A virtual-real simulation system with linkage between a centralized operation console and a 3D model

Abstract

The application relates to the technical field of industrial automation control and dynamic simulation, and discloses a virtual-real simulation system with linkage of a centralized control operation platform and a 3D model, which comprises a control signal sampling interface, a load response modulation unit, a dynamic model evolution unit and a virtual-real closed-loop feedback unit. The load response modulation unit accumulates energy consumption weights of all virtual controlled objects in transient motion to generate a global time scaling factor for real-time correction of the inertial time constant of the system. The dynamic model evolution unit introduces a hard limiting boundary by using a limiting link and calculates process variables. The virtual-real closed-loop feedback unit extracts a control deviation signal and drives physical instruments to generate nonlinear damping actions. The application establishes a global load modulation model based on energy constraints, and under the condition of low computing power without complex fluid network solving, makes the discrete control system spontaneously emerge nonlinear hysteresis characteristics conforming to physical energy conservation.

Description

A virtual-real simulation system that links a centralized control console with a 3D model Technical Field

[0001] This invention belongs to the field of industrial automation control and dynamic simulation technology, and in particular relates to a virtual-real simulation system that links a centralized control console with a 3D model. Background Technology

[0002] In current large-scale industrial scenarios such as coal mining and chemical processes, the centralized control console serves as the output center for production instructions and the hub for sensing the status of field equipment. To improve training effectiveness and verify the reliability of control logic, the industry widely adopts virtual simulation systems. These systems drive virtual equipment models through physical control consoles, simulating real production process actions and technological responses. The core of this approach is to establish a high-fidelity mapping relationship between discrete control logic instructions and continuous physical process variables, ensuring that the response characteristics of virtual environment equipment are consistent with the real physical world.

[0003] Existing industrial simulation technologies generally adopt an idealized state mapping modeling approach. The discrete switching signals output from the control console directly trigger pre-made linear animation clips in the 3D engine, ignoring the inherent inertia, damping, and nonlinear response characteristics of the physical objects. While this can meet basic operational demonstration requirements under steady-state conditions, to improve the distortion of simulations based solely on logical mapping, the technology field is attempting to introduce dynamic models to enhance the accuracy of virtual-real synchronization. For example, Chinese invention patent CN116047889B discloses a control compensation method and device in a virtual-real combined simulation system. By constructing a dynamic model, it utilizes a PID algorithm to control the virtual... Coupled with the acceleration calculation of physical objects, this solution solves the problem of motion trajectory tracking lag. However, this solution focuses on the kinematic parameter correction of a single object under ideal energy supply. In essence, it still treats the controlled object as an isolated load node and does not address the physical nature of the systemic lag caused by the limited bus energy in multi-device concurrent scenarios. Although the existing control compensation strategy can correct trajectory errors, it cannot simulate the nonlinear decay characteristics caused by the sudden drop in hydraulic mains pressure or the drop in grid voltage. It lacks a mechanism to convert the virtual side mechanical jamming or hard limit state into the physical side tactile feedback reverse projection, resulting in the operator still being in a state of visual observation rather than sensory access isolation.

[0004] Therefore, the technical problem to be solved by this invention is how to construct a dynamic simulation system for industrial processes that resolves the spatiotemporal heterogeneity contradiction between discrete logic and continuous processes in a resource-constrained computing environment, and realistically reproduces the multivariable physical coupling effect and the reverse fault feedback mechanism. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A virtual-real simulation system that links a centralized control console with a 3D model, comprising:

[0006] The control signal sampling interface is connected to the physical control port of the centralized control console. It is used to sample and hold discrete operation command signals at a preset control cycle to generate a target state vector.

[0007] The load response modulation unit is used to establish an energy constraint model of the virtual controlled object. In each control cycle, the energy consumption weights of all virtual controlled objects in transient change process are weighted and summed to obtain the real-time load rate of the system. Based on the deviation between the real-time load rate of the system and the preset system capacity limit, a global time scaling factor is generated to characterize the energy redundancy of the system. The global time scaling factor is injected into the dynamic model evolution unit.

[0008] The dynamic model evolution unit is embedded with transfer function parameters describing the dynamic characteristics of the virtual controlled object. The transfer function parameters include at least the inertial time constant and the steady-state gain. The dynamic model evolution unit is used to perform real-time gain scheduling correction on the inertial time constant using a global time scaling factor, and to perform discretized integration on the target state vector based on the corrected inertial time constant to generate instantaneous values ​​of process variables that meet the system energy constraints. The dynamic model evolution unit is also used to introduce hard limit boundary values ​​from the simulation environment, and to clamp the instantaneous values ​​of process variables within the feasible region defined by the hard limit boundary values ​​through a limiting element, thereby generating constrained process variables.

[0009] The virtual-real closed-loop feedback unit is used to map the constrained process variables into driving parameters to drive the pose state evolution of the virtual controlled object. At the same time, it extracts the control deviation signal between the target state vector and the constrained process variables in real time. When the amplitude of the control deviation signal continuously exceeds the preset dead zone threshold, it converts the control deviation signal into a physical reverse driving voltage to drive the electromechanical indicating device on the centralized control console to generate a nonlinear damping action corresponding to the control deviation signal.

[0010] Preferably, the load response modulation unit is configured with nonlinear hysteresis mapping logic; the nonlinear hysteresis mapping logic is used to map the global time scaling factor to a value greater than 1 according to the preset overload-hysteresis characteristic curve when the real-time load rate of the system exceeds the unit reference value, so that the inertial time constant in the dynamic model evolution unit increases monotonically with the increase of the real-time load rate of the system, thereby introducing additional delay characteristics determined by the system load in the step response process of the virtual controlled object.

[0011] Preferably, the dynamic model evolution unit is equipped with environmental parameter drift compensation logic; the system also includes an environmental state setting unit for setting a global environmental scalar that characterizes the physical properties of the simulation environment; the environmental parameter drift compensation logic is used to calculate the environmental deviation between the global environmental scalar and the preset rated operating condition value, and uses the environmental deviation to generate parameter drift coefficients; the dynamic model evolution unit uses the parameter drift coefficients to perform secondary correction on the inertial time constant to simulate the dynamic response characteristic drift of the virtual controlled object under different physical environments; wherein the calculation logic of the parameter drift coefficient β follows the following relationship:

[0012] β=1+K env ×(E curr -E nom );

[0013] Among them, E curr E is a global environmental scalar. nom K is the preset rated operating condition value. env To achieve the preset environmental sensitivity gain, the dynamic model evolution unit uses the parameter drift coefficient β to multiply and correct the inertial time constant.

[0014] Preferably, the dynamic model evolution unit is also equipped with cumulative fatigue decay logic; the cumulative fatigue decay logic is used to integrate and accumulate the absolute value of the rate of change of the instantaneous value of the process variable in the time domain to generate a cumulative travel variable that represents the total amount of motion of the virtual controlled object throughout its entire life cycle; when the cumulative travel variable exceeds the preset aging inflection point threshold, the dynamic model evolution unit superimposes an irreversible decay bias on the steady-state gain or the inertial time constant, so that the dynamic characteristics of the virtual controlled object exhibit a gradual performance decay as the total amount of motion increases.

[0015] Preferably, the virtual-real closed-loop feedback unit is configured with deviation-tactile mapping logic; the deviation-tactile mapping logic is used to map the amplitude of the control deviation signal to the duty cycle parameter of the pulse width modulation signal, and output the pulse width modulation signal as a physical reverse drive voltage to the analog instrument coil of the central control console, driving the pointer of the analog instrument to generate mechanical oscillations with a frequency related to the control deviation signal near the indicated position, so as to physically reproduce the stall condition of the virtual controlled object caused by the hard limit boundary value limitation; and the electromechanical indicating device of the central control console includes an analog pointer instrument and a force feedback actuator; the virtual-real closed-loop feedback unit is used to output the constrained process variable as a feedback quantity to the analog pointer instrument in a closed loop, and use the control deviation signal to control the force feedback actuator to output damping torque, so as to realize visual and tactile dual-channel physical feedback based on the model state.

[0016] Preferably, the dynamic model evolution unit is configured with bidirectional disturbance-free switching logic; when the operation command signal causes a logic flip before the virtual controlled object reaches a steady state, the bidirectional disturbance-free switching logic takes the instantaneous value of the process variable at the current moment as the initial state, switches to the reverse integration path, and performs reverse evolution calculation based on the reset time constant to ensure that the instantaneous value of the process variable remains continuous and without step change at the moment of command switching.

[0017] Preferably, the control signal sampling interface is configured with signal conditioning and clock synchronization logic; the signal conditioning and clock synchronization logic is used to filter out high-frequency noise interference in the operation command signal and align the dejittered valid command signal to the operation clock edge of the dynamic model evolution unit, so as to eliminate sampling jitter error in the discrete control system.

[0018] Preferably, the hard limit boundary value is a variable generated in real time by the geometric constraints or fault preset state of the virtual controlled object in the simulation space; the limiting element is used to introduce nonlinear saturation characteristics into the integral operation loop of the dynamic model evolution unit, so that the numerical trajectory of the constrained process variable cannot exceed the hard limit boundary value, thereby artificially creating and maintaining the control deviation signal in the control loop. The load response modulation unit, the dynamic model evolution unit, and the virtual-real closed-loop feedback unit constitute a deterministic real-time digital control loop; the transfer function parameters and process characteristic curves are stored in a fixed-point lookup table manner, and the integral operation is solved iteratively using a difference equation with a preset step size, without involving any matrix transformation of the graphics rendering pipeline or floating-point physics engine calculation.

[0019] Compared with the prior art, the beneficial effects of the present invention are:

[0020] 1. In the centralized control console and 3D model, a non-physical field energy coupling simulation mechanism is constructed by using the load response modulation unit in conjunction with the dynamic model evolution unit. This eliminates the traditional solution path of complex differential equations in fluid or power network simulation. The weighted aggregation algorithm is used to calculate the real-time total load value and map it to the global time dilation coefficient. The coefficient is used to correct the standard time constant of each controlled object in real time. Under high load conditions with multiple devices operating concurrently, the system automatically exhibits overall action hysteresis or nonlinear decay characteristics that conform to the law of energy conservation. With extremely low computational resource consumption, this solves the problem of isolated islands due to lack of physical correlation between discrete control models. This ensures that the simulation system can reproduce the real physical response of complex pipeline or bus system under transient loads in environments with limited resources such as edge computing devices or industrial gateways.

[0021] 2. By utilizing the limiting element in the dynamic model evolution unit in conjunction with the virtual-real closed-loop feedback unit, a hard logic link is established to transmit virtual constraints to the physical terminal in reverse. When the virtual clamping threshold is triggered during the evolution of the controlled object, the process variable numerical integration path is locked, and a continuous residual is generated with the target logic value of the control command. The virtual-real closed-loop feedback unit captures the residual signal and converts it into a physical reverse drive voltage, which drives the electromechanical indicator device on the centralized control console to generate a slight jitter or abnormal flashing corresponding to the degree of fault. This upgrades the traditional one-way status display to a two-way closed-loop feedback based on control deviation. Operators do not need to rely on visual images; they can perceive mechanical jamming, travel limitation, or overload faults of remote equipment through the dynamic response of the physical terminal, thus improving the depth of human-machine interaction condition perception.

[0022] 3. An embedded response feature description dynamic model evolution unit is introduced to establish an intermediate buffer layer with inertial memory between discrete operation command signals and continuous controlled object states. The step command is integrally processed according to a predefined nonlinear characteristic curve to generate continuously changing normalized process variables. When the control command is interrupted midway, frequently reversed, or undergoes logic jitter and unsteady changes, it smoothly switches to the reset or reverse evolution path according to the current integral state point to avoid instantaneous jumps in state variables. This ensures that the simulation model maintains continuity in accordance with the laws of physical inertia under extreme operation logic, eliminates the tearing phenomenon between logic state and process state under transient conditions, and ensures the rigor of control logic verification. Attached Figure Description

[0023] Figure 1 is a schematic diagram of the virtual-real closed-loop control principle of the linkage between the centralized control console and the 3D model of the present invention.

[0024] Figure 2 is a comparison of load response and hysteresis characteristics under multi-device concurrent operation conditions of the present invention;

[0025] Figure 3 is a diagram of the overall system architecture of the present invention, which combines the physical entity domain and the digital twin domain. Detailed Implementation

[0026] The principles and features of the present invention are described below with reference to the accompanying drawings. The examples given are only for explaining the present invention and are not intended to limit the scope of the present invention.

[0027] This invention provides a virtual-real simulation system that links a centralized control console with a 3D model. Based on the mapping between discrete control logic and continuous physical processes, it solves the technical problem of spatiotemporal decoupling of commands and responses in traditional simulations by constructing a digital closed-loop control loop with energy constraints, environmental sensitivity, and inverse physical feedback. The system consists of a control signal sampling interface, a load response modulation unit, a dynamic model evolution unit, and a virtual-real closed-loop feedback unit. These units interact in real time via a deterministic data bus, forming a pure logic operation kernel independent of the graphics rendering pipeline. The control signal sampling interface, serving as the physical boundary of the system, is directly connected to the electrical output terminals of the centralized control console via hardwiring. Its main function is... The interface converts non-standard, noisy discrete switching signals output from the control panel into standard logic vectors that the system can process. Given the common contact jitter and electromagnetic interference in industrial environments, the interface is equipped with signal conditioning and clock synchronization logic. It performs hardware-level or software-level de-jitter processing on the acquired raw level signals. The specific de-jitter algorithm adopts the time window integration method, that is, within a preset sampling period, such as 10ms, the signal is considered valid only when the level signal remains consistent for more than 80% of the time window. The processed valid command signal is aligned to the operation clock edge of the dynamic model evolution unit to generate a target state vector without timing deviation, thereby eliminating the control logic competition risk caused by asynchronous sampling.

[0028] To address the simulation distortion caused by neglecting the energy coupling effect between controlled objects in existing technologies, the system introduces a load response modulation unit to establish a global energy constraint model. In real industrial hydraulic or electric systems, the limited bus capacity leads to a systematic decrease in driving capability when multiple high-power actuators operate simultaneously. To address this physical fact, the load response modulation unit traverses all virtual controlled objects in a non-steady state (i.e., whose current process variables are not equal to the target state vector) within each control cycle, extracting a preset energy consumption weight parameter W. i This parameter characterizes the degree of system resource consumption of a specific device at the moment of action. This unit performs a weighted summation operation to obtain the real-time system load rate R, which follows the formula: R = (ΣW i ) / W limit ;

[0029] Among them W limitThe system capacity threshold is preset and is calibrated based on the rated power of the actual physical station or the upper limit of the pump station flow. Based on the calculated load rate R, the unit uses nonlinear hysteresis mapping logic to generate a global time scaling factor λ. This mapping logic adopts a piecewise function strategy: when R≤1, λ remains at 1.0; when R>1, λ grows exponentially according to the preset overload-hysteresis characteristic curve. This global time scaling factor λ is injected into the evolution calculation of each controlled object, lengthening the response time. Thus, without running a high-computing-power fluid network solution model, a systemic hysteresis phenomenon that conforms to the law of conservation of energy spontaneously emerges, allowing operators to intuitively perceive the implicit overload condition of the system.

[0030] The dynamic model evolution unit is the core computing engine of this system, responsible for transforming discrete control commands into continuously changing normalized process variables. This unit abandons the dependence on the graphics rendering frame rate and instead uses discretized integration operations based on difference equations with a fixed time step. For each virtual controlled object, this unit internally stores transfer function parameters describing dynamic characteristics, including at least the inertial time constant T. r And the steady-state gain K, in each operation step, this unit receives the global time scaling factor λ from the load response modulation unit, and performs real-time gain scheduling correction on the local inertial time constant, i.e., T r_new =T r ×λ, this unit calculates the increment of the process variable at the current moment based on the corrected time constant and the preset process characteristic curve f(x) (such as an S-shaped velocity curve or an exponential decay curve). When the operation command undergoes a logical reversal midway through the action, such as switching from on to off instantaneously, the bidirectional disturbance-free switching logic configured in this unit immediately intervenes. It does not reset the process variable to the initial point, but instead uses the instantaneous value at the current moment as the new starting point, and calls the reverse reset characteristic curve for reverse integration, thereby ensuring the continuity of data in the time domain and accurately reproducing the dynamic characteristics of the physical entity's inertial gliding. In addition, to simulate the impact of extreme environments on equipment performance, this unit is also configured with environmental parameter drift compensation logic, which reads the global environmental scalar E issued by the environmental state setting unit in real time. curr For example, simulating ambient temperature and calculating the value E under rated operating conditions. nom Environmental bias, expressed by the formula: β = 1 + K env ×(E curr -E nom ), generate parameter drift coefficient β, where K env To obtain the environmental sensitivity gain calibrated by the experiment, the inertial time constant was finally corrected by the coefficient. Under simulated low temperature or high temperature conditions, the response speed of the equipment exhibited a nonlinear drift that conformed to physical laws.

[0031] To enhance the evolutionary characteristics of the simulation system throughout its entire lifecycle, the dynamic model evolution unit further integrates cumulative fatigue decay logic. In each integration operation, it extracts the absolute value of the rate of change of the instantaneous value of the process variable in the time domain, |ΔNPV|, and accumulates it in local non-volatile memory to generate a cumulative travel variable C representing the total historical actions of the virtual device. total The system has a preset aging inflection point threshold based on the physical lifespan model of the device. Once C total If the threshold is exceeded, the system automatically activates the decay mechanism, adding an irreversible decay bias to the steady-state gain K or proportionally increasing the inertial time constant T. r This mechanism allows virtual devices to naturally exhibit aging characteristics such as sluggish movements and increased gaps as simulation running time accumulates. This enables operators to continuously adjust control strategies during long-term training to adapt to the degradation of equipment performance, thereby greatly enhancing the practical value of training. The virtual-real closed-loop feedback unit constitutes the final and crucial link connecting the virtual world and physical perception in this system. It is responsible for upgrading simple visual simulation to a closed loop of tactile and physical logic. This unit not only maps the calculated constrained process variables to the pose driving parameters of the 3D model skeleton, but its core function lies in the configured deviation-tactile mapping logic. This unit monitors the difference between the target state vector (operator intention) and the constrained process variables (actual model state) in real time, i.e., the control deviation signal. When jamming, collision, or hard limit boundary value C occurs in the virtual scene due to malfunction, the unit will respond accordingly. limit When triggered, the clamping effect of the limiting circuit locks the restricted process variable within the feasible region, causing the control deviation signal to persist and fail to converge. Once the amplitude of the deviation signal exceeds the preset dead zone threshold, the unit converts it into a physical reverse drive voltage, mapping the deviation amplitude to the duty cycle parameter of the pulse width modulation (PWM) signal. This physical signal directly drives the analog instrument coil or force feedback actuator on the centralized control console, causing the instrument pointer to generate mechanical oscillations with a frequency related to the magnitude of the deviation near the indicated position, or causing the joystick to generate a damping torque. This design utilizes the residual information in the control loop to achieve the physical projection of fault information, allowing the operator to intuitively judge the mechanical fault of the remote equipment based solely on touch or abnormal dynamics of the instrument pointer without needing to see the alarm on the screen, thus completing a qualitative leap from visual simulation to sensory simulation.

[0032] Example 1: In a complex working condition where the centralized control console performs concurrent column raising and sliding operations on multiple hydraulic supports in a large fully mechanized mining face, this example verifies the dynamic simulation performance of the system under resource-constrained and multivariate coupling conditions. When the centralized control console simultaneously issues a group column raising command for hydraulic support groups 10 to 30, the control signal sampling interface samples the high-frequency pulse command with a period of 10ms. The time window integration method is used to filter out glitches caused by electromagnetic interference, generating a stable target state vector. At this time, the load response modulation unit detects that 20 virtual hydraulic supports are simultaneously in a non-steady-state process transitioning from column lowering to column raising. The system determines the energy consumption weight parameter W for each support. i All values ​​are based on the rated power of a single unit. This unit performs a weighted summation, and the calculated real-time system load rate R is greater than 1.0, indicating that the instantaneous flow demand of the virtual hydraulic power station has exceeded the physical supply capacity limit W. limit Based on the preset overload-hysteresis characteristic curve, the load response modulation unit maps the global time scaling factor λ to a value greater than 1.0, which is injected into the dynamic model evolution unit of each hydraulic support in real time.

[0033] Due to this global factor modulation, the dynamic model evolution unit did not drive the scaffold to rise at an ideal linear speed, but rather according to formula T. r_new =T r The inertial time constant is increased in real time by ×λ, causing the rising motion of all controlled supports to exhibit a consistent lag and slowdown in the time domain. This accurately reproduces the phenomenon of sluggish movement caused by a drop in the main pipe pressure in real physical scenarios. During this process, if the operator discovers an anomaly in support No. 15 and urgently cuts off the column raising command, the dynamic model evolution unit corresponding to that support triggers bidirectional disturbance-free switching logic. The system does not instantaneously reset the process variables of that support, but instead uses the intermediate position value at the current moment, which is not fully raised, as the starting point, and calls the reverse reset characteristic curve to smoothly fall back, avoiding the animation jumps common in traditional simulations. At the same time, for the travel limitation fault caused by the top plate breakage of support No. 18 in the simulation, when the limited process variable reaches the hard limit boundary value C... limit At that time, the dynamic model evolution unit stopped the integration operation, causing the value to be unable to follow the target state vector. The virtual and real closed-loop feedback unit captured the continuous existence of the control deviation signal and exceeded the dead zone threshold in real time, and mapped it into a high duty cycle PWM signal. This signal drove the control handle corresponding to the 18th bracket on the physical operation table to generate high-frequency force feedback oscillation. This physical feedback, before visual confirmation, had already transmitted clear equipment jamming information to the operator through the tactile channel, realizing closed-loop verification from single visual monitoring to multi-dimensional physical perception.

[0034] Example 2: This example constructs a hardware-in-the-loop (HIL) test platform that includes high-frequency electromagnetic interference and extreme load conditions to quantitatively verify the load response modulation accuracy and physical feedback real-time performance of the proposed virtual-real simulation system. The test platform mainly consists of an industrial control computer with a real-time operating system, a programmable signal generator, and a modified physical joystick assembly with force feedback function. The industrial control computer runs the core algorithm of the dynamic model evolution unit, and the processor's main frequency is no less than 3.0 GHz to ensure that the calculation cycle is strictly controlled within 10 ms. The signal generator is used to inject discrete operation commands with simulated field noise into the control signal sampling interface. The physical joystick is directly connected to the output of the virtual-real closed-loop feedback unit and is used to measure the reverse drive voltage and the generated damping torque. Its torque sensor sampling frequency is set to 1 kHz.

[0035] In the experimental parameter setting phase, the system capacity threshold W limit The value is calibrated as a dimensionless value of 100. The logic for setting this value is based on the rated output flow characteristics of a certain type of mining hydraulic pump station, that is, the full-load output of the pump station is defined as the base unit of 100, and the energy consumption weight W of a single virtual hydraulic support is... i The parameter was set to 2.5. This parameter was determined following the flow distribution principle in fluid dynamics and the weighting of the impact of a single unit's action on the pressure drop in the main pipe. Specifically, under rated pressure, a single unit operating at full speed consumes 2.5% of the system's total flow reserve. To verify the signal processing stability of the system in a real industrial electromagnetic environment, Gaussian white noise with a signal-to-noise ratio of 15dB and occasional pulse interference harmonics were actively superimposed onto the standard square wave command output by the signal generator. The time window integration method processing results executed by the control signal sampling interface showed that glitches in the original signal were effectively filtered out, and the generated target state vector maintained the same temporal order as... Strict clock alignment prevented any malfunctions caused by noise. The core phase of the experiment focused on the gradient verification of load response characteristics. The experiment included a prototype sample (enabling the load response modulation unit) and a control sample (using a traditional fixed time constant model). At the input end, a gradient test environment was constructed from light load, critical, overload to saturation by gradually increasing the number of virtual supports with concurrent actions. During the experiment, a high-precision data logger was used to collect the real-time system load rate R, global time scaling factor λ, and actual response time T required for the virtual supports to complete their full stroke under different concurrency levels. act See Table 1, which records the key process data and final response results under different load gradients.

[0036] Table 1: Verification Data Table of Load Response Modulation Characteristics

[0037]

[0038] Table 1 shows that within the light load and critical range (R ≤ 1.0), the measured response times of both the present invention sample and the comparative sample are stable at a baseline value of around 5.0 seconds, indicating that the system maintains an ideal response when resources are sufficient. However, when entering the overload range (R > 1.0), the response time of the comparative sample remains at a non-physically constant value of approximately 5.0 seconds, showing an inability to respond to load changes; while the response time T of the present invention sample... act The response time increases non-linearly with the load rate, especially under the condition of 80 devices operating concurrently, extending to 22.63 seconds. This non-linear hysteresis characteristic is highly consistent with the physical behavior of a real hydraulic network when pressure is depleted. When the load continues to increase to the saturation zone (100 devices operating concurrently), the response time reaches 50.15 seconds, indicating that the system relies on the hard limit boundary value C. limit The clamping logic enters a stable low-speed operating state to avoid calculation divergence. In the verification of the fault projection capability of the virtual-real closed-loop feedback unit, the experimenter artificially set the third virtual support to trigger a mechanical jamming fault when the stroke reaches 50%, that is, to lock its restricted process variable and prevent it from increasing. At this time, the restricted process variable output by the dynamic model evolution unit stagnates at 0.50, while the target state vector remains at 1.00, causing the control deviation signal between the two to accumulate rapidly and stabilize at an amplitude of 0.50. After the virtual-real closed-loop feedback unit detects that the duration of this deviation exceeds the dead zone threshold of 200ms, it activates the deviation-tactile mapping logic. The setting of this dead zone threshold is based on the spectrum analysis of the frequency of human hand operation jitter, which aims to filter out non-faulty transient operation errors. The physical measurement results show that a mechanical vibration with a frequency of 15.2Hz and an amplitude of 2.1mm is generated at the joystick handle, and the magnitude of this damping torque is positively correlated with the amplitude of the deviation signal. The operator can accurately perceive the abnormal jamming of the equipment within 310ms after the fault occurs by touch alone in the blind operation state.

[0039] Example 3: This example, in conjunction with Figures 1 to 3, describes a virtual-real simulation system that links a centralized control console with a 3D model. As shown in Figure 1, the logical architecture begins with the centralized control console in the physical world. The physical control port issues discrete commands, which are conditioned and clock-synchronized via a control signal sampling interface to generate a target state vector. This vector serves as the core input and is fed into the closed loop of the computational logic. The system has a parallel load response modulation unit that receives transient action signals and performs a weighted summation of the energy consumption weights of all virtual objects. Based on energy constraints, a global time scaling factor is generated, which is then injected into the dynamic simulation system. The dynamic model evolution unit is used to correct the inertial time constant in real time. Combined with the environmental parameter drift compensation logic, the dynamic model evolution unit performs discretization integration and hard limit boundary processing on the target state vector, and outputs the restricted process variable. The virtual-real closed-loop feedback unit receives the restricted process variable. On the one hand, it extracts the control deviation signal between the target state vector and the restricted process variable in real time and converts it into a physical reverse driving voltage to drive the electromechanical indicator device to generate nonlinear damping action. On the other hand, it outputs the driving parameters to the 3D model to drive its pose state evolution and visual rendering, thus forming a complete virtual-real closed loop.

[0040] As shown in Figure 2, the horizontal axis represents the number of concurrent supports (units), the left vertical axis represents the response time (units), and the right vertical axis represents the load factor R. The figure contains three main characteristic curves: the theoretical load factor R curve, the response time curve of the comparative sample group, and the response time curve of the sample group of the present invention. The theoretical load factor R increases linearly with the increase of the number of concurrent supports. The response time of the comparative sample group remains constant throughout the entire load range, without reflecting the load effect. However, the response time of the sample group of the present invention maintains the baseline value in the range where the load factor R > 1.0. When the load factor R > 1.0, the response time increases non-linearly and exponentially with the increase of the number of concurrent supports, objectively reflecting the physical hysteresis characteristics of the system under overload conditions.

[0041] As shown in Figure 3, the overall deployment architecture of the system is divided into two major sections: the physical entity domain for actual operation and the digital twin domain for core computing power. The two are connected through a signal synchronization gateway, which is responsible for jitter removal and clock alignment. In the physical entity domain, the signals generated by the operation command sending module, such as buttons / handles, are transmitted to the digital domain through the gateway. At the same time, the human-machine interaction terminal of the central control console includes force feedback and instrument indication modules to receive feedback from the digital domain. In the digital twin domain, the computing control logic center integrates the load response modulation unit for energy constraint calculation, the dynamic model evolution unit for inertia and evolution, and the virtual-real closed-loop feedback unit for deviation extraction. This center calculates and generates constrained process variables and global time scaling factors based on discrete instructions. On the one hand, it drives the synchronization of the virtual device status in the 3D visualization scene. On the other hand, it generates physical reverse drive voltage by calculating control deviations and transmits it back to the force feedback actuator in the physical entity domain, realizing bidirectional real-time mapping between physical operation and digital operation.

[0042] Example 4: This example illustrates the standardized engineering procedures used to determine the key control parameters and core algorithm logic in the system of this invention. For the core discretized integral operation in the dynamic model evolution unit, the system adopts a first-order inertial element iterative algorithm based on backward difference to ensure numerical stability within a 10ms control cycle. For any virtual controlled object, in the k-th control cycle, its normalized process variable NPV... k The update follows the following difference equation:

[0043] ;

[0044] Among them, NPV k-1 The process variable value is from the previous cycle, Δt is a fixed control cycle of 10 ms, and L is the value of the process variable from the previous cycle. target Let T be the target state vector value at the current moment. r_mod The real-time inertial time constant, after global load modulation and environmental correction, is calculated strictly according to the formula: T r_mod =T r ×λ×β, this difference equation clearly defines the numerical evolution path in the discrete domain, ensuring the consistency of the simulation model's response on different computing platforms. When the dynamic model evolution unit performs discrete integration operations, the generation of hard limit boundary values ​​follows a deterministic priority arbitration logic to resolve the timing conflict between geometric constraints and fault states. The logic stipulates that the priority of the fault preset state limit threshold is higher than the inherent geometric travel limit of the equipment. Within a 10ms step-size difference operation cycle, the limiting circuit prioritizes querying the fault state register. When a mechanical jamming or cylinder deformation fault flag is detected, the corresponding current coordinate value or preset jamming position value is called as the hard limit boundary value C. limitThe overwrite geometry determines the feasible domain boundary. The transient overwrite mechanism causes the constrained process variable to stop its numerical evolution when the instruction does not reach the target value. This generates a continuous and stable numerical difference, which forms the basis for the virtual-real closed-loop feedback unit to extract the deterministic data of the control deviation signal. This eliminates the hidden danger of feedback signal instability caused by boundary condition jitter. This relates to the nonlinear hysteresis mapping logic in the load response modulation unit.

[0045] This embodiment uses a piecewise function to precisely define the constraint relationship between the system real-time load rate R and the global time scaling factor λ. When the system real-time load rate R ≤ 1.0, the global time scaling factor λ is constant at 1.0; when R > 1.0, the calculation of the global time scaling factor λ follows the following quadratic growth model: λ = 1.0 + α × (R - 1.0) 2 Where α is the system capacity sensitivity coefficient, in the calibration of a typical fully mechanized mining face hydraulic system in this embodiment, by gradually increasing the load under full load conditions until the system pressure drops to 60% of the rated value, the corresponding action lag time is measured, and the value of α is determined to be 0.8. This functional relationship ensures that when the load rate R reaches 2.0 overload, λ rises to 1.8, extending the equipment action time by 80%, which conforms to the physical characteristics of pressure-flow coupling in fluid networks. For the initialization and quantization of energy constraint model parameters in the load response modulation unit, the system executes a standardized capacity calibration procedure, so that the calculated real-time system load rate R accurately reflects the physical resource occupancy. The preset system capacity limit W limit To determine the dimensionless scalar based on the rated output capacity of the physical station, the method is to read the rated maximum output flow or current value of the hydraulic pump station or power supply bus and define it as the base integer 100; the energy consumption weight W of each virtual controlled object i By determining through single-machine load testing, the ratio of the peak flow or starting current required for the moment the equipment operates at full speed to the rated maximum output is solidified into a lookup table as a weight parameter. Based on the physical rated proportion quantization mapping mechanism, when the energy consumption weights of multiple controlled objects accumulate and exceed the system capacity limit, the global time scaling factor λ is generated based on the scarcity of real drive resources, ensuring that the simulation hysteresis characteristics and the physical main pipe pressure drop phenomenon are strictly corresponded.

[0046] For the environmental sensitivity gain K in the environmental parameter drift compensation logic env In this embodiment, a gradient calibration experiment based on a temperature-controlled environmental chamber is performed. A standard hydraulic support model is selected as the test object, and the rated operating value E is set. nom The ambient temperature was set at 25℃, and the temperature of the environmental chamber was gradually increased from -20℃ to 60℃ in increments of 5℃. After maintaining the temperature at each point for 30 minutes, the support was triggered to perform a full-stroke column lifting action, and the response time T was recorded. meas Through linear regression analysis (T) meas / T base -1) and (E)curr -E nom The slope obtained from the relationship between the two is K. env Experimental data shows that for systems using No. 46 anti-wear hydraulic oil, in the low-temperature range (-20℃ to 0℃), K env The measured value was 0.025, indicating that the response time increased by 2.5% for every 1°C decrease; in both the normal and high temperature ranges, K... env The value was adjusted to 0.012, and this calibration data was written into the parameter lookup table of the dynamic model evolution unit to achieve accurate compensation for the influence of ambient temperature. The deviation-tactile mapping logic in the virtual-real closed-loop feedback unit adopts a piecewise gain control strategy with dead time, defining the control deviation signal amplitude e=|L target -NPV constrained | When e≤0.05 (5% dead zone threshold), the output physical reverse drive voltage V out =0V, to filter out computational noise; when e>0.05, V out The calculation follows the formula: V out =V min +k haptic × (e-0.05), where V min The starting voltage for the force feedback motor is set to 3.5V; k haptic The haptic gain coefficient is set to 18.0. This logic ensures that when the deviation reaches 0.5, i.e., a half-stroke jamming fault, V... out Reaching 11.6V, close to the full-scale drive voltage of 12V, a forced damping torque is generated on the physical joystick, ensuring that the operator can clearly perceive the execution obstacles of the virtual device under complex working conditions. When the virtual-real closed-loop feedback unit generates the physical reverse drive voltage, the deviation-tactile mapping logic adopts a frequency domain modulation signal conversion procedure, so that the nonlinear damping action can be clearly perceived by human touch. The procedure sets the amplitude to 5% of the instrument's full-scale response dead zone. When the amplitude of the control deviation signal exceeds the dead zone threshold, the system starts the pulse width modulation signal output, locks the carrier frequency in the low frequency band of 10Hz to 50Hz, avoids power frequency interference, and drives the analog instrument coil or force feedback motor to produce obvious particle-like mechanical vibration. The duty cycle parameter of the physical reverse drive voltage is not a linear function of the deviation amplitude. Based on the increment of the deviation amplitude above the dead zone, it is mapped to a nonlinear duty cycle by a lookup table, so that the generated damping torque or pointer oscillation intensity is perceptually correlated with the severity of the virtual fault in accordance with the Weber-Fechner law, realizing the intuitive reproduction of the virtual stall condition at the physical operation end.

[0047] Example 5: To ensure the high determinism and reproducibility of the core control parameters and nonlinear characteristic curves in this invention across different hardware platforms and application scenarios, this example constructs a standardized engineering procedure that includes offline calibration and on-site pre-deployment calibration. For the construction of the overload-hysteresis characteristic curve in the load response modulation unit, the system executes an offline calibration and data filling procedure. This procedure is performed on a controlled pressure test bench equipped with a high-precision electronic load and flow meter. The output flow rate of the virtual hydraulic master station is set to the rated value Q. rated and gradually increase the load traffic demand Q demand Calculate the theoretical load factor R=Q demand / Q rated At each R value point, record the minimum response time T required for the system to maintain pressure balance. min By using curve fitting tools, a relationship between R and the time dilation coefficient λ=T is established. min / T base The functional relationship between them, where T base Using the baseline response time, the fitting process employs the least squares method, and the generated characteristic curve parameters are stored in read-only memory as a reference for the lookup table during system operation, thereby eliminating the arbitrariness of parameters caused by empirical settings.

[0048] To address the potential hardware differences and environmental interference that the system may encounter when deployed in different industrial sites, the implementation example establishes a pre-deployment calibration procedure. After the system is initially powered on, the control signal sampling interface automatically executes a baseline calibration program, collects the background noise level under no-operation command conditions, calculates the root mean square value, and uses it to set a dynamic de-jitter threshold. Simultaneously, the virtual-real closed-loop feedback unit performs zero-point drift correction, drives the physical joystick to the mechanical neutral position, reads the sensor voltage value at this time as the zero-point reference, and superimposes a preset dead-zone voltage range on this basis to compensate for mechanical installation errors. The system automatically identifies the actual communication delay and actuator response lag under the current hardware environment by running a set of standard test sequences, including step response and sinusoidal frequency sweep, and fine-tunes the prediction compensation coefficients in the dynamic model evolution unit accordingly to ensure strict synchronization between the simulation model and the physical equipment on the time axis.

[0049] Example 6: This example details a standardized pre-deployment calibration and model building procedure, aiming to ensure that the present invention can quickly establish a simulation environment that matches the actual working conditions when applied to different industrial sites and physical equipment, and solve the problem of response distortion caused by differences in initial conditions. This procedure covers three core stages: on-site identification of physical parameters, adaptive fine-tuning of the algorithm model, and safety verification of boundary conditions. In the on-site identification stage of physical parameters, high-precision laser rangefinders and tilt sensors are used to collect all physical geometric parameters of the centralized control console and the controlled object. For hydraulic supports, the coordinates of key hinge points of the top beam, shield beam, and linkage mechanism are accurately measured, and a high-fidelity geometric mesh model is constructed using 3D scanning technology. Through standardized step response testing, the dynamic characteristics of the actual hydraulic system are identified. Specifically, a series of step commands with different amplitudes are sent to the hydraulic actuator, and its displacement, pressure, and flow response curves are collected simultaneously. Using the system identification algorithm, the inertial time constant T is extracted from the response data. r Key transfer function parameters such as steady-state gain K and pure time delay τ are directly written into the configuration register of the dynamic model evolution unit as the basic physical properties of the simulation model, thereby ensuring that the dynamic behavior of the virtual object is highly consistent with the time-domain characteristics of the real device.

[0050] During the adaptive fine-tuning phase of the algorithm model, online learning and parameter optimization programs are executed for the nonlinear hysteresis mapping logic in the load response modulation unit. During trial operation, the system continuously monitors the dynamic correlation between the main pipe pressure and the actuator response speed. Using recursive least squares (RLS), the system capacity sensitivity coefficient α in the overload-hysteresis characteristic curve is updated in real time. When a deviation between the actual pressure drop and the model prediction is detected, the adaptive algorithm automatically adjusts the α value until the global time scaling factor λ output by the model accurately reflects the current load effect. Furthermore, for the environmental parameter drift compensation logic, the system utilizes temperature and humidity sensors deployed on-site to collect temperature data under actual working conditions. Combined with equipment operation logs, a temperature drift correction model based on on-site data is established. Through long-term accumulation of operating data, the environmental sensitivity gain K is continuously corrected. env This is to eliminate parameter drift caused by oil aging or mechanical wear.

[0051] During the boundary condition safety verification phase, the system performs standardized fault injection and fault tolerance tests. Abnormal instruction sequences exceeding the normal range, such as extremely high-frequency jitter signals or logically conflicting combination instructions, are artificially injected into the control signal sampling interface to verify the effectiveness of the de-jitter algorithm and timing alignment logic. Extreme conditions such as sensor failure or communication interruption are simulated to test the fault perception and processing capabilities of the virtual and real closed-loop feedback unit. The system must ensure that even if any single sensor fails, it can still maintain basic simulation functions through redundant information or state observers and send a clear fault alarm signal to the control panel. Only after all the preset boundary test cases have passed verification is the system allowed to enter the formal operation state. This entire set of pre-calibration and verification procedures constitutes the standardized implementation path of the technical solution of this invention in engineering, ensuring the stability and reliability of the system in complex and ever-changing industrial environments.

[0052] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0053] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A virtual-real simulation system that links a centralized control console with a 3D model, characterized in that, include: The control signal sampling interface is connected to the physical control port of the centralized control console. It is used to sample and hold discrete operation command signals at a preset control cycle to generate a target state vector. The load response modulation unit is used to establish the energy constraint model of the virtual controlled object. In each control cycle, it performs a weighted summation of the energy consumption weights of all virtual controlled objects undergoing transient changes to obtain the real-time load rate of the system. Based on the deviation between the real-time load rate of the system and the preset system capacity limit, it generates a global time scaling factor to characterize the system's energy redundancy and injects the global time scaling factor into the dynamic model evolution unit. The dynamic model evolution unit has embedded transfer function parameters describing the dynamic characteristics of the virtual controlled object. The transfer function parameters include at least the inertial time constant and the steady-state gain. The dynamic model evolution unit is used to perform real-time gain scheduling correction on the inertial time constant using the global time scaling factor. Based on the corrected inertial time constant, it performs discretized integration on the target state vector to generate instantaneous values ​​of process variables that meet the system energy constraints. The dynamic model evolution unit is also used to introduce hard limit boundary values ​​from the simulation environment. Through a limiting element, it clamps the instantaneous values ​​of process variables within the feasible region defined by the hard limit boundary values, generating constrained process variables. The virtual-real closed-loop feedback unit is used to map the constrained process variables into driving parameters to drive the pose state evolution of the virtual controlled object. At the same time, it extracts the control deviation signal between the target state vector and the constrained process variables in real time. When the amplitude of the control deviation signal continuously exceeds the preset dead zone threshold, it converts the control deviation signal into a physical reverse driving voltage to drive the electromechanical indicating device on the centralized control console to generate a nonlinear damping action corresponding to the control deviation signal. Furthermore, the load response modulation unit is configured with nonlinear hysteresis mapping logic; The nonlinear hysteresis mapping logic is used to map the global time scaling factor to a value greater than 1 based on a preset overload-hysteresis characteristic curve when the system's real-time load rate exceeds the unit reference value. This causes the inertial time constant in the dynamic model evolution unit to monotonically increase with the increase of the system's real-time load rate, thereby introducing an additional delay characteristic determined by the system load in the step response process of the virtual controlled object. The control signal sampling interface is configured with signal conditioning and clock synchronization logic. The signal conditioning and clock synchronization logic is used to filter out high-frequency noise interference in the operation command signal and align the debouncing effective command signal to the operation clock edge of the dynamic model evolution unit to eliminate deviations. Sampling jitter error in the distributed control system; hard limit boundary values ​​are variables generated in real time by the geometric constraints or fault preset states of the virtual controlled object in the simulation space; the limiting element is used to introduce nonlinear saturation characteristics into the integral operation loop of the dynamic model evolution unit, so that the numerical trajectory of the constrained process variable cannot exceed the hard limit boundary value, thereby artificially creating and maintaining the control deviation signal in the control loop; the load response modulation unit, the dynamic model evolution unit, and the virtual-real closed-loop feedback unit constitute a deterministic real-time digital control loop; the transfer function parameters and process characteristic curves are stored in a fixed-point lookup table manner, and the integral operation is solved iteratively by the difference equation with a preset step size.

2. The virtual-real simulation system linking a centralized control console and a 3D model according to claim 1, characterized in that, The dynamic model evolution unit is equipped with environmental parameter drift compensation logic; the system also includes an environmental state setting unit, which is used to set global environmental scalars that characterize the physical properties of the simulation environment; The environmental parameter drift compensation logic is used to calculate the environmental deviation between the global environmental scalar and the preset rated operating condition value, and to generate the parameter drift coefficient using the environmental deviation. The dynamic model evolution unit uses the parameter drift coefficient to perform a secondary correction on the inertial time constant to simulate the dynamic response characteristic drift of the virtual controlled object under different physical environments. The calculation logic of the parameter drift coefficient β follows the following relationship: β = 1 + K env ×(E curr -E nom ); Among them, E curr E is a global environmental scalar. nom K is the preset rated operating condition value. env To achieve the preset environmental sensitivity gain, the dynamic model evolution unit uses the parameter drift coefficient β to multiply and correct the inertial time constant.

3. The virtual-real simulation system linking a centralized control console and a 3D model according to claim 1, characterized in that, The dynamic model evolution unit is also equipped with cumulative fatigue decay logic. The cumulative fatigue decay logic is used to integrate and accumulate the absolute value of the rate of change of the instantaneous value of the process variable in the time domain to generate a cumulative travel variable that represents the total amount of motion of the virtual controlled object throughout its entire life cycle. When the cumulative travel variable exceeds the preset aging inflection point threshold, the dynamic model evolution unit superimposes an irreversible decay bias on the steady-state gain or the inertial time constant, so that the dynamic characteristics of the virtual controlled object exhibit a gradual performance decay as the total amount of motion increases.

4. The virtual-real simulation system linking a centralized control console and a 3D model according to claim 1, characterized in that, The virtual-real closed-loop feedback unit is equipped with deviation-tactile mapping logic; the deviation-tactile mapping logic is used to map the amplitude of the control deviation signal to the duty cycle parameter of the pulse width modulation signal, and output the pulse width modulation signal as a physical reverse drive voltage to the analog instrument coil of the central control console, driving the pointer of the analog instrument to generate mechanical oscillations with a frequency related to the control deviation signal near the indicated position, so as to physically reproduce the stall condition of the virtual controlled object caused by the hard limit boundary value limitation; In addition, the electromechanical indicating devices of the centralized control console include analog pointer instruments and force feedback actuators; The virtual-real closed-loop feedback unit is used to output the constrained process variable as a feedback quantity to the analog pointer instrument in a closed loop, and to control the force feedback actuator to output the damping torque using the control deviation signal, so as to realize the visual and tactile dual-channel physical feedback based on the model state.

5. A virtual-real simulation system linking a centralized control console and a 3D model according to claim 1, characterized in that, The dynamic model evolution unit is equipped with bidirectional non-disruptive switching logic. When the operation command signal is not in steady state of the virtual controlled object, the logic flips. The bidirectional non-disruptive switching logic is initialized with the instantaneous value of the process variable at the current moment, switches to the inverse integration path, and performs reverse evolution calculation based on the reset time constant to ensure that the instantaneous value of the process variable remains continuous and without step change at the moment of command switching.

Images