Virtual classroom interaction method and system based on knowledge representation and reasoning

Abstract

The application provides a virtual classroom interaction method and system based on knowledge representation and reasoning, wherein student interaction data flow in a power system training virtual classroom is acquired, knowledge granularity decomposition processing is performed on the student interaction data flow, a knowledge tuple set containing concept nodes and relationship edges is generated, the knowledge tuple set is input into a reasoning engine to perform multi-dimensional reasoning processing, a student knowledge mastery state vector and a learning demand prediction vector are generated, interaction content generation processing is performed based on the knowledge mastery state vector and the learning demand prediction vector, classroom interaction response data with adaptability is generated, the classroom interaction response data is fed back to a virtual classroom interaction interface, and a dynamic interaction process of student learning state and classroom feedback is formed. The application effectively enhances the pertinence and practicality of virtual classroom interaction, thereby improving the learning effect of students in power system training.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of virtual classrooms, in particular to a virtual classroom interaction method and system based on knowledge representation and reasoning. BACKGROUND

[0002] With the development of virtual teaching technology, virtual classrooms have been gradually introduced into the field of power system training to improve the efficiency of knowledge transmission and skill training. Through a digital interactive environment simulating power system operation scenarios, an online learning platform is provided for students to learn knowledge, perform simulation operations, and ask questions for answers. Currently, the virtual classroom interaction method is usually based on the students' instant questions or preset course content for response, such as returning fixed knowledge points according to the students' input question text, or pushing standardized operation demonstrations according to the course progress. However, power system knowledge has strong professional characteristics (such as strict correlation between device principles and fault handling), operation specifications have strong constraints (such as following specific procedures in simulation experiments), and learning processes have strong logicality (such as the progressive relationship from device fundamentals to operation characteristics). The existing interaction method only focuses on a single dimension of information (such as question content or operation results), and cannot comprehensively analyze the students' knowledge mastery, operation standardization, and logical coherence, resulting in a mismatch between the interaction content and the students' actual learning needs, making it difficult to effectively improve the relevance and learning effect of power system training. Based on this, how to build a virtual classroom interaction method that can accurately capture students' learning state and dynamically adapt the interaction content has become a research focus in the current digital transformation of power system training. SUMMARY

[0003] The present application provides a virtual classroom interaction method and system based on knowledge representation and reasoning.

[0004] In a first aspect, the present application embodiment provides a virtual classroom interaction method based on knowledge representation and reasoning, which comprises:

[0005] Obtaining student interaction data stream in power system training virtual classroom, the student interaction data stream contains time-sequentially arranged question text, simulation operation record and historical dialogue log;

[0006] Performing knowledge granularity decomposition processing on the student interaction data stream to generate a knowledge tuple set containing concept nodes and relationship edges, the concept nodes correspond to power system basic concepts, and the relationship edges reflect the logical association between concepts and the context dependence of student interaction behavior;

[0007] Inputting the knowledge tuple set into a reasoning engine to perform multi-dimensional reasoning processing, generating a student knowledge mastery state vector and a learning demand prediction vector, the knowledge mastery state vector contains concept understanding depth, operation specification degree and logical coherence index, and the learning demand prediction vector contains content supplement direction and ability improvement direction;

[0008] performing interaction content generation processing based on the knowledge mastery state vector and the learning demand prediction vector, to generate classroom interaction response data with adaptability, the interaction response data containing knowledge completion information and ability training tasks;

[0009] feedback the classroom interaction response data to the virtual classroom interaction interface to form a dynamic interaction process of student learning state and classroom feedback.

[0010] In a second aspect, an embodiment of the present application provides a computer system, comprising:

[0011] a memory, the memory storing a computer program;

[0012] a processor for loading the computer program to implement the virtual classroom interaction method based on knowledge representation and reasoning as described above.

[0013] The virtual classroom interaction method based on knowledge representation and reasoning provided by the present application can comprehensively capture the knowledge expression, operation behavior and dialogue logic characteristics of students in the learning process by obtaining student interaction data stream (including question text, simulation operation record and historical dialogue log) in the virtual classroom of power system training. The knowledge granularity decomposition processing is performed on the student interaction data stream to generate a knowledge tuple set containing concept nodes and relationship edges, wherein the concept nodes correspond to basic concepts of the power system, and the relationship edges reflect the logical association between the concepts and the context dependence of the student interaction behavior. This processing mode deeply integrates the internal logic of the power system knowledge and the dynamic interaction characteristics of the students, forms an associated structure of knowledge content and behavior mode, and provides a structured information basis for subsequent analysis. The knowledge tuple set is input into a reasoning engine to perform multi-dimensional reasoning processing to generate a student knowledge mastery state vector and a learning demand prediction vector, wherein the knowledge mastery state vector contains concept understanding depth, operation specification degree and logical coherence index, and the learning demand prediction vector contains content supplement direction and ability improvement direction. This multi-dimensional state modeling can comprehensively depict the knowledge mastery level of students in the power system training, avoiding one-sidedness by focusing on only a single dimension. Based on the knowledge mastery state vector and the learning demand prediction vector, classroom interaction response data (containing knowledge completion information and ability training tasks) with adaptability is generated and fed back to the virtual classroom interaction interface to form a dynamic interaction process, so that the interaction content can be dynamically adjusted according to the current learning state of the students, meeting the needs of knowledge supplement and fitting the direction of ability improvement, effectively enhancing the pertinence and practicality of the virtual classroom interaction, thereby improving the learning effect of students in the power system training. BRIEF DESCRIPTION OF DRAWINGS

[0014] Figure 1is a flowchart of a virtual classroom interaction method based on knowledge representation and reasoning provided by an embodiment of the present application.

[0015] Figure 2 is a component schematic diagram of a computer system provided by an embodiment of the present application. DETAILED DESCRIPTION

[0016] Please refer to Figure 1 , Figure 1 is a flowchart of a virtual classroom interaction method based on knowledge representation and reasoning provided by an embodiment of the present application. The method can be executed by a computer system, and the steps are as follows:

[0017] Step S100: Obtain student interaction data stream in power system training virtual classroom, which contains time-sequentially arranged question text, simulation operation record and historical dialogue log.

[0018] The student interaction data stream is a collection of a series of data generated in the process of student interaction with the system in the power system training virtual classroom. These data are arranged in time sequence and can reflect various behaviors and needs of students in the learning process. The question text is the question raised by the student in the virtual classroom. These questions often revolve around the relevant knowledge of the power system, containing device name, operating parameter, fault phenomenon and other information. The simulation operation record is the record of the student's operation in the simulation environment of the virtual classroom, including device selection action, parameter adjustment action, fault simulation action, etc. It records the flow and steps of the student's actual operation, reflecting the student's practical ability and operation specification. The historical dialogue log is the record of past dialogue between the student and the system, including the student's question and the system's feedback. By analyzing these logs, the student's knowledge progression process and problem deepening trajectory can be understood.

[0019] The student interaction data stream can be obtained by setting a corresponding data collection module in the virtual classroom system. For example, for the question text, a data listening program can be set at the input box where the student asks questions. When the student inputs the question and submits, the program automatically saves the question text to the designated data storage location. For the simulation operation record, the simulation system itself will record each operation action of the student and the corresponding timestamp. After arranging these records in time sequence, the simulation operation record can be obtained. For the historical dialogue log, the system will save the content of each student question and system feedback to form a dialogue round. Arranging these dialogue rounds in time sequence will obtain the historical dialogue log.

[0020] Step S200: Perform knowledge granularity decomposition processing on the student interaction data stream to generate a knowledge tuple set containing concept nodes and relationship edges. The concept nodes correspond to the basic concepts of the power system, and the relationship edges reflect the logical association between concepts and the context dependence of student interaction behavior.

[0021] Knowledge granularity decomposition involves refining and analyzing the information in the student interaction data stream, breaking it down into smaller knowledge units and identifying the relationships between these units. Concept nodes are the basic elements in a knowledge tuple set, corresponding to fundamental concepts of the power system, such as generators, transformers, and power loads. These concepts are core components of the power system knowledge system. Relationship edges connect concept nodes, reflecting the logical connections between concepts, such as the binding relationship between devices and attributes and the triggering relationship between devices and behaviors. They also reflect the contextual dependencies of student interactions, such as the semantic progression between subsequent questions and previous questions in a conversation. The process of generating a knowledge tuple set requires comprehensive consideration of the question text, simulation operation records, and historical conversation logs in the student interaction data stream. By processing this data, key concepts and their relationships are extracted and organized into a directed graph, where nodes are concept nodes and edges are relationship edges, forming a knowledge tuple set. This knowledge tuple set more clearly demonstrates the knowledge structure and interaction logic involved in the student's learning process.

[0022] As an implementation manner, step S200 can be specifically implemented as the following steps S210 to S260:

[0023] Step S210: semantic segmentation is performed on the question text in the student interaction data stream to extract semantic segments containing power system terms, where the semantic segments include equipment names, operating parameters, and fault phenomena.

[0024] Semantic segmentation analyzes the question text and breaks it down into semantically specific segments. These segments contain key terms and information about the power system. Equipment names are the specific names of various devices in the power system, such as generators, circuit breakers, and busbars. They are the physical components of the power system. Operating parameters are indicators that describe the operating status of the equipment, such as voltage, current, and power, reflecting the equipment's working status and performance. Fault phenomena are the various abnormal conditions exhibited when equipment fails, such as short circuits, overloads, and power outages.

[0025] As an implementation manner, step S210 can be specifically implemented as the following steps S211 to S215:

[0026] Step S211: Perform word segmentation and filtering processing on the question text, and retain professional vocabulary related to the power system training objectives as candidate terms.

[0027] The word segmentation filtering process is to segment the question text by words and filter out the words related to the power system training target. Professional vocabulary is a term with specific meaning in the field of power system, such as "transformer", "reactive power", "relay protection", etc. These words are important carriers of power system knowledge. Candidate terms are professional vocabulary retained after word segmentation filtering, which are the basis for subsequent semantic analysis.

[0028] The word segmentation filtering process can use word segmentation tools such as NLTK (Natural Language Toolkit) or Jieba segmentation. Taking Jieba segmentation as an example, first input the question text into the word segmentation function of Jieba segmentation, the function will segment the text into individual words. Then, according to the pre-defined power system professional vocabulary dictionary, filter the segmented words, and only keep the words in the dictionary as candidate terms. For example, for the question text "Please ask the voltage of the transformer?", Jieba segmentation will segment it into "please ask" "transformer" "of" "voltage" "is" "how much", and after dictionary filtering, "transformer" and "voltage" are retained as candidate terms.

[0029] Step S212: Perform semantic role labeling processing on the candidate terms to identify the semantic function of each candidate term in the question, including subject function, attribute function, and behavior function, where the subject function corresponds to the device name, the attribute function corresponds to the operating parameter, and the behavior function corresponds to the protection mechanism.

[0030] Semantic role labeling processing is to determine the semantic role played by the candidate term in the question text, i.e. its semantic function. Subject function refers to the candidate term representing a specific device in the power system, such as generator, circuit breaker, etc. These devices are the basis for the operation of the power system. Attribute function refers to the candidate term describing the operating parameters of the device, such as voltage, current, frequency, etc. These parameters reflect the working state of the device. Behavior function refers to the candidate term involving the protection mechanism or operation behavior of the power system, such as relay protection, switch operation, etc. These functions ensure the safe operation of the power system.

[0031] Semantic role labeling processing can use machine learning-based methods such as Conditional Random Fields (CRF) model. First, collect a large amount of power system related text data, and manually annotate the semantic function of the candidate terms in it. Then, use these annotated data to train the CRF model. After training, input the candidate terms to be analyzed into the model, and the model will output the semantic function of each candidate term. For example, for the candidate term "transformer", the model will label its semantic function as subject function; for "voltage", the model will label its semantic function as attribute function.

[0032] Step S213: Adjacency relation identification processing is performed on the candidate terms with semantic functional association to extract direct semantic connections between terms, including device-attribute binding relations and device-behavior triggering relations.

[0033] Adjacency relation identification processing is to find the direct connections between candidate terms with semantic functional association. Device-attribute binding relation refers to the correspondence between a device and a parameter describing its operating state, for example, there is a device-attribute binding relation between "transformer" and "voltage", indicating that voltage is a parameter describing the operating state of the transformer. Device-behavior triggering relation refers to the fact that the operating state or operation of a device will trigger a specific behavior or protection mechanism, for example, "short circuit fault" triggers "relay protection action", which embodies the device-behavior triggering relation.

[0034] Adjacency relation identification processing can use rule-based methods or machine learning methods. Rule-based methods are to judge the adjacency relation between terms according to pre-defined rules. For example, define the rule "if a subject functional term is followed by an attribute functional term without other key separator words in between, then consider them as having a device-attribute binding relation". Machine learning methods can use a graph neural network (GNN) model, which takes candidate terms as nodes and the association between terms as edges to construct a graph structure data. Then, use the GNN model to learn and reason about the graph to identify the direct semantic connections between terms.

[0035] Step S214: Context verification processing is performed on the direct semantic connections to filter out effective semantic connections that conform to the power system knowledge system, combined with the overall semantic direction of the question text.

[0036] Context verification processing is a further check and filtering of the extracted direct semantic connections to ensure that they conform to the power system knowledge system and the overall semantics of the question text. Effective semantic connections are those that have practical significance in the power system knowledge system and are consistent with the semantics of the question text. For example, in a question about transformer fault analysis, the connection between "transformer" and "short circuit fault" is effective because transformers can have short circuit faults in the power system and are consistent with the semantics of the question; if an unreasonable connection such as "transformer" and "car engine fault" appears, it does not conform to the power system knowledge system and the question semantics, and needs to be filtered out.

[0037] The context verification process can be implemented by querying the power system knowledge base. The knowledge base stores various knowledge and rules of the power system, such as the normal operating parameter range of the equipment, fault types and handling methods, etc. The direct semantic connection is compared with the knowledge in the knowledge base. If the connection is consistent with the knowledge in the knowledge base and meets the overall semantic direction of the question text, it is considered that the connection is valid; otherwise, it is filtered out. For example, for the connection of "transformer" and "voltage too high", query the normal voltage range of the transformer in the knowledge base. If the voltage value mentioned in the current question exceeds this range and is consistent with the semantic of the question fault analysis, it is considered that the connection is valid.

[0038] Step S215: Combine the candidate terms and their corresponding valid semantic connections to generate a semantic fragment containing power system terms.

[0039] The semantic fragment is a knowledge unit composed of candidate terms and their valid semantic connections, which can completely express a specific concept or situation in the power system. By combining candidate terms and valid semantic connections, scattered knowledge elements can be integrated to form meaningful semantic information. For example, combining "transformer", "voltage" and the binding relationship between them as equipment and attribute, we can get the semantic fragment "voltage of transformer", which clearly expresses an operating parameter of the transformer.

[0040] Generating semantic fragments can use simple string concatenation and data structure combination methods. Arrange the candidate terms in the order of their appearance in the question text, and then add the corresponding conjunctions or symbols between the terms according to the type of valid semantic connection to form a complete semantic fragment. For example, for the equipment and attribute binding relationship of "transformer" and "voltage", the semantic fragment "voltage of transformer" can be generated. At the same time, store the semantic fragment in a suitable data structure, such as a list or dictionary, for subsequent processing and use.

[0041] Step S220: Perform behavior sequence decomposition processing on the simulation operation record to extract the corresponding relationship between operation actions and operation objects, which includes the binding relationship between equipment selection actions and target equipment, and the constraint relationship between parameter adjustment actions and parameter ranges.

[0042] The behavior sequence disassembly processing is to analyze and disassemble the operation behaviors in the simulation operation record, to decompose them into independent operation actions, and to find the corresponding relationship between the actions and the operation objects. The operation action is the specific behavior performed by the student in the simulation operation, such as selecting a device, adjusting a parameter, simulating a fault, etc. The operation object is the target of the operation action, such as a specific device, a parameter, etc. The binding relationship between the device selection action and the target device means that when the student selects a device for operation, the device is the target of the operation. The constraint relationship between the parameter adjustment action and the parameter range means that when adjusting the parameter, the adjusted value of the parameter must be within the specified range, otherwise the operation may be invalid or cause device failure.

[0043] The behavior sequence disassembly processing can be implemented by analyzing the simulation operation record line by line. First, arrange the operation record in chronological order, then parse each operation action to determine its operation type and operation object. For example, for the operation record "selected the transformer device at 10:00", the operation action can be determined as "select device" and the operation object as "transformer", establishing the binding relationship between the device selection action and the target device. For parameter adjustment actions, refer to the relevant standards of the power system and the parameter setting requirements of the device to determine the reasonable range of the parameter. For example, for the voltage adjustment of the transformer, the voltage value must be within the specified rated voltage range, and when the student performs the voltage adjustment operation, check whether the adjusted value is within this range, thereby establishing the constraint relationship between the parameter adjustment action and the parameter range.

[0044] As an implementation, step S220, the behavior sequence disassembly processing of the simulation operation record is performed to extract the corresponding relationship between the operation action and the operation object, which can be implemented as steps S221-S225:

[0045] Step S221: Time axis alignment processing is performed on the simulation operation record to generate a list of action events arranged in chronological order.

[0046] The time axis alignment processing is to sort the simulation operation record according to the operation time, so that the operation record has a clear chronological order. The action event list is a list formed by arranging each operation action in the simulation operation record as an event in chronological order. Through the time axis alignment processing, the chronological order of the student's operation can be clearly seen, which helps to analyze the logic and flow of the operation.

[0047] The time axis alignment processing can use sorting algorithms such as quicksort or merge sort. First, extract the timestamp information of each operation action from the simulation operation record. Then, use the sorting algorithm to sort these timestamps, and arrange the corresponding operation actions in the order of sorted timestamps to generate the action event list. For example, there are three operations in the simulation operation record: operation A is executed at 10:05, operation B is executed at 10:02, and operation C is executed at 10:08. After sorting, the order of the action event list is operation B, operation A, and operation C.

[0048] Step S222: Type identification processing is performed on each action event in the action event list to distinguish between device operation class actions, parameter adjustment class actions, and fault simulation class actions.

[0049] Type identification processing determines the type of each action event in the action event list. Device operation class actions refer to selecting, starting, stopping, and other operations on devices in the power system, such as selecting a transformer, starting a generator, etc., which directly affect the running state of the device. Parameter adjustment class actions refer to adjusting the operating parameters of the device, such as adjusting voltage, current, power, etc., which changes the working state of the device. Fault simulation class actions refer to simulating fault conditions in the power system, such as short circuit faults, overload faults, etc., to test students' ability to handle faults.

[0050] Type identification processing can be achieved by analyzing the description information of the action event. For example, for the action event "selected circuit breaker", it can be determined as a device operation class action; for "adjusted the voltage of the transformer to 10kV", it can be determined as a parameter adjustment class action; for "simulated a short circuit fault", it can be determined as a fault simulation class action. Machine learning classification algorithms such as support vector machines (SVM) can also be used. First, collect a large number of action event samples and manually label their types, then use these labeled data to train the SVM model, and after training, input the action event to be classified into the model, which will output its type.

[0051] Step S223: Logical dependency analysis processing is performed on adjacent action events to identify the constraint conditions of the previous action on the subsequent action, including the constraint condition that the device cannot be started to adjust the parameters, and the constraint condition that the fault cannot be cleared to execute the recovery operation.

[0052] Logical dependency analysis is to find the logical relationship between adjacent action events and determine the constraint conditions of the previous action on the subsequent action. In power system operation, many operations have a sequence and logical relationship, for example, only after the device is started can the parameter adjustment be performed, and only after the fault is cleared can the recovery operation be performed. Identifying these constraints helps to standardize the student's operation process and improve the correctness and safety of the operation.

[0053] Logical dependency analysis can be implemented by establishing a logical rule library for power system operation. The rule library stores the logical relationships and constraints between various operations. When analyzing adjacent action events, the rule library is queried to determine whether the previous action meets the execution conditions of the subsequent action. For example, if the previous action is "selected transformer but not started" and the subsequent action is "adjust the voltage of the transformer", the rule library query shows that the device cannot be parameter adjusted if it is not started, so the subsequent action does not meet the execution conditions. Knowledge graph-based methods can also be used to construct a knowledge graph of devices, operations, and rules in the power system, and logical dependency relationships can be identified by reasoning in the graph.

[0054] Step S224: Perform pattern abstraction processing on consecutive action events with the same operation target, extract repeated action combinations as operation sub-patterns, and operation sub-patterns include basic operation chains of device startup-parameter setting-operation monitoring and fault handling chains of fault triggering-phenomenon observation-troubleshooting.

[0055] Pattern abstraction processing is to find repeated action combinations with the same operation target from consecutive action events and abstract them as operation sub-patterns. Operation sub-patterns are a summary and induction of a series of operation actions, which can reflect the typical process of power system operation. The basic operation chain of device startup-parameter setting-operation monitoring is a common operation sub-pattern, which describes the basic process of device from startup to normal operation. The fault handling chain of fault triggering-phenomenon observation-troubleshooting is another important operation sub-pattern, which embodies the general steps of handling power system faults.

[0056] Pattern abstraction processing can use sequence mining algorithms such as the Apriori algorithm. First, group the action event list by operation target, then analyze the consecutive action events within each group to find repeated action combinations. For example, in a student's operation record, the action combination "start generator-set generator voltage-monitor generator power" appears multiple times, which can be extracted as an operation sub-pattern through sequence mining algorithms.

[0057] Step S225: Integrate the type information of action events, the constraint conditions of logical dependency, and the combination rules of operation sub-patterns to generate the correspondence between operation actions and operation objects.

[0058] The integration processing is to integrate the type information of action events, the constraint conditions of logical dependencies, and the combination rules of operation sub-patterns to form the corresponding relationship between operation actions and operation objects. The type information of action events clarifies the nature of the operation, the constraint conditions of logical dependencies specify the order of the operation, and the combination rules of operation sub-patterns provide the typical flow of the operation. By integrating these information, the relationship between operation actions and operation objects can be more comprehensively described.

[0059] The integration processing can use data structures to store and organize these information. For example, a dictionary can be used to store the corresponding relationship between operation actions and operation objects, where the key of the dictionary is the operation action, and the value is a list containing information such as operation objects, action types, constraint conditions, and operation sub-patterns. For example, for the operation action "adjust transformer voltage", the corresponding dictionary value may contain information such as operation object "transformer", action type "parameter adjustment class action", constraint condition "transformer has started", and related operation sub-patterns.

[0060] Step S230: Context chain construction processing is performed on the historical dialogue log to extract semantic progression relationships in continuous dialogues, which include concept extension paths and question deepening trajectories.

[0061] The context chain construction processing is to analyze the semantic association between continuous dialogues in the historical dialogue log, construct a context chain, and thus extract semantic progression relationships. Semantic progression relationships reflect the gradual deepening of knowledge and the gradual deepening of questions in the process of students' dialogue with the system. Concept extension paths refer to the expansion process of students' understanding of power system concepts from basic concepts to derived concepts in the dialogue, such as extending from the basic concept of "generator" to the speed regulation system of "generator". Question deepening trajectories refer to the development process of students' questions from surface questions to deep questions, such as from "how does a generator start" to "possible faults in the starting process of a generator and solutions".

[0062] As an implementation, step S230 can be implemented as steps S231-S236 as follows:

[0063] Step S231: Round division processing is performed on the historical dialogue log to generate a set of dialogue rounds composed of student questions and system feedback.

[0064] The round division processing is to divide the historical dialogue log according to the order of student questions and system feedback to form independent dialogue rounds. The set of dialogue rounds is a summary of these dialogue rounds, each containing student questions and system feedback, reflecting a complete interaction process.

[0065] The round division processing can be implemented by searching for separation marks in the historical dialogue log. For example, symbols or keywords can be used in the log to distinguish between student questions and system feedback, such as "Student:" indicating the start of a student question and "System:" indicating the start of system feedback. The log content is segmented according to these marks to form dialogue rounds. Natural language processing methods can also be used to determine whether a student question or system feedback is being asked by analyzing the semantics and tone of the text, thereby performing round division.

[0066] Step S232: Perform semantic vector conversion processing on each dialogue round to generate a semantic vector reflecting the core content of the round.

[0067] Semantic vector conversion processing is the conversion of the text content of a dialogue round into a vector form for subsequent similarity calculation and analysis. The semantic vector is a mathematical representation that can reflect the core content and semantic information of the dialogue round. By converting the dialogue round into a semantic vector, the semantic information of the text can be quantified for computer processing.

[0068] The semantic vector conversion processing can use a pre-trained word vector model such as Word2Vec or GloVe. First, the text of the dialogue round is segmented into individual words through word segmentation processing. Then, the pre-trained word vector model is used to convert each word into a corresponding word vector. Finally, the semantic vector of the entire dialogue round is obtained by weighted averaging or other aggregation operations on these word vectors. For example, for the dialogue round "Student: What is the voltage of the transformer? System: The rated voltage of the transformer is 10kV", after word segmentation, the words "transformer" "voltage" "how much" "rated voltage" "10kV" are obtained, and the Word2Vec model is used to convert these words into word vectors, and then the semantic vector of the dialogue round is obtained by weighted averaging.

[0069] Step S233: Calculate the cosine similarity between the semantic vectors of adjacent dialogue rounds to identify continuous rounds with close semantic association as context chains.

[0070] Cosine similarity is an index that measures the cosine of the angle between two vectors, used to judge the similarity of two vectors. In the context chain construction, by calculating the cosine similarity between the semantic vectors of adjacent dialogue rounds, continuous rounds with close semantic association can be found. The context chain is a chain composed of continuous dialogue rounds with close semantic association, reflecting the progressive process of student knowledge.

[0071] Step S234: Perform differential analysis on the semantic vectors in the context chain to extract the semantic change amount from the initial question to the subsequent question, which includes the concept range expansion amount and the problem depth increase amount.

[0072] The differential analysis process is to analyze the semantic vectors in the context chain, calculate the difference between adjacent semantic vectors, and extract the semantic change amount. The semantic change amount reflects the change of students' knowledge in the dialogue process. The concept range expansion amount refers to the increase of students' understanding range of power system concepts from the initial question to the subsequent question, such as from focusing on "generator" to focusing on "each component and working principle of generator". The problem depth increase amount refers to the change degree of students' questions from simple to complex and from surface to depth, such as from "how to operate the generator" to "safety hazards and preventive measures in the operation process of the generator".

[0073] The differential analysis process can be realized by calculating the difference between adjacent semantic vectors. For example, for two adjacent semantic vectors V1 and V2 in the context chain, calculate their difference DV=V2-V1, and then analyze the difference vector to extract the features related to concept range expansion and problem depth increase. Machine learning methods such as principal component analysis (PCA) can also be used to reduce the dimension of the difference vector and extract the main change features as the semantic change amount.

[0074] Step S235: Direction determination process is performed on the semantic change amount to determine whether the path of semantic progression conforms to the logical structure of the power system knowledge system.

[0075] The direction determination process is to judge whether the direction of semantic change amount conforms to the logical structure of the power system knowledge system. The power system knowledge system has its inherent logical relationship, such as the hierarchical relationship between concepts and the sequence of problem solving. If the path of semantic progression conforms to these logical structures, it means that the student's learning process is reasonable and effective; otherwise, there may be understanding bias or errors.

[0076] The direction determination process can be realized by querying the power system knowledge base. The knowledge base stores various knowledge and logical rules of the power system, and compares the semantic change amount with the logical structure in the knowledge base to determine whether the path of semantic progression is reasonable. For example, if the student suddenly jumps from the concept of "generator" to "lightning protection measures for transmission lines" in the dialogue without going through the relevant transition concepts, this semantic progression path may not conform to the logical structure of the power system knowledge system. Knowledge graph can also be used for direction determination, mapping the semantic change amount to the knowledge graph to see if it progresses along the logical relationship of the knowledge graph.

[0077] Step S236: The associated information of the context chain, the numerical features of the semantic change amount, and the direction of the semantic progression path are integrated to generate the semantic progression relationship in the continuous dialogue.

[0078] The integration processing is to comprehensively integrate the associated information of the context chain, the numerical characteristics of the semantic change amount, and the path direction of the semantic progression to form a complete semantic progression relationship. The associated information of the context chain reflects the semantic association degree between continuous dialogues, the numerical characteristic of the semantic change amount quantifies the change of the student's knowledge, and the path direction of the semantic progression ensures the rationality of the semantic progression.

[0079] The integration processing can use a data structure to store and organize these information. For example, a dictionary can be used to store the semantic progression relationship, where the key of the dictionary is the identification of the dialogue turn, and the value is a list containing the associated information of the context chain, the numerical characteristics of the semantic change amount, and the path direction of the semantic progression. In this way, the semantic progression relationship in continuous dialogues can be completely represented, providing a basis for subsequent analysis and application.

[0080] Step S240: Take the semantic segment, the corresponding relationship of the operation action and the operation object, and the semantic progression relationship as the basic knowledge elements.

[0081] The basic knowledge elements are the key information extracted from the student interaction data stream, including the semantic segment, the corresponding relationship of the operation action and the operation object, and the semantic progression relationship. The semantic segment reflects the student's understanding and focus of power system knowledge in the question, containing device name, operating parameter, fault phenomenon, etc. The corresponding relationship of the operation action and the operation object embodies the student's practical ability and operation specification in the simulation operation, recording the student's operation process and parameter adjustment of the device. The semantic progression relationship shows the student's gradual deepening of knowledge and deepening of problems in the process of dialogue with the system, reflecting the student's learning process and thinking development.

[0082] Taking these information as the basic knowledge elements is to enable subsequent correlation calculation and clustering and merging processing, so as to build a more complete knowledge system and provide a basis for accurately evaluating the student's knowledge mastery and predicting learning needs.

[0083] Step S250: Perform correlation calculation processing on the basic knowledge elements to identify the concept association strength between semantic segments, the content matching degree of operation actions and semantic segments, and the time sequence consistency of the semantic progression relationship and operation actions.

[0084] The correlation degree calculation process is to analyze the correlation degree between the basic knowledge elements, and to evaluate the relationship between them in a quantitative way. The concept correlation strength between semantic segments reflects the closeness between the concepts in different semantic segments, for example, the correlation strength between "transformer" and "voltage" is relatively high, because voltage is an important operating parameter of transformer. The content matching degree of operation action and semantic segment refers to the degree of correspondence between operation action and the knowledge content involved in the semantic segment, for example, the content matching degree of operation action "adjusting the voltage of the transformer" and semantic segment "voltage of the transformer" is high. The time sequence consistency of semantic progression relationship and operation action refers to whether the development of semantic progression relationship and the execution time sequence of operation action are consistent, for example, after the student gradually deepens the understanding of the fault handling knowledge of the generator in the dialogue, he / she then carries out the simulation and handling operation of the generator fault in the simulation operation, in which case the time sequence consistency of semantic progression relationship and operation action is good.

[0085] The correlation degree calculation process can be carried out in various ways. For the concept correlation strength between semantic segments, the standard correlation strength distribution between concepts in the power system knowledge base can be queried, and the matching deviation value of the concept correlation strength in the semantic segment and the standard distribution can be calculated. For the content matching degree of operation action and semantic segment, the historical execution frequency data of operation action can be analyzed, and the correlation coefficient of its content matching degree with semantic segment can be calculated. For the time sequence consistency of semantic progression relationship and operation action, the time span distribution of semantic progression relationship in the historical dialogue log can be analyzed, and the decay coefficient of its time sequence consistency and dialogue round interval can be calculated. Finally, based on these calculation results, a dynamic correlation degree calculation model is constructed to carry out weighted correction processing on the correlation degree of the basic knowledge elements, and the comprehensive correlation degree value is obtained.

[0086] As an implementation, step S250 can be implemented as steps S251-S255 as follows:

[0087] Step S251: Obtain the standard correlation strength distribution of the concept nodes in the power system knowledge system, and the standard correlation strength distribution reflects the hierarchical correlation relationship between the core concepts and the derived concepts.

[0088] The standard correlation strength distribution is a statistical distribution of the correlation strength between the concept nodes in the power system knowledge system, which reflects the hierarchical relationship between the core concepts and the derived concepts. The core concepts are the basic and key concepts of the power system knowledge system, such as generators, transformers, transmission lines, etc., which are the core support of the entire knowledge system. The derived concepts are concepts developed based on core concepts, such as the speed regulation system of the generator, the winding loss of the transformer, etc., which have close correlation with the core concepts.

[0089] The standard association strength distribution can be achieved by analyzing and counting the power system knowledge base. First, a power system knowledge graph is constructed, with concept nodes as nodes in the graph, associations between concepts as edges, and appropriate association strength values assigned to the edges. Then, statistical analysis is performed on the knowledge graph to calculate the association strength of each concept node with other nodes, obtaining the standard association strength distribution. For example, in the knowledge graph, the association strength between "generator" and "rotor of the generator" is high, while the association strength between "generator" and "insulator of the transmission line" is low. By counting these association strength values, the standard association strength distribution can be obtained.

[0090] Step S252: Calculate the matching deviation value of the concept association strength between semantic segments in the basic knowledge element and the standard association strength distribution.

[0091] The matching deviation value is an index that measures the difference between the concept association strength between semantic segments in the basic knowledge element and the standard association strength distribution. If the matching deviation value is small, it means that the concept association between semantic segments is consistent with the standard association of the knowledge system, and the student's understanding and use of concepts are accurate. If the matching deviation value is large, it means that the student's understanding may be biased or incomplete.

[0092] To calculate the matching deviation value, the concept nodes in the semantic segments in the basic knowledge element can be extracted first, and then the association strength between these concept nodes can be calculated according to the knowledge graph. Then, the calculated association strength is compared with the standard association strength distribution, and the difference between the two is calculated.

[0093] Step S253: Extract the historical execution frequency data of the operation action in the simulation operation record, and calculate the correlation coefficient between the content matching degree of the operation action and the semantic segment and the operation execution frequency.

[0094] The historical execution frequency data refers to the statistical information of the number of times each operation action is executed in a certain period of time in the simulation operation record, which reflects the commonality of the operation action and the student's familiarity with the operation. The content matching degree of the operation action and the semantic segment reflects the degree of fit between the knowledge involved in the operation action and the concepts contained in the semantic segment. The correlation coefficient is used to measure the degree of association between the content matching degree of the operation action and the semantic segment and the operation execution frequency. If the correlation coefficient is high, it means that there is a strong association between the operation execution frequency and the content matching degree, i.e., the student tends to perform the operation frequently with high content matching degree of the semantic segment; on the contrary, if the correlation coefficient is low, it means that the association between the two is weak.

[0095] The historical execution frequency data of the operation action in the simulation operation record can be obtained by data statistics on the simulation operation record. First, the operation record is classified according to the type of operation action, and then the execution times of each operation action are counted. For example, in a period of time, the number of times the student performs the "select transformer" operation is 10 times, the number of times the "adjust transformer voltage" operation is performed is 15 times, and so on. The content matching degree of the operation action and the semantic segment can be calculated by analyzing the description of the operation action and the content of the semantic segment. For example, for the operation action "adjust the voltage of the transformer" and the semantic segment "the voltage of the transformer", it can be considered that their content matching degree is high, and a high matching degree score can be given. In order to calculate the correlation coefficient, Pearson correlation coefficient formula can be used. Assuming that the content matching degree of the operation action and the semantic segment is variable X, and the operation execution frequency is variable Y, first calculate the mean value of X and Y, then calculate the difference between each data point and the mean value, then multiply these differences and sum them up, and finally divide the product of the standard deviations of X and Y, that is, the correlation coefficient can be obtained.

[0096] Step S254: Analyze the time span distribution of the semantic progression relationship in the historical dialogue log, and calculate the time sequence consistency of the semantic progression relationship and the operation action and the decay coefficient of the dialogue round interval.

[0097] The time span distribution of the semantic progression relationship refers to the distribution of the time range experienced by the semantic progression relationship from the starting point to the ending point in the historical dialogue log. It reflects the time characteristics of the student's knowledge progression, for example, some students may achieve a large knowledge progression in a short time, while some students need a long time. The time sequence consistency of the semantic progression relationship and the operation action refers to whether the development order of the semantic progression and the execution order of the operation action are consistent, and in the ideal case, the student should have a certain understanding and progression of knowledge before performing the related operation. The dialogue round interval refers to the time interval between two adjacent dialogue rounds, which may affect the time sequence consistency of the semantic progression relationship and the operation action. The decay coefficient is used to measure the degree of weakening of the time sequence consistency of the semantic progression relationship and the operation action with the increase of the dialogue round interval.

[0098] The time span distribution of semantic progression relationships in the historical dialogue log can be analyzed by timestamp analysis of the historical dialogue log. First, the starting dialogue turn and the ending dialogue turn of the semantic progression relationship are determined, and their timestamps are recorded, and then the time span is calculated. Statistical analysis of the time span of multiple semantic progression relationships is performed to obtain the distribution of the time span. The timing consistency of the semantic progression relationship and the operation action can be calculated by comparing the key time points of the semantic progression and the execution time points of the operation action. For example, if the student has a deep semantic progression of the generator fault handling knowledge in the dialogue, and then performs the generator fault handling operation in the simulation operation, it is considered that the timing consistency is good. In order to calculate the decay coefficient, a mathematical model can be established to consider the relationship between the dialogue turn interval and the timing consistency. For example, it can be assumed that the timing consistency decreases exponentially with the increase of the dialogue turn interval, and the value of the decay coefficient is determined by fitting the historical data.

[0099] Step S255: Based on the matching deviation value, the correlation coefficient, and the decay coefficient, a dynamic correlation degree calculation model is constructed to perform weighted correction processing on the correlation degree of the basic knowledge elements, and a comprehensive correlation degree value reflecting the knowledge system structure, operation behavior law, and dialogue time characteristics is generated.

[0100] The dynamic correlation degree calculation model is a mathematical model that comprehensively considers the matching deviation value, the correlation coefficient, and the decay coefficient, and is used to more accurately evaluate the correlation degree between the basic knowledge elements. The matching deviation value reflects the degree of fit between the concept association between semantic segments and the standard knowledge system, the correlation coefficient reflects the association between the operation action and the semantic segment content matching degree and the operation execution frequency, and the decay coefficient measures the influence degree of the semantic progression relationship and the timing consistency of the operation action on the dialogue turn interval. By including these factors in the model, the influence of the knowledge system structure, operation behavior law, and dialogue time characteristics on the correlation degree of the basic knowledge elements can be comprehensively reflected.

[0101] The dynamic correlation degree calculation model can be constructed using a linear weighting method. First, different weights are assigned to the matching deviation value, the correlation coefficient, and the decay coefficient, and the size of the weight is determined according to the importance of each factor in actual application. Then, the values of each factor are multiplied by the corresponding weight, and the results are added to obtain the comprehensive correlation degree value. In the calculation process, the values of each factor can be normalized to ensure the stability and accuracy of the model. For example, the matching deviation value is divided by the maximum possible deviation value, and is normalized to the range of 0 to 1. By performing weighted correction processing on the correlation degree of the basic knowledge elements through the dynamic correlation degree calculation model, a more comprehensive correlation degree value that can reflect the actual situation can be obtained, providing a more reliable basis for subsequent clustering and merging processing.

[0102] Step S260: cluster and merge the basic knowledge elements according to the correlation calculation result to generate a knowledge tuple set with the power system concept as the node and the correlation strength matching consistency as the weight.

[0103] Cluster merging combines highly correlated basic knowledge elements based on correlation calculations to form more meaningful knowledge units. A knowledge tuple set is a collection of these merged knowledge units, with power system concepts serving as nodes and correlation strength matching consistency serving as edge weights. This consistency integrates information such as the strength of conceptual associations between semantic segments, the content matching between operational actions and semantic segments, and the temporal consistency between semantic progression and operational actions, providing a more comprehensive reflection of the relationships between knowledge.

[0104] Cluster merging can be performed using a clustering algorithm, such as a hierarchical clustering algorithm. First, a similarity matrix is ​​constructed based on the correlation calculation results. The elements in the matrix represent the correlation between basic knowledge elements. Then, a hierarchical clustering algorithm is used to process the similarity matrix, gradually merging basic knowledge elements with high correlation into clusters. During this merging process, knowledge tuples are constructed using power system concepts as nodes and the consistency of correlation strength as edge weights. For example, the semantic fragment "transformer voltage" and the action "adjust transformer voltage" are highly correlated and are merged into a single knowledge tuple, with "transformer" as the node and the consistency of correlation strength as the edge weight. Through continuous merging, a set of knowledge tuples is ultimately generated, with power system concepts as nodes and the consistency of correlation strength as the weight. This set of knowledge tuples more clearly demonstrates the structure of power system knowledge and the knowledge connections during student learning.

[0105] Step S300: Input the knowledge tuple set into the inference engine to perform multi-dimensional reasoning processing to generate the student's knowledge mastery state vector and learning needs prediction vector. The knowledge mastery state vector includes the depth of concept understanding, operation standardization and logical coherence indicators, and the learning needs prediction vector includes the content supplement direction and ability improvement direction.

[0106] The reasoning engine is a system capable of analyzing and reasoning on input knowledge, which mines valuable information from the knowledge tuple set based on certain rules and algorithms. Multi-dimensional reasoning processing is to analyze the knowledge tuple set from multiple angles, considering factors such as concept understanding, operation practice and logical thinking. The student knowledge mastery state vector is a vector that comprehensively reflects the student's mastery of power system knowledge. The concept understanding depth index measures the student's understanding of power system concepts, the operation specification index evaluates the student's standard in simulation operation, and the logical coherence index examines the student's logical thinking ability in the process of dialogue and operation. The learning demand prediction vector predicts the direction of knowledge content that needs to be supplemented and the direction of ability that needs to be improved according to the student's knowledge mastery state, providing targeted guidance for subsequent teaching.

[0107] After inputting the knowledge tuple set into the reasoning engine, the reasoning engine will process the information in the knowledge tuple set. For the concept understanding depth index, the reasoning engine will analyze the coverage and correlation of the concept nodes in the knowledge tuple set to calculate the student's understanding of the power system concepts. For the operation specification index, the reasoning engine will check the matching rate of operation actions with standard operation procedures and the compliance rate of constraint conditions in simulation operation records. For the logical coherence index, the reasoning engine will evaluate the consistency of semantic progression relationship and knowledge system logical structure and the rationality of semantic change amount. Through comprehensive calculation and analysis of these indexes, the student knowledge mastery state vector is generated. At the same time, the reasoning engine will analyze the index items in the knowledge mastery state vector that are below the preset threshold to determine the direction of knowledge content that needs to be supplemented and the direction of operation link that needs to be strengthened, thereby generating the learning demand prediction vector.

[0108] As an implementation, step S300 can be implemented as steps S310-S350 as follows:

[0109] Step S310: Process the concept nodes and relationship edges in the knowledge tuple set through the concept understanding reasoning module of the reasoning engine, calculate the coverage and correlation of the student's understanding of the power system concepts, and generate the concept understanding depth index, wherein the coverage is the ratio of the number of mastered concepts to the total number of concepts, and the correlation is the ratio of the number of correct correlation relationships to the total number of correlation relationships.

[0110] The concept understanding reasoning module is a module in the reasoning engine specially used for processing concept understanding related information, which can perform in-depth analysis on concept nodes and relationship edges in the knowledge tuple set. The concept coverage degree reflects the range of concepts that the student has mastered in the power system, the number of mastered concepts refers to the number of concepts that the student has understood in the knowledge tuple set, and the total number of concepts refers to the number of all concepts involved in the power system training target. The concept correlation degree reflects the understanding degree of the student on the correlation between concepts, the number of correct correlation relationships refers to the number of correlation relationships in the knowledge tuple set that meet the power system knowledge system, and the total number of correlation relationships refers to the number of all correlation relationships in the knowledge tuple set. The concept understanding depth index comprehensively reflects the understanding degree of the student on the concept of the power system by comprehensively considering the concept coverage degree and the correlation degree.

[0111] As an implementation manner, the step S310 can be implemented as steps S311-S315.

[0112] Step S311: Extract all concept nodes in the knowledge tuple set, count the number as the number of mastered concepts, and obtain the standard concept set corresponding to the power system training target, and count the number as the total number of concepts.

[0113] The concept node is an element representing the basic concept of the power system in the knowledge tuple set, and all concept nodes can be extracted by traversing the knowledge tuple set. In the traversal process, the concept nodes in each knowledge tuple are extracted and de-duplicated to avoid repeated counting. The number of mastered concepts reflects the number of concepts that the student has contacted and understood in the learning process. The standard concept set corresponding to the power system training target is a set of concepts determined according to the training outline and teaching requirements, which covers all concepts that the student should master in the training process. The number of the standard concept set can be counted by counting the set.

[0114] For example, the knowledge tuple set contains multiple knowledge tuples, such as (“transformer”, “voltage”, correlation relationship), (“generator”, “power”, correlation relationship), etc. Through traversal, the concept nodes “transformer”, “voltage”, “generator”, “power” are extracted, and the number of the de-duplicated concept nodes is 4, which is taken as the number of mastered concepts. The standard concept set corresponding to the power system training target contains 10 concepts such as “transformer”, “voltage”, “generator”, “power”, “transmission line”, “relay protection”, etc. The number of the standard concept set is 10, which is taken as the total number of concepts.

[0115] Step S312: Calculate the ratio of the number of mastered concepts to the total number of concepts as the concept coverage degree.

[0116] Concept coverage is an index to measure the range of concepts mastered by students, which is calculated by the ratio of the number of mastered concepts to the total number of concepts. The larger the ratio, the wider the range of concepts mastered by students; otherwise, the narrower the range of concepts mastered by students.

[0117] The number of mastered concepts and the total number of concepts obtained by statistics in step S311 are used to calculate the concept coverage according to the formula: concept coverage = number of mastered concepts / total number of concepts. For example, if the number of mastered concepts is 4 and the total number of concepts is 10, the concept coverage is 4 / 10 = 0.4.

[0118] Step S313: Extract all relationship edges in the knowledge tuple set and count the number of correct association relationships conforming to the power system knowledge system.

[0119] The relationship edge is an element connecting the concept nodes in the knowledge tuple set, which reflects the logical association between concepts. Extracting all relationship edges can be achieved by traversing the knowledge tuple set and extracting the relationship edges in each knowledge tuple. The correct association relationship conforming to the power system knowledge system refers to those association relationships consistent with the actual knowledge and logic of the power system, such as the binding relationship between "transformer" and "voltage" as a device and attribute, which is a correct association relationship.

[0120] After extracting the relationship edges, it is necessary to compare them with the knowledge in the power system knowledge base to determine whether they conform to the knowledge system. For example, for the relationship edge "transformer-generates-voltage", it is known through the query of the knowledge base that the transformer is a device for changing voltage, not a device for generating voltage, so this relationship edge does not conform to the knowledge system; while for "transformer-has-voltage", it conforms to the knowledge system. Count the number of relationship edges conforming to the knowledge system to obtain the number of correct association relationships.

[0121] Step S314: Count the total number of relationship edges in the knowledge tuple set and calculate the ratio of the number of correct association relationships to the total number of relationship edges as the concept association degree.

[0122] The total number of relationship edges is the number of all relationship edges in the knowledge tuple set, which can be obtained by counting the extracted relationship edges. The concept association degree is an index to measure the degree of understanding of the association relationship between concepts by students, which is calculated by the ratio of the number of correct association relationships to the total number of relationship edges. The larger the ratio, the more accurate the understanding of the association relationship between concepts by students.

[0123] The number of correct association relationships obtained by statistics in step S313 and the total number of relationship edges obtained by statistics are used to calculate the concept association degree according to the formula: concept association degree = number of correct association relationships / total number of relationship edges.

[0124] Step S315: Integrate the concept coverage and the concept correlation to generate a concept understanding depth index.

[0125] The concept understanding depth index is a comprehensive index that reflects the students' understanding of the concept of the power system. It integrates the concept coverage and the concept correlation. The concept coverage reflects the students' mastery of the concept, and the concept correlation reflects the students' understanding of the relationship between concepts. By integrating these two indicators, the students' concept understanding level can be more comprehensively evaluated. The integration of concept coverage and concept correlation can be achieved by weighted average method.

[0126] Step S320: Process the correspondence between operation actions and operation objects in the knowledge tuple set through the operation specification reasoning module of the reasoning engine, calculate the matching rate and the constraint condition compliance rate of the operation sub-pattern and the standard operation process, and generate an operation specification degree index, wherein the matching rate is the ratio of the number of matching sub-patterns to the total number of sub-patterns, and the compliance rate is the ratio of the number of actions that meet the constraint conditions to the total number of actions.

[0127] The operation specification reasoning module is a module in the reasoning engine for processing operation specification related information, which analyzes the correspondence between operation actions and operation objects in the knowledge tuple set. The operation sub-pattern is a repeatedly occurring action combination extracted from the simulation operation record, and the standard operation process is the correct operation process specified in the power system simulation experiment. The matching rate reflects the degree of consistency between the students' operation sub-pattern and the standard operation process, the number of matching sub-patterns refers to the number of operation sub-patterns that match the sub-patterns in the standard operation process, and the total number of sub-patterns refers to the number of all operation sub-patterns in the knowledge tuple set. The constraint condition is the logical dependency relationship between operation actions, such as the device cannot be started for parameter adjustment, etc., and the compliance rate measures the degree of compliance of the students in the operation process, the number of actions that meet the constraint conditions refers to the number of actions that meet the constraint conditions in the operation record, and the total number of actions refers to the number of all operation actions in the knowledge tuple set. The operation specification degree index integrates the matching rate and the compliance rate, and can comprehensively evaluate the students' standard degree in the simulation operation.

[0128] When processing the knowledge tuple set through the operation specification reasoning module, first extract the operation sub-patterns in the knowledge tuple set, compare them with the sub-patterns in the standard operation process, count the number of matching sub-patterns and the total number of sub-patterns, calculate the matching rate according to the formula: matching rate = number of matching sub-patterns / total number of sub-patterns. At the same time, extract the logical dependency constraint conditions in the knowledge tuple set, count the number of actions that meet the constraint conditions and the total number of actions, calculate the compliance rate according to the formula: compliance rate = number of actions that meet the constraint conditions / total number of actions. Finally, integrate the matching rate and the compliance rate to generate the operation specification degree index.

[0129] As an implementation, step S320 can be implemented as steps S321-S325 as follows:

[0130] Step S321: Extract the operation sub-patterns in the knowledge tuple set, count the number as the total sub-pattern number, and obtain the standard operation process corresponding to the power system simulation experiment, extract the standard sub-patterns therein, and count the number as the matching sub-pattern number.

[0131] The operation sub-pattern is a repeated action combination with the same operation target abstracted from the simulation operation record. The operation sub-pattern can be extracted by analyzing the correspondence between the operation action and the operation object in the knowledge tuple set, and finding the repeatedly appearing action sequence. The total sub-pattern number reflects the total number of operation sub-patterns formed by the student in the simulation operation. The standard operation process corresponding to the power system simulation experiment is a verified correct operation process, which contains a series of standard sub-patterns. The standard sub-patterns in the standard operation process are extracted, and the number is counted to obtain the matching sub-pattern number.

[0132] For example, the knowledge tuple set contains operation sub-patterns "start generator-set generator voltage-monitor generator power", "start transformer-adjust transformer oil temperature-check transformer running state", etc., and the number is 5, which is counted as the total sub-pattern number. The standard operation process contains standard sub-patterns "start generator-set generator voltage-monitor generator power", "start transformer-check transformer oil level-adjust transformer voltage", etc., among which 2 match the student operation sub-patterns, and the number is counted as the matching sub-pattern number.

[0133] Step S322: Calculate the ratio of the matching sub-pattern number to the total sub-pattern number as the operation sub-pattern matching rate.

[0134] The operation sub-pattern matching rate is an index for measuring the degree of conformity of the student's operation sub-patterns to the standard operation process, which is calculated by the ratio of the matching sub-pattern number to the total sub-pattern number. The larger the ratio, the closer the student's operation sub-patterns to the standard operation process.

[0135] Using the matching sub-pattern number and the total sub-pattern number counted in step S321, the operation sub-pattern matching rate is calculated according to the formula: operation sub-pattern matching rate = matching sub-pattern number / total sub-pattern number. For example, the matching sub-pattern number is 2, the total sub-pattern number is 5, and the operation sub-pattern matching rate is 2 / 5=0.4.

[0136] Step S323: Extract the logical dependency constraint conditions in the knowledge tuple set, count the number of operation actions that meet the constraint conditions as the number of actions that meet the constraint conditions, and count the total number of action events in the knowledge tuple set.

[0137] Logical dependency constraints are logical relationships between operation actions, such as parameter adjustment cannot be performed if the device is not started, and recovery operation cannot be executed if the fault is not cleared. The logical dependency constraints in the knowledge tuple set can be extracted by analyzing the correspondence between operation actions and operation objects to find the constraint rules. The number of actions meeting the constraint conditions refers to the number of actions meeting the constraint conditions in the operation records, and the total number of action events refers to the total number of operation actions in the knowledge tuple set.

[0138] When counting the number of actions meeting the constraint conditions, each operation action needs to be checked to determine whether it meets the constraint conditions. For example, for the operation action "adjust the voltage of the transformer", it is necessary to check whether the transformer has been started. If it has been started, the action meets the constraint conditions; otherwise, it does not meet the constraint conditions. By checking all operation actions, the number of actions meeting the constraint conditions is counted. At the same time, the total number of action events is counted for all operation actions in the knowledge tuple set.

[0139] Step S324: Calculate the ratio of the number of actions meeting the constraint conditions to the total number of action events as the constraint condition compliance rate.

[0140] The constraint condition compliance rate is an index for measuring the degree to which the student complies with the logical dependency constraints between operation actions in the operation process. It is calculated by the ratio of the number of actions meeting the constraint conditions to the total number of action events. The larger the ratio, the more the student complies with the constraint conditions in the operation process.

[0141] Using the number of actions meeting the constraint conditions and the total number of action events counted in step S323, the constraint condition compliance rate is calculated according to the formula: constraint condition compliance rate = number of actions meeting constraint conditions / total number of action events.

[0142] Step S325: Integrate the operation sub-mode matching rate and the constraint condition compliance rate to generate the operation specification index.

[0143] The operation specification index integrates the operation sub-mode matching rate and the constraint condition compliance rate, and can comprehensively evaluate the specification degree of the student in the simulation operation. The operation sub-mode matching rate reflects the degree of consistency between the operation process of the student and the standard operation process, and the constraint condition compliance rate reflects the situation of the student complying with the logical constraints in the operation process. The integration of the operation sub-mode matching rate and the constraint condition compliance rate can be achieved by weighted average.

[0144] Step S330: Process the semantic progressive relationship in the knowledge tuple set through the logical coherence reasoning module of the reasoning engine, calculate the degree of consistency between the semantic progressive path and the logical structure of the knowledge system and the rationality of the semantic change, and generate a logical coherence index, where the degree of consistency is the ratio of the length of the consistent path to the total path length, and the rationality is the ratio of the number of changes that conform to the law of knowledge deepening to the total number of changes.

[0145] The logical coherence reasoning module is a module within the inference engine that processes information related to logical coherence. It analyzes the semantic progression relationships within a set of knowledge tuples. The semantic progression path is the path along which students progress through knowledge during the conversation, and the logical structure of the knowledge system represents the inherent logical relationships within the power system knowledge system. The degree of consistency reflects the degree of conformity between the semantic progression path and the logical structure of the knowledge system. The consistency path length refers to the length of the portion of the semantic progression path that aligns with the logical structure of the knowledge system, and the total path length is the total length of the semantic progression path. Semantic change refers to the amount of knowledge change from the initial question to the subsequent questions. Rationality measures whether the semantic change conforms to the laws of knowledge deepening. The number of semantic changes that conform to the laws of knowledge deepening refers to the number of semantic changes that conform to the laws of knowledge deepening, and the total number of changes refers to the total number of semantic changes in the set of knowledge tuples. The logical coherence index combines consistency and rationality to assess students' logical thinking abilities during conversation and learning.

[0146] When processing a knowledge tuple set through the logical coherence reasoning module, the semantic progressive paths within the knowledge tuple set are first extracted and compared with the standard logical structure of the power system knowledge system. The length of the matching paths and the total path length are counted, and the degree of consistency is calculated according to the formula: consistency = matching path length / total path length. Simultaneously, the semantic changes within the knowledge tuple set are extracted to determine whether they conform to the knowledge deepening rules. The number of matching changes and the total number of matching changes are counted, and the rationality is calculated according to the formula: rationality = number of matching changes / total number of matching changes. Finally, the degree of consistency and rationality are integrated to generate a logical coherence index.

[0147] As an implementation manner, step S330 may be specifically implemented as the following steps S331 to S335:

[0148] Step S331: extract the semantic progressive path in the knowledge tuple set, count its length as the total path length, obtain the standard logical structure of the power system knowledge system, and count the number of rounds in the semantic progressive path that are consistent with the standard structure as the matching path length.

[0149] The semantic progressive path is the trajectory of student knowledge progression extracted from historical conversation logs. Extracting the semantic progressive path can be achieved by analyzing the context chain and semantic change. The total path length is the total length of the semantic progressive path, which can be determined by counting the number of conversation turns contained in the path. The standard logical structure of the power system knowledge system is the inherent logical relationship of the power system knowledge, such as the hierarchical relationship of concepts, the order of problem solving, etc. Count the number of turns in the semantic progressive path that are consistent with the standard structure, that is, find the conversation turns in the semantic progressive path that conform to the logical structure of the knowledge system, and use their number as the length of the matching path.

[0150] Step S332: Calculate the ratio of the matching path length to the total path length as the logical structure matching degree.

[0151] Logical structure consistency is an indicator that measures the degree to which the semantic progression path matches the logical structure of the knowledge system. It is calculated as the ratio of the length of the matching path to the total path length. The larger the ratio, the more consistent the semantic progression path is with the logical structure of the knowledge system.

[0152] The matching path length and the total path length obtained in step S331 are used to perform calculations according to the formula: logical structure matching degree = matching path length / total path length.

[0153] Step S333: extracting the semantic changes in the knowledge tuple set, counting the number of changes that conform to the knowledge deepening law as the reasonable number of changes, and counting the total number of semantic changes in the knowledge tuple set.

[0154] Semantic change is extracted from differential analysis of semantic vectors in the context chain and reflects the evolution of students' knowledge during the conversation. Changes consistent with knowledge deepening are those in semantic change that align with the logical development and knowledge deepening requirements of the power system knowledge system, such as the expansion from simple to complex concepts and the in-depth study from superficial to deeper issues. Counting the number of changes consistent with knowledge deepening and the total semantic change can be achieved by analyzing and assessing the semantic changes in the knowledge tuple set one by one.

[0155] After extracting semantic changes, they are compared with the knowledge deepening rules in the power system knowledge base to determine whether they conform to the rules. For example, the semantic change "from focusing on the basic structure of the generator to focusing on the principles of generator fault diagnosis" conforms to the knowledge deepening rules; however, if there is no reasonable transition from focusing on the generator to suddenly focusing on the insulators of the transmission line, it does not conform to the knowledge deepening rules. By examining all semantic changes, the number of changes that conform to the rules and the total number of semantic changes are counted.

[0156] Step S334: Calculate the ratio of the number of reasonable changes to the total number of semantic changes as the semantic change reasonableness.

[0157] The semantic change reasonableness is an index for measuring whether the semantic change quantity conforms to the law of knowledge deepening. It is calculated by the ratio of the number of reasonable changes to the total number of semantic changes. The larger the ratio, the more the semantic change quantity conforms to the law of knowledge deepening.

[0158] Using the number of reasonable changes and the total number of semantic changes obtained in step S333, the semantic change reasonableness is calculated according to the formula: semantic change reasonableness = number of reasonable changes / total number of semantic changes.

[0159] Step S335: Integrate the logical structure fit degree and the semantic change reasonableness to generate the logical coherence index.

[0160] The logical coherence index integrates the logical structure fit degree and the semantic change reasonableness, and can comprehensively evaluate the logical thinking ability of students in the dialogue and learning process. The logical structure fit degree reflects the degree of consistency between the semantic progression path and the logical structure of the knowledge system, and the semantic change reasonableness reflects whether the semantic change quantity conforms to the law of knowledge deepening. The integration of the logical structure fit degree and the semantic change reasonableness can be achieved by weighted average.

[0161] Step S340: Normalize the concept understanding depth index, the operation specification index, and the logical coherence index to generate the student knowledge mastery state vector.

[0162] Normalization is to adjust the values of different indexes so that they have the same dimension and range, so as to be analyzed and compared comprehensively. The concept understanding depth index, the operation specification index, and the logical coherence index reflect the students' mastery of power system knowledge from different angles, but their numerical ranges and meanings may be different. Through normalization, these differences can be eliminated, and these indexes can be compared. The student knowledge mastery state vector is a vector that comprehensively reflects the students' knowledge mastery degree, which contains information in multiple aspects such as concept understanding, operation practice, and logical thinking.

[0163] Normalization can be achieved by various methods, such as the minimum-maximum normalization method. First, determine the minimum and maximum values of each index, then convert the values of each index according to the minimum-maximum normalization formula. After normalization, combine the concept understanding depth index, the operation specification index, and the logical coherence index into a vector to generate the student knowledge mastery state vector.

[0164] As an implementation, step S340 can be implemented as steps S341-S345 as follows:

[0165] Step S341: Calculate the first covariance value of the concept understanding depth indicator and the operation specification degree indicator, reflecting the influence degree of concept understanding on operation specification.

[0166] Covariance is a statistical measure of the linear relationship between two variables. The first covariance value measures the correlation between the concept understanding depth indicator and the operation specification degree indicator, reflecting the influence of concept understanding on operation specification. If the first covariance value is positive, it means that an increase in concept understanding depth may lead to an improvement in operation specification degree. If it is negative, there may be a negative correlation between the two. If it is close to 0, the linear relationship between the two is weak.

[0167] Step S342: Calculate the second covariance value of the operation specification degree indicator and the logical coherence indicator, reflecting the promoting effect of operation practice on logical coherence.

[0168] The second covariance value measures the correlation between the operation specification degree indicator and the logical coherence indicator, reflecting the influence of operation practice on logical coherence. Similar to the first covariance value, if the second covariance value is positive, it means that an improvement in operation specification degree may promote the enhancement of logical coherence. If it is negative, there may be a negative correlation. If it is close to 0, the linear relationship between the two is not obvious.

[0169] Step S343: Calculate the third covariance value of the logical coherence indicator and the concept understanding depth indicator, reflecting the strengthening effect of logical coherence on concept mastery.

[0170] The third covariance value measures the correlation between the logical coherence indicator and the concept understanding depth indicator, reflecting the influence of logical coherence on concept mastery. If the third covariance value is positive, it means that the enhancement of logical coherence helps to improve the concept understanding depth. If it is negative, there may be an opposite relationship. If it is close to 0, the linear correlation between the two is weak.

[0171] Step S344: Construct an index correlation matrix based on the first covariance value, the second covariance value, and the third covariance value, which is used to quantify the mutual influence weight between indicators.

[0172] The index correlation matrix is a square matrix whose elements are composed of the first covariance value, the second covariance value, and the third covariance value, which is used to quantify the mutual influence weight between the concept understanding depth indicator, the operation specification degree indicator, and the logical coherence indicator. The diagonal elements of the matrix are usually set to 1, indicating that the weight of each indicator itself is 1.

[0173] Step S345: The concept understanding depth index, the operation specification degree index, and the logical coherence index are weighted and normalized according to the index correlation matrix to generate a student knowledge mastery state vector containing the synergistic relationship between indexes, and the dimensions of the state vector dynamically adjust with the changes in the influence weight between indexes.

[0174] The weighted normalization processing is to weight each index according to the weight in the index correlation matrix on the basis of normalization. First, the concept understanding depth index, the operation specification degree index, and the logical coherence index are normalized by the minimum-maximum normalization processing to make their value range between 0 and 1. Then, the normalized indexes are weighted and summed according to the weight in the index correlation matrix. The weighted index values are combined into a vector to generate a student knowledge mastery state vector containing the synergistic relationship between indexes. Since the weight in the index correlation matrix reflects the mutual influence relationship between indexes, the dimensions of the state vector will dynamically adjust with the changes in the influence weight between indexes.

[0175] Step S350: Analyze the index items in the knowledge mastery state vector that are lower than the preset threshold to determine the knowledge content direction that needs to be supplemented and the operation link direction that needs to be strengthened to generate a learning demand prediction vector.

[0176] The preset threshold is a standard value that is set in advance to judge whether the student's knowledge mastery degree is qualified. By analyzing the index items in the knowledge mastery state vector that are lower than the preset threshold, the deficiencies of the student in knowledge mastery can be found. The knowledge content direction that needs to be supplemented refers to the case where the concept understanding depth index is low, and the knowledge content that the student needs to further learn is determined, for example, if the concept understanding of the power system fault handling in the concept understanding depth index is poor, the related fault handling knowledge needs to be supplemented. The operation link direction that needs to be strengthened refers to the case where the operation specification degree index is low, and the operation link that the student needs to strengthen training is determined, for example, if the operation specification of the device start and parameter adjustment in the operation specification degree index is poor, the training of these operation links needs to be strengthened.

[0177] The learning demand prediction vector is a vector reflecting the learning demand of the student, which contains information such as the knowledge content direction that needs to be supplemented and the operation link direction that needs to be strengthened. When analyzing the knowledge mastery state vector, each index item is compared with the preset threshold, and if a certain index item is lower than the threshold, the knowledge content or operation link corresponding to it is taken as part of the learning demand. For example, if the concept understanding depth index is lower than the threshold and the understanding of the concept of power system relay protection is insufficient, “relay protection knowledge supplement” is taken as the knowledge content direction that needs to be supplemented; if the operation specification degree index is lower than the threshold and the specification degree of the fault simulation operation is poor, “fault simulation operation strengthening” is taken as the operation link direction that needs to be strengthened. These information is combined into the learning demand prediction vector.

[0178] Step S400: Perform interactive content generation processing based on the knowledge mastery state vector and the learning demand prediction vector to generate classroom interactive response data with adaptability, which contains knowledge completion information and ability training tasks.

[0179] Interactive content generation processing is to generate classroom interactive response data suitable for students according to their knowledge mastery state and learning needs. The knowledge mastery state vector reflects the current mastery of students' knowledge of power systems, and the learning demand prediction vector indicates the knowledge content that students need to supplement and the operation links that need to be strengthened. Classroom interactive response data with adaptability refers to response data that can provide targeted knowledge and training tasks for students' specific circumstances. Knowledge completion information is related knowledge content provided to make up for the lack of students' knowledge mastery, and ability training tasks are designed to improve students' practical ability and problem-solving ability.

[0180] Based on the knowledge mastery state vector and the learning demand prediction vector, the system will extract relevant information from the corresponding knowledge base and task library. For the lower index items in the knowledge mastery state vector, according to the direction determined by the learning demand prediction vector, extract the corresponding knowledge content from the power system knowledge base, such as basic definitions, related concepts, typical application cases, etc., to generate knowledge completion information. For the operation links that need to be strengthened in the learning demand prediction vector, select appropriate training tasks from the ability training task library, such as comprehensive application questions, cross-module collaboration questions, system-level analysis questions, etc., to generate ability training tasks. Finally, integrate the knowledge completion information and the ability training tasks to generate classroom interactive response data with adaptability.

[0181] As an implementation, step S400 can be implemented as steps S410-S460 as follows:

[0182] Step S410: For the concept understanding depth index in the knowledge mastery state vector, if the index is lower than the first preset threshold, extract the basic definition, related concept and typical application case of the corresponding concept from the power system knowledge base to generate knowledge completion information.

[0183] The first preset threshold is a standard value for judging whether the concept understanding depth is qualified. When the concept understanding depth index is lower than the first preset threshold, it means that the student's understanding of some power system concepts is insufficient. The power system knowledge base is a database that stores various knowledge of power systems, which contains information such as the basic definition of the concept, the related concept and the typical application case. The basic definition is the basic explanation of the power system concept, the related concept is other concepts related to the concept, and the typical application case is the application instance of the concept in the actual power system.

[0184] When the concept understanding depth indicator is below the first preset threshold, the system determines the specific concept involved in the indicator. For example, if the concept understanding of “transformer” is low, the system extracts the basic definition of “transformer” from the power system knowledge base, such as “transformer is a device that uses electromagnetic induction principle to change alternating voltage”; associated concepts such as “winding” and “core”; typical application cases such as “in power transmission, transformer is used to raise or lower voltage to reduce power transmission loss”. These information is organized and combined to generate knowledge completion information about “transformer”.

[0185] Step S420: For the operation specification degree indicator in the knowledge mastery state vector, if the indicator is below the second preset threshold, the standard process, common errors and correction methods of the corresponding operation are extracted from the simulation operation standard library to generate operation specification reinforcement information.

[0186] The second preset threshold is a standard value for judging whether the operation specification degree is qualified. When the operation specification degree indicator is below the second preset threshold, it means that the student's specification degree in some simulation operations is not enough. The simulation operation standard library is a database that stores power system simulation operation standards and specifications, which contains the standard process, common errors and correction methods of various operations. The standard process is the correct operation steps specified in the power system simulation experiment, the common errors are the errors that students are prone to make in the operation process, and the correction methods are the solutions proposed for these errors.

[0187] When the operation specification degree indicator is below the second preset threshold, the system determines the specific operation involved in the indicator. For example, if the operation specification degree of “generator starting operation” is low, the system extracts the standard process of “generator starting operation” from the simulation operation standard library, such as “check whether the parameters of the generator are normal, close the circuit breaker, start the prime mover, and gradually increase the excitation current”; common errors such as “starting without checking parameters, and increasing excitation current too fast”; correction methods such as “carefully check all parameters before starting, and increase excitation current according to the specified speed”. These information is organized and combined to generate operation specification reinforcement information about “generator starting operation”.

[0188] Step S430: For the logical coherence indicator in the knowledge mastery state vector, if the indicator is below the third preset threshold, the logical deduction process, associated verification method and typical problem chain of the corresponding knowledge module are extracted from the knowledge system logic library to generate logical connection reinforcement information.

[0189] The third preset threshold is a preset standard value for measuring whether the logical coherence meets the standard. When the logical coherence index is lower than the threshold, it means that the student lacks in logical thinking and knowledge connection. The knowledge system logic library is a database specifically storing the logical structure and deduction rules of power system knowledge, which contains the logical deduction process of each knowledge module, the correlation verification method, and the typical problem chain. The logical deduction process shows the reasoning steps from basic concepts to complex conclusions, the correlation verification method is used to check whether the correlation between knowledge is reasonable, and the typical problem chain is a series of interrelated questions that can guide students to deeply understand the logical relationship of knowledge.

[0190] When the logical coherence index is lower than the third preset threshold, the system locates the specific knowledge module involved in the index. For example, if the logical coherence of the "power system fault analysis" knowledge module is poor, the system will extract the logical deduction process of this module from the knowledge system logic library, such as the process of deducing the fault type and location from the fault phenomenon through electrical principles and mathematical models; the correlation verification method, such as using fault recording data to verify the accuracy of fault analysis results; the typical problem chain, such as "what changes will the current and voltage have when a fault occurs? How to determine the fault type according to these changes? What are the differences in handling methods for different fault types?" etc. These information is integrated to generate the logical connection strengthening information about "power system fault analysis".

[0191] Step S440: For the content supplement direction in the learning demand prediction vector, combine the knowledge completion information, operation specification strengthening information, and logical connection strengthening information to generate content-adapted guided explanation data.

[0192] The content supplement direction in the learning demand prediction vector clearly indicates the knowledge field that the student needs to further learn. The knowledge completion information, operation specification strengthening information, and logical connection strengthening information provide resources to fill in the knowledge gaps and improve the ability of the student from different angles. The content-adapted guided explanation data is an organic combination of these information, presented to the student in a guiding way, helping them better understand and master the knowledge.

[0193] The system will filter out knowledge completion information, operation specification reinforcement information, and logic connection reinforcement information related to the content supplement direction. For example, if the content supplement direction is "power system relay protection", the system will select the basic definition, associated concepts, and typical application cases related to relay protection from the previously generated knowledge completion information; extract the standard process, common errors, and correction methods of relay protection device operation from the operation specification reinforcement information; and obtain the logical deduction process, associated verification method, and typical problem chain of the relay protection knowledge module from the logic connection reinforcement information. Then, these information is organized in a certain logical order, for example, first introduce the basic concept of relay protection, then explain the operation specification, and finally guide students to deeply understand through logical deduction and typical problem. In the organization process, guided language will be used, such as "Let's first understand the basic principles of relay protection, which is very important for subsequent operation and analysis. Next, we will see what specifications should be followed in actual operation……", so as to generate content-adapted guided explanation data.

[0194] Step S450: According to the ability improvement direction in the learning demand prediction vector, extract comprehensive application questions, cross-module cooperation questions, and system-level analysis questions that are connected with the mastered knowledge from the ability training task library, and generate ability-adapted training task data.

[0195] The ability improvement direction in the learning demand prediction vector indicates the type of ability that the student needs to focus on improving. The ability training task library is a database that stores various types of training tasks, including comprehensive application questions, cross-module cooperation questions, and system-level analysis questions of different difficulty levels and types, which are related to each knowledge module of the power system. Being connected with the mastered knowledge means that the training task can be expanded and deepened on the basis of the student's existing knowledge, helping the student to apply the learned knowledge to the solution of practical problems and improve the comprehensive ability.

[0196] According to the ability improvement direction, the system will select appropriate questions from the ability training task library. For example, if the ability improvement direction is "fault handling comprehensive ability", the system will select comprehensive application questions related to the mastered knowledge of power system fault analysis, such as "Given that there is a voltage abnormal fluctuation fault in a power system, please analyze the possible fault causes and propose appropriate handling measures"; cross-module cooperation questions, such as "In a power system containing power generation, transmission, and distribution modules, a complex fault occurs, which requires you to coordinate the knowledge of each module to diagnose and repair the fault"; and system-level analysis questions, such as "Evaluate and optimize the emergency plan for the entire power system". Organize and arrange these selected questions to generate ability-adapted training task data.

[0197] Step S460: Content integration and format adaptation processing of the guided explanation data and the training task data are performed to generate classroom interactive response data with adaptability.

[0198] Content integration is the organic integration of guided explanation data and training task data, making them content-related and complementary to each other. Format adaptation processing is the format adjustment of the integrated content to meet the interactive requirements of the virtual classroom, making it easier for students to receive and understand. Classroom interactive response data with adaptability is the response data that can accurately meet the learning needs of students after integration and adaptation.

[0199] During content integration, the system will reasonably match guided explanation data and training task data according to the logical structure of knowledge and the learning law of students. For example, after explaining the guided content of a certain knowledge module, the related training task is arranged to allow students to consolidate the learned knowledge in a timely manner. At the same time, the content will be optimized and adjusted to avoid repeated or contradictory information. In terms of format adaptation processing, the integrated content will be converted into a format suitable for virtual classroom display, such as text, pictures, and videos. If the guided explanation data contains complex principle derivation, it may be presented in the form of text and pictures; for training task data, it will be displayed in the form of a clear list of questions and answer requirements. Through content integration and format adaptation processing, classroom interactive response data with adaptability is finally generated, providing students with a coherent and clear learning experience.

[0200] Step S500: Classroom interactive response data is fed back to the virtual classroom interactive interface to form a dynamic interactive process of student learning state and classroom feedback.

[0201] The virtual classroom interactive interface is a platform for students to interact with the system. Feedback of classroom interactive response data to the interface allows students to obtain feedback information on their own learning situation in a timely manner. The dynamic interactive process of student learning state and classroom feedback is a continuous and mutually influencing process. The learning state of students is reflected through the knowledge mastery state vector and the learning demand prediction vector, and the system generates classroom interactive response data based on these information and feeds it back to the students. After receiving the feedback information, students learn and practice, and their learning state changes. The new learning state is collected and analyzed by the system, and new classroom interactive response data is generated, and the cycle continues.

[0202] When the classroom interaction response data is fed back to the virtual classroom interaction interface, the system will display the data in an intuitive and easy-to-understand manner. For example, the knowledge completion information can be presented in the form of a text box, a pop-up window, or a special topic page, and the ability training tasks can be displayed in the form of a task list, and students can directly answer and submit on the interface. At the same time, the system will provide corresponding navigation and operation instructions to facilitate students to browse and use the feedback information. After the students learn and operate in the virtual classroom, the system will collect the new interaction data of the students in real time, including new questions, simulation operation records and dialogue logs, etc., re-process the knowledge representation and reasoning, update the knowledge mastery state vector and learning demand prediction vector, and generate new classroom interaction response data, thereby forming a dynamic and continuously optimized learning cycle, so that students can gradually improve their mastery of power system knowledge and practical ability in continuous feedback and learning.

[0203] It can be understood that the various algorithms involved in the above introduction of the embodiments of the present application, such as the cosine distance algorithm, the Pearson correlation coefficient algorithm, the clustering algorithm, etc., can be known from the related content in the prior art. In order to save space, the above will not be expanded in the embodiments of the present application. In addition, those skilled in the art can supplement the details according to the common knowledge in the art when implementing the scheme of the present application. For example, according to the common knowledge in the art, the dimensional conflict before feature fusion can be eliminated by normalization, the dimension difference can be eliminated by interpolation, the threshold can be reasonably set based on historical data, experience or business scenario requirements, the model can be trained based on a general model training method, the number of layers in the model structure can be set based on actual needs, the activation function can be selected, etc. The present application will not introduce the redundant implementation process in too much detail.

[0204] Please refer to Figure 2 , Figure 2A structural schematic diagram of a computer system provided by the embodiment of the present application is shown in FIG. 1. The computer system includes at least a processor 101, a communication interface 102 and a memory 103. The processor 101, the communication interface 102 and the memory 103 can be connected through a bus or other means. The processor 101 (also referred to as a central processing unit (CPU)) is the computing core and control core of the computer system, which can parse various instructions in the computer system and process various data of the computer system. The communication interface 102 can optionally include a standard wired interface, a wireless interface (such as WI-FI, a mobile communication interface, etc.), and can be used for transmitting and receiving data under the control of the processor 101; the communication interface 102 can also be used for transmitting and interacting data within the computer system. The memory 103 is a memory device in the computer system, which is used to store programs and data. It can be understood that the memory 103 can include a built-in memory of the computer system, and of course can also include an extended memory supported by the computer system. The memory 103 provides a storage space, which stores an operating system of the computer system, and the present application does not limit this.

[0205] In one embodiment, the processor 101 executes the computer program in the memory 103 to implement the knowledge representation and reasoning-based virtual classroom interaction method provided by the above embodiment of the present application.

Claims

1. A virtual classroom interaction method based on knowledge representation and reasoning, characterized by: The method comprises: Obtaining a student interaction data stream in a power system training virtual classroom, wherein the student interaction data stream includes time-sequentially arranged question texts, simulation operation records, and historical conversation logs; The student interaction data stream is subjected to knowledge granularity decomposition processing to generate a knowledge tuple set containing concept nodes and relationship edges, wherein the concept nodes correspond to basic concepts of the power system, and the relationship edges reflect the logical association between concepts and the contextual dependence of student interaction behaviors, specifically including: semantic segmentation processing of the question text in the student interaction data stream to extract semantic segments containing power system terms, wherein the semantic segments include equipment names, operating parameters, and fault phenomena; behavioral sequence disassembly processing of the simulation operation record to extract the correspondence between the operation action and the operation object, wherein the correspondence includes the binding relationship between the equipment selection action and the target equipment, and the constraint relationship between the parameter adjustment action and the parameter range. ; Context chain construction processing is performed on the historical conversation log to extract the semantic progressive relationship in the continuous conversation, and the semantic progressive relationship includes the concept extension path and the problem deepening trajectory; the semantic fragments, the correspondence between the operation action and the operation object, and the semantic progressive relationship are used as basic knowledge elements; the basic knowledge elements are subjected to correlation calculation processing to identify the conceptual correlation strength between semantic fragments, the content matching degree between the operation action and the semantic fragment, and the temporal consistency between the semantic progressive relationship and the operation action; the basic knowledge elements are clustered and merged according to the correlation calculation result to generate a knowledge tuple set with the power system concept as the node and the correlation strength matching degree consistency as the weight; Input the knowledge tuple set into the inference engine to perform multi-dimensional inference processing to generate a student knowledge mastery state vector and a learning needs prediction vector. The knowledge mastery state vector includes indicators of concept understanding depth, operational standardization, and logical coherence, and the learning needs prediction vector includes content supplementation direction and ability improvement direction. Based on the knowledge mastery state vector and the learning demand prediction vector, interactive content generation processing is performed to generate adaptive classroom interactive response data, wherein the interactive response data includes knowledge completion information and ability training tasks; specifically, for the concept understanding depth index in the knowledge mastery state vector, if the index is lower than a first preset threshold, the basic definition, related concepts and typical application cases of the corresponding concept are extracted from the power system knowledge base to generate knowledge completion information; for the operation standardization index in the knowledge mastery state vector, if the index is lower than a second preset threshold, the standard process, common errors and correction methods of the corresponding operation are extracted from the simulation operation standard library to generate operation specification reinforcement information; for the logical coherence index in the knowledge mastery state vector, If the indicator is lower than the third preset threshold, the logical deduction process, associated verification method and typical problem chain of the corresponding knowledge module are extracted from the knowledge system logic library to generate logical connection reinforcement information; for the content supplement direction in the learning demand prediction vector, the knowledge supplement information, the operation specification reinforcement information and the logical connection reinforcement information are combined to generate content-adapted guided explanation data; for the ability improvement direction in the learning demand prediction vector, comprehensive application questions, cross-module collaboration questions and system-level analysis questions that are connected with the mastered knowledge are extracted from the ability training task library to generate ability-adapted training task data; the guided explanation data and the training task data are content-integrated and format-adapted to generate adaptive classroom interaction response data; The classroom interaction response data is fed back to the virtual classroom interaction interface to form a dynamic interaction process between students' learning status and classroom feedback.

2. The virtual classroom interaction method based on knowledge representation and reasoning according to claim 1, characterized in that: The step of performing semantic segmentation on the question text in the student interaction data stream to extract semantic segments containing power system terms includes: Performing word segmentation and filtering processing on the question text, and retaining professional vocabulary related to the power system training objectives as candidate terms; Performing semantic role labeling on the candidate terms to identify the semantic function of each candidate term in the question, wherein the semantic function includes a subject function, an attribute function, and a behavior function, wherein the subject function corresponds to a device name, the attribute function corresponds to an operating parameter, and the behavior function corresponds to a protection mechanism; Performing adjacency recognition processing on candidate terms with semantic functional associations to extract direct semantic connections between terms, including binding relationships between devices and attributes and triggering relationships between devices and behaviors; Performing context verification on the direct semantic connection, combining the overall semantic orientation of the question text, and screening out valid semantic connections that conform to the power system knowledge system; The candidate terms and their corresponding valid semantic connections are combined to generate semantic segments containing power system terms.

3. The virtual classroom interaction method based on knowledge representation and reasoning according to claim 1, characterized in that: The performing behavior sequence disassembly processing on the simulation operation record to extract the correspondence between the operation action and the operation object includes: Performing time axis alignment processing on the simulation operation record to generate an action event list arranged in the order of operation time; Performing type identification processing on each action event in the action event list to distinguish between device operation actions, parameter adjustment actions, and fault simulation actions; Perform logical dependency analysis on adjacent action events to identify constraints imposed by preceding actions on subsequent actions. These constraints include the inability to adjust parameters if the device is not started and the inability to perform recovery operations if the fault is not cleared. Perform pattern abstraction on continuous action events with the same operation target, extracting recurring action combinations as operation sub-patterns. The operation sub-patterns include a basic operation chain for equipment startup parameter setting and operation monitoring, and a fault handling chain for observing and troubleshooting fault triggering phenomena. The type information of the action event, the constraint conditions of the logical dependency, and the combination rules of the operation sub-mode are integrated to generate a corresponding relationship between the operation action and the operation object.

4. The virtual classroom interaction method based on knowledge representation and reasoning according to claim 1, characterized in that: The process of constructing a context chain on the historical conversation log to extract semantic progressive relationships in the continuous conversation includes: Dividing the historical conversation log into turns to generate a conversation turn set consisting of student questions and system feedback; Perform semantic vector conversion on each conversation turn to generate a semantic vector that reflects the core content of the turn; Calculate the cosine similarity between the semantic vectors of adjacent conversation turns and identify consecutive turns with close semantic connections as context chains; Performing differential analysis on the semantic vectors in the context chain to extract semantic changes from the initial question to subsequent questions, wherein the semantic changes include the expansion of concept scope and the increase in question depth; Performing direction determination processing on the semantic change amount to determine whether the semantic progression path conforms to the logical structure of the power system knowledge system; The association information of the context chain, the numerical characteristics of the semantic change amount, and the path direction of the semantic progression are integrated to generate a semantic progression relationship in a continuous dialogue.

5. The virtual classroom interaction method based on knowledge representation and reasoning according to claim 1, characterized in that: The step of inputting the knowledge tuple set into the inference engine to perform multi-dimensional inference processing to generate a student knowledge mastery state vector and a learning needs prediction vector includes: Processing the concept nodes and relationship edges in the knowledge tuple set through the concept understanding reasoning module of the reasoning engine, calculating the students' coverage and relevance of power system concepts, and generating a concept understanding depth index, where coverage is the ratio of the number of concepts mastered to the total number of concepts, and relevance is the ratio of the number of correct relevance relationships to the total number of relevance relationships; The operation specification reasoning module of the inference engine processes the correspondence between the operation actions and the operation objects in the knowledge tuple set, calculates the matching rate of the operation sub-pattern and the standard operation process and the constraint compliance rate, and generates an operation standardization index, wherein the matching rate is the ratio of the number of matching sub-patterns to the total number of sub-patterns, and the compliance rate is the ratio of the number of actions that meet the constraint conditions to the total number of actions; Processing the semantic progressive relationship in the knowledge tuple set through the logical coherence reasoning module of the reasoning engine, calculating the degree of consistency between the semantic progressive path and the logical structure of the knowledge system and the rationality of the semantic change, and generating a logical coherence index, wherein the degree of consistency is the ratio of the length of the consistent path to the total path length, and the rationality is the ratio of the number of changes that conform to the law of knowledge deepening to the total number of changes; Normalizing the concept understanding depth index, the operation standardization index, and the logical coherence index to generate a student knowledge mastery state vector; Analyze the index items in the knowledge mastery state vector that are lower than the preset threshold, determine the direction of knowledge content that needs to be supplemented and the direction of operation links that need to be strengthened, and generate a learning demand prediction vector.

6. The virtual classroom interaction method based on knowledge representation and reasoning according to claim 5, characterized in that: The concept understanding reasoning module of the reasoning engine processes the concept nodes and relationship edges in the knowledge tuple set and calculates the coverage and relevance of the students' power system concepts, including: Extracting all concept nodes from the knowledge tuple set, counting their number as the number of mastered concepts, and obtaining a standard concept set corresponding to the power system training objective, counting their number as the total number of concepts; Calculate the ratio of the number of concepts mastered to the total number of concepts as concept coverage; Extracting all relationship edges from the knowledge tuple set and counting the number of correct association relationships that conform to the power system knowledge system; Counting the total number of relationship edges in the knowledge tuple set, and calculating the ratio of the number of correct association relationships to the total number of relationship edges as the concept association degree; The concept coverage and the concept association are integrated to generate a concept understanding depth index.

7. The virtual classroom interaction method based on knowledge representation and reasoning according to claim 5, characterized in that: The operation specification reasoning module of the reasoning engine processes the correspondence between the operation actions and the operation objects in the knowledge tuple set, and calculates the matching rate and constraint compliance rate between the operation sub-mode and the standard operation process, including: Extracting the operation sub-patterns in the knowledge tuple set, counting the number of the operation sub-patterns as the total number of sub-patterns, and obtaining the standard operation process corresponding to the power system simulation experiment, extracting the standard sub-patterns therein, and counting the number of the standard sub-patterns as the number of matching sub-patterns; Calculating the ratio of the number of matching sub-patterns to the total number of sub-patterns as the operation sub-pattern matching rate; Extracting logical dependency constraints from the knowledge tuple set, counting the number of operation actions that meet the constraints as the number of actions that meet the constraints, and counting the total number of action events in the knowledge tuple set; Calculating the ratio of the number of actions that meet the constraint condition to the total number of action events as the constraint condition compliance rate; Integrating the operation sub-pattern matching rate and the constraint condition compliance rate to generate an operation standardization index; The processing of the semantic progressive relationship in the knowledge tuple set by the logical coherence reasoning module of the reasoning engine and the calculation of the degree of consistency between the semantic progressive path and the logical structure of the knowledge system and the rationality of the semantic change amount include: Extracting the semantic progressive path from the knowledge tuple set, counting its length as the total path length, obtaining the standard logical structure of the power system knowledge system, and counting the number of rounds in the semantic progressive path that are consistent with the standard structure as the matching path length; Calculating the ratio of the matching path length to the total path length as the logical structure matching degree; Extracting semantic changes in the knowledge tuple set, counting the number of changes that conform to the knowledge deepening law as the reasonable number of changes, and counting the total number of semantic changes in the knowledge tuple set; Calculating the ratio of the reasonable change amount to the total semantic change amount as the semantic change rationality; The logical structure consistency and the semantic change rationality are integrated to generate a logical coherence index.

8. A computer system, characterized in that: include: a memory storing a computer program; A processor, configured to load the computer program to implement the virtual classroom interaction method based on knowledge representation and reasoning according to any one of claims 1 to 7.

Images