Pump station maintenance-oriented dismounting path practical training method and system

Abstract

The invention discloses a pump station maintenance-oriented dismounting path practical training method and system, and relates to the technical field of pump station dismounting. The method comprises the following steps: constructing a virtual-real simulation environment of pump station equipment, carrying out dynamic dismounting monitoring identification on pump station parts in the pump station equipment based on the virtual-real simulation environment, and determining a dismounting path; the method comprises the following steps: acquiring operation track data of a training person in real time through a motion capture device, obtaining a somatosensory interaction motion signal, mapping the somatosensory interaction motion signal to a virtual-real simulation environment for disassembly and assembly analysis, and generating a disassembly and assembly motion deviation parameter set; and performing dynamic optimization based on the disassembly and assembly action deviation parameter set in combination with disassembly and assembly multi-dimensional feedback data, generating multi-stage practical training difficulty, performing adaptation on practical training personnel according to a disassembly and assembly path according to the multi-stage practical training difficulty, and determining a multi-stage practical training scheme. The technical problems that in the prior art, pump station maintenance dismounting practical training lacks real scene simulation, and efficient skill training is difficult to achieve are solved, and the technical effect of improving the practical training effect is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of pump station disassembly and assembly, and particularly to a disassembly and assembly path training method and system for pump station maintenance. Background Art

[0002] As an important infrastructure in the fields of water conservancy, municipal administration, industry, etc., the daily operation and regular maintenance of pump stations are of great significance for ensuring system stability. During the maintenance process of pump station equipment, a large number of complex disassembly and assembly operations are involved, which pose high requirements for the professional skills and practical experience of maintenance personnel. Traditional pump station maintenance training mainly relies on paper materials explanation or on-site guidance methods, with problems such as single training means, low simulation degree, and lagging operation feedback, making it difficult to comprehensively cover complex and variable disassembly and assembly paths and equipment structures. Especially when facing pump station equipment with multiple models and structures, trainees often have low operation efficiency and high error rates due to lack of systematic training, which in turn affects the maintenance quality and safety. Summary of the Invention

[0003] This application provides a disassembly and assembly path training method and system for pump station maintenance, which solves the technical problem that the disassembly and assembly training of pump station maintenance in the prior art lacks real-scene simulation and is difficult to achieve efficient skill training.

[0004] In the first aspect of this application, a disassembly and assembly path training method for pump station maintenance is provided. The method includes: Construct a virtual-real simulation environment of pump station equipment, dynamically monitor and label the pump station components in the pump station equipment based on the virtual-real simulation environment, and determine the disassembly and assembly path. The disassembly and assembly path includes disassembly and assembly multi-dimensional feedback data; real-time collect the operation trajectory data of the trainee through a motion capture device to obtain a somatosensory interaction action signal, map the somatosensory interaction action signal to the virtual-real simulation environment for disassembly and assembly analysis, and generate a set of disassembly and assembly action deviation parameters; perform dynamic optimization based on the set of disassembly and assembly action deviation parameters combined with the disassembly and assembly multi-dimensional feedback data to generate multiple levels of training difficulty, and adapt the trainee according to the multiple levels of training difficulty along the disassembly and assembly path to determine a multi-level training plan.

[0005] In the second aspect of this application, a disassembly and assembly path training system for pump station maintenance is provided. The system includes: A simulation environment construction module for constructing a virtual-real simulation environment of pump station equipment, dynamically disassembling, installing, monitoring and identifying pump station components in the pump station equipment based on the virtual-real simulation environment, determining a disassembly and installation path, where the disassembly and installation path includes multi-dimensional feedback data on disassembly and installation; an operation trajectory acquisition module for real-time collecting operation trajectory data of trainees through a motion capture device, obtaining a somatosensory interaction action signal, mapping the somatosensory interaction action signal to the virtual-real simulation environment for disassembly and installation analysis, and generating a set of disassembly and installation action deviation parameters; a scheme generation module for dynamically optimizing based on the set of disassembly and installation action deviation parameters combined with the multi-dimensional feedback data on disassembly and installation, generating multiple levels of training difficulty, adapting trainees according to the multiple levels of training difficulty along the disassembly and installation path, and determining a multi-level training scheme.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: Construct a virtual-real simulation environment of pump station equipment, dynamically disassemble, install, monitor and identify pump station components in the pump station equipment based on the virtual-real simulation environment, determine a disassembly and installation path, where the disassembly and installation path includes multi-dimensional feedback data on disassembly and installation. Real-time collect operation trajectory data of trainees through a motion capture device, obtain a somatosensory interaction action signal, map the somatosensory interaction action signal to the virtual-real simulation environment for disassembly and installation analysis, and generate a set of disassembly and installation action deviation parameters. Dynamically optimize based on the set of disassembly and installation action deviation parameters combined with the multi-dimensional feedback data on disassembly and installation, generate multiple levels of training difficulty, adapt trainees according to the multiple levels of training difficulty along the disassembly and installation path, and determine a multi-level training scheme. Solve the technical problem that the disassembly and installation training of pump station maintenance in the prior art lacks real scene simulation and is difficult to achieve efficient skill training, and achieve the technical effect of improving the training effect. Description of the Drawings

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0008] Figure 1 It is a schematic flow chart of a disassembly and installation path training method for pump station maintenance provided by an embodiment of this application; Figure 2 It is a schematic structural diagram of a disassembly and installation path training system for pump station maintenance provided by an embodiment of this application.

[0009] Description of the reference numerals: The simulation environment construction module 11, the operation trajectory acquisition module 12, the scheme generation module 13. Detailed Embodiments

[0010] By providing a disassembly and assembly path training method and system for pump station maintenance, the present application solves the technical problem in the prior art that the disassembly and assembly training for pump station maintenance lacks the simulation of real scenarios and it is difficult to achieve efficient skill training.

[0011] Next, the technical solutions in the embodiments of the present application will be clearly and completely described with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts belong to the scope of protection of the present application.

[0012] It should be noted that the terms "including" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product or server including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or are inherent to these processes, methods, products or devices.

[0013] Embodiment 1, as Figure 1 shown, the present application provides a disassembly and assembly path training method for pump station maintenance. Among them, the method includes: Construct a virtual-real simulation environment for pump station equipment, and based on the virtual-real simulation environment, perform dynamic disassembly and assembly monitoring and identification on the pump station components in the pump station equipment to determine the disassembly and assembly path, and the disassembly and assembly path includes disassembly and assembly multi-dimensional feedback data.

[0014] By using a three-dimensional laser scanner or high-precision photogrammetry technology, conduct a full-range scan of the pump station equipment to obtain detailed data such as its geometric dimensions, structural relationships, and component connection methods; import the scan data into three-dimensional modeling software to construct an accurate three-dimensional model of the pump station equipment, and assign physical properties such as mass, material, friction coefficient, and adaptability of disassembly and assembly tools to each component in the model. Verify the mechanical response of key components through finite element analysis to ensure consistency with the actual equipment. Then, import the three-dimensional model into a virtual reality or mixed reality platform, combine augmented reality technology, and use a sensor network to capture the state changes of the actual pump station equipment in real time and synchronize them to the virtual environment to build a virtual-real fusion simulation scenario. On this basis, establish a disassembly and assembly operation rule library based on the maintenance manual, historical disassembly and assembly records, and expert experience of the pump station equipment, clarify the disassembly and assembly sequence of each component, tool usage specifications, tightening torque requirements, and safety precautions, convert them into executable logical judgment conditions and embed them in the control system of the simulation environment. Add a dynamic identification layer to each pump station component in the virtual environment, and use a camera to identify the operation actions of trainees in real time. When a violation is detected, trigger a warning identification. At the same time, generate a disassembly and assembly path tree diagram according to the disassembly and assembly operation rule library and the real-time operations of trainees, record the timestamp, tool selection, operation torque, and component state changes of each step, and collect the mechanical data in the actual operation through a torque sensor. Combine with the path planning algorithm in the virtual environment to generate a disassembly and assembly path containing multi-dimensional feedback data such as component movement trajectory coordinates and rotation angles, operation force magnitudes, torque values, force application point positions, start time, duration, and operation interval of each step, violation operation types, occurrence locations, and corrective suggestions.

[0015] Furthermore, to construct a virtual-real simulation environment for pump station equipment, the method includes: Conduct a laser scan of the target pump station, and construct a three-dimensional structural model of the pump station equipment according to the laser scan data; retrieve historical maintenance logs to extract typical disassembly and assembly sequence data, and set disassembly and assembly node priorities and disassembly and assembly node constraints according to the typical disassembly and assembly sequence data; synchronously embed the disassembly and assembly node priorities into the three-dimensional structural model according to the disassembly and assembly node constraints to construct the virtual-real simulation environment.

[0016] Preferably, a three-dimensional laser scanner is used to scan the structure of the target pump station to obtain the spatial morphology and installation relationship data of each component of the pump station, and the original point cloud data including the boundary contour, dimension parameters and spatial position of the components is generated; based on the laser scanning point cloud data, a modeling software is used to digitally model the pump station equipment to construct a complete three-dimensional structure model, which includes multiple entity structures such as pump bodies, motors, pipelines, valves, and foundation components and their assembly levels. Subsequently, the historical maintenance log database of the pump station is called, and data mining is carried out on the past disassembly and assembly task records to extract information such as typical disassembly and assembly operation sequences, key node operation sequences, and component interlock relationships, and based on this, the priority of each disassembly and assembly node and the operation constraint logic between the nodes are set. Finally, the disassembly and assembly node priorities and their corresponding disassembly and assembly node constraints are synchronously embedded into the constructed three-dimensional structure model to generate a virtual-real simulation environment that supports interactive analysis and dynamic response, providing a basic environment support for subsequent operation path planning, action deviation analysis and training adaptation.

[0017] Furthermore, based on the virtual-real simulation environment, dynamic disassembly and assembly monitoring and identification are carried out on the pump station components in the pump station equipment to determine the disassembly and assembly path, and the disassembly and assembly path includes disassembly and assembly multi-dimensional feedback data. The method includes: The spatial relative position between the capture tool and the component is captured by a binocular camera to generate three-dimensional coordinate point cloud data; based on the virtual-real simulation environment, the typical disassembly and assembly sequence data is extracted to perform dynamic disassembly and assembly monitoring on the pump station components in the pump station equipment to obtain disassembly and assembly multi-dimensional feedback data, and the disassembly and assembly multi-dimensional feedback data includes dynamic disassembly and assembly operation data; the dynamic disassembly and assembly operation data is pre-disassembled and identified according to the disassembly and assembly node constraints to obtain multiple pre-disassembly and assembly nodes; the multiple pre-disassembly and assembly nodes are mapped to the three-dimensional coordinate point cloud data to construct multiple disassembly and assembly operation sequences, and the multiple disassembly and assembly operation sequences are associated with the disassembly and assembly node priorities for segmented marking to construct the disassembly and assembly path.

[0018] By deploying a binocular camera system, the spatial relative position relationship between the disassembly and assembly tools used by the operator and the pump station components is collected in real time, the spatial interaction process between the tool tip and the target component is obtained, and three-dimensional coordinate point cloud data with depth information is generated based on image processing technology to describe the movement trajectories and spatial positions of each interaction object during the operation. Subsequently, in combination with the preset typical disassembly and assembly sequence data in the virtual-real simulation environment, the disassembly and assembly process of the components inside the pump station equipment is dynamically monitored and analyzed to identify the matching degree between the real-time interaction behavior and the preset standards, and then multi-dimensional feedback data for disassembly and assembly, including operation postures, trajectory speeds, tool contact accuracies, etc., is extracted to form a dynamic disassembly and assembly operation data set. Then, according to the disassembly and assembly node constraint conditions set in the construction stage, the dynamic disassembly and assembly operation data is preprocessed to identify multiple feasible pre-disassembly and assembly nodes and mark them. Finally, the pre-disassembly and assembly nodes are mapped to the aforementioned three-dimensional coordinate point cloud data according to their corresponding spatial positions, and then multiple disassembly and assembly operation sequences are constructed, and segmented marking and hierarchical division are carried out according to the priorities of each node in the operation sequence, and finally a complete disassembly and assembly path with clear logic and traceable feedback is formed.

[0019] The operation trajectory data of the training personnel is collected in real time through a motion capture device to obtain a somatosensory interaction action signal, and the somatosensory interaction action signal is mapped to the virtual-real simulation environment for disassembly and assembly analysis to generate a disassembly and assembly action deviation parameter set.

[0020] By deploying a high-precision motion capture device (such as an optical motion capture system or an inertial measurement unit IMU kit), data such as joint angles, limb displacements, and tool operation trajectories of the training personnel during disassembly and assembly operations are collected in real time to form a somatosensory interaction action signal. After the somatosensory interaction action signal is preprocessed (such as filtering and denoising, coordinate system conversion), it is mapped to a pre-constructed virtual-real simulation environment to keep the tool model synchronized with the actual operations of the training personnel. In the simulation environment, the system compares the operation actions of the training personnel with the preset standard disassembly and assembly path in real time, and through computer vision algorithms and kinematic analysis, identifies the deviations in the operations (such as tool angle deviation, incorrect operation sequence, improper torque application timing, etc.) and quantifies them into a disassembly and assembly action deviation parameter set, including key indicators such as spatial position deviation amount, time synchronization deviation value, and mechanical application deviation degree. For example, when disassembling the pump body bolts, if the training personnel do not use the wrench at the standard angle (30°), the system will record the current angle deviation (such as the actual angle of 25°) and calculate the impact on the bolt stress distribution, generating a parameter set containing such deviations to provide data support for subsequent training feedback and skill assessment.

[0021] Furthermore, the operation trajectory data of the training personnel is collected in real time through a motion capture device to obtain a somatosensory interaction action signal, and the method includes: Deploy optical marker points on the hand joints of the training personnel, disassembly and assembly tools, and target components of the pumping station. The optical marker points have three-dimensional spatial coordinates. Capture the optical marker points through a motion capture device to construct three-dimensional spatial coordinates and generate relative pose data. Based on the relative pose data, perform disassembly and assembly motion analysis, construct an operation trajectory sequence, segment and mark the operation trajectory sequence, and associate it with the multiple pre-disassembly and assembly nodes to generate the somatosensory interaction action signal.

[0022] Deploy multiple optical marker points on the hand joints (such as wrist joints, metacarpophalangeal joints) of the training personnel, the disassembly and assembly tools used (such as wrenches, screwdrivers, hex keys, etc.), and the target components of the pumping station equipment. Each optical marker point has a unique code and three-dimensional spatial coordinate information, which is used to achieve high-precision spatial recognition and dynamic tracking during the motion capture process. For example, paste a marker point at each of the three interphalangeal joints of each finger, and paste two marker points at the wrist to accurately capture the rotation and translation movements of the wrist. For example, set marker points at both ends of the jaws and the end of the handle of the wrench to accurately track the spatial position and attitude changes of the tool during disassembly and assembly, ensuring that key actions such as the contact and force application between the tool and the pumping station components can be monitored. On the target components of the pumping station equipment that need to be disassembled and assembled, arrange optical marker points according to the structural characteristics and disassembly and assembly process requirements of the components. For components with regular shapes, marker points can be set at their geometric centers, edge feature points, etc.; for components with complex or irregular shapes, the number of marker points needs to be increased to cover their key structural areas, such as the edges of the inlet and outlet flanges of the pump body, the roots of the blades of the impeller, etc., so as to comprehensively reflect the spatial state changes of the components during disassembly and assembly.

[0023] Continuously capture the above optical marker points through the deployed optical motion capture device, obtain the position change data of each marker point in the three-dimensional space, and generate continuous relative pose data in combination with the time stamp. The relative pose data includes dynamic characteristic information such as the spatial angle, displacement path, and contact angle between the tool and the component. For example, determine the posture and movement direction of the hand by calculating the relative distances and angle changes between the marker points on the hand joints; judge the contact state and relative movement trend between the tool and the component by analyzing the spatial position relationship between the marker points on the tool and the marker points on the pumping station component, and finally generate relative pose data including information such as position, direction, and rotation angle.

[0024] Based on the time series, the relative pose data obtained at each moment is integrated to construct the operation trajectory sequence of the trainee's hand, tool, and the target components of the pump station during disassembly and assembly. The operation trajectory sequence describes the whole process from the tool approaching and contacting the component to completing disassembly or assembly. According to the disassembly and assembly process flow and standard operation specifications of the pump station equipment, the operation trajectory sequence is divided into multiple logical paragraphs, each paragraph corresponding to a specific disassembly or assembly operation step or subtask, and these paragraphs are marked for segmentation. For example, the operation trajectory sequence of disassembling the impeller of a centrifugal pump is divided into paragraphs such as "loosen the impeller locking nut", "remove the impeller locating pin", and "take out the impeller". At the same time, multiple pre-disassembly and assembly nodes are predefined in the 3D model of the pump station equipment, and these nodes correspond to the key positions or states during disassembly and assembly, such as the position where the bolt is tightened in place, the attitude of the component installed in place, etc. The operation trajectory sequence after segmentation marking is associated with these pre-disassembly and assembly nodes, and key operation characteristic parameters are extracted, such as operation speed, acceleration, torque magnitude (indirectly estimated through the motion state of the tool marking point), operation accuracy (deviation from the standard trajectory), etc., and these parameters are integrated into the somatosensory interaction action signal. The somatosensory interaction action signal comprehensively describes the quality, efficiency of the trainee's operation actions and the degree of compliance with the standard operation in a digital form, providing an important basis for subsequent training evaluation, feedback, and optimization.

[0025] Furthermore, mapping the somatosensory interaction action signal to the virtual-real simulation environment for disassembly and assembly analysis to generate a set of disassembly and assembly action deviation parameters, the method includes: Mapping the somatosensory interaction action signal to the virtual-real simulation environment of the pump station equipment to obtain a mapped interaction signal; performing frame-by-frame disassembly and analysis on the mapped interaction signal according to the disassembly and assembly node constraints, and setting multiple frames of disassembly and assembly interaction reference values; according to the multiple frames of disassembly and assembly interaction reference values, synchronously comparing the operation trajectory sequence with the multiple disassembly and assembly operation sequences to generate the disassembly and assembly action deviation parameters.

[0026] First, import the obtained somatosensory interaction action signals into the virtual-real simulation environment constructed by the pumping station equipment. Through coordinate system conversion and motion trajectory fitting operations, map the somatosensory interaction action signals into mapping interaction signals corresponding to the operations of the 3D model. The mapping interaction signals are used to represent the interaction process between the training personnel and the equipment components in the virtual scenario. Subsequently, according to the preset disassembly and assembly node constraint rules, perform frame-by-frame decomposition processing on the mapping interaction signals, extract the operation states at each critical moment, including tool trajectory points, hand postures, contact actions, and path offsets, etc., establish an interaction behavior stream in the form of a time series, and set multi-frame disassembly and assembly interaction reference values accordingly. The multi-frame disassembly and assembly interaction reference values reflect the trajectory accuracy, timing response, and contact behavior indicators that each action unit should achieve under standard operating conditions. Finally, according to the multi-frame disassembly and assembly interaction reference values, perform one-to-one synchronous comparison on the operation trajectory sequence generated by the training personnel and the multiple standard disassembly and assembly operation sequences defined in the virtual-real simulation environment. By calculating indicators such as trajectory offset, timing difference, and contact error, form a quantifiable set of disassembly and assembly action deviation parameters for subsequent training evaluation, difficulty adjustment, and skill feedback output.

[0027] The set of disassembly and assembly action deviation parameters includes trajectory offset error, attitude angle error, operation timing error, etc. Among them, the trajectory offset error refers to the spatial offset between the actual operation trajectory of the training personnel's operation tool or hand during the disassembly and assembly task and the standard trajectory, specifically including the Euclidean distance error of the three-dimensional coordinate points; the attitude angle error reflects the angular difference between the attitude angle (pitch angle, yaw angle, roll angle) of the hand or tool of the training personnel in space and the standard operation attitude during the component disassembly and assembly operation process; the operation timing error refers to the synchronous difference between the time axis progress of the training action and the standard action sequence.

[0028] Furthermore, according to the multi-frame disassembly and assembly interaction reference values, perform synchronous comparison on the operation trajectory sequence and the multiple disassembly and assembly operation sequences to generate the disassembly and assembly action deviation parameters. The method includes: Divide the operation trajectory sequence into multiple first sub-segments according to the disassembly and assembly nodes, and divide the multiple disassembly and assembly operation sequences into multiple second sub-segments according to the disassembly and assembly nodes; align the multiple first sub-segments and the multiple second sub-segments according to the disassembly and assembly nodes to generate multiple sub-segment alignment data groups; perform multi-segment deviation calculation on the multiple sub-segment alignment data groups according to the multi-frame disassembly and assembly interaction reference values to obtain multi-frame deviation values; if the number of frames of the multi-frame deviation values is greater than the preset frame number threshold, mark it as a continuous deviation event and add the continuous deviation event to the disassembly and assembly action deviation parameters.

[0029] Preferably, the operation trajectory sequence generated by the training personnel during the operation process is structurally processed according to the preset disassembly and assembly nodes, divided into multiple first sub-segments, and each first sub-segment corresponds to the actual operation trajectory of a specific disassembly and assembly node; at the same time, the standard disassembly and assembly operation sequence defined in the virtual-real simulation environment is also divided into multiple second sub-segments according to the same disassembly and assembly nodes, respectively representing the standard operation paths under these nodes. The multiple first sub-segments and the second sub-segments are paired and compared one by one according to the disassembly and assembly nodes to generate multiple sub-segment alignment data groups, which are used to reflect the dynamic correspondence between the training operation and the standard operation under each operation node. Based on the set multi-frame disassembly and assembly interaction reference values, multi-segment deviation calculations are performed on the multiple sub-segment alignment data groups, and the deviation amounts in dimensions such as attitude, path, contact, and timing are calculated for each frame respectively, and multi-frame deviation values corresponding to the time series are formed; the multi-frame deviation values can reflect the stability and accuracy of the training personnel's operation during the duration. If in any sub-segment, the number of consecutive frames with deviation values exceeding the reference threshold reaches or exceeds the preset frame number threshold, it is determined that a continuous deviation event has occurred in this sub-segment. The continuous deviation event indicates that there are systematic deviations or skill deficiencies in the training personnel's operation at this node, and it is added to the disassembly and assembly action deviation parameters as a key evaluation result for subsequent training difficulty adjustment, path optimization, or personalized training strategy configuration.

[0030] Based on the set of disassembly and assembly action deviation parameters and combined with the multi-dimensional feedback data of disassembly and assembly, dynamic optimization is performed to generate multiple levels of training difficulty, and according to the multiple levels of training difficulty, the training personnel are adapted according to the disassembly and assembly path to determine a multi-level training plan.

[0031] By extracting disassembly and assembly action deviation parameters such as trajectory deviation, attitude error, and operation timing offset generated by the training personnel during the disassembly and assembly process, and combining multi-dimensional feedback data such as response time, operation stability, and fault handling performance, a three-dimensional skill evaluation matrix for ability quantification is constructed; based on this matrix, the abilities of the training personnel are divided into multiple evaluation levels, and multiple levels of training difficulty corresponding to the basic level, advanced level, and expert level are generated. Subsequently, difficulty grading mapping is performed on the node tasks in the disassembly and assembly path to establish a node difficulty model; then, according to the evaluation level to which the training personnel belong, the disassembly and assembly path subsets corresponding to the corresponding level are dynamically matched to form a personalized multi-level training plan. If insufficient operation accuracy, fluctuating response efficiency, or abnormal fault handling performance is detected during the training process, the training difficulty upgrade or downgrade mechanism is triggered to achieve dynamic training adjustment to improve the training effect and adaptability.

[0032] Furthermore, based on the set of disassembly and assembly action deviation parameters and combined with the multi-dimensional feedback data of disassembly and assembly, dynamic optimization is performed to generate multiple levels of training difficulty. The method includes: Construct a three-dimensional skill evaluation matrix based on the disassembly and assembly action deviation parameter set in combination with the disassembly and assembly multi-dimensional feedback data; set multiple evaluation levels according to the disassembly and assembly action deviation parameter set through the three-dimensional skill evaluation matrix; map the disassembly and assembly path based on the multiple evaluation levels to generate the multi-level training difficulty.

[0033] First, extract the disassembly and assembly action deviation parameter set and the disassembly and assembly multi-dimensional feedback data formed during the training process. The disassembly and assembly action deviation parameter set includes indicators such as trajectory offset error, attitude angle deviation, contact error, and timing error. The disassembly and assembly multi-dimensional feedback data includes training performance information such as operation time consumption, response speed, number of repeated operations, and node completion rate. Normalize the data in the above multiple dimensions and construct a three-dimensional skill evaluation matrix in combination with the weight rules set by the system. The three-dimensional skill evaluation matrix takes three types of core ability dimensions as the axes: the operation accuracy dimension (such as trajectory accuracy and contact deviation), the timeliness response dimension (such as operation duration and response delay), and the fault handling dimension (such as node completion efficiency and error correction ability), which is used to comprehensively evaluate the disassembly and assembly ability performance of the training personnel. Subsequently, based on the scoring results of each dimension in the three-dimensional skill evaluation matrix, combined with the key performance fluctuation sections recorded in the disassembly and assembly action deviation parameter set, quantitatively evaluate the comprehensive ability level of the training personnel, and divide the evaluation results into multiple evaluation levels according to the set grading standards. For example, the basic level, the advanced level, and the expert level. Among them, the basic level evaluation conditions include loose node constraint conditions, large tolerance ranges, only including the standard disassembly and assembly process of a single path, and all fault points are explicitly prompted; on the basis of the basic level, the advanced level evaluation conditions add hidden path branch selection, shorten the operation time window, and introduce a single fault concurrent disturbance scenario to increase the training complexity; the expert level evaluation conditions further enable strict tolerance thresholds, construct a simulation scenario with multiple faults intertwined, and at the same time integrate dynamically changing physical simulation parameters (such as gravity change, resistance simulation, inertia delay, etc.) to evaluate the response ability and operation stability of the training personnel under complex dynamic conditions. Finally, based on the above multiple evaluation levels, perform logical mapping and structural adjustment on the constructed disassembly and assembly path, that is, map the ability requirements at each level to the node structure and interaction rules in the disassembly and assembly path, and complete the dynamic configuration of path depth, operation freedom, prompt strategy, and task tolerance, so as to generate a multi-level training difficulty model suitable for training personnel of different ability levels, which is used as the core basis for subsequent training plan adaptation.

[0034] Furthermore, constructing a three-dimensional skill evaluation matrix based on the disassembly and assembly action deviation parameter set in combination with the disassembly and assembly multi-dimensional feedback data, the method includes: Based on the disassembly and assembly action deviation parameters, conduct operation accuracy analysis to determine the operation accuracy index; based on the disassembly and assembly multi-dimensional feedback data, conduct response recording to generate the response speed index; construct a disassembly and assembly fault knowledge base, and based on the disassembly and assembly fault knowledge base, conduct processing analysis on the pump station equipment to formulate the fault handling ability index; normalize the operation accuracy index, the response speed index, and the fault handling ability index to construct the three-dimensional skill evaluation matrix.

[0035] For the disassembly and assembly action deviation parameters recorded during the training process, carry out operation accuracy analysis, calculate the average error values and fluctuation ranges in dimensions such as trajectory deviation, attitude angle error, and contact position deviation generated by the training personnel at each disassembly and assembly node, and comprehensively form the operation accuracy index, which is used to measure the precise control ability of the training personnel when executing the standard path. Secondly, based on the disassembly and assembly multi-dimensional feedback data, extract and statistically analyze information such as the response duration, task completion time, and delay between actions for each operation node during the training process, and construct the response speed index, which is used to evaluate the time control ability and execution efficiency of the training personnel during task advancement. Further, construct a set of disassembly and assembly fault knowledge bases, which record common pump station equipment fault types, fault manifestation characteristics, corresponding standard handling strategies and response paths; during the training process, compare and analyze the response behaviors of the training personnel at multiple typical fault points according to the handling steps, judgment accuracy rate, and repair completion degree when the training personnel face simulated fault situations, and quantify their problem recognition and handling abilities, so as to form the fault handling ability index. Finally, normalize the index values of the above three dimensions, namely the operation accuracy index, the response speed index, and the fault handling ability index, eliminate the dimension difference and unify the scoring interval, and integrate them through a matrix structure to construct the three-dimensional skill evaluation matrix.

[0036] Furthermore, adapt the training personnel according to the multi-level training difficulty according to the disassembly and assembly path to determine the multi-level training plan. The method includes: Analyze the multi-level training difficulty to determine the basic training difficulty, advanced training difficulty, and expert training difficulty; traverse the disassembly and assembly path to match the basic training difficulty, the advanced training difficulty, and the expert training difficulty to obtain multiple disassembly and assembly node difficulty data; extract the multiple disassembly and assembly node difficulty data for the basic training difficulty, and allocate a basic training plan for the training personnel according to the basic training difficulty. If the continuous disassembly and assembly accuracy fails to meet the standard during the execution of the basic training plan, trigger a difficulty upgrade operation to upgrade the basic training difficulty to the advanced training difficulty; extract the multiple disassembly and assembly node difficulty data for the advanced training difficulty, and allocate an advanced training plan for the training personnel according to the advanced training difficulty. During the execution of the advanced training plan, extract the fault repair efficiency, and trigger a difficulty upgrade operation according to the fault repair efficiency to upgrade the advanced training difficulty to the expert training difficulty; extract the multiple disassembly and assembly node difficulty data for the expert training difficulty, and allocate an expert training plan for the training personnel according to the expert training difficulty. During the execution of the expert training plan, perform disassembly and assembly accuracy detection. If there is an upward trend in the accuracy index, trigger an automatic downgrade operation.

[0037] Preferably, analyze the multi-level training difficulty generated according to the three-dimensional skill evaluation matrix, identify and divide it into three levels: basic training difficulty, advanced training difficulty, and expert training difficulty. Each level corresponds to different evaluation index ranges and simulation interaction complexities. The basic difficulty focuses on standard path training and explicit prompt guidance. The advanced difficulty introduces time constraints, path branches, and single-fault perturbations. The expert difficulty incorporates multi-fault scenarios and dynamic perturbation factors. Then, traverse and analyze the constructed disassembly and assembly path, and combine the multi-level training difficulty levels to match each node in the path to obtain disassembly and assembly node difficulty data covering multiple difficulty levels. The node difficulty data is used to indicate the operation complexity and interaction requirements of each disassembly and assembly task unit at different training levels.

[0038] For the training personnel at the basic training difficulty level, extract the node subset applicable to the basic difficulty from the disassembly and assembly node difficulty data, generate the corresponding basic training plan, and deploy it to the simulation system for the training personnel to operate. If during the execution of the basic training plan, it is detected that the training personnel have continuous disassembly and assembly accuracy that fails to meet the standard at multiple consecutive nodes, trigger a difficulty upgrade mechanism to raise the current difficulty level to the advanced training difficulty.

[0039] For the personnel at the advanced training difficulty level, the system will allocate an advanced training plan that includes hidden path branches, shortened time windows, and single-fault perturbations. During the training execution, the system extracts the fault repair efficiency in real time, including indicators such as fault identification delay, processing accuracy rate, and repair success rate. If the fault repair efficiency is better than the set threshold, trigger the upgrade mechanism again and transfer to the expert training difficulty.

[0040] For the training personnel who enter the expert training difficulty level, select a node group containing intertwined multiple faults, extremely strict tolerance requirements, and dynamic interference factors from the node difficulty data to construct an expert training plan. During the execution of the expert plan, the system continuously detects the changing trend of the disassembly and assembly accuracy index of the training personnel. If it detects that the accuracy index shows a downward trend or the fluctuation increases, it triggers an automatic downgrading operation to ensure the training quality and safety.

[0041] In summary, the embodiments of the present application have at least the following technical effects: Construct a virtual-real simulation environment for pumping station equipment, and based on the virtual-real simulation environment, perform dynamic disassembly and assembly monitoring and identification on the pumping station components in the pumping station equipment to determine the disassembly and assembly path, and the disassembly and assembly path includes disassembly and assembly multi-dimensional feedback data. Real-time collect the operation trajectory data of the training personnel through an action capture device to obtain a somatosensory interaction action signal, map the somatosensory interaction action signal to the virtual-real simulation environment for disassembly and assembly analysis, and generate a set of disassembly and assembly action deviation parameters. Based on the set of disassembly and assembly action deviation parameters combined with the disassembly and assembly multi-dimensional feedback data for dynamic optimization, generate multiple levels of training difficulty, and adapt the training personnel according to the multiple levels of training difficulty according to the disassembly and assembly path to determine a multi-level training plan. It solves the technical problem that the disassembly and assembly training for pumping station maintenance in the prior art lacks real-scene simulation and is difficult to achieve efficient skill training, and achieves the technical effect of improving the training effect.

[0042] Embodiment 2, based on the same inventive concept as the disassembly and assembly path training method for pumping station maintenance in the foregoing embodiment, as Figure 2 shown, the present application provides a disassembly and assembly path training system for pumping station maintenance, wherein the system includes: A simulation environment construction module 11, configured to construct a virtual-real simulation environment for pumping station equipment, perform dynamic disassembly and assembly monitoring and identification on the pumping station components in the pumping station equipment based on the virtual-real simulation environment, and determine the disassembly and assembly path, and the disassembly and assembly path includes disassembly and assembly multi-dimensional feedback data; an operation trajectory acquisition module 12, configured to real-time collect the operation trajectory data of the training personnel through an action capture device to obtain a somatosensory interaction action signal, map the somatosensory interaction action signal to the virtual-real simulation environment for disassembly and assembly analysis, and generate a set of disassembly and assembly action deviation parameters; a plan generation module 13, configured to perform dynamic optimization based on the set of disassembly and assembly action deviation parameters combined with the disassembly and assembly multi-dimensional feedback data, generate multiple levels of training difficulty, and adapt the training personnel according to the multiple levels of training difficulty according to the disassembly and assembly path to determine a multi-level training plan.

[0043] Further, the simulation environment construction module 11 is used to execute the following method: By performing laser scanning on the target pumping station, a three-dimensional structural model of the pumping station equipment is constructed based on the laser scanning data; historical maintenance logs are retrieved to extract typical disassembly and assembly sequence data, and disassembly and assembly node priorities and disassembly and assembly node constraints are set according to the typical disassembly and assembly sequence data; the disassembly and assembly node priorities are synchronously embedded into the three-dimensional structural model according to the disassembly and assembly node constraints to construct the virtual-real simulation environment.

[0044] Further, the simulation environment construction module 11 is used to execute the following method: The spatial relative positions of the capture tool and the component are captured by a binocular camera to generate three-dimensional coordinate point cloud data; based on the virtual-real simulation environment, the typical disassembly and assembly sequence data is extracted to perform dynamic disassembly and assembly monitoring on the pumping station components in the pumping station equipment, and disassembly and assembly multi-dimensional feedback data is obtained, where the disassembly and assembly multi-dimensional feedback data includes dynamic disassembly and assembly operation data; the dynamic disassembly and assembly operation data is pre-disassembled and assembled marked according to the disassembly and assembly node constraints to obtain a plurality of pre-disassembly and assembly nodes; the plurality of pre-disassembly and assembly nodes are mapped to the three-dimensional coordinate point cloud data to construct a plurality of disassembly and assembly operation sequences, and the plurality of disassembly and assembly operation sequences are associated with the disassembly and assembly node priorities for segmented marking to construct the disassembly and assembly path.

[0045] Further, the operation trajectory acquisition module 12 is used to execute the following method: Optical marker points are deployed on the hand joints of the training personnel, the disassembly and assembly tools, and the pumping station target components, and the optical marker points have three-dimensional spatial coordinates; the optical marker points are captured by a motion capture device to construct three-dimensional spatial coordinates and generate relative pose data; based on the relative pose data, disassembly and assembly motion analysis is performed to construct an operation trajectory sequence, and the operation trajectory sequence is segmented and marked and associated with the plurality of pre-disassembly and assembly nodes to generate the somatosensory interaction action signal.

[0046] Further, the operation trajectory acquisition module 12 is used to execute the following method: The somatosensory interaction action signal is mapped to the virtual-real simulation environment of the pumping station equipment to obtain a mapped interaction signal; the mapped interaction signal is disassembled and analyzed frame by frame according to the disassembly and assembly node constraints to set multi-frame disassembly and assembly interaction reference values; according to the multi-frame disassembly and assembly interaction reference values, the operation trajectory sequence and the plurality of disassembly and assembly operation sequences are synchronously compared to generate the disassembly and assembly action deviation parameters.

[0047] Further, the operation trajectory acquisition module 12 is used to execute the following method: Divide the operation trajectory sequence into multiple first sub - segments according to disassembly and assembly nodes, and divide the multiple disassembly and assembly operation sequences into multiple second sub - segments; align the multiple first sub - segments and the multiple second sub - segments according to disassembly and assembly nodes to generate multiple sub - segment alignment data groups; perform multi - segment deviation calculation on the multiple sub - segment alignment data groups according to the multi - frame disassembly and assembly interaction reference value to obtain a multi - frame deviation value; if the number of frames of the multi - frame deviation value is greater than a preset frame number threshold, mark it as a continuous deviation event and add the continuous deviation event to the disassembly and assembly action deviation parameter.

[0048] Further, the solution generation module 13 is used to execute the following method: Construct a three - dimensional skill evaluation matrix based on the disassembly and assembly action deviation parameter set in combination with the disassembly and assembly multi - dimensional feedback data; set multiple evaluation levels according to the disassembly and assembly action deviation parameter set through the three - dimensional skill evaluation matrix; map the disassembly and assembly path based on the multiple evaluation levels to generate the multi - level training difficulty.

[0049] Further, the solution generation module 13 is used to execute the following method: Perform operation accuracy analysis based on the disassembly and assembly action deviation parameter to determine the operation accuracy index; perform response record based on the disassembly and assembly multi - dimensional feedback data to generate the response speed index; construct a disassembly and assembly fault knowledge base, perform processing analysis on the pumping station equipment based on the disassembly and assembly fault knowledge base, and formulate a fault handling ability index; perform normalization processing on the operation accuracy index, the response speed index, and the fault handling ability index to construct the three - dimensional skill evaluation matrix.

[0050] Further, the solution generation module 13 is used to execute the following method: Analyze the multi-level training difficulty to determine the basic training difficulty, advanced training difficulty, and expert training difficulty; traverse the disassembly and assembly path to match the basic training difficulty, advanced training difficulty, and expert training difficulty to obtain multiple disassembly and assembly node difficulty data; extract the multiple disassembly and assembly node difficulty data for the basic training difficulty, and allocate a basic training plan for the training personnel according to the basic training difficulty. If the continuous disassembly and assembly accuracy fails to meet the standard when executing the basic training plan, trigger a difficulty upgrade operation to upgrade the basic training difficulty to the advanced training difficulty; extract the multiple disassembly and assembly node difficulty data for the advanced training difficulty, allocate an advanced training plan for the training personnel according to the advanced training difficulty, execute the advanced training plan to extract the fault repair efficiency, and trigger a difficulty upgrade operation according to the fault repair efficiency to upgrade the advanced training difficulty to the expert training difficulty; extract the multiple disassembly and assembly node difficulty data for the expert training difficulty, allocate an expert training plan for the training personnel according to the expert training difficulty, execute the expert training plan to perform disassembly and assembly accuracy detection, and if there is an upward trend in the accuracy index, trigger an automatic downgrade operation.

[0051] It should be noted that the above-mentioned order of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above describes specific embodiments of this specification. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0052] The above are only the preferred embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0053] This specification and the drawings are only exemplary descriptions of the present application and are considered to have covered any and all modifications, variations, combinations, or equivalents within the scope of the present application. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalent technologies, the present application is intended to include these changes and modifications.

Claims

1. A disassembly and assembly path training method for pump station maintenance, characterized in that: The method comprises: Constructing a virtual-reality simulation environment for pump station equipment, and dynamically disassembling and monitoring the pump station components in the pump station equipment based on the virtual-reality simulation environment to determine a disassembly and assembly path, wherein the disassembly and assembly path includes disassembly and assembly multi-dimensional feedback data; The operation trajectory data of the trainees are collected in real time by a motion capture device to obtain a somatosensory interaction action signal, the somatosensory interaction action signal is mapped to the virtual-reality simulation environment for disassembly and assembly analysis, and a disassembly and assembly action deviation parameter set is generated; Based on the disassembly and assembly action deviation parameter set combined with the disassembly and assembly multi-dimensional feedback data, dynamic optimization is performed to generate multi-level training difficulty, and trainees are adapted according to the disassembly and assembly path according to the multi-level training difficulty to determine a multi-level training plan.

2. The disassembly and assembly path training method for pump station maintenance according to claim 1 is characterized in that: Construct a virtual-real simulation environment for pump station equipment, including: By laser scanning the target pumping station, a three-dimensional structural model of the pumping station equipment is constructed based on the laser scanning data; Retrieving historical maintenance logs to extract typical disassembly and assembly sequence data, and setting disassembly and assembly node priorities and disassembly and assembly node constraints according to the typical disassembly and assembly sequence data; The assembly and disassembly node priorities are synchronously embedded into the three-dimensional structure model according to the assembly and disassembly node constraints to construct the virtual-reality simulation environment.

3. The disassembly and assembly path training method for pump station maintenance as claimed in claim 2 is characterized in that: Based on the virtual-reality simulation environment, the pump station components in the pump station equipment are dynamically disassembled and monitored to identify, and a disassembly path is determined, wherein the disassembly path includes disassembly multi-dimensional feedback data, and the method includes: The binocular camera captures the spatial relative position of the tool and the component to generate three-dimensional coordinate point cloud data; Extracting the typical assembly and disassembly sequence data based on the virtual-reality simulation environment to dynamically monitor the assembly and disassembly of pump station components in the pump station equipment, and obtaining assembly and disassembly multi-dimensional feedback data, wherein the assembly and disassembly multi-dimensional feedback data includes dynamic assembly and disassembly operation data; Performing pre-disassembly marking on the dynamic disassembly and assembly operation data according to the disassembly and assembly node constraints to obtain a plurality of pre-disassembly and assembly nodes; The multiple pre-disassembly nodes are mapped to three-dimensional coordinate point cloud data, multiple disassembly operation sequences are constructed, the multiple disassembly operation sequences are associated with the disassembly node priorities for segmentation marking, and the disassembly path is constructed.

4. The disassembly and assembly path training method for pump station maintenance as claimed in claim 3 is characterized in that: The motion capture device is used to collect the operation trajectory data of the trainee in real time to obtain the somatosensory interaction motion signal, and the method includes: Deploy optical marking points on the hand joints of the trainees, disassembly and assembly tools, and target components of the pump station, wherein the optical marking points have three-dimensional spatial coordinates; The optical markers are captured by a motion capture device to construct three-dimensional space coordinates and generate relative posture data; Based on the relative posture data, disassembly and assembly motion analysis is performed to construct an operation trajectory sequence, the operation trajectory sequence is segmented and labeled and associated with the multiple pre-disassembly and assembly nodes, and the somatosensory interaction action signal is generated.

5. The disassembly and assembly path training method for pump station maintenance as claimed in claim 4 is characterized in that: The somatosensory interaction action signal is mapped to the virtual-reality simulation environment for disassembly and assembly analysis to generate a disassembly and assembly action deviation parameter set, the method comprising: Mapping the somatosensory interactive action signal to the virtual-reality simulation environment of the pump station equipment to obtain a mapped interactive signal; According to the disassembly and assembly node constraints, the mapping interaction signal is disassembled and analyzed frame by frame, and a multi-frame disassembly and assembly interaction benchmark value is set; According to the multi-frame assembly and disassembly interaction benchmark values, the operation trajectory sequence is synchronously compared with the multiple assembly and disassembly operation sequences to generate the assembly and disassembly action deviation parameter.

6. The disassembly and assembly path training method for pump station maintenance as claimed in claim 5, characterized in that: According to the multi-frame assembly and disassembly interaction benchmark values, the operation trajectory sequence is synchronously compared with the multiple assembly and disassembly operation sequences to generate the assembly and disassembly action deviation parameter, the method comprising: Dividing the operation trajectory sequence into a plurality of first sub-segments according to the assembly and disassembly nodes, and dividing the plurality of assembly and disassembly operation sequences into a plurality of second sub-segments according to the assembly and disassembly nodes; Aligning the plurality of first sub-segments with the plurality of second sub-segments according to the disassembly and assembly nodes to generate a plurality of sub-segment alignment data groups; Perform multi-segment deviation calculation on the plurality of sub-segment alignment data groups according to the multi-frame assembly and disassembly interaction reference values ​​to obtain multi-frame deviation values; If the number of frames of the multi-frame deviation value is greater than a preset frame number threshold, it is marked as a continuous deviation event, and the continuous deviation event is added to the disassembly and assembly action deviation parameter.

7. The disassembly and assembly path training method for pump station maintenance according to claim 1 is characterized in that: Based on the assembly and disassembly action deviation parameter set combined with the assembly and disassembly multi-dimensional feedback data, dynamic optimization is performed to generate multi-level training difficulty, the method includes: Constructing a three-dimensional skill evaluation matrix based on the disassembly and assembly action deviation parameter set and the disassembly and assembly multi-dimensional feedback data; Setting a plurality of evaluation levels according to the assembly and disassembly action deviation parameter set through the three-dimensional skill evaluation matrix; The disassembly and assembly paths are mapped based on the multiple evaluation levels to generate the multi-level practical training difficulty.

8. The disassembly and assembly path training method for pump station maintenance as claimed in claim 7, characterized in that: The method comprises: constructing a three-dimensional skill evaluation matrix based on the disassembly and assembly action deviation parameter set and the disassembly and assembly multi-dimensional feedback data; Performing an operation accuracy analysis based on the assembly and disassembly action deviation parameters to determine an operation accuracy index; Perform response recording based on the disassembly and assembly multi-dimensional feedback data to generate a response speed index; Construct a disassembly and assembly fault knowledge base, perform processing and analysis on pump station equipment based on the disassembly and assembly fault knowledge base, and formulate fault processing capability indicators; The operation accuracy index, the response speed index, and the fault handling capability index are normalized to construct the three-dimensional skill evaluation matrix.

9. The disassembly and assembly path training method for pump station maintenance as claimed in claim 1, characterized in that: According to the multi-level training difficulty, trainees are adapted according to the disassembly and assembly path to determine a multi-level training plan, the method comprising: Analyze the multi-level training difficulty to determine the basic training difficulty, advanced training difficulty, and expert training difficulty; Traversing the disassembly and assembly path to match the basic training difficulty, the advanced training difficulty, and the expert training difficulty to obtain multiple disassembly and assembly node difficulty data; Extracting the basic training difficulty from the plurality of disassembly and assembly node difficulty data, assigning a basic training plan to the trainees according to the basic training difficulty, and triggering a difficulty upgrade operation to upgrade the basic training difficulty to the advanced training difficulty if the disassembly and assembly accuracy does not meet the standard when executing the basic training plan; Extracting the advanced training difficulty from the plurality of disassembly and assembly node difficulty data, assigning an advanced training plan to the trainees according to the advanced training difficulty, extracting the fault repair efficiency by executing the advanced training plan, triggering a difficulty upgrade operation according to the fault repair efficiency, and upgrading the advanced training difficulty to the expert training difficulty; The difficulty data of the multiple disassembly and assembly nodes are extracted as the expert training difficulty, an expert training plan is assigned to the trainees according to the expert training difficulty, and the expert training plan is executed to perform disassembly and assembly accuracy detection. If there is an upward trend in the accuracy index, an automatic downgrade operation is triggered.

10. The disassembly and assembly path training system for pump station maintenance is characterized by: The system is used to implement the disassembly and assembly path training method for pump station maintenance according to any one of claims 1 to 9, and comprises: A simulation environment construction module is used to construct a virtual-real simulation environment for pump station equipment, and based on the virtual-real simulation environment, dynamically disassemble and monitor the pump station components in the pump station equipment to identify the disassembly and assembly path, wherein the disassembly and assembly path includes disassembly and assembly multi-dimensional feedback data; An operation trajectory acquisition module is used to collect the operation trajectory data of the trainee in real time through a motion capture device, obtain a somatosensory interaction action signal, map the somatosensory interaction action signal to the virtual-reality simulation environment for disassembly and assembly analysis, and generate a disassembly and assembly action deviation parameter set; A scheme generation module is used to dynamically optimize based on the disassembly and assembly action deviation parameter set combined with the disassembly and assembly multi-dimensional feedback data, generate multi-level training difficulty, adapt trainees according to the disassembly and assembly path according to the multi-level training difficulty, and determine the multi-level training scheme.

Images