenon belongs, and the event description corresponding to the failure phenomenon.

[0061] Since there is a certain logical relationship between each event element, when the content of the event element is extracted from the nuclear power experience report according to the above method, more comprehensive and logically related event element content can be extracted, thereby improving the accuracy of the extracted event element content.

[0062] Of course, in actual situations, other event elements may also be set, such as time, location, etc., which are not limited here.

[0063] The content of the event element here refers to the specific event content corresponding to the event element. For example, if the event occurs on December 16, 2024, when the event element is "time", the event element corresponding to the "time" is "2024.12.16".

[0064] In the embodiment of the present application, the content of the event element corresponding to the preset event element in the experience feedback report can be determined manually, or the content of the event element corresponding to the preset event element can be extracted from the experience feedback report through a pre-trained neural network model, which is not limited here.

[0065] S13, obtaining a work instruction corresponding to the above experience feedback report.

[0066] Among them, one experience feedback report may correspond to one or more work instructions. The work instruction corresponding to the experience feedback report refers to the work instruction corresponding to the event corresponding to the experience feedback report. For example, the work instruction at the time when the event is generated can be used as the work instruction corresponding to the experience feedback report; for another example, all work instructions between the preset time before the event corresponding to the experience feedback report is generated and the time when the event is generated can be used as the work instruction corresponding to the experience feedback report, and so on.

[0067] Specifically, considering that the working processes of nuclear power plants are standardized, that is, before each work is implemented, the preparation engineer calls the appropriate work instructions to carry out pre-work preparations, and then determines the work steps, spare parts required, reference materials, etc., that is, there is a certain relationship between the work instructions and the experience feedback reports. Therefore, according to the information of the event corresponding to each experience feedback report (such as time, place, equipment where the event occurred, etc.), the work instructions corresponding to the experience feedback report can be accurately found.

[0068] S14, training the nuclear power experience feedback model to be trained according to the above-mentioned event elements, the content of the above-mentioned event elements, the above-mentioned experience feedback report and the work instructions corresponding to the above-mentioned experience feedback report to obtain a trained nuclear power experience feedback model, wherein the above-mentioned trained nuclear power experience feedback model is used to output corresponding experience feedback according to the received work instructions.

[0069] Among them, the nuclear power experience feedback model to be trained is a neural network model.

[0070] Optionally, considering that the experience feedback report is usually in the form of text, a suitable model architecture that performs well in processing natural language tasks, such as Transformer, GPT, BERT, etc., can be selected as the nuclear power experience feedback model to be trained. Among them, Transformer is a neural network architecture for processing sequence data. It introduces a self-attention mechanism and can effectively model the dependencies between sequences; GPT (Generative Pre-trained Transformer) is a generative pre-trained model based on Transformer. BERT (Bidirectional Encoder Representations from Transformers) is a pre-trained language model based on the Transformer structure. It learns rich language representations through unsupervised training on large-scale corpus.

[0071] After determining the nuclear power experience feedback model to be trained, a sample for training the nuclear power experience feedback model to be trained is constructed according to event elements, content of event elements, experience feedback reports, and work instructions corresponding to the experience feedback reports.

[0072] For example, positive samples for training are constructed based on the event elements with corresponding relationships, the content of the event elements, the experience feedback report, and the work instructions. Further, to improve the accuracy of the constructed positive samples, after determining that the experience feedback report is a correct report, positive samples for training are constructed based on the event elements with corresponding relationships, the content of the event elements, the experience feedback report, and the work instructions.

[0073] For another example, negative samples for training are constructed based on event elements, content of event elements, experience feedback reports, and work instructions that have no corresponding relationship; or negative samples for training are constructed based on erroneous experience feedback reports, event elements, content of event elements, and work instructions.

[0074] After constructing the samples (or positive samples and negative samples) for training, the constructed samples with labeled information are used to iteratively train the nuclear power experience feedback model to be trained until the nuclear power experience feedback model obtained after training meets the preset evaluation indicators. The iterative training is stopped. Among them, the preset evaluation indicators include accuracy (Precision), recall rate (Recall), F1 score (the F1 score is an indicator used in machine learning to measure the accuracy of the binary classification (or multi-task binary classification) model, which takes into account the accuracy and recall rate of the classification model at the same time). In practical applications, corresponding evaluation indicators can be selected according to different task requirements to comprehensively measure the performance of the trained nuclear power experience feedback model. For example, when the task requirement pays more attention to accuracy, select accuracy as the preset evaluation indicator, and when the task requirement pays more attention to recall rate, select recall rate as the preset evaluation indicator.

[0075] During the training process, if it is determined that the nuclear power experience feedback model in training does not meet the preset evaluation index, the nuclear power experience feedback model in training is optimized or the training strategy is adjusted; or, if it is determined that the feedback experience output by the nuclear power experience feedback model in training is wrong, the wrong feedback experience is analyzed, and the nuclear power experience feedback model in training is optimized or the training strategy is adjusted according to the analysis result to improve the performance of the nuclear power experience feedback model in training. Optimizing the nuclear power experience feedback model in training includes adjusting the model architecture, and adjusting the training strategy includes increasing the training samples or improving the feature extraction method.

[0076] Optionally, in order to improve the accuracy of the trained nuclear power experience feedback model, after obtaining the trained nuclear power experience feedback model that meets the preset evaluation indicators through the above method, the hyperparameters (such as learning rate, batch size, number of iterations, etc.) of the trained nuclear power experience feedback model that meets the preset evaluation indicators can be adjusted through cross-validation and other methods, and the best set of hyperparameter combinations are selected as the hyperparameters of the trained nuclear power experience feedback model that meets the preset evaluation indicators to obtain the final trained nuclear power experience feedback model.

[0077] Since the trained nuclear power experience feedback model has learned the basic structure of work instructions, event elements, the content of event elements, and experience feedback reports, and has learned the semantic rules between the above work instructions, event elements, the content of event elements, and experience feedback report data, after obtaining the trained nuclear power experience feedback model, the trained nuclear power experience feedback model can be used to predict which experience feedback (the experience feedback includes event elements and / or experience feedback reports) a work instruction is compatible with. For example, assuming that the preset event elements include "work type" and "event cause", and a certain work content (process) is a work instruction for valve dismantling and maintenance (or standard work instruction), then the trained nuclear power experience feedback model includes the following processing flow for the standard work instruction: according to the work instruction for valve dismantling and maintenance, searching for the event element "work type" in history, and the content of its event element is "valve maintenance", and the event element is "event cause", and the content of the event element corresponding to the "event cause" involves the experience feedback event of "valve dismantling", and outputting the experience feedback corresponding to the found experience feedback event.

[0078] In an embodiment of the present application, the nuclear power experience feedback model to be trained is trained according to the event elements, the content of the event elements, the experience feedback report, and the work instructions corresponding to the experience feedback report to obtain a trained nuclear power experience feedback model, and the trained nuclear power experience feedback model is used to output the corresponding experience feedback according to the received work instructions. Since the nuclear power experience feedback model can output the corresponding experience feedback according to the input work instructions, that is, there is no need to perform manual search, thus improving the efficiency of obtaining experience feedback. In addition, since the experience feedback report, the preset event elements, the content of the event elements, and the work instructions are used in the training process, the trained nuclear power experience feedback model can perform a more comprehensive experience feedback search based on the experience feedback report, or perform a faster experience feedback search based on the content of the event elements, thereby improving the comprehensiveness of the output experience feedback, and improving the efficiency of the output experience feedback, thereby improving the utilization rate of the experience feedback report.

[0079] In some embodiments, in order to improve the complexity and speed of training the nuclear power experience feedback model to be trained, before the above S14, training the nuclear power experience feedback model to be trained according to the above event elements, the content of the above event elements, the above experience feedback report and the work instructions corresponding to the above experience feedback report, it also includes:

[0080] The experience feedback report is segmented, and the segmented experience feedback report includes at least two text units, and the amount of data contained in each of the text units is smaller than the amount of data contained in the experience feedback report.

[0081] Correspondingly, the above S14, training the nuclear power experience feedback model to be trained according to the above event elements, the content of the above event elements, the above experience feedback report and the work instructions corresponding to the above experience feedback report, includes:

[0082] The nuclear power experience feedback model to be trained is trained according to the above event elements, the content of the above event elements, the segmented experience feedback reports and the work instructions corresponding to the above experience feedback reports.

[0083] Specifically, the experience feedback report is segmented into at least two text units, and all text units are combined to obtain the above experience feedback report. When segmenting, the segmentation can be performed based on sentences, paragraphs, or directories, etc. For example, if the experience feedback report has 1,000 sentences and is segmented based on sentences, the experience feedback report can be segmented into 1,000 text units.

[0084] Since the experience feedback report is divided into one or more text units containing less data, and the text units containing less data are conducive to reducing the computational burden of the nuclear power experience feedback model in training when processing, it is conducive to reducing the computational complexity. At the same time, since the trained nuclear power experience feedback model can process each text unit containing less data more quickly, the processing efficiency of the nuclear power experience feedback model in training for the overall experience feedback report is improved.

[0085] In some embodiments, before the above S14, training the nuclear power experience feedback model to be trained according to the above event elements, the content of the above event elements, the segmented experience feedback reports and the work instructions corresponding to the above experience feedback reports, it also includes:

[0086] The segmented experience feedback report and the work instructions corresponding to the experience feedback report are vectorized to obtain vectorized data.

[0087] Among them, the purpose of vectorization is to convert text into a numerical form that can be processed by the machine learning model (i.e., the nuclear power experience feedback model to be trained).

[0088] Specifically, according to a preset vectorization method, vectorization is performed on each segmented experience feedback report and the work instruction corresponding to the experience feedback report to obtain corresponding vectorized data.

[0089] Optionally, during the vectorization process, corresponding weights may be assigned to the experience feedback report and the work instruction according to their importance, and then corresponding vectorized data may be obtained according to the assigned weights.

[0090] Correspondingly, the above S14, training the nuclear power experience feedback model to be trained according to the above event elements, the content of the above event elements, the segmented experience feedback reports, and the work instructions corresponding to the above experience feedback reports, includes:

[0091] The nuclear power experience feedback model to be trained is trained according to the above event elements, the content of the above event elements and the above vectorized data.

[0092] Specifically, samples can be constructed based on event elements, the content of event elements, and vectorized data, and the nuclear power experience feedback model to be trained can be trained based on the constructed samples. Alternatively, after extracting corresponding features from event elements, the content of event elements, and the vectorized data, the extracted features are used to construct samples for training the nuclear power experience feedback model to be trained. Since the nuclear power experience feedback model to be trained is trained after the features are extracted, the complexity of training is reduced. Optionally, after extracting corresponding features from event elements, the content of event elements, and the vectorized data, corresponding weights can be assigned to the extracted corresponding features according to the importance of the event elements, the content of event elements, and the vectorized data, and then the extracted features and the assigned weights are used to construct samples for training the nuclear power experience feedback model to be trained.

[0093] In an embodiment of the present application, considering that the recognition obtained by the nuclear power experience feedback model to be trained of the vectorized data is more accurate than the original text, therefore, after the experience feedback report and work instructions are vectorized, the nuclear power experience feedback model to be trained is trained based on the event elements, the content of the event elements and the vectorized data, which is beneficial to improve the accuracy of the training results.

[0094] Optionally, the number of the event elements is greater than 1, and the vectorizing of the segmented experience feedback report and the work instructions corresponding to the experience feedback report includes:

[0095] A vectorization method that can reflect the semantic order is selected to vectorize the segmented experience feedback report and the work instructions corresponding to the experience feedback report.

[0096] The above-mentioned semantic order refers to the arrangement order of semantic components reflected in the syntactic structure.

[0097] Optionally, vectorization methods that can reflect semantic order include Doc2Vec and BERT. Doc2Vec, also known as Paragraph2vec or Sentence embeddings, is an unsupervised algorithm that can obtain vectors of sentences, paragraphs, and documents. Its expression is an extension of Word2Vec and is used to learn document-level vector representations to capture the semantic information of documents. BERT (Bidirectional Encoder Representations from Transformers) is a pre-trained deep bidirectional representation model that generates vector representations of words by considering contextual information.

[0098] In the embodiment of the present application, considering that the purpose of training is to enable the trained nuclear power experience feedback model to learn the basic structure and semantic rules of event elements, the content of event elements, experience feedback reports and work instructions, so as to predict which experience feedbacks a certain standard work instruction is adaptable to, therefore, selecting a vectorization method that can reflect the semantic order is beneficial to improving the accuracy of the obtained vectorized data.

[0099] Optionally, determining the content of the event elements corresponding to each preset event element is equivalent to performing a multi-dimensional logical decomposition of the experience feedback report to obtain data of the experience feedback report in multiple dimensions. After obtaining the data in multiple dimensions, the obtained data in multiple dimensions can be structured. Among them, structuring the data refers to the process of converting raw data into structured data. The structured data obtained by structuring the data can be stored in the form of a table. The structured data can improve the accessibility, queryability and analyzability of the data, thereby improving the training efficiency of the nuclear power experience feedback model to be trained.

[0100] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0101] After the trained nuclear power experience feedback model is obtained, corresponding experience feedback can be output for the work instructions in the current work according to the trained nuclear power experience feedback model.

[0102] Figure 3 A schematic diagram of a process for obtaining experience feedback provided in an embodiment of the present application is shown, and is described in detail as follows:

[0103] S31, obtaining the work instruction to be evaluated.

[0104] The work order to be evaluated is usually a work order to be executed, or a work order currently being executed. The work order can be expressed in the form of a character string, or in the form of work content or process, which is not limited here.

[0105] S32, using the work instruction to be evaluated as the input of the trained nuclear power experience feedback model as described above, to obtain the experience feedback output by the trained nuclear power experience feedback model.

[0106] Specifically, after determining the input work instruction, the trained nuclear power experience feedback model searches for the content of the event element that matches the work instruction in the event element. If found, the experience feedback including the content of the event element is output. If not found, the model searches for the matching content in the nuclear power experience report according to the work instruction. If found, the output experience feedback is determined according to the found nuclear power experience report. Optionally, if no content matching the work instruction is found, a message indicating that there is no experience feedback can be output.

[0107] In an embodiment of the present application, the trained nuclear power experience feedback model can perform a more comprehensive experience feedback search based on the experience feedback report, or perform a faster experience feedback search based on the content of the event elements, thereby improving both the comprehensiveness of the output experience feedback and the efficiency of the output experience feedback, thereby improving the utilization rate of the experience feedback report.

[0108] In some embodiments, the above S31 includes:

[0109] A work instruction package is obtained, wherein the work instruction package includes at least one of the work instructions to be evaluated. For each of the work instructions to be evaluated, the work instruction to be evaluated is used as an input of the trained nuclear power experience feedback model to obtain experience feedback output by the trained nuclear power experience feedback model.

[0110] Specifically, considering that the working processes of nuclear power plants are standardized, for example, a nuclear power plant presents multiple work instructions in the form of a task list (i.e., a standard work instruction package). Therefore, when obtaining work instructions, a standard work instruction package may be obtained, that is, one acquisition action will obtain multiple work instructions. At this time, for each work instruction, these work instructions can be used as inputs of the trained nuclear power experience feedback model respectively. In this way, the experience feedback corresponding to these work instructions output by the above-mentioned trained nuclear power experience feedback model will be obtained.

[0111] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0112] Corresponding to the training method of the nuclear power experience feedback model described in the above embodiment, Figure 4 A structural block diagram of a training device for a nuclear power experience feedback model provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown.

[0113] Reference Figure 4 The nuclear power experience feedback model training device 4 includes: an experience feedback report acquisition module 41, an event element extraction module 42, a work instruction acquisition module 43, and a nuclear power experience feedback model training module 44. Among them:

[0114] The experience feedback report acquisition module 41 is used to acquire the experience feedback report of the nuclear power plant.

[0115] The event element extraction module 42 is used to extract corresponding content from the above-mentioned experience feedback report according to preset event elements to obtain the content of the above-mentioned event elements.

[0116] The work instruction acquisition module 43 is used to acquire the work instruction corresponding to the above experience feedback report.

[0117] The nuclear power experience feedback model training module 44 is used to train the nuclear power experience feedback model to be trained according to the above-mentioned event elements, the content of the above-mentioned event elements, the above-mentioned experience feedback report and the work instructions corresponding to the above-mentioned experience feedback report to obtain the trained nuclear power experience feedback model, wherein the above-mentioned trained nuclear power experience feedback model is used to output corresponding experience feedback according to the received work instructions.

[0118] In an embodiment of the present application, the nuclear power experience feedback model to be trained is trained according to event elements, the content of the event elements, an experience feedback report, and a work instruction corresponding to the experience feedback report, to obtain a trained nuclear power experience feedback model. The trained nuclear power experience feedback model is used to output corresponding experience feedback according to the received work instruction. Since the nuclear power experience feedback model can output corresponding experience feedback according to the input work instruction, that is, there is no need for manual search, the efficiency of obtaining experience feedback is improved. In addition, since the experience feedback report, preset event elements, the content of the event elements, and the work instruction are used in the training process, the trained nuclear power experience feedback model can perform a more comprehensive search for experience feedback according to the experience feedback report, or perform a faster search for experience feedback according to the content of the event elements, thereby improving both the comprehensiveness of the output experience feedback and the efficiency of the output experience feedback, and further improving the utilization rate of the experience feedback report.

[0119] In some embodiments, the above-mentioned experience feedback report acquisition module 41 is specifically configured to:

[0120] Acquire a plurality of the above-mentioned experience feedback reports of the nuclear power plant, wherein, among the plurality of the above-mentioned experience feedback reports, at least two of the above-mentioned experience feedback reports have different contents of the above-mentioned event elements corresponding to the above-mentioned preset event elements.

[0121] In some embodiments, the training device 4 of the nuclear power experience feedback model provided by the embodiment of the present application further includes:

[0122] An experience feedback report segmentation module, configured to segment the above-mentioned experience feedback report before training the nuclear power experience feedback model to be trained according to the above-mentioned event elements, the content of the above-mentioned event elements, the above-mentioned experience feedback report, and the work instruction corresponding to the above-mentioned experience feedback report. The segmented above-mentioned experience feedback report includes at least two text units, and the data volume contained in each of the above-mentioned text units is less than the data volume contained in the above-mentioned experience feedback report.

[0123] Correspondingly, when the above-mentioned nuclear power experience feedback model training module 44 trains the nuclear power experience feedback model to be trained according to the above-mentioned event elements, the content of the above-mentioned event elements, the above-mentioned experience feedback report, and the work instruction corresponding to the above-mentioned experience feedback report, it is specifically configured to:

[0124] Train the nuclear power experience feedback model to be trained according to the above-mentioned event elements, the content of the above-mentioned event elements, the segmented above-mentioned experience feedback report, and the work instruction corresponding to the above-mentioned experience feedback report.

[0125] In some embodiments, the training device 4 of the nuclear power experience feedback model provided by the embodiment of the present application further includes:

[0126] The text vectorization module is used to vectorize the segmented experience feedback report and the work instructions corresponding to the experience feedback report before training the nuclear power experience feedback model to be trained based on the event elements, the content of the event elements, the segmented experience feedback report and the work instructions corresponding to the experience feedback report to obtain vectorized data.

[0127] Correspondingly, when the nuclear power experience feedback model training module 44 trains the nuclear power experience feedback model to be trained according to the event elements, the content of the event elements, the segmented experience feedback reports, and the work instructions corresponding to the experience feedback reports, it is specifically used to:

[0128] The nuclear power experience feedback model to be trained is trained according to the above event elements, the content of the above event elements and the above vectorized data.

[0129] In some embodiments, the number of the event elements is greater than 1, and the segmented experience feedback report and the work instructions corresponding to the experience feedback report are vectorized to obtain vectorized data including:

[0130] A vectorization method that can reflect the semantic order is selected to vectorize the segmented experience feedback report and the work instructions corresponding to the experience feedback report to obtain vectorized data.

[0131] In some embodiments, the above-mentioned event elements include: equipment, work type, occurrence process, consequences, cause of the event, and corrective actions.

[0132] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0133] Corresponding to the method for obtaining experience feedback described in the above embodiment, Figure 5 A structural block diagram of an experience feedback acquisition device provided in an embodiment of the present application is shown. For ease of explanation, only the parts related to the embodiment of the present application are shown.

[0134] Reference Figure 5 The experience feedback acquisition device 5 comprises: a work instruction acquisition module 51 and an experience feedback output module 52. Wherein:

[0135] The work instruction acquisition module 51 is used to acquire the work instruction to be evaluated.

[0136] The experience feedback output module 52 is used to use the above work instructions as the input of the trained nuclear power experience feedback model as described in the training method of the above nuclear power experience feedback model, and obtain the experience feedback output by the above trained nuclear power experience feedback model.

[0137] In an embodiment of the present application, the trained nuclear power experience feedback model can perform a more comprehensive experience feedback search based on the experience feedback report, or perform a faster experience feedback search based on the content of the event elements, thereby improving both the comprehensiveness of the output experience feedback and the efficiency of the output experience feedback, thereby improving the utilization rate of the experience feedback report.

[0138] In some embodiments, the work instruction acquisition module 51 is specifically configured to include:

[0139] Obtain a work instruction package, the work instruction package including at least one of the work instructions to be evaluated. For each of the work instructions to be evaluated, use the work instruction as an input of the trained nuclear power experience feedback model according to any one of claims 1 to 6, and obtain experience feedback output by the trained nuclear power experience feedback model.

[0140] It should be noted that the information interaction, execution process, etc. between the above-mentioned devices / units are based on the same concept as the method embodiment of the present application. Their specific functions and technical effects can be found in the method embodiment part and will not be repeated here.

[0141] Figure 6 This is a schematic diagram of the structure of an electronic device provided by an embodiment of the present application. Figure 6 As shown, the electronic device 6 of this embodiment includes: at least one processor 60 ( Figure 6 Only one processor is shown in the figure), a memory 61, and a computer program 62 stored in the memory 61 and executable on the at least one processor 60, wherein the processor 60 implements the steps of any of the above-mentioned method embodiments when executing the computer program 62.

[0142] The electronic device 6 may be a computing device such as a desktop computer, a notebook, a PDA, or a cloud server. The electronic device may include, but is not limited to, a processor 60 and a memory 61. Those skilled in the art will appreciate that Figure 6 It is only an example of the electronic device 6 and does not constitute a limitation on the electronic device 6. It may include more or fewer components than shown in the figure, or a combination of certain components, or different components. For example, it may also include input and output devices, network access devices, etc.

[0143] The processor 60 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc.

[0144] In some embodiments, the memory 61 may be an internal storage unit of the electronic device 6, such as a hard disk or memory of the electronic device 6. In other embodiments, the memory 61 may also be an external storage device of the electronic device 6, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 6. Further, the memory 61 may also include both an internal storage unit of the electronic device 6 and an external storage device. The memory 61 is used to store an operating system, an application program, a boot loader (BootLoader), data, and other programs, such as the program code of the computer program, etc. The memory 61 may also be used to temporarily store data that has been output or is to be output.

[0145] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0146] An embodiment of the present application also provides a network device, which includes: at least one processor, a memory, and a computer program stored in the memory and executable on the at least one processor, wherein the processor implements the steps in any of the above-mentioned method embodiments when executing the computer program.

[0147] An embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps in the above-mentioned method embodiments can be implemented.

[0148] An embodiment of the present application provides a computer program product. When the computer program product runs on an electronic device, the electronic device can implement the steps in the above-mentioned method embodiments when executing the computer program product.

[0149] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present application implements all or part of the processes in the above-mentioned embodiment method, which can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may at least include: any entity or device that can carry the computer program code to the camera / electronic device, recording medium, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, RandomAccess Memory), electric carrier signal, telecommunication signal and software distribution medium. For example, a USB flash drive, a mobile hard disk, a magnetic disk or an optical disk. In some jurisdictions, according to legislation and patent practice, computer-readable media cannot be electric carrier signals and telecommunication signals.

[0150] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0151] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0152] In the embodiments provided in the present application, it should be understood that the disclosed devices / network equipment and methods can be implemented in other ways. For example, the device / network equipment embodiments described above are merely schematic. For example, the division of the modules or units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0153] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0154] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A training method for a nuclear power experience feedback model, characterized in that: include: Obtain experience feedback reports from nuclear power plants; Extracting corresponding content from the experience feedback report according to preset event elements to obtain the content of the event elements; Obtaining work instructions corresponding to the experience feedback report; The nuclear power experience feedback model to be trained is trained according to the event elements, the content of the event elements, the experience feedback report and the work instructions corresponding to the experience feedback report to obtain a trained nuclear power experience feedback model, wherein the trained nuclear power experience feedback model is used to output corresponding experience feedback according to the received work instructions.

2. The training method of the nuclear power experience feedback model according to claim 1, characterized in that: The experience feedback report obtained from nuclear power plants includes: A plurality of the experience feedback reports of the nuclear power plant are obtained, wherein among the plurality of the experience feedback reports, at least two of the experience feedback reports have different contents of the event elements corresponding to the preset event elements.

3. The training method of the nuclear power experience feedback model according to claim 1 or 2, characterized in that: Before training the nuclear power experience feedback model to be trained according to the event elements, the content of the event elements, the experience feedback report, and the work instructions corresponding to the experience feedback report, the method further includes: Segmenting the experience feedback report, wherein the segmented experience feedback report includes at least two text units, and the amount of data contained in each of the text units is smaller than the amount of data contained in the experience feedback report; The training of the nuclear power experience feedback model to be trained according to the event element, the content of the event element, the experience feedback report, and the work instruction corresponding to the experience feedback report includes: The nuclear power experience feedback model to be trained is trained according to the event elements, the content of the event elements, the segmented experience feedback reports, and the work instructions corresponding to the experience feedback reports.

4. The training method of the nuclear power experience feedback model according to claim 3, characterized in that: Before training the nuclear power experience feedback model to be trained according to the event elements, the content of the event elements, the segmented experience feedback reports, and the work instructions corresponding to the experience feedback reports, the method further includes: Vectorizing the segmented experience feedback report and the work instructions corresponding to the experience feedback report to obtain vectorized data; The training of the nuclear power experience feedback model to be trained according to the event elements, the content of the event elements, the segmented experience feedback reports, and the work instructions corresponding to the experience feedback reports includes: The nuclear power experience feedback model to be trained is trained according to the event elements, the content of the event elements and the vectorized data.

5. The training method of the nuclear power experience feedback model according to claim 4, characterized in that: The number of event elements is greater than 1, and the segmented experience feedback report and the work instructions corresponding to the experience feedback report are vectorized to obtain vectorized data including: A vectorization method that can reflect the semantic order is selected to vectorize the segmented experience feedback report and the work instructions corresponding to the experience feedback report to obtain vectorized data.

6. The training method of the nuclear power experience feedback model according to claim 5, characterized in that: The event elements include: equipment, work type, occurrence process, consequences, cause of the event, and corrective actions.

7. A method for obtaining experience feedback, characterized in that: include: Obtain work orders to be evaluated; The work instruction to be evaluated is used as an input of the trained nuclear power experience feedback model as described in any one of claims 1 to 6 to obtain the experience feedback output by the trained nuclear power experience feedback model.

8. The method for obtaining experience feedback according to claim 7, characterized in that: The obtaining of the work instruction to be evaluated comprises: Obtaining a work instruction package, wherein the work instruction package includes at least one work instruction to be evaluated; For each of the work instructions to be evaluated, the work instruction to be evaluated is used as the input of the trained nuclear power experience feedback model as described in any one of claims 1 to 6 to obtain the experience feedback output by the trained nuclear power experience feedback model.

9. A training device for a nuclear power experience feedback model, characterized in that: include: An experience feedback report acquisition module is used to obtain experience feedback reports from nuclear power plants; An event element extraction module, used to extract corresponding content from the experience feedback report according to preset event elements to obtain the content of the event elements; A work instruction acquisition module, used to acquire a work instruction corresponding to the experience feedback report; The nuclear power experience feedback model training module is used to train the nuclear power experience feedback model to be trained according to the event elements, the content of the event elements, the experience feedback report and the work instructions corresponding to the experience feedback report to obtain the trained nuclear power experience feedback model, wherein the trained nuclear power experience feedback model is used to output corresponding experience feedback according to the received work instructions.

10. An experience feedback acquisition device, characterized in that: include: A work instruction acquisition module, used to acquire the work instructions to be evaluated; An experience feedback output module is used to use the work instruction to be evaluated as an input of the trained nuclear power experience feedback model as described in any one of claims 1 to 6 to obtain the experience feedback output by the trained nuclear power experience feedback model.

11. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented, or the method according to claim 7 or 8 is performed.

12. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented, or the method according to claim 7 or 8 is performed.

13. A computer program product, characterized in that The method comprises a computer program, which, when being executed, executes the method according to any one of claims 1 to 6, or executes the method according to claim 7 or 8.

Images