A power metering device practical training method, system, computer device and medium

Abstract

The application relates to the technical field of device practical training, and discloses a power metering device practical training method, a power metering device practical training system, computer equipment and a medium, which comprises the following steps: performing knowledge and ability evaluation, learning style analysis and interest entity extraction on each practical training student to obtain the ability value, the knowledge weak point set, the learning style and the interest vector of the practical training student, and constructing the student portrait corresponding to the practical training student; extracting a learning subgraph from a power metering device knowledge graph according to the student portrait, sorting the knowledge nodes of the learning subgraph by using a KgRank algorithm, and generating a learning path; generating learning content based on AR interaction according to the learning path; and in response to completing interactive learning, updating the knowledge graph and the student portrait according to the learning behavior of the practical training student, and adjusting the learning path. Through immersive interaction of AR technology and structured reasoning of the knowledge graph, and in combination with dynamic evaluation and path recommendation, the safety, efficiency and individualization level of power practical training are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of equipment training technology, and in particular to a training method, system, computer equipment, and medium for power metering equipment. Background Technology

[0002] Traditional training for electricity metering equipment primarily relies on paper-based materials, on-site explanations, and practical demonstrations. These conventional methods suffer from several drawbacks, including low information acquisition efficiency, high comprehension barriers, poor training effectiveness, and significant safety hazards. Specifically, low acquisition efficiency stems from the abundance of paper-based materials and the difficulty in retrieval, making it challenging for new employees to quickly obtain accurate information (such as the internal structure of electricity meters and wiring specifications). High comprehension barriers arise from the complexity of equipment principles and operating procedures, making it difficult to visually demonstrate hidden components (such as the internal coils of current transformers) and their dynamic operating mechanisms through simple explanations. Poor training effectiveness is characterized by a lack of personalized support and interactive design, leading to insufficient skill mastery due to varying learning paces, and traditional assessment methods (written tests + practical exercises) failing to accurately quantify skill gaps. Significant safety hazards include the risk of electric shock during high-voltage operations and high wear and tear costs due to frequent equipment use (e.g., a wear and tear rate ≥15% during electricity meter disassembly and assembly training).

[0003] Therefore, there is an urgent need for a new training method that can meet the safety, efficiency, and personalization requirements of power metering equipment training. Summary of the Invention

[0004] To address the aforementioned technical problems, this invention provides a training method, system, computer equipment, and medium for power metering equipment, which can solve the problem that traditional power metering equipment training cannot meet the requirements for safety, efficiency, and personalization, thereby achieving the technical effect of improving training efficiency and safety.

[0005] In a first aspect, the present invention provides a training method for power metering equipment, the method comprising:

[0006] For each trainee, knowledge and ability assessments, learning style analysis, and interest entity extraction are conducted to obtain the trainee's ability value, set of knowledge weaknesses, learning style, and interest vector.

[0007] Based on the ability value, the set of knowledge weaknesses, the learning style, and the interest vector, construct a student profile corresponding to the practical training student;

[0008] Based on the student profile, a learning subgraph is extracted from the pre-constructed knowledge graph of power metering equipment. The KgRank algorithm is used to sort the knowledge nodes of the learning subgraph to generate the learning path for the trainees.

[0009] AR-based interactive learning content is generated according to the learning path, so that trainees can interactively learn the learning content.

[0010] In response to the completion of the interactive learning, the knowledge graph and the student profile are updated based on the learning behavior of the trainees, and the learning path is adjusted until the power metering equipment training ends.

[0011] Furthermore, the steps of conducting knowledge and ability assessments, learning style analysis, and interest entity extraction for each trainee to obtain their ability scores, set of knowledge weaknesses, learning styles, and interest vectors include:

[0012] Determine whether the trainee is taking the assessment for the first time. If so, generate initial test questions randomly based on the trainee's registration information to conduct an initial assessment. If not, generate personalized test questions based on the previous assessment results to conduct a dynamic assessment of the trainee.

[0013] Based on the accuracy rate of the dynamic assessment, the trainees' ability values ​​are obtained, and the knowledge points that were not passed in the assessment are identified as knowledge weaknesses, forming a set of knowledge weaknesses.

[0014] Based on a pre-set scale, the explicit learning styles of the trainees are obtained, and the explicit learning styles are modified according to the trainees' learning behaviors to obtain the trainees' learning styles.

[0015] Using a pre-defined interest recognition model, interest entities are extracted from the learning behavior logs of trainees to obtain their interest vectors. The interest recognition model is constructed based on a bidirectional long short-term memory neural network and a stochastic conditional field.

[0016] Furthermore, the step of extracting a learning subgraph from a pre-constructed knowledge graph of power metering equipment based on the student profile, and using the KgRank algorithm to sort the knowledge nodes of the learning subgraph to generate the learning path for the trainees includes:

[0017] Based on the set of knowledge gaps, a learning subgraph is extracted from the knowledge graph of power metering equipment.

[0018] Based on the ability value, determine the mastery level of each knowledge node in the learning subgraph;

[0019] Based on the matching degree between the learning style and the resource type, calculate the resource type weight of each knowledge node;

[0020] Calculate the interest matching degree of each knowledge node based on the similarity between the interest vector and the knowledge node;

[0021] Based on the mastery level, the resource type weight, and the interest matching degree, the KgRank algorithm is used to calculate the dynamic weight value of each knowledge node.

[0022] Based on the dynamic weight values, the knowledge nodes of the learning subgraph are sorted to obtain the learning path.

[0023] Furthermore, after the step of calculating the dynamic weight value of each knowledge node using the KgRank algorithm based on the mastery level, the resource type weight, and the interest matching degree, the method further includes:

[0024] The dynamic weight value is corrected based on the PageRank value of each knowledge node to obtain the corrected dynamic weight value.

[0025] The corrected dynamic weight value is expressed by the following formula:

[0026]

[0027] In the formula, n i Score(n) represents the i-th knowledge node. i ) represents the corrected dynamic weight value of the i-th knowledge node, PR(n) i ) represents the PageRank value of the i-th knowledge node, and Similarity(n) represents the similarity value of the i-th knowledge node. i ) represents the interest matching degree between the i-th knowledge node and the interest vector, mi represents the ability value of the i-th knowledge node, and ω k S represents the initial weight of the k-th resource type corresponding to the i-th knowledge node. uk α represents the matching degree between the learning style and the k-th resource type, where α, β, γ, and δ all represent weighting coefficients.

[0028] Furthermore, the steps of updating the knowledge graph and the student profile based on the learning behavior of the trainees, and adjusting the learning path, include:

[0029] For each trainee, a new knowledge and ability assessment, learning style analysis, and interest entity extraction were conducted to obtain the latest ability value, the latest set of knowledge weaknesses, the latest learning style, and the latest interest vector.

[0030] Statistical analysis was conducted on each knowledge weakness in the latest set of knowledge weaknesses, using the number of trainees as the unit.

[0031] If the number of students corresponding to the knowledge gap is greater than the number threshold, the association weight of the predecessor knowledge points of the knowledge gap in the knowledge graph will be adjusted.

[0032] Based on each trainee's latest ability value, latest set of knowledge weaknesses, latest learning style, and latest interest vector, the trainee profile is adjusted, and a corresponding learning path is regenerated based on the adjusted trainee profile and the adjusted knowledge graph.

[0033] Furthermore, the steps for constructing the knowledge graph of the power metering equipment include:

[0034] Knowledge data on power metering equipment is obtained from professional model libraries and learning resource libraries, and the knowledge data is preprocessed to obtain a cleaned structured dataset.

[0035] Using knowledge points, questions, students, and majors as entities and hierarchical and mapping relationships as relations, knowledge is extracted from the structured dataset to obtain a knowledge graph of power metering equipment.

[0036] Furthermore, the step of extracting knowledge from the structured dataset to obtain a knowledge graph of power metering equipment includes:

[0037] Entity relation tuples are extracted from the structured dataset to obtain explicit knowledge;

[0038] By using a pre-defined knowledge extraction model, potential relationships between entities are extracted from historical failure cases, and the logic between knowledge points is inferred through a rule engine to obtain implicit knowledge. The knowledge extraction model is constructed based on a bidirectional long short-term memory neural network.

[0039] By fusing the explicit and implicit knowledge, a knowledge graph of power metering equipment is obtained.

[0040] Secondly, the present invention provides a power metering equipment training system, the system comprising:

[0041] The trainee profile building module is used to assess the knowledge and ability of each trainee, analyze their learning style, and extract their interest entities, resulting in the trainee's ability value, set of knowledge weaknesses, learning style, and interest vector.

[0042] Based on the ability value, the set of knowledge weaknesses, the learning style, and the interest vector, construct a student profile corresponding to the practical training student;

[0043] The learning path generation module is used to extract learning subgraphs from the pre-built knowledge graph of power metering equipment based on the student profile, and to sort the knowledge nodes of the learning subgraphs using the KgRank algorithm to generate the learning path of the trainees.

[0044] The AR interactive learning module is used to generate AR-based interactive learning content according to the learning path, so that trainees can interactively learn the learning content.

[0045] The learning path update module is used to update the knowledge graph and the student profile based on the learning behavior of the trainees in response to the completion of the interactive learning, and to adjust the learning path until the power metering equipment training is completed.

[0046] Thirdly, embodiments of the present invention also provide a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0047] Fourthly, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-described method.

[0048] This invention provides a training method, system, computer equipment, and medium for power metering equipment. Through AR interactive 3D visualization, this invention improves trainees' understanding of complex structures, ensures training safety, reduces equipment wear and tear, accurately portrays trainee profiles through personalized dynamic assessments, enhances the systematization of training content by constructing a hierarchical knowledge network, and improves trainees' learning efficiency through personalized learning path recommendations. This invention effectively improves the safety, efficiency, and personalization of power metering equipment training by combining immersive interaction with AR technology and structured reasoning from knowledge graphs with dynamic assessments and path recommendations. Attached Figure Description

[0049] Figure 1 This is a flowchart illustrating the training method for power metering equipment according to an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the structure of the power metering equipment training system according to an embodiment of the present invention;

[0051] Figure 3 This is an internal structural diagram of the computer device in an embodiment of the present invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] Please see Figure 1 The first embodiment of the present invention proposes a training method for power metering equipment, which includes steps S10 to S50:

[0054] Step S10: Conduct knowledge and ability assessments, learning style analysis, and interest entity extraction for each trainee to obtain the trainee's ability value, set of knowledge weaknesses, learning style, and interest vector.

[0055] Step S20: Construct a student profile corresponding to the practical training student based on the ability value, the set of knowledge weaknesses, the learning style, and the interest vector;

[0056] Step S30: Based on the student profile, extract the learning subgraph from the pre-constructed knowledge graph of power metering equipment, and use the KgRank algorithm to sort the knowledge nodes of the learning subgraph to generate the learning path of the trainee.

[0057] Step S40: Generate AR-based interactive learning content according to the learning path, so that trainees can interactively learn the learning content;

[0058] In step S50, in response to the completion of the interactive learning, the knowledge graph and the student profile are updated according to the learning behavior of the trainees, and the learning path is adjusted until the power metering equipment training ends.

[0059] This invention provides a training method for power metering equipment based on a combination of AR interaction and knowledge graph. By analyzing the learning situation of trainees and the knowledge graph of power metering equipment, personalized learning paths are planned for trainees, and AR interaction technology is used for training to improve the learning efficiency of trainees.

[0060] In this embodiment, the individual circumstances of each trainee are first analyzed to construct a trainee profile. The specific analysis steps include:

[0061] Determine whether the trainee is taking the assessment for the first time. If so, generate initial test questions randomly based on the trainee's registration information to conduct an initial assessment. If not, generate personalized test questions based on the previous assessment results to conduct a dynamic assessment of the trainee.

[0062] Based on the accuracy rate of the dynamic assessment, the trainees' ability values ​​are obtained, and the knowledge points that were not passed in the assessment are identified as knowledge weaknesses, forming a set of knowledge weaknesses.

[0063] Based on a pre-set scale, the explicit learning styles of the trainees are obtained, and the explicit learning styles are modified according to the trainees' learning behaviors to obtain the trainees' learning styles.

[0064] Using a pre-defined interest recognition model, interest entities are extracted from the learning behavior logs of trainees to obtain their interest vectors. The interest recognition model is constructed based on a bidirectional long short-term memory neural network and a stochastic conditional field.

[0065] In this embodiment, the analysis of trainees' situations includes three aspects: knowledge mastery analysis, learning style analysis, and personal interest analysis. Specifically, for the knowledge mastery analysis, this embodiment analyzes trainees' current knowledge mastery through answering questions. For trainees who have never been assessed, initial test questions are randomly generated based on their registration information, and the trainees answer these questions to obtain their initial knowledge mastery. The registration information includes name, age, gender, education, job position, hobbies, career plans, professional title, skill level, etc. Knowledge mastery includes ability value and knowledge weaknesses. The ability value refers to the accuracy rate of the assessment, that is, the ratio of the number of correct answers to the total number of questions. Each knowledge point corresponds to several questions. The accuracy rate of answering questions for each knowledge point determines whether the trainee has mastered the knowledge point. If the assessment is not passed, that is, the accuracy rate is less than the threshold, then the knowledge point is identified as a knowledge weakness. The initial assessment questions are randomly selected from the corresponding question bank based on the trainee's education, job position, skill level, and other information related to the professional knowledge required for the job.

[0066] After the initial assessment, a formal dynamic assessment is conducted based on the trainees' performance to accurately estimate their abilities and diagnose their knowledge. In this embodiment, a knowledge graph of power metering equipment is pre-constructed; the specific construction steps will be explained later. Based on the results of the initial assessment and the predecessor-successor relationships of knowledge points in the knowledge graph, dynamic assessment questions are selected. First, it checks whether the nearest predecessor and farthest successor knowledge points of the knowledge point to be assessed already have assessment data. If the nearest predecessor knowledge point of the knowledge point has assessment data (the number of questions exceeds a certain threshold) and the mastery status is proficient (the correct answer rate exceeds the accuracy rate threshold), then easy questions are skipped during assessment, and questions of medium difficulty are selected directly. Conversely, if the mastery status is poor, difficult questions are skipped, and questions are selected from the easy ones. If the farthest successor knowledge point of the knowledge point has assessment data and the mastery status is proficient, then it is inferred that the mastery status of the knowledge point is proficient. If the mastery status of the farthest successor knowledge point of the knowledge point is poor, then it is considered that the knowledge point has not yet been mastered.

[0067] Preferably, a maximum information entropy question selection strategy can be adopted to dynamically select test questions, prioritizing those that can best differentiate learners' ability levels and avoiding repeated testing of already mastered content. Furthermore, a cognitive diagnostic model can be constructed based on item response theory or Bayesian networks to quantify the match between learners' learning abilities and question difficulty, thereby achieving more targeted question selection.

[0068] Finally, based on the dynamic assessment results, the student's ability score and knowledge gaps are determined. In this embodiment, the ability score includes the ability score for mastering each knowledge point and the ability score for mastering all knowledge points. Only when the ability score for mastering each knowledge point meets the standard is the ability score for all knowledge points considered to be up to standard.

[0069] For the analysis of learning styles, this embodiment adopts a scale approach. Based on the content filled in by the learners on the scale, the learners' explicit learning styles are determined, such as video-based learning preferences, book-based learning preferences, auditory learning preferences, etc. Different weights are assigned to different types of learning according to the degree of preference. The explicit learning style will be used as the initial learning style. After the learning begins, the explicit learning style will be modified according to the learners' learning behavior. For example, if the learners frequently operate AR animations, the weight of video-based learning will be increased, etc., thereby obtaining the learners' learning styles.

[0070] For the analysis of interest vectors, this embodiment employs an interest recognition model. Interest entities are extracted from the trainees' learning behavior logs to obtain their interest vectors. The interest recognition model is constructed based on a bidirectional long short-term memory neural network and a random conditional field, specifically a Bi-LSTM+CRF model. A BERT-based pre-trained model is also incorporated into the Bi-LSTM+CRF model. The Bi-LSTM+CRF model analyzes the trainees' learning behavior logs, outputting their interest entities, which are then mapped to a vector space to obtain the interest vectors. Finally, based on ability scores, sets of knowledge weaknesses, learning styles, and interest vectors, a trainee profile is constructed.

[0071] Then, based on the student profile, the corresponding learning path is extracted from the knowledge graph. Before explaining the steps for generating the learning path, the process of constructing the knowledge graph for power metering equipment is first explained:

[0072] Knowledge data on power metering equipment is obtained from professional model libraries and learning resource libraries, and the knowledge data is preprocessed to obtain a cleaned structured dataset.

[0073] Using knowledge points, questions, students, and majors as entities and hierarchical and mapping relationships as relations, knowledge is extracted from the structured dataset to obtain a knowledge graph of power metering equipment.

[0074] In this embodiment, a structured network of knowledge about power metering equipment is established through multi-source data fusion and dynamic reasoning mechanisms. First, the relevant data of power metering equipment is cleaned. This data includes power industry standard documents, equipment principle manuals, and fault case libraries in the professional model library. For example, knowledge data such as professional skills assessments and business processing for power marketing departments, including electricity inspection, meter reading and verification, meter installation and connection, electricity meter repair and calibration, line loss investigation, and terminal operation and maintenance. The learning material library includes textbook texts, AR 3D model metadata, and test question banks (including the mapping relationship between questions and knowledge points). For example, textbooks, teaching audio and video, VR training courseware, teaching PPTs, test questions, and test papers are included in the learning material library, covering the principles, operation procedures, and fault repair of power metering equipment.

[0075] Then, the data in the professional model library and learning resource library are preprocessed. For unstructured text, NLP tools (such as Spacy) are used to perform word segmentation and entity recognition (e.g., extracting entities such as "electricity meter," "current transformer," and "wiring principle"). For structured data, data alignment is performed, associating questions in the question bank with knowledge point IDs (e.g., question Q001 corresponds to knowledge points K001 and K002). Finally, the cleaned structured dataset is output.

[0076] Then, the entities and relations in the knowledge graph are defined, where the entity definition includes:

[0077] Key knowledge points: Equipment principles, operating procedures, safety regulations, etc. (e.g., "K001-Structure of Three-Phase Energy Meter" and "K002-Live Wiring Operation Specifications").

[0078] Title: Assessment questions and their related knowledge points, difficulty level (e.g., "Q001 - Wiring Correctness Judgment Question, Difficulty L2").

[0079] Trainee: Trainee ID, Job Type, Historical Assessment Records.

[0080] Specialization: Electrical inspection, meter installation and connection, and other electrical-related professional categories.

[0081] The relation definition includes:

[0082] Hierarchical relationship:

[0083] Precursor-successor dependencies between knowledge points (e.g., "K001→K002" means that you must master K001 before you can learn K002).

[0084] The relationship between majors and knowledge points (e.g., "meter installation and electrical connection major → K001, K002").

[0085] Mapping relationship:

[0086] The questions are related to the knowledge points (e.g., "Q001→K001,K003").

[0087] The student's mastery of the knowledge points (e.g., "S001→K001: Ability value 80%").

[0088] After defining entities and relationships, knowledge extraction is performed on the structured dataset to extract entity-relation tuples, thereby obtaining the knowledge graph of the power metering equipment. In practice, the directly extracted entity-relation tuples represent explicit knowledge. However, some implicit knowledge exists within fault cases. Therefore, in a preferred embodiment, this invention uses a knowledge extraction model to extract potential associations between entities from historical fault cases (e.g., "wiring error → electricity meter burnout"), and uses a rule engine to infer the logic between knowledge points (e.g., "if knowledge point A is a predecessor of B, and B is a predecessor of C, then A→C is an indirect dependency") to obtain implicit knowledge. Optionally, the knowledge extraction model is constructed based on a bidirectional long short-term memory neural network. Finally, the explicit and implicit knowledge are fused to obtain the knowledge graph of the power metering equipment.

[0089] After obtaining the student profile and the knowledge graph of power metering equipment, the learning path is extracted from the knowledge graph based on the student profile. The specific steps include:

[0090] Based on the set of knowledge gaps, a learning subgraph is extracted from the knowledge graph of power metering equipment.

[0091] Based on the ability value, determine the mastery level of each knowledge node in the learning subgraph;

[0092] Based on the matching degree between the learning style and the resource type, calculate the resource type weight of each knowledge node;

[0093] Calculate the interest matching degree of each knowledge node based on the similarity between the interest vector and the knowledge node;

[0094] Based on the mastery level, the resource type weight, and the interest matching degree, the KgRank algorithm is used to calculate the dynamic weight value of each knowledge node.

[0095] Based on the dynamic weight values, the knowledge nodes of the learning subgraph are sorted to obtain the learning path.

[0096] In this embodiment, based on the student's set of knowledge weaknesses K = {K001, K002, ..., KN}, a connected subgraph G centered on K is extracted from the knowledge graph. The connected subgraph G contains the predecessor and successor knowledge points of each knowledge weakness. Specifically, the predecessor knowledge point is obtained by tracing backwards along the knowledge graph to the root node, thus ensuring the completeness of the basic data. The successor knowledge point is obtained by extending forwards along the knowledge graph to the leaf nodes, thus preventing knowledge gaps. For example, taking the knowledge weakness K003 - electricity theft detection as an example, the connected subgraph contains the predecessor knowledge points: K001 - electricity meter structure → K002 - wiring principle → K003, and the successor knowledge point: K003 → K004 - troubleshooting. Furthermore, for knowledge points involving high-risk operations (such as live wiring), a safety specification node is forcibly inserted. In this embodiment, whether an operation is considered high-risk is determined by the preset operational risk value of each knowledge point.

[0097] The mastery level of each knowledge node in the subgraph is determined based on the ability value. For example, if the ability value of a knowledge node is 30%, it is converted into a non-linear weight using the Sigmoid function, thus obtaining the mastery level of that knowledge node.

[0098]

[0099] In the formula, mi represents the capability value for the i-th knowledge node.

[0100] Based on the level of mastery, we can prioritize recommending knowledge points that we haven't mastered, thus avoiding repetitive training on content that we have already mastered.

[0101] Based on the matching degree between learning style and resource type, the resource type weight of each knowledge node is calculated. Specifically, assuming that the resource types associated with a certain knowledge point include videos, texts, and test questions, different initial weights are preset for different resource types, such as video = 0.6, text = 0.3, and test questions = 0.1. Then, according to the rule table, the matching degree between learning style and resource type is determined. For example, if the student's style is video-oriented, then the matching degree of video resources is 1, and the matching degree of other resource types is 0. Then, the products of the initial weights of all resource types and their corresponding matching degrees are summed to obtain the overall matching degree between the learning resource types associated with the knowledge point and the student's learning style, and this is used as the resource type weight. Based on the resource type weight, resources that are suitable for the student's learning style can be recommended.

[0102] Based on the similarity between interest vectors and knowledge nodes, the interest matching degree of each knowledge node in the learning subgraph is calculated. Preferably, cosine similarity can be used to calculate the interest matching degree. By quantifying the interest relevance through the cosine similarity between the student's interest vector and the knowledge point representation vector, student engagement can be improved by increasing the scores of knowledge points with preferences (such as "the principle of electricity theft"). Of course, other similarity calculation methods such as Euclidean distance can also be used, and no further restrictions are imposed here.

[0103] Then, the KgRank algorithm is used to fuse multi-dimensional features to calculate the dynamic weight value of each knowledge node:

[0104]

[0105] In the formula, n i Score′(n) represents the i-th knowledge node. i Similarity(n) represents the dynamic weight value of the i-th knowledge node. i ) represents the interest matching degree between the i-th knowledge node and the interest vector, mi represents the ability value of the i-th knowledge node, and ω k S represents the initial weight of the k-th resource type corresponding to the i-th knowledge node. uk This represents the matching degree between the learning style and the k-th resource type. β, γ, and δ all represent weighting coefficients, which are preset values ​​that can be set according to different job requirements.

[0106] After calculating the dynamic weight values ​​of each knowledge node, they are sorted in descending order to obtain the learning path of the trainees.

[0107] In a preferred embodiment, the present invention, based on the aforementioned dynamic weight values, uses the PageRank values ​​of each knowledge node to correct the dynamic weight values. The PageRank value is an indicator obtained by calculating the importance of web pages using the PageRank algorithm, reflecting the logical importance of knowledge points in the overall knowledge network. The specific calculation process is detailed in the PageRank algorithm calculation process and will not be elaborated upon here. Correcting the dynamic weight values ​​using PageRank values ​​ensures the logical coherence of the learning path, prioritizing important knowledge points (such as basic principles). The corrected dynamic weight values ​​are expressed as follows:

[0108]

[0109] In the formula, n i Score(n) represents the i-th knowledge node. i ) represents the corrected dynamic weight value of the i-th knowledge node, PR(n) i ) represents the PageRank value of the i-th knowledge node, and Similarity(n) represents the similarity value of the i-th knowledge node. i ) represents the interest matching degree between the i-th knowledge node and the interest vector, mi represents the ability value of the i-th knowledge node, and ω k S represents the initial weight of the k-th resource type corresponding to the i-th knowledge node. uk α represents the matching degree between the learning style and the k-th resource type, where α, β, γ, and δ all represent weighting coefficients.

[0110] Assume the weighting coefficients are preset as follows: α = 0.4, β = 0.3, γ = 0.2, δ = 0.1. In practical applications, these values ​​can be adjusted according to different scenarios. For example, for trainees in electricity inspection, the β value can be increased; for high-risk operation knowledge points, the α value can be increased, etc.

[0111] The learning path generation method in this embodiment ensures the logical integrity, personalized adaptation, and adaptive dynamic optimization of the learning path through collaborative analysis of multi-dimensional parameters, while also guaranteeing safety and compliance, thus achieving efficient, safe, and personalized path recommendation.

[0112] During practical training, students use AR glasses for interactive learning. They scan the QR code on the device to access the training course on the power metering equipment. The system also accesses the student's current learning path to generate an AR interactive course. By overlaying 3D disassembly animations and virtual operation guides, the system supports multimodal interaction via gestures and voice to achieve AR learning.

[0113] After the current stage of learning is completed, the knowledge graph and student profile are updated based on the learners' learning behavior, and the learning path is adjusted. Specific steps include:

[0114] For each trainee, a new knowledge and ability assessment, learning style analysis, and interest entity extraction were conducted to obtain the latest ability value, the latest set of knowledge weaknesses, the latest learning style, and the latest interest vector.

[0115] Statistical analysis was conducted on each knowledge weakness in the latest set of knowledge weaknesses, using the number of trainees as the unit.

[0116] If the number of students corresponding to the knowledge gap is greater than the number threshold, the association weight of the predecessor knowledge points of the knowledge gap in the knowledge graph will be adjusted.

[0117] Based on each trainee's latest ability value, latest set of knowledge weaknesses, latest learning style, and latest interest vector, the trainee profile is adjusted, and a corresponding learning path is regenerated based on the adjusted trainee profile and the adjusted knowledge graph.

[0118] In this embodiment, each trainee undergoes a new knowledge and ability assessment, learning style analysis, and interest entity extraction. The ability value, knowledge weakness set, learning style, and interest vector are updated to update the trainee profile.

[0119] Then, the knowledge weaknesses of all trainees are statistically analyzed. If the number of trainees with a particular knowledge weakness exceeds a threshold, meaning that a certain knowledge point is a weakness for the majority of trainees, the association weight of the predecessor knowledge points of that knowledge weakness in the knowledge graph is adjusted. By increasing the association weight, the associative reinforcement learning ability of that knowledge point is enhanced. Finally, based on the latest trainee profiles and the latest knowledge graph, a new learning path is generated, thereby achieving dynamic optimization and closed-loop updates of the learning path. This closed-loop feedback mechanism enables the self-optimization of the learning path, continuously improving training effectiveness.

[0120] The learning path is iteratively optimized according to the above steps, and AR-based interactive training is conducted for trainees according to the learning path until trainees master all knowledge points and complete the training course on power metering equipment.

[0121] This embodiment provides a training method for power metering equipment. This invention improves trainees' understanding of complex structures through AR interactive 3D visualization, ensures training safety, reduces equipment wear and tear, accurately portrays trainee profiles through personalized dynamic assessments, enhances the systematization of training content by constructing a hierarchical knowledge network, and improves trainees' learning efficiency through personalized learning path recommendations. This invention effectively improves the safety, efficiency, and personalization of power metering equipment training by combining immersive AR interaction with structured reasoning from knowledge graphs, along with dynamic assessments and path recommendations.

[0122] Please see Figure 2 Based on the same inventive concept, the second embodiment of the present invention proposes a power metering equipment training system, comprising:

[0123] The trainee profile construction module 10 is used to assess the knowledge and ability of each trainee, analyze their learning style, and extract their interest entities, thereby obtaining the trainee's ability value, set of knowledge weaknesses, learning style, and interest vector.

[0124] Based on the ability value, the set of knowledge weaknesses, the learning style, and the interest vector, construct a student profile corresponding to the practical training student;

[0125] The learning path generation module 20 is used to extract a learning subgraph from a pre-built knowledge graph of power metering equipment based on the student profile, and to sort the knowledge nodes of the learning subgraph using the KgRank algorithm to generate the learning path of the trainee.

[0126] AR interactive learning module 30 is used to generate AR-based interactive learning content according to the learning path, so that trainees can interactively learn the learning content.

[0127] The learning path update module 40 is used to update the knowledge graph and the student profile according to the learning behavior of the trainees in response to the completion of the interactive learning, and to adjust the learning path until the power metering equipment training ends.

[0128] The technical features and effects of the power metering equipment training system proposed in this embodiment of the invention are the same as those of the method proposed in this embodiment of the invention, and will not be repeated here. Each module in the above-mentioned power metering equipment training system can be implemented entirely or partially through software, hardware, or a combination thereof. Each module can be embedded in or independent of the processor in a computer device in hardware form, or it can be stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0129] Furthermore, embodiments of the present invention also propose a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described method.

[0130] Please see Figure 3 The diagram illustrates the internal structure of a computer device in one embodiment. This computer device can specifically be a terminal or a server. The computer device includes a processor, memory, network interface, display, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a training method for power metering equipment. The display screen of the computer device can be a liquid crystal display (LCD) or an e-ink display. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0131] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computing devices may include more or fewer components than those shown in the figure, or combine certain components, or have the same component arrangement.

[0132] Furthermore, embodiments of the present invention also propose a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0133] In summary, the present invention provides a training method, system, computer equipment, and medium for power metering equipment. The method involves assessing the knowledge and abilities of each trainee, analyzing their learning style, and extracting interest entities to obtain their ability values, set of knowledge weaknesses, learning style, and interest vectors. Based on these values, the set of knowledge weaknesses, the learning style, and the interest vectors, a trainee profile is constructed. Based on the trainee profile, a learning subgraph is extracted from a pre-constructed knowledge graph of power metering equipment. The KgRank algorithm is used to rank the knowledge nodes of the learning subgraph, generating a learning path for the trainee. AR-based interactive learning content is generated based on the learning path, enabling trainees to interactively learn the content. In response to the completion of the interactive learning, the knowledge graph and the trainee profile are updated based on the trainee's learning behavior, and the learning path is adjusted until the power metering equipment training ends. This invention enhances trainees' understanding of complex structures through AR-interactive 3D visualization, while ensuring training safety and reducing equipment wear and tear. Personalized dynamic assessments accurately portray trainee profiles, and the hierarchical knowledge network improves the systematization of training content. Furthermore, personalized learning path recommendations enhance trainees' learning efficiency. This invention effectively improves the safety, efficiency, and personalization of power metering equipment training through immersive AR interaction, structured reasoning using knowledge graphs, combined with dynamic assessment and path recommendations.

[0134] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0135] The embodiments described above are merely preferred embodiments of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various improvements and substitutions without departing from the technical principles of this invention, and these improvements and substitutions should also be considered within the scope of protection of this application. Therefore, the scope of protection of this patent application should be determined by the scope of the claims.

Claims

1. A power metering device practical training method, characterized by, include: For each trainee, knowledge and ability assessments, learning style analysis, and interest entity extraction are conducted to obtain the trainee's ability value, set of knowledge weaknesses, learning style, and interest vector. Based on the ability value, the set of knowledge weaknesses, the learning style, and the interest vector, construct a student profile corresponding to the practical training student; Based on the student profile, a learning subgraph is extracted from the pre-constructed knowledge graph of power metering equipment. The KgRank algorithm is used to calculate the dynamic weight value of each knowledge node in the learning subgraph. The dynamic weight value is then corrected based on the PageRank value of each knowledge node to obtain the corrected dynamic weight value. Based on the corrected dynamic weight value, the knowledge nodes in the learning subgraph are sorted to generate the learning path for the trainees. AR-based interactive learning content is generated according to the learning path, so that trainees can interactively learn the learning content. In response to the completion of the interactive learning, the knowledge graph and the student profile are updated based on the learning behavior of the trainees, and the learning path is adjusted until the power metering equipment training ends. The corrected dynamic weight value is represented by the following formula: In the formula, This represents the i-th knowledge node. This represents the corrected dynamic weight value of the i-th knowledge node. This represents the PageRank value of the i-th knowledge node. This represents the interest matching degree between the i-th knowledge node and the interest vector. This represents the capability value of the i-th knowledge node. This represents the initial weight of the k-th resource type corresponding to the i-th knowledge node. α represents the matching degree between the learning style and the k-th resource type, where α, β, γ, and δ all represent weighting coefficients.

2. The training method for power metering equipment according to claim 1, characterized in that, The steps of conducting knowledge and ability assessments, learning style analysis, and interest entity extraction for each trainee to obtain their ability scores, set of knowledge weaknesses, learning styles, and interest vectors include: Determine whether the trainee is taking the assessment for the first time. If so, generate initial test questions randomly based on the trainee's registration information to conduct an initial assessment. If not, generate personalized test questions based on the previous assessment results to conduct a dynamic assessment of the trainee. Based on the accuracy rate of the dynamic assessment, the trainees' ability values ​​are obtained, and the knowledge points that were not passed in the assessment are identified as knowledge weaknesses, forming a set of knowledge weaknesses. Based on a pre-set scale, the explicit learning styles of the trainees are obtained, and the explicit learning styles are modified according to the trainees' learning behaviors to obtain the trainees' learning styles. Using a pre-defined interest recognition model, interest entities are extracted from the learning behavior logs of trainees to obtain their interest vectors. The interest recognition model is constructed based on a bidirectional long short-term memory neural network and a stochastic conditional field.

3. The training method for power metering equipment according to claim 1, characterized in that, The steps of extracting a learning subgraph from a pre-constructed knowledge graph of power metering equipment based on the student profile, and using the KgRank algorithm to sort the knowledge nodes of the learning subgraph to generate the learning path for the trainees include: Based on the set of knowledge gaps, a learning subgraph is extracted from the knowledge graph of power metering equipment. Based on the ability value, determine the mastery level of each knowledge node in the learning subgraph; Based on the matching degree between the learning style and the resource type, calculate the resource type weight of each knowledge node; Calculate the interest matching degree of each knowledge node based on the similarity between the interest vector and the knowledge node; Based on the mastery level, the resource type weight, and the interest matching degree, the KgRank algorithm is used to calculate the dynamic weight value of each knowledge node; Based on the dynamic weight values, the knowledge nodes of the learning subgraph are sorted to obtain the learning path.

4. The training method for power metering equipment according to claim 1, characterized in that, The steps of updating the knowledge graph and the student profile based on the learning behavior of the trainees, and adjusting the learning path, include: For each trainee, a new knowledge and ability assessment, learning style analysis, and interest entity extraction were conducted to obtain the latest ability value, the latest set of knowledge weaknesses, the latest learning style, and the latest interest vector. Statistical analysis was conducted on each knowledge weakness in the latest set of knowledge weaknesses, using the number of trainees as the unit. If the number of students corresponding to the knowledge gap is greater than the number threshold, the association weight of the predecessor knowledge points of the knowledge gap in the knowledge graph will be adjusted. Based on each trainee's latest ability value, latest set of knowledge weaknesses, latest learning style, and latest interest vector, the trainee profile is adjusted, and a corresponding learning path is regenerated based on the adjusted trainee profile and the adjusted knowledge graph.

5. The training method for power metering equipment according to claim 1, characterized in that, The steps for constructing the knowledge graph of the power metering equipment include: Knowledge data on power metering equipment is obtained from professional model libraries and learning resource libraries, and the knowledge data is preprocessed to obtain a cleaned structured dataset. Using knowledge points, questions, students, and majors as entities and hierarchical and mapping relationships as relations, knowledge is extracted from the structured dataset to obtain a knowledge graph of power metering equipment.

6. The training method for power metering equipment according to claim 5, characterized in that, The step of extracting knowledge from the structured dataset to obtain a knowledge graph of power metering equipment includes: Entity relation tuples are extracted from the structured dataset to obtain explicit knowledge; By using a pre-defined knowledge extraction model, potential relationships between entities are extracted from historical failure cases, and the logic between knowledge points is inferred through a rule engine to obtain implicit knowledge. The knowledge extraction model is constructed based on a bidirectional long short-term memory neural network. By fusing the explicit and implicit knowledge, a knowledge graph of power metering equipment is obtained.

7. A training system for power metering equipment, characterized in that, The system is applied to the method as described in any one of claims 1 to 6, comprising: The trainee profile building module is used to assess the knowledge and ability of each trainee, analyze their learning style, and extract their interest entities, resulting in the trainee's ability value, set of knowledge weaknesses, learning style, and interest vector. Based on the ability value, the set of knowledge weaknesses, the learning style, and the interest vector, construct a student profile corresponding to the practical training student; The learning path generation module is used to extract learning subgraphs from the pre-built knowledge graph of power metering equipment based on the student profile, and to sort the knowledge nodes of the learning subgraphs using the KgRank algorithm to generate the learning path of the trainees. The AR interactive learning module is used to generate AR-based interactive learning content according to the learning path, so that trainees can interactively learn the learning content. The learning path update module is used to update the knowledge graph and the student profile based on the learning behavior of the trainees in response to the completion of the interactive learning, and to adjust the learning path until the power metering equipment training is completed.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images