Deep-sea exploration parallel propeller technical framework and method based on digital twinning

Abstract

The invention relates to a deep sea exploration parallel propeller technical framework and method based on digital twinning, the technical framework comprises a physical layer, an interaction layer, a virtual layer, a decision-making layer and a communication layer, the physical layer is composed of a three-degree-of-freedom parallel propeller and a physical sensor, the interaction layer is used for man-machine interaction and displaying exploration operation content, and the virtual layer is used for virtual decision-making. The virtual layer comprises a digital twin model and a virtual control unit, the decision-making layer comprises an autonomous decision-making engine and a predictive maintenance center, and the communication layer performs data transmission and interaction; the method comprises the following steps of: synchronizing sensor data to a Unity digital twinborn body on the basis of an OPC protocol, triggering an instruction through man-machine interaction, inversely resolving joint parameters through kinematics, driving a parallel propeller to execute a task by a virtual PLC (Programmable Logic Controller) through an OPC, and autonomously optimizing a dynamic trajectory and responding to a fault on the basis of a TCN (Triggered Crystal Network). According to the invention, digital twinning and virtual PLCs are deeply integrated and deployed to physical equipment, so that the debugging risk and cost in a deep sea environment can be greatly reduced, and the problems of low pose precision and fault response lag of a propeller are effectively solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the fields of deep sea exploration and digital twin technology, and particularly relates to a deep sea exploration parallel thruster technical framework and method based on digital twin. BACKGROUND

[0002] With the exhaustion of land resources, deep sea exploration, as a key field of human cognition of the ocean, development of resources and maintenance of ecological security, has become increasingly important. The energy, mineral and biological resources in the deep sea have become an important support for sustainable development, and the exploration operation in extreme environments also faces unprecedented challenges. The complex working conditions of strong ocean currents, low temperature and near zero visibility in the deep sea result in inherent defects of traditional thrusters, such as high development and maintenance costs, inaccurate pose control and weak anti-interference ability. These defects not only restrict the reliability and accuracy of the equipment, but also greatly limit the efficiency and safety of exploration due to the high risk and cost of deep sea debugging. As a frontier field of marine science and technology, deep sea exploration needs to rely on the integration of digital twin and new equipment technologies to achieve new breakthroughs in the field of exploration.

[0003] Digital twin, with the help of digital technology, can build a virtual mapping of physical entities, realize virtual-real interaction, dynamic simulation and intelligent decision-making. Through real-time data driving, the state, behavior and life cycle of physical entities in the virtual space are reproduced, and the state monitoring, fault prediction and decision optimization are realized by combining real-time physical world data from sensors to form a closed-loop optimization system, which can significantly improve the accuracy and reliability of the system and provide favorable technical support for the field of deep sea exploration.

[0004] The application of digital twin technology in the field of deep sea exploration can realize the closed-loop life cycle management of deep sea exploration parallel thruster design simulation, real-time control and performance optimization, and the virtual and real synchronous operation, which helps to reduce the development and debugging cost of thrusters and improve the efficiency and safety of exploration. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a deep sea exploration parallel thruster technical framework and method based on digital twin. The digital twin technology can realize real-time mapping and control of the physical world and the virtual model, and has outstanding advantages in solving the problem of thruster pose accuracy in complex fluid environment in the deep sea. By combining digital twin with virtual PLC control strategy and deploying it to physical equipment, the present application can greatly reduce the debugging risk and cost in the deep sea environment.

[0006] The present application adopts the following technical solutions:

[0007] A deep sea exploration parallel thruster technical framework based on digital twin, characterized by comprising a physical layer, an interaction layer, a virtual layer, a decision layer and a communication layer.

[0008] The physical layer includes a three-degree-of-freedom parallel thruster and a physical sensor; the three-degree-of-freedom parallel thruster includes a moving platform, a fixed platform, a main driven branch chain, a constraint branch chain and a propeller, the main driven branch chain includes a first, a second and a third main driven branch chain, the first, the second and the third main driven branch chain are circumferentially distributed between the moving platform and the fixed platform, the first, the second and the third main driven branch chain are connected by a hook joint, a servo-driven moving pair and a hook joint in sequence, the constraint branch chain is arranged along the central axis of the moving platform and the fixed platform, and the constraint branch chain is connected by a constraint rotating pair and a hook joint in sequence, so that the moving platform realizes three-rotating-degree-of-freedom movement; the physical sensor is installed on the three-degree-of-freedom parallel thruster and includes a water pressure gauge, a temperature sensor, a sonar sensor, an inertial measurement unit (IMU) and an encoder;

[0009] The interaction layer is a man-machine interaction interface in a Unity environment, including a virtual-real synchronous monitoring area and a control instruction issuing module; the virtual-real synchronous monitoring area includes a digital twin visualization interface and a monitoring interface displaying the deep-sea exploration working values of the parallel thruster, the working driving values of the digital twin and the working value difference between the two; the control instruction issuing module controls the parallel thruster in an instruction mode, providing a manual control mode and an automatic task mode;

[0010] The virtual layer includes a digital virtual world and a virtual control unit; the digital virtual world includes a digital twin and a virtual fluid environment; the digital twin uses SolidWorks to build a thruster three-dimensional geometric model, converts the format, and builds a digital twin based on Unity; the virtual fluid environment generates dynamic fluid particles according to the underwater pressure, temperature and obstacle position of the parallel thruster detected by the physical sensor, analyzes the current velocity profile and the seafloor topography, and is driven by the Unity particle system; the virtual control unit receives and processes the sensing parameters of the physical sensor and the task instructions issued by the interaction layer, generates control signals based on the virtual PLC, drives the digital twin to move, executes the working instructions, and synchronizes the actions to the parallel thruster with the help of the communication layer;

[0011] The decision layer includes an autonomous decision engine and a predictive maintenance center, and the decision layer receives the simulation data output by the virtual layer; the autonomous decision engine is based on a time series convolution network (TCN) and analyzes sonar obstacle data and historical path sequences to generate an anti-flow compensation matrix and path re-planning instructions, and outputs optimized instructions to the virtual layer; the predictive maintenance center relies on a TCN model, receives encoder vibration frequency spectrum and temperature time series data, trains a prediction model according to a historical data set, and is directly connected to the physical layer;

[0012] The communication layer includes a bidirectional data interaction channel and a shared memory partition; the bidirectional data interaction channel is based on an OPC protocol to realize bidirectional data synchronization between a virtual environment and a physical device, including virtual layer control instruction downward and sensor data feedback upward; the shared memory partition is an OPC shared memory partition, and the virtual layer and the decision layer interact with each other through data.

[0013] Further limitation, the three-degree-of-freedom parallel thruster has a symmetrical configuration, and the central axis of the Hooke joint connecting shaft passes through the intersection of two perpendicular lines of the centers of the moving platform and the static platform.

[0014] Further limitation, all the functional modules of the interaction layer are integrated in the same human-computer interaction interface, the creation of the human-computer interaction interface follows the Canvas layered rendering in Unity, C# scripts with function functions are written in the Unity platform, the function functions are mounted under UI elements, the function buttons of the virtual-real synchronization monitoring area and the control instruction issuing module are set, and the operation instructions are unfolded by clicking the function buttons; the function buttons of the virtual-real synchronization monitoring area provide value monitoring function buttons, including viewing of the deep sea exploration working value of the parallel thruster, the working driving value of the digital twin and the difference between the two working values, and the unfolding operation of the parallel thruster deep sea exploration working value monitoring function button includes a deep sea resource detection data display interface.

[0015] A deep sea exploration parallel thruster method based on digital twinning, characterized in that the method comprises the following steps:

[0016] S1: building a digital twin of the parallel thruster in Unity, the physical sensor collecting fluid dynamics parameters and thruster pose information in real time, deploying an OPC server in the physical layer edge controller, configuring a communication interface with the PLC, binding the physical sensor data address to the server, and transmitting the fluid dynamics parameters and pose information to the Unity digital twin through OPC uplink transmission;

[0017] S2: triggering the digital twin working instruction by the human-computer interaction interface, driving the digital twin to plan the path and the working pose, and uploading the planning instruction to the OPC shared memory partition;

[0018] S3: realizing the linkage of the B&R PLC and Unity by means of the S7-PLCSIM Advanced software, adding a C# script for receiving B&R data in Unity, and transmitting the data to the parallel thruster through OPC downlink to execute physical operation;

[0019] S4: Real-time display of virtual and real synchronization monitoring area during task execution to promote the working information of the parallel manipulator and the digital twin, and the autonomous decision engine dynamically generates PWM optimization instructions based on the anti-flow compensation matrix according to the original planning instructions uploaded to the OPC shared memory partition, to correct the deviation between the current virtual pose and the target virtual pose;

[0020] S5: Repeat step S3 to correct the virtual pose and actual pose of the parallel manipulator;

[0021] S6: The OPC protocol further synchronizes the environment and motion parameters obtained by the physical sensor to the digital twin model, optimizes the control parameters in a closed loop, and realizes real-time data updating.

[0022] Further limited, the motion path planning includes the following steps:

[0023] S21: Import the same parallel manipulator FBX format model as the digital twin in Unity again, hide the mesh renderer, realize the stealth function of the model, and make it exist as a shadow model of the digital twin;

[0024] S22: Add the CCD IK plug-in to the shadow model to perform inverse kinematics analysis of the parallel manipulator, automatically calculate the joint position parameters of the manipulator according to the target position and attitude input by the human-computer interaction interface, and make the shadow model of the parallel manipulator plan the path according to the set point;

[0025] S23: Integrate the OPC client library in the Unity virtual layer, push the control instructions sent by the digital twin to the physical device to the server through asynchronous writing, and store them in the shared memory partition;

[0026] S24: Perform PLC ladder programming in Bo Plot, transmit the target pose to Unity through the S7.NetPlus plug-in, and drive the virtual parallel manipulator to move.

[0027] Further limited, the anti-flow compensation matrix PWM control signal is based on the motion coupling analysis of the dynamic model in the virtual fluid environment, extracts the ocean current disturbance pattern through the time convolution network TCN dilated convolution layer, outputs the thrust correction parameter matrix, and performs pose compensation, including the following steps:

[0028] S41: The time convolution network TCN acquires the fluid dynamics parameters in advance according to the physical sensor, establishes a learning model according to the historical parameters, solves the rigid body motion equation of the manipulator connecting rod and the fluid resistance equation based on CFD, and forms a coupled differential equation group:

[0029]

[0030] wherein q is a joint pose, M(q) is a mechanism mass matrix, is a Coriolis force matrix, G(q) is a static torque generated by gravity, τ is an effective dynamic torque input, D is a drag torque matrix generated by the fluid velocity field v, J(q) is a Jacobian matrix, and T denotes the transpose of a matrix;

[0031] S42: calling the MKL mathematical library for linear algebra solution, real-time solving the coupling effect, and outputting the joint pose q containing fluid interference;

[0032] S43: mapping dynamic parameters by means of the communication layer OPC protocol, and writing the fluid drag torque D in the solving result and the joint pose q containing fluid interference into the OPC shared memory address;

[0033] S44: based on the calculated pose deviation and the mapped dynamic parameters, real-time updating the motion planning, and generating a corresponding PWM optimized compensation signal by the virtual control end to correct the cumulative error between the current pose and the target pose of the digital twin.

[0034] Further limitation, in the step S4, relying on the TCN model predictive maintenance center, for the working data monitored by the virtual-real synchronous monitoring area that may cause abnormal parallel thruster, predicting the parallel thruster failure in advance, generating a predictive maintenance scheme, and forcibly interrupting the parallel thruster operation when exceeding the maintenance range.

[0035] Compared with the prior art, the advantages of the present application are as follows: (1) the virtual-real synchronous control precision is improved, the virtual PLC intelligent control physical actuator is adopted through the linkage of PLC and Unity virtual environment, the action synchronization of digital twin and physical parallel thruster ("virtual control real") is realized, and the transmission precision and execution fidelity of control instruction are significantly improved; (2) the fault prediction and operation and maintenance cost are reduced, the human-machine interaction interface (HMI) is collected in the Unity end, the equipment working state is visualized in real time, the time series data is fault predicted based on the TCN time series convolution network, the equipment failure rate is effectively reduced, and the development debugging and maintenance cost of the entity prototype is reduced; (3) the dynamic anti-interference ability is enhanced, the joint level action planning is realized through the inverse kinematics algorithm, the pose deviation caused by water flow disturbance is autonomously corrected through the anti-flow compensation solution in the fluid environment, and the motion stability of the thruster in the complex deep sea environment is ensured; (4) the reliability of the closed-loop optimization system is improved, the real-time feedback of the physical sensor and the dynamic adjustment of the digital twin form a "perception-decision-execution" closed-loop optimization system, and the motion trajectory accuracy and the reliability of the overall system are significantly improved; (5) the development efficiency and economy are optimized, the digital twin technology replaces the traditional entity prototype repeated test, the development cycle is shortened, and the research and development and trial and error cost of the deep sea exploration equipment is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a deep sea exploration parallel thruster technical framework and method based on digital twinning of the present application;

[0037] Figure 2 is a deep sea exploration parallel thruster technical framework based on digital twinning of the present application;

[0038] Figure 3 is an isometric view of a deep sea exploration parallel thruster mechanical structure based on digital twinning of the present application;

[0039] Figure 4 is a deep sea exploration parallel thruster structure diagram based on digital twinning of the present application;

[0040] Figure 5 is a deep sea exploration parallel thruster method based on digital twinning of the present application;

[0041] Figure 6 is a deep sea exploration parallel thruster path planning diagram based on digital twinning of the present application;

[0042] Figure 7 is a deep sea exploration parallel thruster anti-flow compensation solution flowchart based on digital twinning of the present application;

[0043] Figure 8 is a sensor data link diagram of a deep sea exploration parallel thruster based on digital twinning of the present application;

[0044] Figure 9 is a human-machine interaction interface of a deep sea exploration parallel thruster based on digital twinning of the present application.

[0045] BRIEF DESCRIPTION OF DRAWINGS

[0046] Figure 4 In the figure: 1, moving platform; 2, fixed platform; 3, first active branch chain; 4, second active branch chain; 5, third active branch chain; 6, constraint branch chain; 7, propeller; 31, hooke joint; 32, moving unit; 33, hooke joint; 61, composite spherical hinge. DETAILED DESCRIPTION

[0047] To further understand the deep sea exploration parallel thruster technical framework and method based on digital twinning and virtual PLC provided by the present application, the following is a specific implementation description of the present application. It should be particularly pointed out that the orientation description involved in the present application is only based on the orientation relationship shown in the drawings.

[0048] In the detailed embodiments section, it should be specifically noted that the directional terms used in this invention include, but are not limited to, spatial descriptions such as "upper part," "lower part," "top," "bottom," "inner edge," and "outer edge," all of which are set with reference to the relative positional relationships presented in the accompanying drawings and are intended to more intuitively describe the technical solutions of this invention. They should not be construed as limiting the scope of protection of this invention. Furthermore, the sequential descriptions such as "first," "second," and "third" are only used to distinguish technical features and do not represent any priority or importance.

[0049] Example 1:

[0050] like Figures 1-2 As shown, the present invention provides a deep-sea exploration parallel thruster technology architecture based on digital twins, including a physical layer, an interaction layer, a virtual layer, a decision layer, and a communication layer.

[0051] The physical layer includes a three-degree-of-freedom parallel thruster and physical sensors. The three-degree-of-freedom parallel thruster performs physical operations, and various physical sensors collect information on the working environment and equipment status data. With the help of the communication layer, it can provide data support for the virtual layer to issue subsequent work instructions and for the decision layer to further optimize instructions, and update monitoring data in real time to achieve closed-loop optimization.

[0052] like Figures 3-4 As shown, a three-degree-of-freedom parallel thruster is a mechanical device with an ingenious structure and a clearly defined function. Its overall structure consists of a moving platform 1, a fixed platform 2, a first active branch 3, a second active branch 4, a third active branch 5, a constraint branch 6, and a propeller 7. In the structural layout of this device, the high-precision Hooke joints 31 connected to the moving platform 1 are distributed in an equilateral triangle pattern. This equilateral triangle distribution allows the high-precision Hooke joints 31 to be evenly stressed on the moving platform 1, thus providing a stable foundation for subsequent motion transmission. Similarly, the Hooke joints 33 connected to the fixed platform 2 are also distributed in an equilateral triangle pattern on the fixed platform 2, which echoes the distribution of the high-precision Hooke joints 31 on the moving platform 1, further ensuring the structural symmetry and stability of the entire device. The propeller 7 is directly fixed to the center of the moving platform 1, and the central axis of the connecting shaft of the Hooke joint 61 coincides with the intersection of the two perpendicular lines passing through the centers of the moving and fixed platforms, making the entire three-degree-of-freedom parallel thruster form a symmetrical configuration. This symmetrical configuration can achieve stable motion and effectively reduce the additional stress and interference caused by structural asymmetry during motion.

[0053] The motion control of the device is realized by the first active branch chain 3, the second active branch chain 4, the third active branch chain 5, the constraint branch chain 6 and the propeller 7; wherein the first active branch chain 3, the second active branch chain 4 and the third active branch chain 5 are provided with moving pairs, and by accurately controlling the extension and contraction of these moving pairs, the entire motion platform can be driven to complete the motion of three rotational degrees of freedom; during the motion process, the fixed platform 2 serves as the basic component of the entire device and provides a stable installation and support reference for other components; the moving platform 1 is the execution component of the motion and moves according to the predetermined trajectory and attitude under the driving of the first active branch chain 3, the second active branch chain 4 and the third active branch chain 5; the first active branch chain 3, the second active branch chain 4 and the third active branch chain 5 transmit power to the moving platform 1 through the extension and contraction of the servo-driven moving pairs, and the constraint branch chain 6 plays a role in limiting unnecessary motion, ensuring the motion accuracy and stability of the moving platform 1 during the motion process, so that the moving platform 1 can only move within the set three-rotational-degree-of-freedom range, avoiding excessive motion interference and ensuring that the entire device can accurately complete the expected motion task; the propeller 7 adjusts the thrust direction synchronously with the attitude change of the moving platform 1, directly converts the three-degree-of-freedom motion of the moving platform into dynamic adjustment of the thrust direction of the propeller, realizes vector propulsion, and meets the requirements of flow compensation and precise pose control.

[0054] The physical sensor is arranged in the parallel thruster, the active branch chain 3, the active branch chain 4 and the active branch chain 5 of the parallel thruster are driven by a waterproof brushless DC motor, an encoder is embedded in the output shaft of the motor, and the joint rotation angle is fed back to the main controller in real time; the moving platform 1 is installed with a high-precision IMU and is strictly aligned with the mass center of the moving platform, and is used for continuously collecting attitude angle and angular velocity signals; the thruster is installed with a water pressure gauge in the direction close to the axis of the propeller 7 for detecting the working environment water pressure; the sonar sensor includes a front-mounted sonar, a side-mounted sonar and a bottom-mounted sonar, the front-mounted sonar is installed on the static platform of the thruster and faces the forward direction, the side-mounted sonar is asymmetrically distributed and installed on the side of the thruster and faces the outside of the thruster, and the optimal installation position of the bottom-mounted sonar is the center of the bottom of the moving platform and is aligned with the IMU; the sonar provides data support for the Unity to build a dynamic seabed digital twin environment by emitting sound waves and receiving seabed reflected echoes, and adopts a multi-directional sonar arrangement to accurately complete real-time obstacle recognition, target positioning and motion tracking.

[0055] As Figure 8As shown, the data collected by the physical sensors are used for the model construction of virtual fluid environment building, fluid-rigid body motion coupling analysis and time series convolution network learning, the real-time pressure provided by the water pressure gauge is used to calculate the fluid dynamic parameters, the water temperature of the temperature sensor is used to calibrate the fluid viscosity characteristics, the obstacle position and seabed topography of the sonar array are used to define the fluid simulation boundary, the IMU and encoder are used to calculate the thruster pose, the rigid body dynamics equation is established, the fluid resistance equation based on CFD is combined with the rigid body dynamics equation to form the coupling differential equation, and the theoretical basis for flow compensation is provided.

[0056] The virtual layer includes a digital virtual world and a virtual control unit, and through high-fidelity modeling and real-time simulation, dynamic mapping of the physical world and pre-rehearsal of control instructions can be realized, and with the help of the communication layer OPC protocol, control of the parallel thruster of the physical layer can be completed.

[0057] The digital virtual world includes a digital twin and a virtual fluid environment. The digital twin creation process can refer to the following steps: first, based on SolidWorks, a precise parameterized three-dimensional model of the thruster is built, and a three-dimensional dynamics model is exported in STL format. The STL format model is imported into Blender software, and the conversion of the three-dimensional dynamics model from STL format to FBX format is completed with the help of Blender. The FBX format of the three-dimensional dynamics model is imported into Unity, and the digital twin of the parallel thruster can be built.

[0058] The degree of fit between the virtual fluid environment and the real underwater environment mainly depends on the underwater environment parameters transmitted in real time by the physical sensors. According to the data of underwater pressure, temperature, obstacle position and other data of the parallel thruster detected by various physical sensors, the current velocity profile and seabed topography are analyzed, the non-uniform flow field boundary conditions are provided for the particle system of Unity, the dynamic fluid particles are driven to generate, and the virtual simulation of the fluid environment is realized.

[0059] As shown in Figure 9 The interactive layer, i.e. all functional modules of the human-computer interaction interface, are integrated into the same interface. Specific function instructions can be expanded by the operator clicking the corresponding module button, so as to perform the next step of work monitoring or instruction issuing. Specifically, it includes a virtual-real synchronous monitoring area and a control instruction issuing module. The virtual-real synchronous monitoring area includes a digital twin visualization interface and a monitoring interface displaying the deep sea exploration work value of the parallel thruster, the digital twin work driving value and the difference between the two work values, so that the operator can view the work progress and visualize the fault in real time. The control instruction issuing module provides a manual control mode and an automatic task mode, which jointly act on the virtual control unit to control the parallel thruster.

[0060] The human-computer interaction interface in the creation process follows the layered rendering of Canvas in Unity. First, a C# script containing function functions is written on the Unity platform. By mounting UI elements under Canvas and mounting function functions under C# scripts under UI elements, function buttons are set to have functions such as automatic monitoring of work values and work instruction issuing. The virtual-real synchronous monitoring area function button provides value function buttons, including viewing of parallel thruster deep sea exploration work values, digital twin work driving values and differences between the two work values. The expansion operation of the parallel thruster deep sea exploration work value function button includes a deep sea resource detection display interface. The manual control mode supports horizontal, vertical and vertical displacement, i.e. "forward", "backward", "left", "right", "up" and "down" directions. If abnormal work data is sensed in the virtual-real synchronous monitoring area, the operator can directly manually control the parallel thruster to stop operation.

[0061] The virtual control unit receives and processes joint pose angles transmitted by physical sensors, responds to task instructions issued by the virtual layer human-computer interaction interface, determines the next action, and completes work pose calculation. With the help of PLC ladder diagram programming of Baidu software, C# scripts are written on the Unity side, control signals are generated based on virtual PLC, digital twin motion is driven and work instructions are executed, and the action of the digital twin is synchronized to the parallel thruster with the help of the communication layer.

[0062] The decision layer includes an autonomous decision engine and a predictive maintenance center, which is used to receive simulation data output by the virtual layer, unify management, processing and supply of multi-source heterogeneous data through data driving, and support real-time decision optimization of the virtual layer. It has the function of directly connecting and turning on and off the parallel thruster. The decision layer time convolution network TCN trains the prediction model through three core mechanisms of causal convolution, dilated convolution and residual connection, which can realize efficient modeling of time series data. The autonomous decision engine is based on the time convolution network TCN, analyzes sonar obstacle data and historical path sequence, predicts safe trajectory through causal convolution, generates flow-resistant compensation matrix and path re-planning instruction, and outputs optimization instruction to the virtual layer; the predictive maintenance center receives encoder vibration spectrum and temperature time series data based on the time convolution network TCN model, and gives early warning and maintenance for equipment failure. If necessary, bypass the virtual layer rehearsal link, force the parallel thruster to stop operation, and reduce the probability of parallel thruster failure.

[0063] The communication layer selects an OPC protocol for communication, and the OPC protocol is a general communication standard in the field of industrial automation, and can solve the interoperability problem between devices and software of different manufacturers. Specifically, an OPC UA protocol can be selected, the OPC UA protocol supports cross-platform information transmission, realizes real-time transmission of physical sensor data and control instructions through a high-reliability protocol and memory optimization, and is suitable for complex environments of deep sea exploration. The communication layer specifically includes a bidirectional data interaction channel and a shared memory area. The bidirectional data interaction channel is based on the OPC protocol to realize bidirectional data synchronization between the virtual environment and the physical device, so as to realize the downward transmission of the control instruction of the virtual layer to the parallel thruster and the upward transmission of the physical sensor feedback to the virtual layer. The shared memory area, that is, the OPC shared memory area, can ensure that the virtual layer provides high-fidelity simulation results to the decision layer, and realizes the feedback of the re-planned path of the decision layer to the virtual layer for pre-rehearsal verification, without causing congestion of the communication layer.

[0064] Embodiment 2

[0065] As shown in the figure, the underwater parallel thruster technical framework and method based on digital twinning provided by the application realize the following steps: Figures 1-5

[0066] S1: a digital twin of the parallel thruster is constructed in Unity, the physical sensor collects fluid dynamics parameters and thruster pose information in real time, an OPC server is deployed on the edge controller of the physical layer, a communication interface with the PLC is configured, the server binds the physical sensor data address, and the fluid dynamics parameters and the pose information are transmitted to the Unity digital twin through the OPC uplink;

[0067] S2: the human-computer interaction interface triggers the working instruction of the digital twin, drives the digital twin to plan the path and the working pose, and uploads the planning instruction to the OPC shared memory partition;

[0068] As shown in the figure, the specific sub-steps of step S2 are as follows: Figure 6

[0069] S21: the path planning uses a CCD IK plug-in of Unity, which can automatically calculate the movement data according to the target pose, so that the parallel thruster can reach the target pose in an instant, but the process cannot be displayed, so the same parallel thruster FBX format model as the digital twin needs to be imported again in Unity, a mesh renderer is hidden, the function of model invisibility is realized, and the model exists as a shadow model of the digital twin;

[0070] ​​S22: Add CCD IK plug-in to the shadow model, perform parallel thruster inverse kinematics analysis, automatically calculate thruster joint position parameters according to the target position and attitude input by the human-computer interaction interface, so that the parallel thruster shadow model can follow the set point planning path;

[0071] S23: Integrate the OPC client library in the Unity virtual layer, push the control instructions sent by the digital twin to the physical device to the server through asynchronous writing, and store them in the shared memory partition;

[0072] S24: Perform PLC ladder programming in BoT, transmit the target pose to Unity through the S7.NetPlus plug-in, and drive the virtual parallel thruster to move;

[0073] S3: Use S7-PLCSIM Advanced software to realize the linkage between BoT PLC and Unity, add a C# script in Unity to receive BoT data, and transmit it to the parallel thruster through OPC downward transmission to perform physical work;

[0074] S4: Real-time display of virtual and real synchronization monitoring area during task execution Parallel thruster and digital twin working information, autonomous decision engine dynamically generates PWM optimization instructions based on anti-flow compensation matrix to correct the deviation between current virtual pose and target virtual pose;

[0075] In step S4, relying on the predictive maintenance center established by the time convolution network TCN, if a fault prediction is made based on the current working data of the parallel thruster, the autonomous decision engine stops generating optimization instructions, and the predictive maintenance center generates a predictive maintenance scheme for the predicted fault. If the maintenance range is exceeded, the parallel thruster work will be forced to stop, if the fault can be maintained, after successful fault maintenance, the parallel thruster will continue to work;

[0076] As shown in Figure 7 The anti-flow compensation matrix PWM control signal is based on motion coupling analysis of the dynamic model in the virtual fluid environment for pose compensation, which includes the following steps:

[0077] S41: The time convolution network TCN obtains fluid dynamics parameters from the physical sensor, and simultaneously solves the rigid body motion equation of the thruster connecting rod and the fluid resistance equation based on CFD to form a coupled differential equation group:

[0078]

[0079] Where q is the joint pose, M(q) is the mechanism mass matrix, G(q) is a static torque matrix generated by gravity, τ is an effective dynamic torque input, D is a resistance torque matrix generated by the fluid velocity field v, J(q) is a Jacobian matrix, and T represents the transpose of a matrix;

[0080] S42: Call the MKL mathematical library for linear algebra solution, real-time solve the coupling effect, output the joint pose q containing fluid disturbance;

[0081] S43: With the help of the communication layer OPC protocol, map the dynamic parameters, write the fluid resistance torque D in the solution result and the joint pose q containing fluid disturbance into the OPC shared memory address;

[0082] S44: Based on the calculated pose deviation and the mapped dynamic parameters, real-time update the motion planning, the virtual control end generates the corresponding PWM optimization compensation signal, and corrects the cumulative error between the current pose of the digital twin and the target pose;

[0083] S5: Repeat step S3 to correct the virtual pose and the actual pose of the parallel thruster;

[0084] S6: The OPC protocol further synchronizes the environment and motion parameters obtained by the physical sensor to the digital twin model, optimizes the control parameters in a closed loop, and realizes real-time updating of data.

[0085] The specific implementation of the deep-sea exploration parallel thruster technical framework and method based on digital twinning disclosed in the patent is not limited to the implementation forms listed in the foregoing. For those skilled in the art, based on the technical concept and core method disclosed in the present application, the same technical effect can be achieved through other equivalent or alternative implementation forms without departing from the essential content of the present application. The foregoing specific embodiments are only exemplary descriptions given to facilitate understanding of the present application, and are not intended to limit the scope of protection of the present application.

Claims

1. A deep-sea exploration parallel thruster technical framework based on digital twinning, characterized in that: The physical layer includes a three-degree-of-freedom parallel thruster and a physical sensor; the three-degree-of-freedom parallel thruster includes a moving platform, a fixed platform, a main driven branch chain, a constraint branch chain and a propeller, the main driven branch chain includes first, second and third main driven branch chains which are structurally identical and circumferentially distributed between the moving platform and the fixed platform, the first, second and third main driven branch chains are sequentially connected by a hook joint, a servo-driven moving pair and a hook joint, the constraint branch chain is arranged along the central axis of the moving platform and the fixed platform, and the constraint branch chain is sequentially connected by a constraint rotating pair and a hook joint, so that the moving platform realizes three-rotating-degree-of-freedom movement; the physical sensor is installed on the three-degree-of-freedom parallel thruster and includes a water pressure gauge, a temperature sensor, a sonar sensor, an inertial measurement unit (IMU) and an encoder; The interactive layer is a human-computer interaction interface in a Unity environment, and includes a virtual-real synchronous monitoring area and a control instruction issuing module; the virtual-real synchronous monitoring area includes a digital twin visualization interface and a monitoring interface for displaying deep-sea exploration work values of the parallel thruster, digital twin work driving values and work value differences between the two; the control instruction issuing module controls the parallel thruster in an instruction mode, and provides a manual control mode and an automatic task mode; The virtual layer includes a digital virtual world and a virtual control unit; the digital virtual world includes a digital twin and a virtual fluid environment; the digital twin uses SolidWorks to build a thruster three-dimensional geometric model, converts the format, and builds the digital twin based on Unity; the virtual fluid environment generates dynamic fluid particles based on the underwater pressure, temperature and obstacle position of the parallel thruster detected by the physical sensor, the current velocity profile and the seafloor topography, and is driven by the Unity particle system; the virtual control unit receives and processes the sensing parameters of the physical sensor and the task instructions issued by the interactive layer, generates control signals by the virtual PLC through the linkage of B&R and Unity, drives the digital twin to move, executes the work instructions, and synchronously moves to the parallel thruster by means of the communication layer; The decision layer includes an autonomous decision engine and a predictive maintenance center, and receives simulation data output by the virtual layer; the autonomous decision engine is based on a time series convolution network (TCN), analyzes sonar obstacle data and historical path sequences, generates an anti-flow compensation matrix and a path re-planning instruction, and outputs the optimized instruction to the virtual layer by means of the communication layer shared memory partition; the predictive maintenance center relies on the TCN model, receives the encoder vibration spectrum and temperature time series data, trains the prediction model according to the historical data set, and is directly connected with the physical layer; The communication layer includes a bidirectional data interaction channel and a shared memory partition; the bidirectional data interaction channel synchronizes bidirectional data between the virtual environment and the physical device based on the OPC protocol, and includes virtual layer control instruction downward and sensor data feedback upward; the shared memory partition is an OPC shared memory partition, and the virtual layer and the decision layer interact with each other. ​ 2. The deep sea exploration parallel thruster technical framework based on digital twinning of claim 1, wherein: The three-degree-of-freedom parallel thruster has a symmetrical configuration, and the central axis of the shaft is connected by a Hooke joint, and passes through the intersection of two perpendicular lines of the center of the moving platform and the static platform.

3. The deep sea exploration parallel thruster technical framework based on digital twinning of claim 1, wherein: All function modules of the interaction layer are integrated in the same human-computer interaction interface, the creation of the human-computer interaction interface follows the Canvas layered rendering in Unity, a C# script with a function function is written in the Unity platform, and the function function is mounted under the UI element, the function buttons of the virtual-real synchronous monitoring area and the control instruction issuing module are set, and the operation instruction is unfolded by clicking the function button; the function button of the virtual-real synchronous monitoring area provides a numerical monitoring function button, including viewing of parallel thruster deep sea exploration work values, digital twin work driving values and differences between the two work values, and the unfolding operation of the parallel thruster deep sea exploration work value monitoring function button includes a deep sea resource detection data display interface.

4. A method for deep sea exploration parallel thruster based on digital twinning, characterized in that: The method implementation steps are as follows: S1: build the digital twin of the parallel thruster in Unity, the physical sensor collects fluid dynamics parameters and thruster pose information in real time, deploy an OPC server in the physical layer edge controller, configure the communication interface with the PLC, bind the physical sensor data address to the server, and transmit the fluid dynamics parameters and pose information to the Unity digital twin through OPC uplink transmission; S2: the human-computer interaction interface triggers the digital twin work instruction to drive the digital twin to plan the path and the operation pose, and uploads the planning instruction to the OPC shared memory partition at the same time; S3: use S7-PLCSIM Advanced software to realize the linkage of B&R PLC and Unity, add a C# script for receiving B&R data in Unity, and transmit it to the parallel thruster through OPC downlink to execute physical operation; S4: the virtual-real synchronous monitoring area displays the parallel thruster and digital twin work information in real time during task execution, and the autonomous decision engine dynamically generates PWM optimization instructions based on the anti-flow compensation matrix according to the original planning instruction uploaded to the OPC shared memory partition, to correct the deviation between the current virtual pose and the target virtual pose; S5: repeat step S3 to correct the virtual pose and actual pose of the parallel thruster; S6: further synchronize the environment and motion parameters obtained by the physical sensor to the digital twin model through the OPC protocol, optimize the control parameters in a closed loop, and realize real-time data updating.

5. The method according to claim 4, wherein the method is a method for deep sea exploration based on digital twinning, characterized in that: The motion path planning includes the following steps: S21: import the FBX format model of the parallel thruster which is exactly the same as the digital twin in Unity again, hide the mesh renderer, realize the function of model invisibility, and make it exist as a shadow model of the digital twin; S22: add the CCD IK plug-in to the shadow model, perform inverse kinematics analysis of the parallel thruster, automatically calculate the joint position parameters of the thruster according to the target position and attitude input by the human-computer interaction interface, and make the shadow model of the parallel thruster plan the path according to the set point position. S23: Integrate the OPC client library in the Unity virtual layer, push the control instructions sent by the digital twin to the physical device through asynchronous write instructions to the server, and store them in the shared memory partition; S24: Perform PLC ladder programming in the blog, transmit the target pose to Unity through the S7.NetPlus plug-in, and drive the virtual parallel thruster to move.

6. The method of claim 4, wherein: The anti-flow compensation matrix PWM control signal is based on the motion coupling analysis of the dynamic model in the virtual fluid environment, extracts the ocean current disturbance pattern through the time convolution network TCN inflation convolution layer, outputs the thrust correction parameter matrix, and performs pose compensation, specifically including the following steps: S41: The time convolution network TCN acquires fluid dynamics parameters in advance according to the physical sensor, establishes a learning model according to historical parameters, and simultaneously solves the rigid body motion equation of the thruster connecting rod and the fluid resistance equation based on CFD to form a coupled differential equation group: where q is the joint pose, M(q) is the mechanism mass matrix, is the Coriolis matrix, G(q) is the static torque due to gravity, τ is the effective dynamic torque input, D is the drag matrix generated by the fluid velocity field v, J(q) is the Jacobian, and T denotes the transpose of a matrix; S42: Call the MKL mathematical library to solve linear algebra, real-time solve the coupling effect, and output the joint pose q containing fluid disturbance; S43: With the help of the communication layer OPC protocol, map the dynamic parameters, and write the fluid resistance moment D in the solution result and the joint pose q containing fluid disturbance into the OPC shared memory address; S44: Based on the calculated pose deviation and the mapped dynamic parameters, real-time update the motion planning, the virtual control end generates the corresponding PWM optimized compensation signal, and corrects the cumulative error between the current pose and the target pose of the digital twin.

7. The method of claim 4, wherein: In step S4, relying on the TCN model predictive maintenance center, the working data that may cause abnormal parallel thruster is monitored in the virtual-real synchronous monitoring area, the parallel thruster fault is predicted in advance, the predictive maintenance scheme is generated, and when the maintenance range is exceeded, the parallel thruster operation will be forced to stop.

Images