Virtual-real fusion intelligent electric power practical training method and system for power distribution system, and medium

Abstract

The invention discloses a virtual-real fusion intelligent electric power practical training method and system for a power distribution system and a medium, and the method comprises the steps: evaluating the content matching degree of teaching practical training content and an electric power maintenance operation practical training target through historical data and a matching degree algorithm, and carrying out the multi-dimensional classification of the teaching content based on the student feedback and the content matching degree; the pushing frequency is predicted in combination with keyword association and feedback data, the practical training scheduling level is determined by comprehensively considering the pushing frequency and priority, personalized pushing and scientific scheduling of teaching content are achieved, the learning rhythms and requirements of different students are met, and the knowledge absorption effect and learning efficiency are improved; teaching content is enabled to accurately meet operation and maintenance requirements of the novel power distribution system, students can learn practical knowledge, training efficiency and quality are improved, and resource waste is avoided.

Description

Technical Field

[0001] The present invention relates to electric power training technology, and in particular to a virtual-reality integrated smart electric power training method, system and medium for a power distribution system. Background Art

[0002] The determination of teaching content in traditional power practical training often relies on experience, and does not fully incorporate new features such as intelligence, digitization, and distributed energy access in new distribution systems, resulting in the knowledge learned by trainees not meeting actual work needs. Currently, practical training adopts a unified teaching schedule and content arrangement, ignoring individual differences among trainees. Different trainees have different speeds and degrees of mastery of knowledge, and some trainees feel that the learning content lacks challenge, which reduces the trainees' enthusiasm for learning and learning outcomes. Power practical training mostly relies on physical equipment or simple virtual simulations, which cannot truly restore the complex operating scenarios and fault conditions of new distribution systems. Trainees find it difficult to come into contact with the various complex problems that may be encountered in actual work during practical training, resulting in insufficient practical ability and an inability to quickly adapt to their jobs. Summary of the Invention

[0003] Purpose of the invention: The present invention aims to provide a virtual-reality integrated smart power training method for distribution systems that takes into account individual differences of trainees and complex scenarios and faults of power systems; another purpose of the present invention is to provide a virtual-reality integrated smart power training system and medium for distribution systems.

[0004] Technical solution: The virtual-reality integrated smart power training method for distribution systems described in the present invention includes the following steps:

[0005] Based on the power maintenance operation labels, a matching algorithm is used to calculate the content matching degree between the teaching training content set and the power maintenance operation training objectives;

[0006] Generate feedback trend analysis results for the training feedback dataset of students on the teaching training content set. Based on the content matching degree and feedback trend analysis results, perform multi-dimensional classification of the teaching training content set according to the preset classification rules to generate an index category set containing priority identifiers.

[0007] Extract keyword sequences from the teaching and training content set, input the correlation distance values ​​between keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process;

[0008] According to the push frequency and the priority identifier in the index category set, the training scheduling level of the teaching and training content set is determined through a preset level mapping table. Based on the training scheduling level, the teaching and training content set is dynamically weighted and evaluated to obtain the content level label of the teaching and training content during the scheduling process.

[0009] The teaching and training content collection is optimized and sorted according to the content level labels, a training scheduling plan for the virtual-reality fusion scenario is generated, and the training scheduling plan is synchronized with the operation instruction flow of the digital training platform in real time.

[0010] Furthermore, based on the power maintenance operation label, a matching algorithm is used to calculate the content matching degree between the teaching training content set and the power maintenance operation training objectives, as follows:

[0011] According to a preset similarity threshold, a set of similar training content items having the same power maintenance operation label as the current teaching training content item is screened out from the historical training content items of the historical teaching training data;

[0012] For each similar training content item in the set of similar training content items, the same vocabulary matching algorithm is used to extract a common keyword sequence between the similar training content item and the current teaching training content item, and the single-category confidence between the two is calculated based on the length and distribution density of the common keyword sequence;

[0013] Traversing all similar training content items in the similar training content item set to generate a single-category confidence set corresponding to the current teaching training content item;

[0014] Inputting the single-category confidence set into a preset confidence aggregation model, and generating a content matching quantification value between the current teaching training content item and the power maintenance operation training objective through weighted normalization processing, wherein the weight coefficient is dynamically adjusted according to the operation complexity and execution success rate of the historical training content item;

[0015] Each similar training item in the set of similar training content items includes the number of historical trainees, operation type labels and confidence parameters;

[0016] The preset entropy value algorithm is used to calculate the student quantity entropy based on the historical distribution characteristics of the number of students. The student quantity entropy is used to characterize the dispersion of student participation in similar training projects.

[0017] According to the confidence parameters corresponding to the set of similar training content items, a content adjustment degree of the current teaching training content item is generated through a preset adjustment degree calculation model. The content adjustment degree is used to reflect the influence coefficient of the fluctuation of the confidence degree of a single category on the adaptability of the training target;

[0018] Input the student quantity entropy and the content adjustment degree into the dynamic weight allocation strategy, and dynamically adjust the weight ratio of the two in the matching degree calculation according to the discrete degree of the student quantity entropy;

[0019] Based on the weighted student population entropy and content adjustment degree, a content matching quantization value between the current teaching training content item and the power maintenance operation training objective is generated through linear normalization processing, wherein the matching quantization value is negatively correlated with the student population entropy and positively correlated with the content adjustment degree.

[0020] Furthermore, for the training feedback dataset of the trainees on the teaching training content set, feedback trend analysis results are generated. Based on the content matching degree and feedback trend analysis results, the teaching training content set is classified into multiple dimensions according to the preset classification rules, and an index category set containing priority identifiers is generated, as follows:

[0021] The training feedback data set of trainees on the teaching training content set is divided into a positive feedback data subset and a negative feedback data subset;

[0022] Using a preset entropy calculation model, a dispersion analysis is performed on the score distribution characteristics in the positive feedback data subset to generate a positive feedback entropy value representing the concentration of positive opinions; a volatility calculation is performed on the frequency of occurrence of question types in the negative feedback data subset to generate a negative feedback entropy value representing the diffusion of negative opinions;

[0023] According to a preset weight allocation strategy, combined with the index category characteristics of the teaching and training content, the weight coefficients of the positive feedback entropy value and the negative feedback entropy value are dynamically allocated;

[0024] The weighted positive feedback entropy value and the negative feedback entropy value are input into the trend quantification model, and a feedback trend quantification index is generated through a preset trend mapping function. The feedback trend quantification index includes: when the proportion of positive entropy values ​​exceeds a first threshold, it is determined to be a positive trend; when the proportion of negative entropy values ​​exceeds a second threshold, a content optimization alarm is triggered;

[0025] Based on the feedback trend quantitative indicators and the preset trend level comparison table, a feedback trend analysis result including a trend strength indicator and a confidence interval is generated. The confidence interval is dynamically calculated according to the amount of historical feedback data and the classification accuracy;

[0026] Obtaining the teaching time parameters of each content item in the teaching and training content set, and performing timeliness correction on the content matching degree based on a preset time decay factor to generate a dynamic content matching degree weighted value;

[0027] Inputting the dynamic content matching degree weighted value and the trend strength indicator in the feedback trend analysis result into a preset effectiveness calculation model, and generating a quantitative value of the practical training effectiveness of the teaching and training content through a nonlinear fitting algorithm, wherein the effectiveness calculation model integrates a compensation function for the attenuation of the operation effect of the teaching time parameter;

[0028] According to the numerical interval of the quantified value of the training effectiveness and the confidence interval range in the feedback trend analysis result, a dynamic classification matrix is ​​constructed, and based on the dynamic classification matrix, an index category set containing priority identifiers is generated.

[0029] Furthermore, we extract keyword sequences from the teaching and training content set, input the correlation distance values ​​between keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process, as follows:

[0030] Extracting keyword sequences from the teaching and training content set, and calculating a content gap index of the teaching and training content using a preset gap model based on association distance values ​​between the keywords, wherein the content gap index is negatively correlated with the semantic association strength between the keywords;

[0031] Based on the distribution of student ratings and operational error types in the training feedback dataset, a preset discrimination algorithm is used to generate a quantitative value for content discrimination.

[0032] According to the numerical range of the content gap indicator and the content resolution quantification value, the influence weight of the content gap indicator and the content resolution quantification value on the push frequency is adjusted through a dynamic weight allocation strategy. When the content gap indicator exceeds a preset threshold, the weight allocation is tilted towards the content resolution quantification value.

[0033] The weighted content gap index and content discrimination quantization value and the initial push frequency reference value are input into the frequency correction function, and a corrected push frequency intermediate value is generated through a nonlinear mapping relationship. The corrected push frequency intermediate value is negatively correlated with the content gap index and positively correlated with the content discrimination quantization value;

[0034] The intermediate value of the corrected push frequency is dynamically smoothed according to the recent student activity parameters in the training feedback data set to generate a final push frequency. The window size of the dynamic smoothing process is adaptively matched to the fluctuation amplitude of the student activity.

[0035] The virtual-reality integrated smart power training system for power distribution systems of the present invention includes:

[0036] A content matching calculation module is used to calculate the content matching degree between the teaching and training content set and the power maintenance operation training objectives based on the power maintenance operation label and using a matching algorithm;

[0037] An index category set generation module is used to generate feedback trend analysis results based on the training feedback data set of trainees on the teaching and training content set. Based on the content matching degree and feedback trend analysis results, the teaching and training content set is classified into multiple dimensions according to preset classification rules to generate an index category set containing priority identifiers.

[0038] The push frequency calculation module is used to extract the keyword sequence in the teaching and training content set, input the correlation distance value between the keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process;

[0039] A content level label generation module is used to determine the training scheduling level of the teaching and training content set based on the push frequency and the priority identifier in the index category set through a preset level mapping table, and dynamically evaluate the teaching and training content set based on the training scheduling level to obtain the content level label of the teaching and training content during the scheduling process;

[0040] The training scheduling plan generation and synchronization module is used to optimize the sorting of teaching and training content collections according to content level labels, generate training scheduling plans for virtual and real fusion scenarios, and synchronize the training scheduling plans with the operation instruction stream of the digital training platform in real time.

[0041] Furthermore, in the content matching calculation module, based on the power maintenance operation label, a matching algorithm is used to calculate the content matching between the teaching training content set and the power maintenance operation training objective, as follows:

[0042] According to a preset similarity threshold, a set of similar training content items having the same power maintenance operation label as the current teaching training content item is screened out from the historical training content items of the historical teaching training data;

[0043] For each similar training content item in the set of similar training content items, the same vocabulary matching algorithm is used to extract a common keyword sequence between the similar training content item and the current teaching training content item, and the single-category confidence between the two is calculated based on the length and distribution density of the common keyword sequence;

[0044] Traversing all similar training content items in the similar training content item set to generate a single-category confidence set corresponding to the current teaching training content item;

[0045] Inputting the single-category confidence set into a preset confidence aggregation model, and generating a content matching quantification value between the current teaching training content item and the power maintenance operation training objective through weighted normalization processing, wherein the weight coefficient is dynamically adjusted according to the operation complexity and execution success rate of the historical training content item;

[0046] Each similar training item in the set of similar training content items includes the number of historical trainees, operation type labels and confidence parameters;

[0047] The preset entropy value algorithm is used to calculate the student quantity entropy based on the historical distribution characteristics of the number of students. The student quantity entropy is used to characterize the dispersion of student participation in similar training projects.

[0048] According to the confidence parameters corresponding to the set of similar training content items, a content adjustment degree of the current teaching training content item is generated through a preset adjustment degree calculation model. The content adjustment degree is used to reflect the influence coefficient of the fluctuation of the confidence degree of a single category on the adaptability of the training target;

[0049] Input the student quantity entropy and the content adjustment degree into the dynamic weight allocation strategy, and dynamically adjust the weight ratio of the two in the matching degree calculation according to the discrete degree of the student quantity entropy;

[0050] Based on the weighted student population entropy and content adjustment degree, a content matching quantization value between the current teaching training content item and the power maintenance operation training objective is generated through linear normalization processing, wherein the matching quantization value is negatively correlated with the student population entropy and positively correlated with the content adjustment degree.

[0051] Furthermore, in the index category set generation module, feedback trend analysis results are generated for the training feedback data set of the trainees on the teaching training content set. Based on the content matching degree and the feedback trend analysis results, the teaching training content set is multi-dimensionally classified according to the preset classification rules to generate an index category set containing priority identifiers, as follows:

[0052] The training feedback data set of trainees on the teaching training content set is divided into a positive feedback data subset and a negative feedback data subset;

[0053] Using a preset entropy calculation model, a dispersion analysis is performed on the score distribution characteristics in the positive feedback data subset to generate a positive feedback entropy value representing the concentration of positive opinions; a volatility calculation is performed on the frequency of occurrence of question types in the negative feedback data subset to generate a negative feedback entropy value representing the diffusion of negative opinions;

[0054] According to a preset weight allocation strategy, combined with the index category characteristics of the teaching and training content, the weight coefficients of the positive feedback entropy value and the negative feedback entropy value are dynamically allocated;

[0055] The weighted positive feedback entropy value and the negative feedback entropy value are input into the trend quantification model, and a feedback trend quantification index is generated through a preset trend mapping function. The feedback trend quantification index includes: when the proportion of positive entropy values ​​exceeds a first threshold, it is determined to be a positive trend; when the proportion of negative entropy values ​​exceeds a second threshold, a content optimization alarm is triggered;

[0056] Based on the feedback trend quantitative indicators and the preset trend level comparison table, a feedback trend analysis result including a trend strength indicator and a confidence interval is generated. The confidence interval is dynamically calculated according to the amount of historical feedback data and the classification accuracy;

[0057] Obtaining the teaching time parameters of each content item in the teaching and training content set, and performing timeliness correction on the content matching degree based on a preset time decay factor to generate a dynamic content matching degree weighted value;

[0058] Inputting the dynamic content matching degree weighted value and the trend strength indicator in the feedback trend analysis result into a preset effectiveness calculation model, and generating a quantitative value of the practical training effectiveness of the teaching and training content through a nonlinear fitting algorithm, wherein the effectiveness calculation model integrates a compensation function for the attenuation of the operation effect of the teaching time parameter;

[0059] According to the numerical interval of the quantified value of the training effectiveness and the confidence interval range in the feedback trend analysis result, a dynamic classification matrix is ​​constructed, and based on the dynamic classification matrix, an index category set containing priority identifiers is generated.

[0060] Furthermore, in the push frequency calculation module, the keyword sequence in the teaching and training content set is extracted, and the correlation distance value between the keywords and the training feedback data set are input into the preset frequency prediction model to obtain the push frequency of the teaching and training content in the power training process, as follows:

[0061] Extracting keyword sequences from the teaching and training content set, and calculating a content gap index of the teaching and training content using a preset gap model based on association distance values ​​between the keywords, wherein the content gap index is negatively correlated with the semantic association strength between the keywords;

[0062] Based on the distribution of student ratings and operational error types in the training feedback dataset, a preset discrimination algorithm is used to generate a quantitative value for content discrimination.

[0063] According to the numerical range of the content gap indicator and the content resolution quantification value, the influence weight of the content gap indicator and the content resolution quantification value on the push frequency is adjusted through a dynamic weight allocation strategy. When the content gap indicator exceeds a preset threshold, the weight allocation is tilted towards the content resolution quantification value.

[0064] The weighted content gap index and content discrimination quantization value and the initial push frequency reference value are input into the frequency correction function, and a corrected push frequency intermediate value is generated through a nonlinear mapping relationship. The corrected push frequency intermediate value is negatively correlated with the content gap index and positively correlated with the content discrimination quantization value;

[0065] The intermediate value of the corrected push frequency is dynamically smoothed according to the recent student activity parameters in the training feedback data set to generate a final push frequency. The window size of the dynamic smoothing process is adaptively matched to the fluctuation amplitude of the student activity.

[0066] The computer device of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0067] The computer-readable storage medium of the present invention stores a computer program thereon, and the computer program implements the steps of the above method when executed by a processor.

[0068] Beneficial effects: Compared with the existing technology, the present invention has the following significant advantages: 1. The present invention uses historical data and matching algorithms to evaluate the content matching between teaching training content and power maintenance operation training objectives, and conducts multi-dimensional classification of teaching content based on student feedback and content matching. It combines keyword association and feedback data to predict push frequency, and comprehensively considers push frequency and priority to determine training scheduling level, so as to realize personalized push and scientific scheduling of teaching content, meet the learning rhythm and needs of different students, and improve knowledge absorption effect and learning efficiency; 2. The present invention enables teaching content to be accurately connected with the operation and maintenance needs of the new distribution system, so that students can learn practical knowledge, improve training efficiency and quality, and avoid waste of resources; 3. The present invention matches teaching content with the virtual-real fusion scenario of the new distribution system, and synchronizes with the digital training platform in real time, providing students with a training experience that is highly close to the actual work scenario, enhancing students' ability to solve practical problems, and enabling students to adapt to their job. DETAILED DESCRIPTION

[0069] The virtual-real integration smart power training method for power distribution systems of the present invention comprises the following steps:

[0070] (1) Based on the power maintenance operation labels, a matching algorithm is used to calculate the content matching degree between the teaching training content set and the power maintenance operation training objectives.

[0071] Determine a set of teaching and training content corresponding to power maintenance operations, which includes power equipment maintenance procedures, fault handling procedures and safe operation standards.

[0072] In this embodiment, the process of determining the set of teaching and training content corresponding to power maintenance operations is summarized, specifically including: establishing a key technology analysis model for power maintenance operations to identify the core elements in power equipment maintenance procedures, fault handling processes and safety operation standards; building a power maintenance operation case library to collect and organize historical power maintenance cases, and screen out typical scenarios that meet the needs of the new distribution system; and determining the teaching and training content set by cross-comparing the key technology analysis model with the case library.

[0073] It should be noted that, based on actual application scenarios and industry standards, the teaching content is reversely deduced to construct a knowledge system that fits the operation and maintenance needs of the new distribution system. The benefit of this step is to provide a clear and professional content framework for subsequent training, avoid the disconnection between teaching content and actual work, enable trainees to systematically master the core knowledge and skills of power maintenance operations, significantly improve the pertinence and practicality of training, and ensure that the knowledge learned by trainees can be directly applied to actual work scenarios. This embodiment can combine the key technology analysis model with the case library, and through cross-comparison, it can accurately identify the teaching and training content set. This cross-comparison is not a simple comparison, but uses advanced semantic analysis and pattern recognition technology to dig out the technical points and operating specifications hidden in the case, thereby ensuring that the determined teaching and training content set not only covers the necessary content of traditional power maintenance operations, but also keeps up with the development trend of the new distribution system.

[0074] Virtual reality scenarios can be configured as needed. The VR scenario adaptability assessment mechanism simulates the actual operating environment of new power distribution systems to preview teaching and training content, ensuring its effective implementation within VR scenarios and providing a solid foundation for a virtual-reality integrated training model. The industry dynamics tracking system monitors the latest technological developments and standard updates in the power industry in real time. When new power equipment or maintenance technologies are detected, it automatically triggers content updates and incorporates relevant information into the teaching and training content, ensuring it remains current and cutting-edge.

[0075] A key technology analysis module for power maintenance operations can also be developed. This module, jointly developed by domain experts and data scientists, utilizes natural language processing technology to conduct in-depth analysis of technical documents and industry standards related to power maintenance. Key elements, such as key technical terms, operational process nodes, and safety regulations, are extracted and stored in a structured database to form a comprehensive knowledge graph. A virtual reality scenario adaptability assessment module can also be introduced. Using T3D or Laya, a virtual model of a new power distribution system, including substations, distribution lines, and smart terminal devices, can be constructed. The operational processes within the preliminarily identified set of teaching and training content are simulated in a virtual scenario. Interaction data, operation duration, error rates, and other metrics are collected during the simulation. An evaluation algorithm is then used to assess the content's operability and teaching effectiveness. In addition to precisely identifying the set of teaching and training content relevant to power maintenance operations, innovative elements such as virtual reality scenario adaptability assessment and industry trend tracking are introduced to effectively address the issues inherent in existing technologies, such as content lag and disconnection from practical applications.

[0076] Obtain historical teaching and training data, and calculate the content matching degree between each content item in the teaching and training content set and the power maintenance operation training objective according to a preset matching degree algorithm. Calculate the content matching degree between each content item in the teaching and training content set and the power maintenance operation training objective according to a preset matching degree algorithm, specifically as follows:

[0077] Acquire multiple historical training content items and their associated power maintenance operation labels in the historical teaching training data; screen out a set of similar training content items with the same operation label as the current teaching training content item from the historical training content items according to a preset similarity threshold; for each similar training content item in the set of similar training content items, use the same vocabulary matching algorithm to extract the common keyword sequence between it and the current teaching training content item, and calculate the single-category confidence between the two based on the length and distribution density of the common keyword sequence; traverse all similar training content items in the set of similar training content items to generate a single-category confidence set corresponding to the current teaching training content item; input the single-category confidence set into a preset confidence aggregation model, and generate a content matching quantization value between the current teaching training content item and the power maintenance operation training target through weighted normalization processing, wherein the weight coefficient is dynamically adjusted according to the operation complexity and execution success rate of the historical training content item.

[0078] The content matching degree between each content item in the teaching and training content set and the power maintenance operation training objective is calculated according to a preset matching algorithm, further comprising: extracting all similar training project data sets associated with the current teaching and training content item based on the historical teaching and training data, wherein each similar training project includes the number of historical trainees, an operation type label, and a confidence parameter; using a preset entropy value algorithm to calculate the student number entropy according to the distribution characteristics of the number of historical trainees in the similar training project data set, the student number entropy is used to characterize the student participation dispersion of similar training projects; and calculating the student number entropy according to the confidence parameter of each project in the similar training project data set. The content adjustment degree of the current teaching and training content item is generated by a preset adjustment degree calculation model, and the content adjustment degree reflects the influence coefficient of the fluctuation of the same type of confidence on the adaptability of the training target; the student quantity entropy and the content adjustment degree are input into the dynamic weight allocation strategy, and the weight ratio of the two in the matching degree calculation is dynamically adjusted according to the discrete degree of the student quantity entropy; based on the weighted student quantity entropy and content adjustment degree, a content matching quantization value between the current teaching and training content item and the power maintenance operation training target is generated through linear normalization processing, wherein the matching quantization value is negatively correlated with the student quantity entropy and positively correlated with the content adjustment degree.

[0079] In this embodiment, historical teaching and training data is obtained, and the content matching degree between each content item in the teaching and training content set and the power maintenance operation training objectives is calculated based on a preset matching degree algorithm. This may include past student learning records, practical assessment data, and student learning feedback.

[0080] The pre-set matching algorithm uses the Analytic Hierarchy Process (AHP) and incorporates expert experience to assign different weights to different types of teaching content, such as power equipment maintenance procedures, troubleshooting processes, and safe operating standards. Furthermore, considering the training objectives' requirements for depth of knowledge, skill proficiency, and safety awareness, an evaluation index system is constructed to quantitatively assess each teaching content item from multiple dimensions, ultimately calculating its matching degree with the training objectives.

[0081] As can be seen, by leveraging empirical information from historical data and using scientific algorithmic models, we can objectively assess the degree of alignment between teaching content and training objectives. The benefit of this step is that it can screen out teaching content that is highly relevant to the training objectives, remove redundant or irrelevant content, optimize the allocation of teaching resources, prevent students from wasting time and energy on unnecessary knowledge, improve training efficiency, and enable teaching resources to be precisely invested in the development of key knowledge and skills.

[0082] (2) Generate feedback trend analysis results for the students' training feedback data set on the teaching training content set. Based on the content matching degree and feedback trend analysis results, perform multi-dimensional classification of the teaching training content set according to the preset classification rules, and generate an index category set containing priority identification.

[0083] Obtain a training feedback data set of trainees on the teaching training content set, generate feedback trend analysis results based on the training feedback data set, combine the content matching degree and feedback trend analysis results, perform multi-dimensional classification on the teaching training content set according to preset classification rules, and generate an index category set containing priority identifiers.

[0084] In this embodiment, a training feedback data set of students on the teaching training content set is obtained, and a feedback trend analysis result is generated based on the training feedback data set. In combination with the content matching degree and the feedback trend analysis result, the teaching training content set is multi-dimensionally classified according to the preset classification rules to generate an index category set containing a priority identifier. The specific details are that the training feedback data set is obtained through multiple channels, and students can evaluate the difficulty, practicality, and learning experience of the teaching content they are learning at any time; after each training course, an electronic questionnaire is distributed to collect students' systematic feedback on the overall teaching content; at the same time, the teacher's observation and evaluation during the training process is recorded, and these feedback data are cleaned and analyzed using data mining technology. Feedback trend analysis results are generated through methods such as time series analysis and sentiment analysis. In combination with the content matching degree obtained in step S102, according to the preset classification rules, the teaching content is divided into different categories such as high-priority key learning category, medium-priority regular learning category, and low-priority extended learning category from multiple dimensions such as knowledge difficulty, importance, application frequency, and student acceptance, and each category is assigned a corresponding priority identifier to form an index category set.

[0085] This example objectively categorizes teaching content from multiple perspectives based on students' real-world learning feedback and data analysis to reflect its importance and applicability in actual teaching. This step offers the advantage of rationally arranging the teaching sequence and resource allocation of teaching content based on students' actual needs and learning situations, prioritizing the reinforcement of important content that students find difficult to learn, increasing student motivation and effectiveness, achieving precise instruction, and improving overall teaching quality.

[0086] Extract the keyword sequence from the teaching and training content set, use the co-occurrence distance algorithm to calculate the correlation distance value between each keyword, input the correlation distance value and the training feedback data set into the preset frequency prediction model, and generate the push frequency parameters of each teaching and training content in the power maintenance operation process. In the specific operation, natural language processing (NLP) technology, such as lexical analysis, named entity recognition, etc., is used to extract key knowledge points and concepts from the text materials of the teaching and training content to form a keyword sequence. For example, keywords such as "transformer oil chromatography analysis" and "circuit breaker contact inspection" are extracted from the power equipment maintenance regulations. The co-occurrence distance algorithm calculates the degree of correlation between each keyword by analyzing the frequency of occurrence, adjacent distance, number of co-occurrences and other information of the keywords in the teaching content text. The smaller the correlation distance value, the closer the connection between the keywords. The association distance value and the training feedback dataset are input into a frequency prediction model based on deep learning, such as the long short-term memory network (LSTM) model. This model predicts the reasonable push frequency parameters of each teaching and training content during the power maintenance operation process by learning the relationship between keyword associations, student feedback and teaching content push frequency in historical data. For example, "transformer oil chromatography analysis" is pushed twice a week and "circuit breaker contact inspection" is pushed three times a week.

[0087] This embodiment can reasonably arrange the push rhythm of teaching content according to the importance of the teaching content and the actual needs of the students, avoid blind push or over-concentrated push of teaching content, enable students to learn appropriate content at the right time, improve knowledge absorption effect, reduce learning fatigue and knowledge forgetting, and improve learning efficiency.

[0088] Furthermore, a semantic analysis is performed on the training feedback data set through a preset feedback polarity classifier, and the trainees' opinions are divided into a positive feedback data subset and a negative feedback data subset, wherein the classification threshold is dynamically adjusted according to the frequency of historical feedback words; a preset entropy value calculation model is used to perform a discrete degree analysis on the score distribution characteristics in the positive feedback data subset, and generate a feedback positive entropy value representing the concentration of positive opinions. At the same time, a volatility calculation is performed on the frequency of occurrence of the problem types in the negative feedback data subset, and a feedback negative entropy value representing the diffusion of negative opinions is generated; according to the preset weight distribution strategy, combined with the index category characteristics of the teaching training content, the feedback is dynamically distributed. The weight coefficient of the positive entropy value and the negative entropy value of the feedback, among which the weight of the negative entropy value of the security operation content is increased by 30%-50%; the weighted positive entropy value of the feedback and the negative entropy value of the feedback are input into the trend quantification model, and the feedback trend quantification index is generated through the preset trend mapping function. The feedback trend quantification index includes: when the proportion of positive entropy value exceeds the first threshold, it is judged as a positive trend; when the proportion of negative entropy value exceeds the second threshold, a content optimization alarm is triggered; based on the feedback trend quantification index and the preset trend level comparison table, a feedback trend analysis result including a trend strength identifier and a confidence interval is generated, and the confidence interval is dynamically calculated according to the historical feedback data volume and classification accuracy.

[0089] Combined with the content matching degree and feedback trend analysis results, the teaching and training content set is multi-dimensionally classified according to preset classification rules, the teaching time parameter of each content item in the teaching and training content set is obtained, and the content matching degree is time-effectively corrected based on a preset time decay factor to generate a dynamic content matching degree weighted value; the dynamic content matching degree weighted value and the trend strength identifier in the feedback trend analysis result are input into a preset effectiveness calculation model, and a training effectiveness quantification value of the teaching and training content is generated through a nonlinear fitting algorithm, wherein the effectiveness calculation model integrates a compensation function for the attenuation of the teaching time parameter on the operation effect; according to the numerical range of the training effectiveness quantification value and the confidence interval range in the feedback trend analysis result, a dynamic classification matrix is ​​constructed, and the dynamic classification matrix includes:

[0090] When the quantified value of the training effectiveness is higher than the first effectiveness threshold and the lower limit of the confidence interval meets the preset credibility, it is marked as a high priority index category;

[0091] When the quantified value of the practical training effectiveness is lower than the second effectiveness threshold and the feedback trend analysis result triggers a content optimization alarm, it is marked as an index category to be optimized;

[0092] The remaining cases are assigned a standard priority index category based on the matching relationship between the teaching duration parameter and the trend strength indicator;

[0093] Based on the output result of the dynamic classification matrix, an index category set including a classification basis traceability code is generated, and the traceability code is associated with the original data chain of the content matching quantification value, feedback trend quantification indicator and teaching time parameter.

[0094] (3) Extract the keyword sequence from the teaching and training content set, input the correlation distance value between the keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process.

[0095] Extract keyword sequences from the teaching and training content set, use the co-occurrence distance algorithm to calculate the association distance value between each keyword, input the association distance value and the training feedback data set into the preset frequency prediction model, and generate the push frequency parameters of each teaching and training content during the power maintenance operation process.

[0096] This embodiment extracts keyword sequences from a set of teaching and training content, uses a co-occurrence distance algorithm to calculate the correlation distance between each keyword, and then inputs this correlation distance value and the training feedback dataset into a preset frequency prediction model to generate push frequency parameters for each teaching and training content during power maintenance operations. In specific operations, natural language processing (NLP) techniques, such as lexical analysis and named entity recognition, are used to extract key knowledge points and concepts from the text of the teaching and training content to form a keyword sequence. For example, keywords such as "transformer oil chromatographic analysis" and "circuit breaker contact inspection" are extracted from the power equipment maintenance procedures. The co-occurrence distance algorithm calculates the degree of correlation between each keyword by analyzing the frequency of occurrence, proximity, and number of co-occurrences of the keywords in the teaching content text. A smaller correlation distance value indicates a closer connection between the keywords. The correlation distance value and the training feedback dataset are then input into a deep learning-based frequency prediction model, such as a long short-term memory network (LSTM) model, to predict the appropriate push frequency parameters for each teaching and training content during power maintenance operations. For example, "transformer oil chromatographic analysis" should be pushed twice a week, and "circuit breaker contact inspection" should be pushed three times a week.

[0097] By exploring the inherent logical relationships within teaching content and the patterns of student feedback, machine learning models are used to predict the appropriate frequency of content delivery to suit students' learning pace and knowledge acquisition needs. This step benefits from arranging the delivery cadence of teaching content based on its importance and students' actual needs, avoiding blind or overly concentrated delivery of content. This allows students to learn the right content at the right time, improving knowledge absorption, reducing learning fatigue and knowledge forgetting, and ultimately increasing learning efficiency.

[0098] The step of inputting the associated distance value and the training feedback data set into the preset frequency prediction model to generate the push frequency parameter specifically includes: based on the co-occurrence distance of all keywords in the keyword sequence, calculating the content gap index of the teaching training content through a preset gap model, and the content gap index is negatively correlated with the semantic association strength between keywords; extracting the student score distribution and operation error type data in the training feedback data set, and using a preset resolution algorithm to generate a content discrimination quantization value, and the content discrimination quantization value reflects the completeness of the training content's coverage of the operation points; obtaining a preset initial push frequency benchmark value, and according to the numerical range of the content gap index and the content discrimination quantization value, The dynamic weight allocation strategy adjusts the weight of the influence of the two on the push frequency, wherein when the content gap index exceeds the preset threshold, the weight allocation is tilted towards the content discrimination quantization value; the weighted content gap index and the content discrimination quantization value and the initial push frequency reference value are input into the frequency correction function, and a corrected push frequency intermediate value is generated through a nonlinear mapping relationship. The corrected push frequency intermediate value is negatively correlated with the content gap index and positively correlated with the content discrimination quantization value; the corrected push frequency intermediate value is dynamically smoothed according to the recent student activity parameters in the training feedback data set to generate the final push frequency parameter, and the window size of the dynamic smoothing process is adaptively matched with the fluctuation amplitude of the student activity.

[0099] In this embodiment, the historical scheduling records corresponding to each priority identifier in the index category set are obtained, and a preset proportion model is used to calculate the dynamic category proportion value of the teaching and training content in the power maintenance operation scenario. The dynamic category proportion value reflects the resource allocation ratio of similar index content in historical scheduling; the time series data of the push frequency parameter and the trend intensity change rate in the feedback trend analysis result are extracted, and a training influence quantization value is generated by a preset influence algorithm. The influence algorithm integrates the nonlinear association rule between the fluctuation amplitude of the push frequency and the change direction of the feedback trend; a dynamic scheduling matrix is ​​constructed, and the dynamic category proportion value and the training influence quantization value are mapped to a preset two-dimensional decision space, where:

[0100] When the dynamic category ratio is higher than the preset benchmark ratio and the training impact quantification value is in the positive growth range, it is marked as the first-level scheduling level;

[0101] When the dynamic category ratio is lower than the preset benchmark ratio but the training impact quantification value triggers the alarm threshold, it is marked as the third-level scheduling level and the content emergency optimization instruction is activated;

[0102] In other cases, the secondary scheduling level is assigned based on the weighted combination of the dynamic category proportion and the quantified value of the practical training impact;

[0103] Based on the output results of the dynamic scheduling matrix, the scheduling level is verified in combination with the traceability code in the index category set. When a conflict is detected in the historical data chain associated with the traceability code, the adaptive calibration mechanism is activated to dynamically correct the training scheduling level; the calibrated training scheduling level is matched with the preset scheduling policy library to generate a final training scheduling level set including a scheduling priority identifier and a resource allocation weight, and the resource allocation weight is exponentially positively correlated with the training scheduling level.

[0104] (4) According to the push frequency and the priority identifier in the index category set, the training scheduling level of the teaching training content set is determined through a preset level mapping table, and the teaching training content set is dynamically weighted based on the training scheduling level to obtain the content level label of the teaching training content during the scheduling process.

[0105] According to the push frequency parameters and the priority identifier in the index category set, the training scheduling level of the teaching and training content set is determined through a preset level mapping table. Based on the training scheduling level, a dynamic weight evaluation is performed on the teaching and training content set, and the content level label of each content item in the scheduling process is output.

[0106] This embodiment determines the training scheduling level of the teaching training content set through a preset level mapping table based on the push frequency parameter and the priority identifier in the index category set, performs dynamic weight evaluation on the teaching training content set based on the training scheduling level, and outputs the content level label of each content item in the scheduling process. In detailed operation, the preset level mapping table pre-sets the training scheduling level corresponding to different combinations of push frequency and priority identifier, for example, high priority and high push frequency correspond to the first-level training scheduling level, and low priority and low push frequency correspond to the third-level training scheduling level. The push frequency parameter and the priority identifier in the index category set are substituted into the level mapping table to determine the training scheduling level of each teaching training content. Then, based on the training scheduling level, a dynamic weight allocation algorithm is adopted to assign different weights to teaching contents of different levels, and the content level label of each content item in the scheduling process is obtained through weight calculation to quantify its importance in the overall training arrangement.

[0107] Taking into account the frequency and importance of teaching content, and using pre-set rules and algorithms, we determine the priority and weight of teaching content in practical training scheduling, providing a quantitative basis for practical training arrangements. This step makes practical training arrangements more scientific, reasonable, and orderly, ensuring that important and frequently used teaching content is prioritized and focused on, optimizing the allocation of practical training resources, improving the overall effectiveness and quality of practical training, and enabling trainees to more efficiently learn key knowledge and skills.

[0108] (5) Optimize and sort the teaching and training content collection according to the content level labels, generate a training scheduling plan for the virtual-reality fusion scenario, and synchronize the training scheduling plan with the operation instruction flow of the digital training platform in real time.

[0109] The teaching training content set is optimized and sorted according to the content level labels, a training scheduling plan adapted to the virtual-reality fusion scenario of the new distribution system is generated, and the training scheduling plan is synchronized with the operation instruction flow of the digital training platform in real time.

[0110] This embodiment optimizes and sorts the teaching and training content set according to the content level label, generates a training scheduling plan adapted to the virtual-reality fusion scenario of the new distribution system, and synchronizes the training scheduling plan with the operation instruction stream of the digital training platform in real time. In actual operation, the teaching and training content set is optimized and sorted from high to low according to the level of the content level label, and the teaching content with high-level labels is placed at the forefront as the key teaching content. At the same time, combined with the characteristics of the virtual-reality fusion scenario of the new distribution system, such as the three-dimensional modeling of power equipment and operation process simulation in virtual simulation software, as well as the physical operation environment and operation specifications of the actual distribution equipment, the teaching content is accurately matched and arranged with the operation links, equipment status changes, etc. in the virtual-reality fusion scenario.

[0111] The optimized sorting is based on the content level label and the adaptability of the teaching and training content set to the virtual-reality fusion scenario of the new distribution system, and the training scheduling plan also includes a real-time synchronization strategy that matches the operation instruction stream of the digital training platform; the real-time synchronization is achieved by establishing a data communication interface between the training scheduling plan and the digital training platform, and using an incremental update algorithm to ensure the consistency of the operation instruction stream with the training scheduling plan, thereby improving the response speed and training effect of the digital virtual-reality fusion smart power training prototype system.

[0112] In building an intelligent sorting system, the system receives a collection of teaching and training content with content-level labels and integrates a compatibility assessment module to evaluate the compatibility of each teaching content with the virtual-reality integration scenario of the new distribution system. This assessment is based on factors such as whether the teaching content addresses key technologies of the new distribution system and whether it aligns with the training model for virtual-reality integration. The content-level labels and compatibility assessment results are input into a comprehensive sorting algorithm, which uses a weighted sorting approach to optimize the ranking of the teaching and training content collection based on pre-set weight coefficients, generating a preliminary training scheduling framework.

[0113] This embodiment combines the importance ranking of teaching content with practical training scenarios for the new power distribution system, achieving precise matching and real-time interaction between teaching and training through technical means. The benefit of this step is that it provides students with a coherent, orderly, and highly realistic training experience. This allows students to efficiently learn power maintenance operation skills in a virtual and real-world environment, improving the authenticity and effectiveness of practical training, cultivating students' ability to solve practical problems, and better meeting the skill requirements of the new power distribution system for operation and maintenance personnel.

[0114] The virtual-reality integrated smart power training system for power distribution systems of the present invention includes:

[0115] A content matching calculation module is used to calculate the content matching degree between the teaching and training content set and the power maintenance operation training objectives based on the power maintenance operation label and using a matching algorithm;

[0116] An index category set generation module is used to generate feedback trend analysis results based on the training feedback data set of trainees on the teaching and training content set. Based on the content matching degree and feedback trend analysis results, the teaching and training content set is classified into multiple dimensions according to preset classification rules to generate an index category set containing priority identifiers.

[0117] The push frequency calculation module is used to extract the keyword sequence in the teaching and training content set, input the correlation distance value between the keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process;

[0118] A content level label generation module is used to determine the training scheduling level of the teaching and training content set based on the push frequency and the priority identifier in the index category set through a preset level mapping table, and dynamically evaluate the teaching and training content set based on the training scheduling level to obtain the content level label of the teaching and training content during the scheduling process;

[0119] The training scheduling plan generation and synchronization module is used to optimize the sorting of teaching and training content collections according to content level labels, generate training scheduling plans for virtual and real fusion scenarios, and synchronize the training scheduling plans with the operation instruction stream of the digital training platform in real time.

[0120] Furthermore, in the content matching calculation module, based on the power maintenance operation label, a matching algorithm is used to calculate the content matching between the teaching training content set and the power maintenance operation training objective, as follows:

[0121] According to a preset similarity threshold, a set of similar training content items having the same power maintenance operation label as the current teaching training content item is screened out from the historical training content items of the historical teaching training data;

[0122] For each similar training content item in the set of similar training content items, the same vocabulary matching algorithm is used to extract a common keyword sequence between the similar training content item and the current teaching training content item, and the single-category confidence between the two is calculated based on the length and distribution density of the common keyword sequence;

[0123] Traversing all similar training content items in the similar training content item set to generate a single-category confidence set corresponding to the current teaching training content item;

[0124] Inputting the single-category confidence set into a preset confidence aggregation model, and generating a content matching quantification value between the current teaching training content item and the power maintenance operation training objective through weighted normalization processing, wherein the weight coefficient is dynamically adjusted according to the operation complexity and execution success rate of the historical training content item;

[0125] Each similar training item in the set of similar training content items includes the number of historical trainees, operation type labels and confidence parameters;

[0126] The preset entropy value algorithm is used to calculate the student quantity entropy based on the historical distribution characteristics of the number of students. The student quantity entropy is used to characterize the dispersion of student participation in similar training projects.

[0127] According to the confidence parameters corresponding to the set of similar training content items, a content adjustment degree of the current teaching training content item is generated through a preset adjustment degree calculation model. The content adjustment degree is used to reflect the influence coefficient of the fluctuation of the confidence degree of a single category on the adaptability of the training target;

[0128] Input the student quantity entropy and the content adjustment degree into the dynamic weight allocation strategy, and dynamically adjust the weight ratio of the two in the matching degree calculation according to the discrete degree of the student quantity entropy;

[0129] Based on the weighted student population entropy and content adjustment degree, a content matching quantization value between the current teaching training content item and the power maintenance operation training objective is generated through linear normalization processing, wherein the matching quantization value is negatively correlated with the student population entropy and positively correlated with the content adjustment degree.

[0130] In actual implementation, a data fusion module is also required to receive and integrate push frequency parameters and priority identifiers from the index category set. This module ensures data accuracy and completeness by performing necessary data cleaning and preprocessing. Next, a generated level mapping table is used. This mapping table is regularly updated and maintained by a dedicated data analysis system based on the complexity level of power maintenance operations and the historical application of the teaching and training content set. Each time a training scheduling level needs to be determined, the fused data is entered into the mapping table for search and matching, resulting in a preliminary scheduling level. Next, a dynamic weighted evaluation mechanism is activated, employing a multi-factor comprehensive evaluation algorithm, which can be run in a separate evaluation calculation unit. This algorithm retrieves update frequency data for the teaching and training content set from the data repository and quantitatively analyzes it along with the preliminary scheduling level and other relevant factors. These new factors are introduced to avoid the prior art approach of focusing solely on push frequency and priority identifiers, which may overlook other important influencing factors. By setting appropriate weight coefficients and performing a weighted calculation on each factor, a comprehensive weight value is ultimately determined for each teaching and training content item, thereby determining its content level label during the scheduling process. The entire process must ensure the security of data transmission and the efficiency of processing, and a corresponding monitoring and logging system must be established to trace and optimize the evaluation process.

[0131] Furthermore, in the index category set generation module, feedback trend analysis results are generated for the training feedback data set of the trainees on the teaching training content set. Based on the content matching degree and the feedback trend analysis results, the teaching training content set is multi-dimensionally classified according to the preset classification rules to generate an index category set containing priority identifiers, as follows:

[0132] The training feedback data set of trainees on the teaching training content set is divided into a positive feedback data subset and a negative feedback data subset;

[0133] Using a preset entropy calculation model, a dispersion analysis is performed on the score distribution characteristics in the positive feedback data subset to generate a positive feedback entropy value representing the concentration of positive opinions; a volatility calculation is performed on the frequency of occurrence of question types in the negative feedback data subset to generate a negative feedback entropy value representing the diffusion of negative opinions;

[0134] According to a preset weight allocation strategy, combined with the index category characteristics of the teaching and training content, the weight coefficients of the positive feedback entropy value and the negative feedback entropy value are dynamically allocated;

[0135] The weighted positive feedback entropy value and the negative feedback entropy value are input into the trend quantification model, and a feedback trend quantification index is generated through a preset trend mapping function. The feedback trend quantification index includes: when the proportion of positive entropy values ​​exceeds a first threshold, it is determined to be a positive trend; when the proportion of negative entropy values ​​exceeds a second threshold, a content optimization alarm is triggered;

[0136] Based on the feedback trend quantitative indicators and the preset trend level comparison table, a feedback trend analysis result including a trend strength indicator and a confidence interval is generated. The confidence interval is dynamically calculated according to the amount of historical feedback data and the classification accuracy;

[0137] Obtaining the teaching time parameters of each content item in the teaching and training content set, and performing timeliness correction on the content matching degree based on a preset time decay factor to generate a dynamic content matching degree weighted value;

[0138] Inputting the dynamic content matching degree weighted value and the trend strength indicator in the feedback trend analysis result into a preset effectiveness calculation model, and generating a quantitative value of the practical training effectiveness of the teaching and training content through a nonlinear fitting algorithm, wherein the effectiveness calculation model integrates a compensation function for the attenuation of the operation effect of the teaching time parameter;

[0139] According to the numerical interval of the quantified value of the training effectiveness and the confidence interval range in the feedback trend analysis result, a dynamic classification matrix is ​​constructed, and based on the dynamic classification matrix, an index category set containing priority identifiers is generated.

[0140] Furthermore, in the push frequency calculation module, the keyword sequence in the teaching and training content set is extracted, and the correlation distance value between the keywords and the training feedback data set are input into the preset frequency prediction model to obtain the push frequency of the teaching and training content in the power training process, as follows:

[0141] Extracting keyword sequences from the teaching and training content set, and calculating a content gap index of the teaching and training content using a preset gap model based on association distance values ​​between the keywords, wherein the content gap index is negatively correlated with the semantic association strength between the keywords;

[0142] Based on the distribution of student ratings and operational error types in the training feedback dataset, a preset discrimination algorithm is used to generate a quantitative value for content discrimination.

[0143] According to the numerical range of the content gap indicator and the content resolution quantification value, the influence weight of the content gap indicator and the content resolution quantification value on the push frequency is adjusted through a dynamic weight allocation strategy. When the content gap indicator exceeds a preset threshold, the weight allocation is tilted towards the content resolution quantification value.

[0144] The weighted content gap index and content discrimination quantization value and the initial push frequency reference value are input into the frequency correction function, and a corrected push frequency intermediate value is generated through a nonlinear mapping relationship. The corrected push frequency intermediate value is negatively correlated with the content gap index and positively correlated with the content discrimination quantization value;

[0145] The intermediate value of the corrected push frequency is dynamically smoothed according to the recent student activity parameters in the training feedback data set to generate a final push frequency. The window size of the dynamic smoothing process is adaptively matched to the fluctuation amplitude of the student activity.

[0146] The computer device of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0147] The computer-readable storage medium of the present invention stores a computer program thereon, and the computer program implements the steps of the above method when executed by a processor.

[0148] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiment of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal translation scripting language JavaScript, etc.

[0149] The present application is described with reference to the flow chart and / or block diagram of the method, device (system), and computer program product according to the embodiment of the present application. It should be understood that each flow process and / or box in the flow chart and / or block diagram and the combination of the flow process and / or box in the flow chart and / or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processing machine or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one flow chart flow or multiple flows and / or one box or multiple boxes of the block diagram.

[0150] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0151] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

Claims

1. A virtual-real integration smart power training method for distribution systems, characterized by: The following steps are involved: Based on the power maintenance operation labels, a matching algorithm is used to calculate the content matching degree between the teaching training content set and the power maintenance operation training objectives; Generate feedback trend analysis results for the training feedback dataset of students on the teaching training content set. Based on the content matching degree and feedback trend analysis results, perform multi-dimensional classification of the teaching training content set according to the preset classification rules to generate an index category set containing priority identifiers. Extract keyword sequences from the teaching and training content set, input the correlation distance values ​​between keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process; According to the push frequency and the priority identifier in the index category set, the training scheduling level of the teaching and training content set is determined through a preset level mapping table. Based on the training scheduling level, the teaching and training content set is dynamically weighted and evaluated to obtain the content level label of the teaching and training content during the scheduling process. The teaching and training content collection is optimized and sorted according to the content level labels, a training scheduling plan for the virtual-reality fusion scenario is generated, and the training scheduling plan is synchronized with the operation instruction flow of the digital training platform in real time.

2. The virtual-reality integrated smart power training method for power distribution systems according to claim 1 is characterized in that: Based on the power maintenance operation label, a matching algorithm is used to calculate the content matching degree between the teaching training content set and the power maintenance operation training objectives, as follows: According to a preset similarity threshold, a set of similar training content items having the same power maintenance operation label as the current teaching training content item is screened out from the historical training content items of the historical teaching training data; For each similar training content item in the set of similar training content items, the same vocabulary matching algorithm is used to extract a common keyword sequence between the similar training content item and the current teaching training content item, and the single-category confidence between the two is calculated based on the length and distribution density of the common keyword sequence; Traversing all similar training content items in the similar training content item set to generate a single-category confidence set corresponding to the current teaching training content item; Inputting the single-category confidence set into a preset confidence aggregation model, and generating a content matching quantification value between the current teaching training content item and the power maintenance operation training objective through weighted normalization processing, wherein the weight coefficient is dynamically adjusted according to the operation complexity and execution success rate of the historical training content item; Each similar training item in the set of similar training content items includes the number of historical trainees, operation type labels and confidence parameters; The preset entropy value algorithm is used to calculate the student quantity entropy based on the historical distribution characteristics of the number of students. The student quantity entropy is used to characterize the dispersion of student participation in similar training projects. According to the confidence parameters corresponding to the set of similar training content items, a content adjustment degree of the current teaching training content item is generated through a preset adjustment degree calculation model. The content adjustment degree is used to reflect the influence coefficient of the fluctuation of the confidence degree of a single category on the adaptability of the training target; Input the student quantity entropy and the content adjustment degree into the dynamic weight allocation strategy, and dynamically adjust the weight ratio of the two in the matching degree calculation according to the discrete degree of the student quantity entropy; Based on the weighted student population entropy and content adjustment degree, a content matching quantization value between the current teaching training content item and the power maintenance operation training objective is generated through linear normalization processing, wherein the matching quantization value is negatively correlated with the student population entropy and positively correlated with the content adjustment degree.

3. The virtual-reality integrated smart power training method for distribution systems according to claim 2 is characterized in that: For the training feedback dataset of students on the teaching and training content set, feedback trend analysis results are generated. Based on the content matching degree and feedback trend analysis results, the teaching and training content set is classified into multiple dimensions according to the preset classification rules, and an index category set containing priority identifiers is generated, as follows: The training feedback data set of trainees on the teaching training content set is divided into a positive feedback data subset and a negative feedback data subset; Using a preset entropy calculation model, a dispersion analysis is performed on the score distribution characteristics in the positive feedback data subset to generate a positive feedback entropy value representing the concentration of positive opinions; a volatility calculation is performed on the frequency of occurrence of question types in the negative feedback data subset to generate a negative feedback entropy value representing the diffusion of negative opinions; According to a preset weight allocation strategy, combined with the index category characteristics of the teaching and training content, the weight coefficients of the positive feedback entropy value and the negative feedback entropy value are dynamically allocated; The weighted positive feedback entropy value and the negative feedback entropy value are input into the trend quantification model, and a feedback trend quantification index is generated through a preset trend mapping function. The feedback trend quantification index includes: when the proportion of positive entropy values ​​exceeds a first threshold, it is determined to be a positive trend; when the proportion of negative entropy values ​​exceeds a second threshold, a content optimization alarm is triggered; Based on the feedback trend quantitative indicators and the preset trend level comparison table, a feedback trend analysis result including a trend strength indicator and a confidence interval is generated. The confidence interval is dynamically calculated according to the amount of historical feedback data and the classification accuracy; Obtaining the teaching time parameters of each content item in the teaching and training content set, and performing timeliness correction on the content matching degree based on a preset time decay factor to generate a dynamic content matching degree weighted value; Inputting the dynamic content matching degree weighted value and the trend strength indicator in the feedback trend analysis result into a preset effectiveness calculation model, and generating a quantitative value of the practical training effectiveness of the teaching and training content through a nonlinear fitting algorithm, wherein the effectiveness calculation model integrates a compensation function for the attenuation of the operation effect of the teaching time parameter; According to the numerical interval of the quantified value of the training effectiveness and the confidence interval range in the feedback trend analysis result, a dynamic classification matrix is ​​constructed, and based on the dynamic classification matrix, an index category set containing priority identifiers is generated.

4. The virtual-reality integrated smart power training method for distribution systems according to claim 3 is characterized in that: Extract the keyword sequence from the teaching and training content set, input the correlation distance value between the keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process, as follows: Extracting keyword sequences from the teaching and training content set, and calculating a content gap index of the teaching and training content using a preset gap model based on association distance values ​​between the keywords, wherein the content gap index is negatively correlated with the semantic association strength between the keywords; Based on the distribution of student ratings and operational error types in the training feedback dataset, a preset discrimination algorithm is used to generate a quantitative value for content discrimination. According to the numerical range of the content gap indicator and the content resolution quantification value, the influence weight of the content gap indicator and the content resolution quantification value on the push frequency is adjusted through a dynamic weight allocation strategy. When the content gap indicator exceeds a preset threshold, the weight allocation is tilted towards the content resolution quantification value. The weighted content gap index and content discrimination quantization value and the initial push frequency reference value are input into the frequency correction function, and a corrected push frequency intermediate value is generated through a nonlinear mapping relationship. The corrected push frequency intermediate value is negatively correlated with the content gap index and positively correlated with the content discrimination quantization value; The intermediate value of the corrected push frequency is dynamically smoothed according to the recent student activity parameters in the training feedback data set to generate a final push frequency. The window size of the dynamic smoothing process is adaptively matched to the fluctuation amplitude of the student activity.

5. A virtual-real integration smart power training system for distribution systems, characterized by: include: A content matching calculation module is used to calculate the content matching degree between the teaching and training content set and the power maintenance operation training objectives based on the power maintenance operation label and using a matching algorithm; An index category set generation module is used to generate feedback trend analysis results based on the training feedback data set of trainees on the teaching and training content set. Based on the content matching degree and feedback trend analysis results, the teaching and training content set is classified into multiple dimensions according to preset classification rules to generate an index category set containing priority identifiers. The push frequency calculation module is used to extract the keyword sequence in the teaching and training content set, input the correlation distance value between the keywords and the training feedback data set into the preset frequency prediction model, and obtain the push frequency of the teaching and training content in the power training process; A content level label generation module is used to determine the training scheduling level of the teaching and training content set based on the push frequency and the priority identifier in the index category set through a preset level mapping table, and dynamically evaluate the teaching and training content set based on the training scheduling level to obtain the content level label of the teaching and training content during the scheduling process; The training scheduling plan generation and synchronization module is used to optimize the sorting of teaching and training content collections according to content level labels, generate training scheduling plans for virtual and real fusion scenarios, and synchronize the training scheduling plans with the operation instruction stream of the digital training platform in real time.

6. The virtual-reality integrated smart power training system for power distribution systems according to claim 5 is characterized in that: In the content matching calculation module, based on the power maintenance operation label, a matching algorithm is used to calculate the content matching between the teaching and training content set and the power maintenance operation training objectives, as follows: According to a preset similarity threshold, a set of similar training content items having the same power maintenance operation label as the current teaching training content item is screened out from the historical training content items of the historical teaching training data; For each similar training content item in the set of similar training content items, the same vocabulary matching algorithm is used to extract a common keyword sequence between the similar training content item and the current teaching training content item, and the single-category confidence between the two is calculated based on the length and distribution density of the common keyword sequence; Traversing all similar training content items in the similar training content item set to generate a single-category confidence set corresponding to the current teaching training content item; Inputting the single-category confidence set into a preset confidence aggregation model, and generating a content matching quantification value between the current teaching training content item and the power maintenance operation training objective through weighted normalization processing, wherein the weight coefficient is dynamically adjusted according to the operation complexity and execution success rate of the historical training content item; Each similar training item in the set of similar training content items includes the number of historical trainees, operation type labels and confidence parameters; The preset entropy value algorithm is used to calculate the student quantity entropy based on the historical distribution characteristics of the number of students. The student quantity entropy is used to characterize the dispersion of student participation in similar training projects. According to the confidence parameters corresponding to the set of similar training content items, a content adjustment degree of the current teaching training content item is generated through a preset adjustment degree calculation model. The content adjustment degree is used to reflect the influence coefficient of the fluctuation of the confidence degree of a single category on the adaptability of the training target; Input the student quantity entropy and the content adjustment degree into the dynamic weight allocation strategy, and dynamically adjust the weight ratio of the two in the matching degree calculation according to the discrete degree of the student quantity entropy; Based on the weighted student population entropy and content adjustment degree, a content matching quantization value between the current teaching training content item and the power maintenance operation training objective is generated through linear normalization processing, wherein the matching quantization value is negatively correlated with the student population entropy and positively correlated with the content adjustment degree.

7. The virtual-reality integrated smart power training system for power distribution systems according to claim 6 is characterized in that: In the index category set generation module, feedback trend analysis results are generated for the training feedback dataset of students on the teaching and training content set. Based on the content matching degree and feedback trend analysis results, the teaching and training content set is classified into multiple dimensions according to the preset classification rules to generate an index category set with priority identification, as follows: The training feedback data set of trainees on the teaching training content set is divided into a positive feedback data subset and a negative feedback data subset; Using a preset entropy calculation model, a dispersion analysis is performed on the score distribution characteristics in the positive feedback data subset to generate a positive feedback entropy value representing the concentration of positive opinions; a volatility calculation is performed on the frequency of occurrence of question types in the negative feedback data subset to generate a negative feedback entropy value representing the diffusion of negative opinions; According to a preset weight allocation strategy, combined with the index category characteristics of the teaching and training content, the weight coefficients of the positive feedback entropy value and the negative feedback entropy value are dynamically allocated; The weighted positive feedback entropy value and the negative feedback entropy value are input into the trend quantification model, and a feedback trend quantification index is generated through a preset trend mapping function. The feedback trend quantification index includes: when the proportion of positive entropy values ​​exceeds a first threshold, it is determined to be a positive trend; when the proportion of negative entropy values ​​exceeds a second threshold, a content optimization alarm is triggered; Based on the feedback trend quantitative indicators and the preset trend level comparison table, a feedback trend analysis result including a trend strength indicator and a confidence interval is generated. The confidence interval is dynamically calculated according to the amount of historical feedback data and the classification accuracy; Obtaining the teaching time parameters of each content item in the teaching and training content set, and performing timeliness correction on the content matching degree based on a preset time decay factor to generate a dynamic content matching degree weighted value; Inputting the dynamic content matching degree weighted value and the trend strength indicator in the feedback trend analysis result into a preset effectiveness calculation model, and generating a quantitative value of the practical training effectiveness of the teaching and training content through a nonlinear fitting algorithm, wherein the effectiveness calculation model integrates a compensation function for the attenuation of the operation effect of the teaching time parameter; According to the numerical interval of the quantified value of the training effectiveness and the confidence interval range in the feedback trend analysis result, a dynamic classification matrix is ​​constructed, and based on the dynamic classification matrix, an index category set containing priority identifiers is generated.

8. The virtual-reality integrated smart power training system for power distribution systems according to claim 7 is characterized in that: In the push frequency calculation module, the keyword sequence in the teaching and training content set is extracted, and the correlation distance value between the keywords and the training feedback data set are input into the preset frequency prediction model to obtain the push frequency of the teaching and training content during the power training process, as follows: Extracting keyword sequences from the teaching and training content set, and calculating a content gap index of the teaching and training content using a preset gap model based on association distance values ​​between the keywords, wherein the content gap index is negatively correlated with the semantic association strength between the keywords; Based on the distribution of student ratings and operational error types in the training feedback dataset, a preset discrimination algorithm is used to generate a quantitative value for content discrimination. According to the numerical range of the content gap indicator and the content resolution quantification value, the influence weight of the content gap indicator and the content resolution quantification value on the push frequency is adjusted through a dynamic weight allocation strategy. When the content gap indicator exceeds a preset threshold, the weight allocation is tilted towards the content resolution quantification value. The weighted content gap index and content discrimination quantization value and the initial push frequency reference value are input into the frequency correction function, and a corrected push frequency intermediate value is generated through a nonlinear mapping relationship. The corrected push frequency intermediate value is negatively correlated with the content gap index and positively correlated with the content discrimination quantization value; The intermediate value of the corrected push frequency is dynamically smoothed according to the recent student activity parameters in the training feedback data set to generate a final push frequency. The window size of the dynamic smoothing process is adaptively matched to the fluctuation amplitude of the student activity.

9. A computer 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 steps of the method according to any one of claims 1 to 4 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.