Virtual simulation and intelligent decision-making method based on mechanical-electrical-hydraulic coupling dynamic model

Abstract

The invention belongs to the technical field of virtual simulation and control of a coal mine fully mechanized coal mining face supporting system, and provides a virtual simulation and intelligent decision making method based on an electromechanical-hydraulic coupling dynamic model in order to solve the problems that high-precision modeling of dynamics and hydraulic characteristics is lacked at present, and decision making optimization cannot be achieved in a virtual environment. The method comprises the following steps: constructing a high-fidelity hydraulic support multi-field dynamic model in a virtual environment; performing simulated working condition simulation on the model, and constructing a dynamic response model of the hydraulic support; constructing a hydraulic support digital twin system, performing real-time data exchange with the hydraulic support dynamic response model, and calculating and feeding back support actions in real time; a state evaluation and optimization strategy mechanism is introduced, posture abnormity is automatically recognized, corresponding bracket adjustment strategy suggestions are provided, and dynamic evaluation and intelligent regulation and control of the bracket state under the complex working condition are achieved through closed-loop iteration of man-machine interaction and data driving.

Description

Technical Field

[0001] The present invention belongs to the technical field of virtual simulation and control of a fully mechanized coal mining working face support system, and in particular relates to a virtual simulation and intelligent decision-making method based on an electromechanical-hydraulic coupling dynamics model. Background Art

[0002] With the continuous deepening of coal resource development, the complex phenomena of movement, fracture, delamination and other surrounding rocks of the roof of the fully-mechanized working face during the mining process have put forward higher requirements on the support performance of hydraulic supports. Traditional research is mostly based on static or simplified dynamic assumptions, which makes it difficult to reflect the complex force process and dynamic interaction of the support-surrounding rock system under actual working conditions. In addition, the existing virtual simulation methods are mostly based on kinematics, lacking high-precision modeling of dynamics and hydraulic characteristics, and it is difficult to truly reproduce the changes in the support posture, the mutual coupling between the hydraulic circuit response and the electronic control system. At the same time, the lack of dynamic feedback mechanism and real-time interaction makes it impossible to adjust the support support strategy in time according to the changes in transient working conditions.

[0003] With the development of digital twin technology, combining high-fidelity dynamic simulation with machine learning-driven data analysis and realizing decision optimization and implementation interaction in a virtual environment has become a new direction for intelligent research on hydraulic supports. By quickly evaluating and optimizing support strategies in a virtual environment, the risks and costs of actual underground testing can be significantly reduced. Summary of the invention

[0004] In order to solve at least one of the above-mentioned technical problems existing in the prior art, the present invention provides a virtual simulation and intelligent decision-making method based on a mechatronic-hydraulic coupling dynamics model.

[0005] The present invention is implemented by the following technical solution: a virtual simulation and intelligent decision-making method based on a mechanical, electrical and hydraulic coupling dynamics model, comprising the following steps:

[0006] Acquire the mechanical structure transmission system parameters and hydraulic fluid power system parameters of the hydraulic support, and build a multi-domain dynamic model of the hydraulic support based on the mechanical structure transmission system parameters and the hydraulic fluid power system parameters, and simultaneously acquire the operating environment data of the hydraulic support to build a digital twin system of the hydraulic support;

[0007] Acquire the entity operation data of the hydraulic support, and optimize the parameters of the multi-domain dynamic model of the hydraulic support based on the entity operation data to obtain a high-fidelity multi-domain dynamic model of the hydraulic support;

[0008] Performing working condition simulation based on the high-fidelity hydraulic support multi-domain dynamics model, and constructing a hydraulic support dynamic response model according to the simulation results;

[0009] The hydraulic support dynamic response model outputs the support action characteristics of the virtual hydraulic support in the hydraulic support digital twin system according to the user instructions received by the hydraulic support digital twin system, and feeds it back to the hydraulic support digital twin system. The hydraulic support digital twin system generates a corresponding digital twin system control signal according to the support action characteristics, and feeds the digital twin system control signal back to the hydraulic support dynamic response model for optimization and adjustment.

[0010] Preferably, the mechanical structure transmission system parameters include at least mechanical component mass parameters and density parameters, and motion constraints are established according to joint components in the mechanical components;

[0011] The hydraulic fluid power system parameters at least include component parameters corresponding to the hydraulic source, the metering pump, the relief valve, the reversing valve, and the double-acting hydraulic cylinder;

[0012] The operating environment data at least include dynamic environmental parameters corresponding to the stroke, pressure, flow, and posture of the hydraulic support.

[0013] Preferably, the working condition simulation based on the high-fidelity hydraulic support multi-domain dynamics model includes:

[0014] Based on the masonry beam theory, the parameter values ​​corresponding to the surrounding rock settlement, rotation angle and rotation rate are set;

[0015] The corresponding load is set according to the actual working condition, and the load is applied to the high-fidelity hydraulic support multi-domain dynamics model to simulate the working condition.

[0016] Preferably, it also includes:

[0017] A three-dimensional model of the hydraulic support is constructed according to the mechanical structure transmission system parameters and the hydraulic fluid power system parameters of the hydraulic support, and the three-dimensional model of the support is subjected to surface reduction processing and size adjustment to obtain a virtual hydraulic support in the digital twin system of the hydraulic support.

[0018] Preferably, before constructing the dynamic response model of the hydraulic support according to the simulation results, the method further includes:

[0019] The simulation results are cleaned, feature selected and normalized, and then the machine learning model is used to train and verify the processed simulation results. Finally, the dynamic response model of the hydraulic support is constructed based on the verification results.

[0020] Preferably, it also includes:

[0021] The stroke change of the hydraulic support cylinder in the simulation results is obtained, and the corresponding neural network model is established. At the same time, the corresponding neural network model is trained and converged by using standardization processing, mean square error loss function and Adam optimizer.

[0022] Preferably, it also includes:

[0023] Generate a frame adjustment strategy suggestion based on the optimization adjustment result, sort and filter the frame adjustment strategy suggestion to obtain the optimal adjustment plan, save the optimal adjustment plan, optimize the frame adjustment strategy suggestion, and iteratively update the dynamic response model of the hydraulic support according to the optimal adjustment plan.

[0024] Compared with the prior art, the present invention has the following beneficial effects:

[0025] The high-fidelity hydraulic support multi-domain dynamic model constructed by the present invention can more accurately simulate the operating behavior of the actual hydraulic support under actual working conditions. At the same time, by setting different modules to simulate a variety of different working conditions, run simulations and collect data, the operating data of the hydraulic support under different loads and different action positions are obtained, and a hydraulic support dynamic response model is constructed to realize the prediction of the hydraulic support. At the same time, a state evaluation and optimization strategy mechanism is introduced into the digital twin system of the hydraulic support to monitor and evaluate the operating data of the hydraulic support, timely discover the abnormal state of the hydraulic support, and automatically generate adjustment strategy recommendations, so as to realize dynamic evaluation and intelligent regulation of the state of the hydraulic support under complex working conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0027] Figure 1 A schematic diagram of the architecture of a virtual simulation and intelligent decision-making method based on a mechatronic-hydraulic coupling dynamics model provided by an embodiment of the present invention;

[0028] Figure 2 A schematic diagram of a mechanical model of a hydraulic support provided in an embodiment of the present invention;

[0029] Figure 3 A schematic diagram of a hydraulic system model of a hydraulic support provided by an embodiment of the present invention;

[0030] Figure 4 A schematic diagram of the cross-domain coupling portion of the hydraulic support dynamics model provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] In conjunction with the drawings in the embodiments of the present invention, the technical solutions in the embodiments of the present invention are clearly and completely described. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other implementation methods obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0032] It should be noted that the structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so that people familiar with this technology can understand and read them. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportional relationship or adjustment of size should fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention. It should be noted that in this specification, relational terms such as first and second are only used to distinguish one entity from several other entities, and do not necessarily require or imply any actual relationship or order between these entities.

[0033] like Figure 1 As shown, the present invention provides a virtual simulation and intelligent decision-making method based on a mechanical, electrical and hydraulic coupling dynamics model, comprising the following steps:

[0034] S1: Acquire the mechanical structure transmission system parameters and hydraulic fluid power system parameters of the hydraulic support, and build a multi-domain dynamic model of the hydraulic support based on the mechanical structure transmission system parameters and the hydraulic fluid power system parameters. At the same time, acquire the operating environment data of the hydraulic support to build a digital twin system of the hydraulic support.

[0035] Optionally, the mechanical structure transmission system parameters include at least mechanical component mass parameters and density parameters, and motion constraints are established based on the joint components in the mechanical components; the hydraulic fluid power system parameters include at least component parameters corresponding to the hydraulic source, metering pump, overflow valve, reversing valve, and double-acting hydraulic cylinder; the operating environment data include at least dynamic environment parameters corresponding to the stroke, pressure, flow, and posture of the hydraulic support.

[0036] In this embodiment, the Simulink tool module in MATLAB is used to establish the electromechanical-hydraulic coupling dynamic model of the hydraulic support through Simulink and its SimscapeMultibody and SimscapeFluid modules, and then the three-dimensional model of the support constructed by the CAD modeling software is imported into SimscapeMultibody, and at the same time, parameters such as mass and density are added to each component, and motion constraints between the support components are established through joint components; a hydraulic fluid power system model is established in SimscapeFluid, including component parameters such as hydraulic source, metering pump, overflow valve, reversing valve, double-acting hydraulic cylinder, etc., to realize the coupling modeling of the mechanical structure and the hydraulic system; the dynamic environmental parameters such as the stroke, pressure, flow, posture, etc. of the virtual hydraulic support are monitored through the virtual sensor module.

[0037] In this embodiment, the DesignOptimization function is used to automatically optimize the parameters of the hydraulic cylinder, valve body, etc. for the constructed electromechanical-hydraulic coupling dynamics model to improve the accuracy and reliability of the simulation model.

[0038] In this embodiment, Figure 2 As shown, the virtual hydraulic support established in the CAD modeling software SolidWorks is exported as ".STL" files according to different components, including the base, front connecting rod, rear connecting rod, quantitative components, etc.;

[0039] Add the exported model file to the "FileSolid" module of the virtual hydraulic support and define the object's mass, density and other parameters;

[0040] Add a "RigidTransform" coordinate system transformation relationship module between the components of the virtual hydraulic support to connect the joints and define the coordinate transformation parameters;

[0041] Analyze the motion relationship between the components of the virtual hydraulic support, and add constraint modules between the components. Add joint components such as "PrismaticJoint" prismatic joint and "RevoluteJoint" revolute joint to connect the components of the hydraulic support, and define the spring stiffness, damping and other parameters of the joint components. Among them, the base is connected to the world coordinate system, and then the base is connected to the two column outer cylinders, two front connecting rods and two rear connecting rods respectively with revolute joints; the two column outer cylinders are connected to the column middle cylinder with a prismatic joint, and the column middle cylinder is connected to the column piston with a prismatic joint; the shielding beam is connected to the two front connecting rods, two rear connecting rods, the balancing jack cylinder, and the top beam through a revolute joint; the top beam is connected to the two column pistons, the balancing jack push rod, and the shielding beam through a revolute joint;

[0042] like Figure 3As shown, a Simulink simulation model control block diagram of the hydraulic system of the hydraulic support based on Simscape Fluid is established to simulate the oil circuit structure, valve control, etc. in the hydraulic fluid power system to ensure that the dynamic response of the hydraulic system in the virtual hydraulic support can reflect the actual working conditions;

[0043] Add the hydraulic source module "Reservoir" under "IsothermalFluid" in SimscapeFluid to represent the oil tank;

[0044] Add the "Fixed-DisplacementPump" module under "IsothermalFluid" to represent the fixed displacement pump in the hydraulic system, and adjust the displacement, nominal speed and other parameters of the fixed displacement pump;

[0045] Add the "PressureReliefValve" module under "IsothermalFluid" to represent the relief valve in the hydraulic system, and adjust the control mode, opening pressure and other parameters of the relief valve;

[0046] Add the "4-Way3-PositionDirectionalValve" module under "IsothermalFluid" to represent the directional valve in the hydraulic system, and set parameters such as flow area and connection mode;

[0047] Add the "Double-ActingActuator" module under "IsothermalFluid" to represent the double-acting hydraulic cylinder in the hydraulic system, and set parameters such as stroke, rod cavity and rodless cavity area;

[0048] Add the "MechanicalTranslationalReference" module under "TranslationalElements" as the displacement motion reference of the hydraulic cylinder;

[0049] The input end of the metering pump module is connected to the oil tank module, and the output end is connected to an input end of the reversing valve module; the other three input ends of the reversing valve module are respectively connected to the input port of the hydraulic cylinder, the output port of the hydraulic cylinder, and the port of the oil tank module; the input port of the pump station overflow valve is connected to the output port of the metering pump, and the output port is connected to the oil tank module; the input port of the hydraulic cylinder overflow valve is connected to the loop connecting the reversing valve and the output port of the hydraulic cylinder, and the output port is connected to the oil tank module; each group of hydraulic cylinders is connected with the r end and the c end, indicating a double-telescopic hydraulic cylinder, and the other r end and c end are respectively connected to the port of the displacement reference module;

[0050] like Figure 4As shown, the interaction between the hydraulic fluid power system and the mechanical structure transmission system is realized through the coupling of mechanical and hydraulic modules. The hydraulic fluid power system is connected to the mechanical structure transmission system, and the "TranslationalMultibodyInterface" module under "MultibodyInterface" is added to represent the cross-domain signal conversion, which transmits the power signal of the hydraulic system to the mechanical structure transmission system;

[0051] Add virtual sensor modules to the multi-domain dynamics model of hydraulic supports to monitor the hydraulic fluid power system and mechanical structure transmission system in real time;

[0052] Add the "FlowRateSensor" module under "IsothermalFluid" to monitor the hydraulic flow rate in the hydraulic fluid power system and adjust the flow rate monitoring parameters;

[0053] Add the "PressureSensor" module under "IsothermalFluid" to monitor the hydraulic pressure in the hydraulic fluid power system and adjust the pressure monitoring parameters;

[0054] Add monitoring function in the "PrismaticJoint" module, turn on options such as "Velocity" and "Position" to monitor the velocity, displacement, acceleration and other operating data of the joint motion;

[0055] Add the "TransformSensor" module to connect the "FileSolid" of different parts of the bracket to monitor the relative movement between two different parts and read the operating parameters such as position and angle.

[0056] S2: Acquire entity operation data of the hydraulic support, and optimize parameters of the multi-domain dynamics model of the hydraulic support based on the entity operation data to obtain a high-fidelity multi-domain dynamics model of the hydraulic support.

[0057] In this embodiment, the model is corrected by comparing with the measured data of the physical prototype and using parameter optimization tools, so that the multi-domain dynamic model of the hydraulic support reaches high fidelity, and a high-fidelity multi-domain dynamic model of the hydraulic support is obtained.

[0058] In this embodiment, in order to ensure that the established multi-domain dynamic model of the hydraulic support has practical significance and high accuracy, it needs to be verified with the operating data of the actual support. In the model settings, change the solver to a variable step solver, set the simulation time and step size for dynamic simulation, collect the simulation data of the hydraulic support, including hydraulic flow rate, column pressure, column stroke, etc., and check whether the simulation results accurately reflect the operating characteristics of the actual support;

[0059] In this embodiment, the operation data of the hydraulic support under actual working conditions is collected by installing monitoring equipment on the physical prototype of the hydraulic support, and the monitoring equipment includes but is not limited to a stroke sensor, a pressure sensor, a hydraulic flow rate sensor, etc.;

[0060] Based on the sensor monitoring data in the simulation and the operating data collected by the actual sensors, the DesignOptimization function is used to automatically tune the model's parameter variables, and the hydraulic cylinder, valve body and other parameters in the model are optimized through the optimization algorithm. The model parameters are optimized to generate results that match the measured data, completing the construction of a high-fidelity hydraulic support dynamic model. The tuned model will have a higher accuracy and can more accurately simulate the movement behavior of the actual support under various working conditions.

[0061] S3: Perform working condition simulation based on the high-fidelity hydraulic support multi-domain dynamics model, and construct a hydraulic support dynamic response model according to the simulation results.

[0062] Optionally, parameter values ​​corresponding to surrounding rock settlement, rotation angle and rotation rate are set based on masonry beam theory; corresponding loads are set according to actual working conditions, and the loads are applied to the high-fidelity hydraulic support multi-domain dynamics model to perform working condition simulation.

[0063] In this embodiment, based on the masonry beam theory, the top plate separation area of ​​the hydraulic support is divided into three key rock blocks A, B, and C. By setting the sinking amount, rotation angle, and rotation rate of the rock block, different types of loads act on the hydraulic support. Simulink is used to run multi-scenario simulations to collect the travel change, attitude inclination, force distribution, and system pressure fluctuation data of the hydraulic support under various load conditions.

[0064] In this embodiment, based on the verified high-fidelity hydraulic support multi-domain dynamics model, simulations under different load conditions are performed, the "ExternalForce" module is added to simulate the load stress of the hydraulic support, the "RigidTransform" module is added to change the load action position, and the "SignalBuilder" module is added to simulate the change curve of the load force value. Set a variety of different working conditions, run the simulation and collect data, obtain the operating data of the hydraulic support under different loads and different action positions, including but not limited to column pressure, column stroke, balance jack stroke, etc., and save and store the data as a ".csv" file.

[0065] Optionally, before constructing the dynamic response model of the hydraulic support according to the simulation results, it also includes: cleaning, feature selection and normalization of the simulation results, then using a machine learning model to train and verify the processed simulation results, and finally constructing the dynamic response model of the hydraulic support according to the verification results.

[0066] Optionally, it also includes: obtaining the stroke change of the hydraulic support cylinder in the simulation results, and establishing a corresponding neural network model, and using standardization processing, mean square error loss function and Adam optimizer to train and convergence control the corresponding neural network model.

[0067] In this embodiment, the dynamic response model of the hydraulic support can perform online learning and parameter update according to the real-time data of the underground sensor to cope with the changes of uncertain factors in the actual working conditions and improve the system's adaptability and decision-making flexibility to the complex mine environment.

[0068] In this embodiment, based on the simulation data, a deep neural network model is used to train the hydraulic support cylinder stroke change prediction data model to achieve the prediction ability of the hydraulic support model. The input features of the model are the load position and load value, and the output is the predicted values ​​of the three cylinder positions.

[0069] (Position1, Position2, Position3).

[0070] First load the data, read the simulation data file, and extract the input features from it

[0071] (LoadForce, LoadPosition) and output labels (Position1, Position2, Position3), the output labels correspond to the stroke changes of the cylinders in the hydraulic support; split the data into training sets and test sets, using a ratio of 8:2 to ensure the generalization ability of the data; standardize the data, and use StandardScaler to standardize the input data and output data so that the model can converge faster during training and avoid the influence of data of different dimensions. The standardized data is converted to PyTorch's FloatTensor type to adapt to the input of the neural network model; prepare features and labels, the input features include load size (LoadForce) and load position (LoadPosition), and the output labels are the changes in the positions of the three cylinders (Position1, Position2, Position3). Split the position label of each cylinder into three independent y_train_tensors and y_test_tensors so that a separate model can be trained for each cylinder.

[0072] The corresponding model is structurally designed: using a neural network architecture, a multi-layer fully connected neural network is selected as the basic model, which contains multiple hidden layers, and the nonlinear expression ability is increased by activation functions. The output of each network in the output layer is a predicted value, and separate neural network models are trained for different oil cylinder positions. Use activation functions to enhance the nonlinear ability of the network and avoid the gradient vanishing problem; each model uses mean square error loss to evaluate the error between the predicted value and the true value, and the optimization goal is to minimize the error. Using the Adam optimizer, the learning rate of the model can be adaptively adjusted; in the training stage, for each prediction model of the oil cylinder position, 200 training cycles are performed, and the output is calculated by forward propagation in each epoch, and the model parameters are updated by back propagation. The current training loss (MSE) is output every 10 epochs to track the training progress and determine whether the model has converged.

[0073] In each training, the input features are first passed into the corresponding network to obtain the predicted value, and then the loss between the predicted value and the true label is calculated. The optimizer function is used to clear the gradient, perform backpropagation, calculate the gradient, and update the parameters; after the training is completed, each model is used to predict the test set data. When predicting, the gradient calculation is turned off to increase the inference speed. The output of each model is denormalized to obtain the actual cylinder position prediction result.

[0074] Optionally, a three-dimensional model of the hydraulic support is constructed according to the mechanical structure transmission system parameters and the hydraulic fluid power system parameters of the hydraulic support, and the three-dimensional model of the support is subjected to surface reduction processing and size adjustment to obtain a virtual hydraulic support in the hydraulic support digital twin system.

[0075] In this embodiment, the three-dimensional model of the bracket is imported into 3dsMax for preprocessing, the three-dimensional model of the bracket is reduced in surface area, and the size unit is adjusted. The adjusted three-dimensional model file of the bracket is then imported into the digital twin environment of the hydraulic support in the ".fbx" format, and physical constraints are added to each structural model to achieve virtual assembly of the hydraulic support in the digital twin.

[0076] In this embodiment, a digital twin scene of a hydraulic support is constructed in Unity, and the digital mapping of mine tunnels, roofs, coal walls and hydraulic supports is reproduced through 3D models and physical engine rendering; a UI display panel of the digital twin system of the hydraulic support is constructed, and a UI display module for the operation monitoring data of the hydraulic support is designed. The user can intuitively observe the dynamic response of the hydraulic support in the visual interface, and continuously adjust the input conditions or call the adjustment strategy according to the actual situation; a UI control panel of the digital twin system of the hydraulic support is constructed, and a signal input module for the hydraulic support control is designed. The user sets the load conditions or roof movement parameters in Unity, and the virtual scene inputs the information into the trained machine learning model; a data communication channel is constructed, and a TCP communication interface is developed to connect the Unity scene with the constructed dynamic response model of the hydraulic support, realize real-time data exchange, and construct a dynamic response model of the hydraulic support to quickly calculate the support stroke changes and posture adjustments according to user instructions, and feed back the prediction results to Unity in real time, thereby driving the action of the virtual support.

[0077] In this embodiment, a control script is written on the digital twin end of the hydraulic support to control the action of the hydraulic support so that the hydraulic support can realize the basic actions required for virtual simulation;

[0078] Design an interactive interface, design the hydraulic support operation panel through UGUI, and the user inputs the working condition parameters of the support through the graphical interface. Simulate the manual control simulation of the actual hydraulic support controller, and also input different working conditions through the interface to communicate with the Python end model;

[0079] Develop a TCP communication interface to realize real-time data exchange between the hydraulic support digital twin end and the hydraulic support dynamic response model. It is responsible for sending transmission working conditions to the hydraulic support dynamic response model end, receiving the data output by the hydraulic support dynamic response model, and controlling the action of the hydraulic support in the virtual environment. Through this interface, Unity can send working conditions, the hydraulic support dynamic response model calculates according to the input conditions, and returns the stroke change data to Unity;

[0080] In the digital twin environment of hydraulic supports, the virtual simulation of hydraulic supports achieves real-time feedback by interacting with the trained dynamic response model of hydraulic supports:

[0081] The Unity digital twin terminal inputs the working condition information into the trained dynamic response model of the hydraulic support. The dynamic response model of the hydraulic support processes the input conditions and feeds back the change trend of the hydraulic support under these conditions.

[0082] The stroke change data output by the hydraulic support dynamic response model will be fed back to the Unity digital twin in real time, driving the virtual hydraulic support to complete the corresponding action. Through the visualization interface of the hydraulic support digital twin system, users can observe the dynamic response of the support;

[0083] After the virtual simulation is completed, the actual hydraulic support test data can be accessed for analysis. For example, the actual movement of the hydraulic support under different operating conditions can be monitored in real time and compared with the virtual simulation results to further optimize the simulation model.

[0084] S4: The hydraulic support dynamic response model outputs the support action characteristics of the virtual hydraulic support in the hydraulic support digital twin system according to the user instructions received by the hydraulic support digital twin system, and feeds it back to the hydraulic support digital twin system. The hydraulic support digital twin system generates a corresponding digital twin system control signal according to the support action characteristics, and feeds the digital twin system control signal back to the hydraulic support dynamic response model for optimization and adjustment.

[0085] Optionally, it also includes: generating frame adjustment strategy suggestions based on the optimization adjustment results, sorting and screening the frame adjustment strategy suggestions to obtain the optimal adjustment plan, saving the optimal adjustment plan, and optimizing the frame adjustment strategy suggestions, and iteratively updating the hydraulic support dynamic response model according to the optimal adjustment plan.

[0086] In this embodiment, a state evaluation and optimization strategy mechanism is introduced into the digital twin system of the hydraulic support. When the prediction results show that the hydraulic support may be unstable or have abnormal stroke, the digital twin system of the hydraulic support can automatically propose adjustment strategy suggestions, such as increasing the corresponding hydraulic cylinder stroke, adjusting the valve setting, etc. Users can verify the effect of the strategy execution in the digital twin system of the hydraulic support and conduct repeated experiments to optimize the control strategy. Through this closed-loop iteration of human-computer interaction and data-driven, dynamic evaluation and intelligent regulation of the support status under complex working conditions can be achieved.

[0087] In this embodiment, a state evaluation and optimization strategy mechanism is introduced into the hydraulic support digital twin system to monitor and evaluate the operation data of the hydraulic support; the hydraulic support digital twin system collects real-time operation data of the hydraulic support, such as hydraulic cylinder pressure and support inclination, and uses machine learning algorithms to classify and evaluate the operation status of the hydraulic support in combination with the user input working conditions; when the virtual simulation results show that the hydraulic support may have abnormal states such as posture instability or stroke abnormality, the hydraulic support digital twin system can automatically propose adjustment strategy suggestions, such as increasing the corresponding hydraulic cylinder stroke, adjusting the valve setting, etc.; verify the effect of the strategy after execution in the virtual environment, and repeatedly test to optimize the control strategy. The hydraulic support digital twin system presets a variety of adjustment strategies (such as support inclination adjustment, hydraulic cylinder pressure balance optimization), and matches them based on the simulation results of the historical knowledge base and the high-fidelity hydraulic support multi-domain dynamic model. Through reinforcement learning methods, such as deep Q network, the optimal adjustment strategy is calculated according to the real-time working conditions, and specific parameters such as hydraulic cylinder pressure and adjustment frequency are recommended for adjustment. Through this closed-loop iteration of human-computer interaction and data-driven, dynamic evaluation and intelligent regulation of the support state under complex working conditions are realized.

[0088] The embodiment of the present invention constructs a high-fidelity hydraulic support multi-domain dynamic model in the Simulink environment; compares it with the measured data of the physical hydraulic support, optimizes the parameters to ensure its high accuracy and authenticity; runs multi-condition simulations to collect dynamic response data of the hydraulic support under various load conditions; uses machine learning algorithms to construct a hydraulic support dynamic response model for the simulation data, and performs model training to establish a mapping relationship between the working condition input and the dynamic response of the support; constructs a hydraulic support digital twin system in the Unity platform, and realizes real-time data exchange with the machine learning model through a communication interface. Users can input working conditions in a virtual environment, and the system calculates and feeds back the support action in real time. At the same time, a state evaluation and optimization strategy mechanism is introduced to automatically identify possible posture anomalies of the support and propose corresponding frame adjustment strategy suggestions. Users verify and optimize the control strategy in a virtual environment, and through human-computer interaction and data-driven closed-loop iteration, dynamic evaluation and intelligent regulation of the support state under complex working conditions are realized.

[0089] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A virtual simulation and intelligent decision-making method based on a mechanical, electrical and hydraulic coupling dynamics model, characterized in that: The steps include: Acquire the mechanical structure transmission system parameters and hydraulic fluid power system parameters of the hydraulic support, and build a multi-domain dynamic model of the hydraulic support based on the mechanical structure transmission system parameters and the hydraulic fluid power system parameters, and simultaneously acquire the operating environment data of the hydraulic support to build a digital twin system of the hydraulic support; Acquire the entity operation data of the hydraulic support, and optimize the parameters of the multi-domain dynamic model of the hydraulic support based on the entity operation data to obtain a high-fidelity multi-domain dynamic model of the hydraulic support; Performing working condition simulation based on the high-fidelity hydraulic support multi-domain dynamics model, and constructing a hydraulic support dynamic response model according to the simulation results; The hydraulic support dynamic response model outputs the support action characteristics of the virtual hydraulic support in the hydraulic support digital twin system according to the user instructions received by the hydraulic support digital twin system, and feeds it back to the hydraulic support digital twin system. The hydraulic support digital twin system generates a corresponding digital twin system control signal according to the support action characteristics, and feeds the digital twin system control signal back to the hydraulic support dynamic response model for optimization and adjustment.

2. The virtual simulation and intelligent decision-making method based on the electromechanical-hydraulic coupling dynamics model according to claim 1 is characterized in that: The mechanical structure transmission system parameters at least include mechanical component mass parameters and density parameters, and motion constraints are established according to joint components in the mechanical components; The hydraulic fluid power system parameters at least include component parameters corresponding to the hydraulic source, the metering pump, the relief valve, the reversing valve, and the double-acting hydraulic cylinder; The operating environment data at least include dynamic environmental parameters corresponding to the stroke, pressure, flow, and posture of the hydraulic support.

3. The virtual simulation and intelligent decision-making method based on the electromechanical-hydraulic coupling dynamics model according to claim 1 is characterized in that: The working condition simulation based on the high-fidelity hydraulic support multi-domain dynamics model includes: Based on the masonry beam theory, the parameter values ​​corresponding to the surrounding rock settlement, rotation angle and rotation rate are set; The corresponding load is set according to the actual working condition, and the load is applied to the high-fidelity hydraulic support multi-domain dynamics model to simulate the working condition.

4. The virtual simulation and intelligent decision-making method based on the electromechanical-hydraulic coupling dynamics model according to claim 1 is characterized in that: Also includes: A three-dimensional model of the hydraulic support is constructed according to the mechanical structure transmission system parameters and the hydraulic fluid power system parameters of the hydraulic support, and the three-dimensional model of the support is subjected to surface reduction processing and size adjustment to obtain a virtual hydraulic support in the digital twin system of the hydraulic support.

5. The virtual simulation and intelligent decision-making method based on the electromechanical-hydraulic coupling dynamics model according to claim 1 is characterized in that: Before building the dynamic response model of hydraulic support according to the simulation results, it also includes: The simulation results are cleaned, feature selected and normalized, and then the machine learning model is used to train and verify the processed simulation results. Finally, the dynamic response model of the hydraulic support is constructed based on the verification results.

6. The virtual simulation and intelligent decision-making method based on the electromechanical-hydraulic coupling dynamics model according to claim 5 is characterized in that: Also includes: The stroke change of the hydraulic support cylinder in the simulation results is obtained, and the corresponding neural network model is established. At the same time, the corresponding neural network model is trained and converged by using standardization processing, mean square error loss function and Adam optimizer.

7. The virtual simulation and intelligent decision-making method based on the electromechanical-hydraulic coupling dynamics model according to claim 1 is characterized in that: Also includes: Generate a frame adjustment strategy suggestion based on the optimization adjustment result, sort and filter the frame adjustment strategy suggestion to obtain the optimal adjustment plan, save the optimal adjustment plan, optimize the frame adjustment strategy suggestion, and iteratively update the dynamic response model of the hydraulic support according to the optimal adjustment plan.

Images