A real-time transmission method and system for Internet of Things perception data in a digital twin scenario

Abstract

The present invention provides a real-time transmission method and system for Internet of Things perception data in a digital twin scenario. By collecting and processing vibration spectrum data, pressure gradient change data, and temperature distribution data of equipment in an industrial production line, a preprocessed data packet is generated. The equipment state fingerprint features and environmental working condition labels are analyzed, and the dynamic parameter correction loop of the digital twin body is used to generate target virtual control parameters and target error compensation amounts. Combining with the material flow rate data, the actual material flow state is generated. An optimization strategy sequence for multi-device collaborative operation is generated using a distributed computing architecture and converted into trajectory compensation parameters of the end effector in the equipment. Combining with the target error compensation amount, data closed-loop transmission is performed. The present invention realizes the millisecond-level time domain matching of vibration, pressure, and temperature data with the material flow rate through the dynamic timestamp alignment mechanism of the 5G edge computing node, solving the problem of cumulative synchronization errors of traditional Internet of Things perception data.

Description

Technical Field

[0001] The present invention relates to the technical field of data transmission, and particularly to a real-time transmission method and system for Internet of Things perception data in a digital twin scenario. Background Art

[0002] Real-time data transmission has become a key link to ensure production efficiency and product quality. With the popularization of industrial Internet of Things, the demand for data collection and transmission of industrial equipment is increasing day by day. Real-time data transmission not only requires the efficiency of data transmission, but also must ensure the accuracy and security of data to support application scenarios such as industrial automation, intelligent manufacturing, and data-driven decision-making.

[0003] Currently, existing solutions include a digital twin multi-source data synchronization framework based on edge computing, etc., which have some defects. For example, the resampling mechanism with a fixed time window cannot adapt to the scenario of sudden change in material flow rate, resulting in insufficient data synchronization accuracy; the offline model update period is long, and the compensation delay is significant when the equipment suddenly fails, resulting in data transmission delay, etc. Summary of the Invention

[0004] Embodiments of the present invention provide a real-time transmission method and system for Internet of Things perception data in a digital twin scenario, so as to solve the problems in the prior art that it cannot adapt to the scenario of sudden change in material flow rate, resulting in insufficient data synchronization accuracy; the compensation delay is significant when the equipment suddenly fails, resulting in data transmission delay, etc.

[0005] In a first aspect, an embodiment of the present invention provides a real-time transmission method for Internet of Things perception data in a digital twin scenario, including:

[0006] Collect the vibration spectrum data, pressure gradient change data, and temperature distribution data of the equipment in the industrial production line, and construct a timestamp alignment mechanism in the 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet;

[0007] Analyze the preprocessing data packet to obtain the equipment state fingerprint feature and the environmental working condition label;

[0008] According to the equipment state fingerprint feature and the environmental working condition label, use the dynamic parameter correction loop of the digital twin body to generate the target virtual control parameter of the equipment and the target error compensation amount of the clearance of the transmission chain in the equipment, wherein the digital twin body is constructed based on the equipment topology relationship network and process constraint conditions corresponding to the equipment;

[0009] Generate the actual material flow state based on the device state fingerprint features, environmental condition labels, and material flow rate data, and use a distributed computing architecture to perform a coupled analysis on the target virtual control parameters, target error compensation amount, and actual material flow state to generate an optimization strategy sequence for multi-device collaborative operation;

[0010] Convert the optimization strategy sequence into trajectory compensation parameters for the end effector in the device, and perform data closed-loop transmission in combination with the target error compensation amount.

[0011] Optionally, based on the device state fingerprint features and environmental condition labels, use the dynamic parameter correction loop of the digital twin to generate the target virtual control parameters of the device and the target error compensation amount for the clearance of the transmission chain in the device. The digital twin is constructed based on the device topology relationship network and process constraint conditions corresponding to the device, including:

[0012] Construct a digital twin based on the device topology relationship network and process constraint conditions. The device topology relationship network uses a graph theory modeling method to establish the mechanical connection attributes and signal transmission paths between device nodes;

[0013] Based on the dynamic parameter correction loop in the digital twin, decompose the device state fingerprint features into axial vibration mode features, radial thermal expansion coefficient features, and tangential stress distribution features, and combine the maximum allowable deformation threshold in the process constraint conditions and the temperature sensitivity coefficient of the material property parameters in the digital twin to generate a multi-source fusion feature vector;

[0014] Generate an initial error compensation amount based on the multi-source fusion feature vector, and dynamically adjust the initial error compensation amount using a sliding window algorithm to obtain the target error compensation amount, where the window length is adaptively adjusted according to the change rate of the material flow rate data;

[0015] Generate initial virtual control parameters based on the working parameter feasible region of each device node in the digital twin and the environmental condition labels;

[0016] Dynamically and collaboratively correct the initial virtual control parameters according to the coupling relationship of the mechanical connection attributes of the device and the target error compensation amount to obtain the target virtual control parameters.

[0017] Optionally, combine the maximum allowable deformation threshold in the process constraint conditions and the temperature sensitivity coefficient of the material property parameters in the digital twin to generate a multi-source fusion feature vector, including:

[0018] Based on the axial vibration mode characteristics, the vibration energy value is obtained by integrating the spectral energy of the vibration signal in the direction of the main axis of the device. Based on the radial thermal expansion coefficient characteristics, the deformation gradient value is calculated through the product relationship between the temperature distribution data and the material thermal expansion coefficient. Based on the tangential stress distribution characteristics, the cross product result is calculated through the pressure gradient data and the contact surface curvature;

[0019] According to the ratio of the vibration energy value to the maximum allowable deformation threshold, the first deformation contribution rate is calculated; according to the ratio of the deformation gradient value to the maximum allowable deformation threshold, the second deformation contribution rate is calculated. According to the deviation degree of the ambient temperature from the standard operating temperature of the material, the second deformation contribution rate is dynamically compensated to generate a corrected second deformation contribution rate;

[0020] According to the cross product result, the projection components in the tangent direction of the device movement trajectory are weighted and accumulated to calculate the third deformation contribution rate;

[0021] According to the corrected second deformation contribution rate and the temperature sensitivity coefficient of the material property parameters in the digital twin, the second reference weight is calculated;

[0022] According to the first deformation contribution rate, the third deformation contribution rate, the vibration energy attenuation coefficient, and the friction coefficient correction factor, the first reference weight and the third reference weight are calculated. Combining the second reference weight, the first quantization value, the second quantization value, and the third quantization value, a multi-source fusion feature vector is generated.

[0023] Optionally, based on the device state fingerprint characteristics, the environmental operating condition labels, and the material flow rate data, the actual material flow state is generated. Using the distributed computing architecture, the target virtual control parameters, the target error compensation amount, and the actual material flow state are coupled and analyzed to generate an optimization strategy sequence for multi-device collaborative operation, including:

[0024] Based on the vibration energy distribution parameters in the device state fingerprint characteristics and the temperature and humidity parameters in the environmental operating condition labels, the material flow rate data is processed to generate a discretized flow rate interval and the corresponding material accumulation density distribution parameters. Combining the vibration energy distribution parameters, the material transfer efficiency attenuation factor between adjacent devices is calculated;

[0025] According to the non-linear relationship between the temperature and humidity parameters and the material viscosity coefficient, a flow state correction coefficient is generated. Using the distributed computing architecture, the target virtual control parameters are mapped to the kinematic constraint boundaries of the device nodes, the target error compensation amount is converted into the dynamic tolerance interval of the transmission chain clearance, and the flow state correction coefficient is embedded as a time-varying weight factor into the matrix dimension to construct a device collaborative analysis matrix;

[0026] Based on the connection topology in the device topology relationship network and the device collaborative analysis matrix, the energy transfer function and the material conservation equation between device nodes are constructed;

[0027] An initial optimization strategy is generated by iteratively solving the joint solution set of the energy transfer function and the material conservation equation, and the initial optimization strategy is optimized by combining the material transfer efficiency decay factor and the equipment safe operation threshold in the process constraint conditions to obtain a preliminarily optimized optimization strategy;

[0028] Based on the distributed computing architecture and the equipment topology relationship network, the preliminarily optimized optimization strategy is verified distributively to generate an optimization strategy sequence for multi-device collaborative operation.

[0029] Optionally, according to the non-linear relationship between the temperature and humidity parameters and the material viscosity coefficient, a flow state correction coefficient is generated. Using the distributed computing architecture, the target virtual control parameters are mapped to the kinematic constraint boundaries of the equipment nodes, the target error compensation amount is converted into the dynamic tolerance interval of the transmission chain clearance, and the flow state correction coefficient is embedded as a time-varying weight factor into the matrix dimension to construct an equipment collaborative analysis matrix, including:

[0030] Based on the non-linear relationship between the temperature and humidity parameters and the material viscosity coefficient, a three-dimensional interpolation grid is established at the temperature and humidity sampling points of the environmental condition label to calculate the change gradient of the material viscosity coefficient under continuous working conditions, and combined with the material flow rate data, a flow state correction coefficient is generated;

[0031] The target virtual control parameters are decomposed into the displacement control amount, speed control amount and acceleration control amount of the equipment nodes. Based on the kinematic constraint parameters in the equipment topology relationship network, the displacement control amount is converted into a position boundary threshold, the speed control amount is converted into a rate change envelope, and the acceleration control amount is converted into an inertia compensation amount by using the distributed computing architecture to generate the kinematic constraint boundaries of the equipment nodes;

[0032] Based on the spatial distribution characteristics of the transmission chain clearance, the target error compensation amount is decomposed into an axial compensation component and a tangential compensation component. Using the distributed computing architecture, the axial compensation component is converted into an axial tolerance interval, the tangential compensation component is converted into an angular tolerance interval, and combined with the axial tolerance interval and the angular tolerance interval, a dynamic tolerance interval of the transmission chain clearance is generated;

[0033] The flow state correction coefficient is embedded as a time-varying weight factor into the matrix weight, and combined with the kinematic constraint boundaries, the dynamic tolerance interval and the connection topology, an equipment collaborative analysis matrix is constructed.

[0034] Optionally, the optimization strategy sequence is converted into the trajectory compensation parameters of the end effector in the equipment, and combined with the target error compensation amount, data closed-loop transmission is performed, including:

[0035] Decompose the optimized strategy sequence into motion control instruction sets for each end effector. Based on the motion control instruction sets, generate the coordinates of the trajectory reference points, the interpolation path function between adjacent reference points, and the smoothing coefficient of the trajectory transition section to obtain the initial trajectory compensation parameters;

[0036] Based on the spatial distribution characteristics of the transmission chain clearance and the initial trajectory compensation parameters, use a distributed computing architecture to decompose the target error compensation amount into an axial position compensation amount and an angular offset compensation amount. Combine the axial position compensation amount, the angular offset compensation amount, the trajectory reference point coordinates, the interpolation path function, and the smoothing coefficient to generate the fused trajectory compensation parameters;

[0037] Verify the fused trajectory compensation parameters to obtain the verified trajectory compensation parameters. Use the timestamp alignment mechanism to process the verified trajectory compensation parameters, generate a transmission data packet, and distribute the transmission data packet to each end effector controller through the downlink of the 5G edge computing node to synchronously collect the residual between the actual motion trajectory data of each end effector and the predicted trajectory of the digital twin;

[0038] When the residual exceeds the dynamic tolerance interval of the transmission chain clearance, trigger the dynamic parameter correction loop to recalculate the compensation amount to achieve closed-loop data transmission.

[0039] Optionally, construct a timestamp alignment mechanism in the 5G edge computing node to perform time-domain matching of vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet, including:

[0040] Construct a timestamp alignment mechanism in the 5G edge computing node. Based on the pulse signal of the material flow rate data, trigger the timestamp alignment mechanism to perform time-domain matching processing on the vibration spectrum data, the pressure gradient change data, and the temperature distribution data to obtain the vibration spectrum data, pressure gradient change data, and temperature distribution data after time-domain matching;

[0041] According to the spatial coordinates of the device topology relationship network, decompose the vibration spectrum data after time-domain matching into axial vibration components in the direction of the device main axis, convert the pressure gradient change data after time-domain matching into a normal pressure distribution matrix of the device contact surface, and map the temperature distribution data after time-domain matching into an equivalent heat source intensity parameter;

[0042] Perform a convolution operation on the axial vibration component and the time series of the material flow rate data to generate vibration and flow rate correlation features. Perform a tensor product operation on the normal pressure distribution matrix and the equivalent heat source intensity parameter to generate pressure and temperature coupling features. Combine the vibration and flow rate correlation features to generate a preprocessing data packet.

[0043] In a second aspect, an embodiment of the present invention provides a real-time transmission system based on Internet of Things perception data in a digital twin scenario, including:

[0044] An acquisition module, configured to acquire vibration spectrum data, pressure gradient change data, and temperature distribution data of devices in an industrial production line, and construct a timestamp alignment mechanism in a 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet;

[0045] An analysis module, configured to analyze the preprocessing data packet to obtain device state fingerprint features and environmental working condition labels;

[0046] A generation module, configured to generate a target virtual control parameter of the device and a target error compensation amount of the clearance of the transmission chain in the device according to the device state fingerprint features and environmental working condition labels, by using a dynamic parameter correction loop of the digital twin body, where the digital twin body is constructed based on the device topology relationship network and process constraint conditions corresponding to the device;

[0047] A determination module, configured to generate an actual material flow state based on the device state fingerprint features, environmental working condition labels, and material flow rate data, and perform coupled analysis on the target virtual control parameter, target error compensation amount, and actual material flow state by using a distributed computing architecture to generate an optimization strategy sequence for multi-device collaborative operation;

[0048] A conversion module, configured to convert the optimization strategy sequence into trajectory compensation parameters of an end effector in the device, and perform data closed-loop transmission in combination with the target error compensation amount.

[0049] In a third aspect, an embodiment of the present invention provides a computing device, including a processor and a memory, where a computer program is stored in the memory, and the processor is configured to run the computer program to execute a real-time transmission method based on Internet of Things perception data in a digital twin scenario according to any one of the first aspects.

[0050] In a fourth aspect, an embodiment of the present invention provides a computer storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, a real-time transmission method based on Internet of Things perception data in a digital twin scenario according to any one of the first aspects is implemented.

[0051] In the embodiments of the present invention, vibration spectrum data, pressure gradient change data, and temperature distribution data of equipment in an industrial production line are collected, and a timestamp alignment mechanism is constructed in a 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line to generate a preprocessing data packet; the preprocessing data packet is analyzed to obtain equipment state fingerprint features and environmental working condition labels; according to the equipment state fingerprint features and environmental working condition labels, a dynamic parameter correction loop of a digital twin is used to generate target virtual control parameters of the equipment and a target error compensation amount for the clearance of the transmission chain in the equipment, where the digital twin is constructed based on the equipment topology relationship network and process constraint conditions corresponding to the equipment; based on the equipment state fingerprint features, environmental working condition labels, and material flow rate data, an actual material flow state is generated, and a distributed computing architecture is used to perform coupled analysis on the target virtual control parameters, target error compensation amount, and actual material flow state to generate an optimization strategy sequence for multi-device collaborative operation; the optimization strategy sequence is converted into trajectory compensation parameters of the end effector in the equipment, and combined with the target error compensation amount, data closed-loop transmission is performed. The technical solution provided by the present invention eliminates the time reference difference of heterogeneous sensor data through time-domain matching of multi-dimensional industrial data (vibration, pressure, temperature) with the material flow rate, ensures the precise synchronization of the time axes of the physical space and the digital twin, and provides a timing consistency guarantee for multi-source data fusion; extracts equipment state fingerprint features from complex working condition data, combines with environmental parameter tagging, realizes the characteristic representation of the equipment operation state and the quantitative isolation of external interference factors, and provides an analyzable input feature space for the digital twin model; a dynamic parameter correction loop constructed based on the equipment topology network and process constraints dynamically associates the physical equipment characteristics with the target virtual model parameters, realizes the multi-physical field coupling calculation of the transmission chain clearance compensation amount and the online iterative optimization of the control parameters, and breaks through the accuracy limitation of traditional static compensation; through the distributed coupling analysis of the target virtual control parameters, target error compensation amount, and actual material flow state, a spatio-temporal constraint relationship model for multi-device collaborative operation is established, and an optimization strategy sequence that matches the process rhythm and equipment dynamic performance is generated, avoiding system-level conflicts caused by single-device optimization; the optimization strategy is converted into trajectory compensation parameters of the end effector and forms a closed-loop feedback with the target error compensation amount, realizing the lossless mapping of digital twin instructions to physical actuators, and ensuring the dynamic following accuracy of control instructions under time-varying working conditions. Among them, through topology-driven multi-physical field feature fusion and working condition adaptive dynamic compensation mechanism, the accuracy and response speed of transmission chain clearance compensation are significantly improved; the collaborative correction of target virtual control parameters combined with equipment node parameter constraints effectively solves the motion interference problem in multi-device collaborative operation and realizes the precise matching of digital twin control instructions and physical equipment dynamic characteristics under complex working conditions.

[0052] These aspects or other aspects of the present invention will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0054] Figure 1 It is a flowchart of a real-time transmission method for Internet of Things perception data in a digital twin scenario provided by an embodiment of the present invention;

[0055] Figure 2 It is a schematic structural diagram of a real-time transmission system for Internet of Things perception data in a digital twin scenario provided by an embodiment of the present invention;

[0056] Figure 3 It is a schematic structural diagram of a computing device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0057] In order to enable those skilled in the art to better understand the solutions of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention.

[0058] In some processes described in the specification and claims of the present invention and the above-mentioned drawings, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear in this article or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0059] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present invention.

[0060] Figure 1The figure below is a flowchart of a real-time transmission method for Internet of Things perception data in a digital twin scenario provided by an embodiment of the present invention. As Figure 1 shown, the method includes:

[0061] Aiming at the technical bottlenecks of traditional digital twin systems in multi-source data synchronization, non-linear working condition modeling, and real-time closed-loop control in intelligent manufacturing scenarios, the present invention proposes a dynamic Internet of Things perception data transmission and processing method. Aiming at the data mismatch problem caused by sudden changes in material flow rate, a pulse-triggered dynamic timestamp alignment mechanism of 5G edge computing nodes is adopted to achieve microsecond-level precise matching of vibration, pressure, and temperature data with the flow rate. To solve the problem of state characterization under complex working conditions such as temperature-vibration coupling, a dynamic coupling model of device state fingerprint features and environmental working condition labels is constructed to achieve adaptive fusion of multi-physical field data. To overcome the defect of large closed-loop response delay in traditional solutions, a coupling analysis engine is developed based on a distributed computing architecture, which greatly shortens the response time of control parameter generation and error compensation. Through the collaborative optimization of time series alignment, feature fusion, and real-time decision-making, this solution effectively solves the key technical problems such as data synchronization inaccuracy, state characterization deviation, and control response lag in industrial digital twin systems, providing a high-precision and highly reliable real-time transmission and control solution for intelligent manufacturing. Based on this, the present invention provides a real-time transmission method for Internet of Things perception data in a digital twin scenario, as Figure 1 follows:

[0062] Step 101: Collect the vibration spectrum data, pressure gradient change data, and temperature distribution data of the equipment in the industrial production line, and construct a timestamp alignment mechanism in the 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessed data packet;

[0063] In this step, the vibration spectrum data refers to the vibration energy distribution data of industrial equipment at different frequencies collected by MEMS sensors, which characterizes the mechanical vibration characteristics during equipment operation. The pressure gradient change data refers to the change rate data of the pressure on the equipment contact surface in the spatial dimension, which reflects the dynamic change characteristics of the pressure distribution during the material processing process. The temperature distribution data refers to the set of temperature measurement values of the equipment surface or the internal heat source area, which is used to analyze the thermodynamic state of the equipment and its impact on mechanical properties. The timestamp alignment mechanism refers to a data synchronization method based on 5G edge nodes, which realizes the time-domain matching of sensing data and material flow rate through pulse signal-triggered dynamic time window resampling of multi-source data. The material flow rate data refers to the real-time monitoring data of the material flow rate in the industrial production line, which is used as the reference quantity for the time series matching of multi-source data. The preprocessed data packet refers to a structured data set after time-domain matching, feature extraction, and data fusion, which contains vibration-flow rate correlation features and pressure-temperature coupling features.

[0064] In an embodiment of the present invention, by deploying MEMS vibration sensors, piezoelectric sensor arrays, and infrared thermal imagers, device vibration spectrum data (three-axis acceleration signals), pressure gradient change data (contact surface normal / tangential pressure difference), and temperature distribution data (surface temperature field) are collected; in a 5G edge computing node, a dynamic timestamp alignment mechanism is triggered based on the pulse signal of the material flow rate data for processing. The processing process includes resampling and aligning the vibration spectrum data using the dynamic time warping algorithm to generate an axial vibration component matching the time axis of the material flow rate; using the adaptive interpolation algorithm to match the time domain characteristics of the flow rate for the pressure gradient change data to generate a normal pressure distribution matrix; establishing a heat conduction delay compensation model for the temperature distribution data and generating an equivalent heat source intensity parameter after correcting the sensor measurement delay; finally, encapsulating the axial vibration component, normal pressure distribution matrix, and equivalent heat source intensity parameter into a preprocessing data packet, which contains vibration, pressure, and temperature three-source data after timestamp alignment.

[0065] Step 102: Analyze the preprocessing data packet to obtain device state fingerprint features and environmental condition labels;

[0066] In this step, the device state fingerprint features refer to the unique characterization parameters of the device operating state extracted from the preprocessing data, including physical quantities such as vibration modes and stress distributions, which are used to represent the real-time performance degradation degree of the device. The environmental condition labels refer to the quantitative identifiers representing production environment parameters (temperature, humidity, air pressure), which are used to dynamically correct the device control strategy.

[0067] In an embodiment of the present invention, multi-physical field feature analysis is performed on the preprocessing data packet, including vibration spectrum analysis, pressure field analysis, and temperature field modeling. The specific process is to perform wavelet packet transform on the axial vibration component of the preprocessing data packet to extract the axial vibration energy density feature in the main axis direction of the device and the vibration mode offset in the radial and tangential directions; perform stress gradient calculation on the normal pressure distribution matrix of the preprocessing data packet to generate the normal stress gradient feature of the contact surface; combine the equivalent heat source intensity parameter of the preprocessing data packet with the infrared thermal imaging data to inversely generate a dynamic temperature field label through the thermodynamic diffusion equation; use a convolutional neural network to fuse the vibration energy density feature and the normal stress gradient feature to generate device state fingerprint features; analyze the time series of the dynamic temperature field label through a long short-term memory network to generate environmental condition labels.

[0068] Step 103: According to the device state fingerprint features and environmental condition labels, use the dynamic parameter correction loop of the digital twin to generate the target virtual control parameters of the device and the target error compensation amount of the clearance of the transmission chain in the device, where the digital twin is constructed based on the device topology relationship network and process constraint conditions corresponding to the device;

[0069] In this step, the dynamic parameter correction loop of the digital twin refers to the control parameter optimization module in the digital twin based on real-time feedback, which generates an error compensation amount and virtual control parameters through multi-physical field coupling analysis. The target virtual control parameters refer to the optimized device motion control instruction set after being optimized by the dynamic correction loop, including the constraint boundaries of displacement, velocity, and acceleration. The target error compensation amount refers to the correction value for the dynamic change of the transmission chain clearance, and the axial position and angular offset compensation amounts adaptively adjusted by the sliding window algorithm. The device topology relationship network refers to the industrial device connection relationship diagram modeled by graph theory, including mechanical degrees of freedom, signal transmission paths, and kinematic constraint parameters between nodes. The process constraint conditions refer to the safety thresholds and physical limitations for the operation of devices in the production process, including boundary conditions such as the maximum allowable deformation and temperature tolerance range.

[0070] In the embodiment of the present invention, based on the device topology relationship network (adjacency matrix modeled by graph theory) and process constraint conditions (including deformation thresholds and material properties), a dynamic parameter correction loop is constructed in the digital twin to decompose the device state fingerprint features into axial vibration modes (FFT spectrum integration), radial thermal expansion coefficients (product of temperature - material expansion rate), and tangential stress distribution (vector projection) features by using this correction loop. Combining the maximum allowable deformation threshold and material temperature sensitivity coefficient in the process constraints, a non-linear weight allocation model is constructed, and a multi-source fusion feature vector is generated through the non-linear weight fusion model. Based on the multi-source fusion feature vector, an initial error compensation amount is generated; the sliding window algorithm (the window length of which is adaptively adjusted according to the flow rate change rate) is used to iteratively optimize the initial error compensation amount to generate the target error compensation amount. Combining the feasible region constraints of the device working parameters and the dynamic boundary constraints of the environmental working condition labels in the digital twin, the target virtual control parameters are generated through the parameter space search algorithm.

[0071] Step 104: Based on the device state fingerprint features, environmental working condition labels, and material flow rate data, generate the actual material flow state, and use the distributed computing architecture to perform coupling analysis on the target virtual control parameters, target error compensation amount, and actual material flow state to generate an optimization strategy sequence for multi-device collaborative operation;

[0072] In this step, the actual material flow state refers to the dynamic material flow model generated based on the device state features and environmental working conditions, including discretized flow rate intervals and bulk density distribution parameters. The optimization strategy sequence refers to the set of timing control instructions for multi-device collaborative operation, which is an action priority sorting scheme generated through coupling analysis by the distributed computing architecture.

[0073] In an embodiment of the present invention, in a distributed computing architecture, a target virtual control parameter is mapped to the kinematic constraint boundary of a device node, a target error compensation amount is converted into a dynamic tolerance interval of the transmission chain clearance, and material flow rate data is used to generate a discretized flow rate interval and a bulk density distribution parameter; an energy transfer function between device nodes (generated based on a difference model of the vibration energy density feature and the constraint boundary) and a material conservation equation (discretized flow rate interval × bulk density distribution parameter) are constructed, and the dynamic tolerance interval is used as a relaxation variable for equation solving; a multi-objective optimization algorithm is used to solve the energy-material joint equation, an initial optimization strategy set is generated, and feasibility screening and dynamic priority ranking are performed based on process constraint conditions (device safety threshold) and device state fingerprint features, and a multi-device collaborative operation optimization strategy sequence is output.

[0074] Step 105: Convert the optimization strategy sequence into trajectory compensation parameters of the end effector in the device, and combine with the target error compensation amount to perform data closed-loop transmission;

[0075] In this step, the trajectory compensation parameter refers to the correction amount of the motion trajectory of the end effector, including the Cartesian space coordinate compensation value and the tangent direction adjustment parameter of the path interpolation function.

[0076] In an embodiment of the present invention, the optimization strategy sequence is parsed into displacement increment parameters, speed adjustment curves, and acceleration constraint thresholds of the end effector, and the trajectory reference point coordinates in the Cartesian space are generated through an inverse kinematic model; the target error compensation amount is decomposed in space to obtain an axial position compensation amount and an angular offset compensation amount, and the compensation amount is superimposed on the trajectory reference point coordinates and the path tangent direction through a homogeneous coordinate transformation matrix to generate compensated trajectory parameters; the compensated trajectory parameters are transmitted to the end effector controller through a 5G downlink, and the actual motion trajectory data is synchronously collected, and the residual between it and the digital twin predicted trajectory is calculated. When the residual exceeds the dynamic tolerance interval of the transmission chain clearance in step 104, a dynamic parameter correction loop is triggered to recalculate the error compensation amount to form a closed-loop control link.

[0077] The embodiment of the present invention solves the technical bottlenecks of traditional digital twin systems in aspects such as data synchronization misalignment, rigid non-linear working condition modeling, and lagging closed-loop control response through multi-source data dynamic alignment, multi-physical field feature fusion, and distributed collaborative optimization, and significantly improves the state perception accuracy, control real-time performance, and dynamic error suppression ability of multi-device collaborative operations in industrial production lines.

[0078] The present invention provides a specific embodiment. In step 103, according to the device state fingerprint features and environmental working condition labels, using the dynamic parameter correction loop of the digital twin, the target virtual control parameters of the device and the target error compensation amount of the transmission chain clearance in the device are generated. The digital twin is constructed based on the device topology relationship network corresponding to the device and process constraint conditions, and specifically includes the following steps:

[0079] Step 301: Based on the device topology relationship network and process constraint conditions, construct a digital twin. The device topology relationship network uses a graph theory modeling method to establish the mechanical connection attributes and signal transmission paths between device nodes;

[0080] In this step, the mechanical connection attributes between device nodes refer to the mechanical coupling relationships formed between industrial device nodes through physical structures (such as bearings, couplings, gear sets, etc.), including parameters such as connection stiffness, degree-of-freedom constraints, and transmission ratios, which are used to describe the power transmission characteristics between devices. The signal transmission path refers to the communication link topology for transmitting data such as control instructions and status feedback between device nodes, defining the signal transmission direction, delay characteristics, and bandwidth limitations, and reflecting the logical channels for information interaction in the digital twin.

[0081] In the embodiment of the present invention, when constructing the digital twin, first establish a device topology relationship network based on the graph theory modeling method. The construction process includes defining device nodes, mechanical connection attribute modeling, signal transmission path modeling, and process constraint integration. The process of defining device nodes includes abstracting each device in the production line (such as motors, conveyor belts, robotic arms) as nodes in graph theory and marking their geometric dimensions (such as shaft length, contact surface curvature) and kinematic parameters (such as degrees of freedom, maximum rotational speed); the process of mechanical connection attribute modeling includes describing the physical connection relationship between devices through an adjacency matrix. For example, if device A is connected to device B through a gear set, its adjacency weight is the transmission stiffness (the square of the shaft diameter multiplied by the shear modulus of the material and then divided by the shaft length) and the transmission ratio (the number of teeth of the driving gear divided by the number of teeth of the driven gear); the process of signal transmission path modeling includes using an incidence matrix to define the control signal transmission logic between devices. For example, the path delay for device A to send a control instruction to device B is the communication distance between devices divided by the signal transmission rate, and the bandwidth limitation is the minimum communication channel capacity; then, embed the maximum allowable deformation threshold (yield strength of the material divided by the safety factor) and material property parameters (such as thermal expansion coefficient, elastic modulus) in the process constraint conditions into the parametric model of the digital twin to form a physical-information fusion digital twin architecture. The device topology relationship network completely describes the mechanical and signal coupling relationships between devices through a node attribute table, adjacency matrix, and incidence matrix, providing a topological basis for the dynamic simulation of the digital twin.

[0082] Step 302: Based on the dynamic parameter correction loop in the digital twin, decompose the device state fingerprint features into axial vibration mode features, radial thermal expansion coefficient features, and tangential stress distribution features. Combine the maximum allowable deformation threshold in the process constraint conditions and the temperature sensitivity coefficient of the material property parameters in the digital twin to generate a multi-source fusion feature vector;

[0083] In this step, the axial vibration mode features refer to the vibration energy distribution characteristics in the direction of the device spindle. The natural frequency, damping ratio, and amplitude decay coefficient extracted through spectrum analysis characterize the dynamic stability of the device in the axial direction. The radial thermal expansion coefficient features refer to the deformation gradient parameters caused by temperature changes in the radial direction (perpendicular to the spindle direction) of the device, which are obtained by multiplying the material thermal expansion coefficient and the temperature distribution data, reflecting the influence of thermal stress on the geometric accuracy of the device. The tangential stress distribution features refer to the set of stress vectors in the tangential direction of the device contact surface, which are generated through the convolution operation of the pressure gradient and the curvature tensor and are used to quantify the interference of tangential forces such as friction and shear on the device motion trajectory. The maximum allowable deformation threshold refers to the safety limit value of the device deformation specified in the process constraint conditions. Exceeding this threshold may lead to mechanical failure or out-of-tolerance machining accuracy, usually calibrated jointly by the material yield strength and the working condition load. The temperature sensitivity coefficient refers to the gradient factor characterizing the thermodynamic response in the material property parameters, defined as the ratio of the change in the material deformation rate or mechanical properties (such as elastic modulus) caused by a unit temperature change. The multi-source fusion feature vector refers to a composite vector generated by weighted fusion of the axial vibration, radial thermal expansion, and tangential stress features. Each dimension corresponds to a type of physical field feature and is normalized to a dimensionless scalar to support cross-domain analysis.

[0084] In the embodiment of the present invention, in the dynamic parameter correction loop, the physical field decoupling of the device state fingerprint features is first performed. The specific process includes extracting the axial vibration mode features by performing frequency-domain integration of the vibration spectrum data in the main axis direction (total energy = sum of the squares of the amplitudes of each frequency), which characterizes the core vibration energy distribution of the device; calculating the radial thermal expansion coefficient features by multiplying the temperature distribution data point by point with the material thermal expansion coefficient (deformation gradient = temperature value × thermal expansion coefficient) to quantify the influence of thermal deformation; obtaining the tangential stress distribution features through the outer product operation of the pressure gradient data and the contact surface curvature vector (tangential stress = pressure gradient × curvature vector modulus length) to reflect the friction and shear effects; subsequently, feature fusion is performed in combination with process constraints and material characteristics, specifically including normalizing each eigenvalue by dividing it by its corresponding maximum allowable deformation threshold (such as radial eigenvalue / radial deformation threshold); performing dynamic compensation on the radial thermal expansion coefficient features by superimposing the temperature sensitivity coefficient (compensated feature = original feature × temperature sensitivity coefficient × current temperature offset rate) to obtain the compensated radial thermal expansion coefficient features; finally, the three types of features (including the compensated radial thermal expansion coefficient features, axial vibration mode features, and tangential stress distribution features) are spliced into a multi-source fusion feature vector according to the normalized weights, where the axial weight ratio is the ratio of the vibration energy to the total energy, the radial weight is the ratio of the compensated feature, and the tangential weight is determined by the product of the stress vector modulus length and the friction coefficient.

[0085] Step 303: Based on the multi-source fusion feature vector, generate an initial error compensation amount, and dynamically adjust the initial error compensation amount using a sliding window algorithm to obtain a target error compensation amount, where the window length is adaptively adjusted according to the change rate of the material flow rate data;

[0086] In this step, the initial error compensation amount refers to a set of compensation parameters generated based on the difference between the multi-source fusion feature vector and the maximum allowable deformation threshold, and includes preliminary estimated values of the axial displacement compensation and angular offset compensation of the transmission chain clearance.

[0087] In the embodiment of the present invention, each eigenvalue in the multi-source fusion feature vector is compared with the corresponding threshold (such as the maximum allowable deformation threshold), the difference is calculated, and the differences are weighted and summed to generate an initial error compensation vector including axial displacement compensation and angular offset compensation; according to the change rate of the material flow rate (such as shortening the window when the flow rate suddenly changes), the number of data points in the sliding window is adjusted; the data points in the window are weighted according to a time decay function (such as exponential decay), and the weight of the latest data is higher; based on the error trend after weighted averaging within the window, the initial compensation amount is iteratively corrected to generate a target error compensation amount.

[0088] Step 304: Based on the feasible region of the working parameters of each device node in the digital twin and the environmental condition label, generate an initial virtual control parameter;

[0089] In this step, the initial virtual control parameters refer to the device motion control reference values (such as target displacement, speed, acceleration) preset according to the environmental condition labels within the feasible region of the device operating parameters, representing the original control instruction set that has not considered the influence of mechanical coupling and errors.

[0090] In the embodiment of the present invention, first, based on the preset feasible regions of the operating parameters of each device node in the digital twin (such as physical constraint boundaries like displacement limits, speed upper limits, acceleration thresholds, etc.), a set of control variable parameters to be screened is extracted. This set of control variable parameters includes displacement candidate parameters, speed candidate parameters, acceleration candidate parameters, etc.; Subsequently, the environmental condition labels (such as temperature grade, humidity value) are converted into environmental correction factors through a non - linear mapping relationship, including thermal expansion correction factor, friction correction factor, and dynamic load correction factor. For example, the thermal expansion correction factor is the ratio of the environmental temperature to the standard thermal expansion coefficient of the material, which is used to scale the displacement parameter (such as compensating for thermal elongation of the displacement at high temperatures), and the friction correction factor is the output value of the mapping function between humidity and the friction coefficient (such as when the humidity increases and the friction coefficient rises, the speed parameter needs to be adjusted downward according to the reciprocal relationship); The dynamic load correction factor is the proportional coefficient of the inertial load caused by the change in the material flow rate, which is used to adjust the acceleration parameter (such as when the flow rate suddenly increases, the acceleration needs to be reduced accordingly); Then, within this feasible region, with the process constraint conditions as the boundary, a parameter space traversal search method is used to find the control parameter combination that meets the working condition adaptability. For example, within the range of acceleration candidate parameters, the product of each acceleration candidate parameter and the dynamic load correction factor is iteratively calculated step - by - step to screen out the set of corrected acceleration parameters that minimizes the deviation of the device motion trajectory. Similarly, the corrected displacement parameter set = displacement candidate parameter × (1 + thermal expansion correction factor), and the corrected speed parameter set = speed candidate parameter × (1 / friction correction factor). Finally, the initial virtual control parameters that conform to both physical constraints and the current environmental conditions are generated.

[0091] Step 305: Dynamically and cooperatively correct the initial virtual control parameters according to the coupling relationship of the mechanical connection attributes of the device and the target error compensation amount to obtain the target virtual control parameters;

[0092] In the embodiments of the present invention, during the dynamic collaborative correction process, first, a multi-device dynamics coupling equation is constructed based on the mechanical connection attributes (such as gear transmission ratio, coupling stiffness, etc.) defined in the device topology relationship network. The axial displacement compensation component in the target error compensation amount is converted into a position correction term (the displacement adjustment amount is obtained by dividing the compensation amount by the equivalent stiffness of the transmission chain), and the angular offset compensation component is converted into an attitude correction matrix (the angular adjustment amount is obtained by multiplying the compensation amount by the inverse matrix of the inertia tensor). Then, the initial virtual control parameters (such as target displacement, velocity) are substituted into the multi-device dynamics coupling equation, and collaborative optimization is performed through a distributed iterative algorithm: in each iteration, first, the theoretical motion state of each device node is calculated according to the current control parameters, and then compared with the expected state after injecting the error compensation. The difference is multiplied by a dynamic weight coefficient (obtained by normalizing the reciprocal of the mechanical connection stiffness) to generate a control parameter correction amount. Finally, when the motion state deviation of all device nodes is less than the preset convergence threshold, the target virtual control parameters that satisfy the mechanical coupling constraint and compensate for the clearance error of the transmission chain are output. This process ensures the physical executability of control instructions and system stability by coordinating the mechanical interaction and error compensation among multiple devices in real time.

[0093] The embodiments of the present invention significantly improve the control accuracy and real-time performance of the industrial digital twin system through dynamic feature fusion, adaptive error compensation, and multi-device collaborative correction. Aiming at problems such as data synchronization misalignment, rigid feature fusion, and closed-loop response lag in the prior art, it realizes the precise matching of multi-physical field data, adaptive modeling of non-linear working conditions, and dynamic optimization of multi-device collaborative control, effectively suppressing the influence of complex factors such as mechanical wear and thermal deformation on the stability of the production line, and providing a highly reliable real-time control solution for the intelligent manufacturing scenario.

[0094] The present invention provides a specific embodiment. Step 302: Generate a multi-source fusion feature vector by combining the maximum allowable deformation threshold in the process constraint conditions and the temperature sensitivity coefficient of the material property parameters in the digital twin body, specifically including the following steps:

[0095] Step 311: Based on the axial vibration mode characteristics, obtain the vibration energy value through the spectral energy integration of the vibration signal in the main axis direction of the device. Based on the radial thermal expansion coefficient characteristics, calculate the deformation gradient value through the product relationship between the temperature distribution data and the material thermal expansion coefficient. Based on the tangential stress distribution characteristics, calculate the vector product result through the pressure gradient data and the contact surface curvature.

[0096] In this step, the spectral energy integration refers to the cumulative calculation of the energy of the device vibration signal within a specific frequency domain range (such as the natural frequency bandwidth in the spindle direction), which is obtained by integrating the square of the spectral amplitude after Fourier transform with respect to frequency, and represents the overall intensity of the vibration energy. The vibration energy value refers to the scalar value obtained through spectral energy integration, which quantifies the vibration energy level of the device in a specific direction (such as the axial direction) and is used to evaluate the dynamic stability of the mechanical structure. The coefficient of thermal expansion of the material refers to the material property parameter, which is defined as the linear expansion rate of the material caused by a unit temperature change (unit: 1 / °C) and is used to calculate the deformation amount under a temperature gradient. The deformation gradient value refers to the parameter generated by multiplying the temperature distribution data by the coefficient of thermal expansion of the material, which represents the thermal expansion deformation amount per unit length (such as the number of micrometers of deformation per millimeter in the radial direction). The pressure gradient data refers to the rate of change of the normal pressure on the contact surface measured by the pressure sensor array (unit: Pa / mm), which describes the spatial distribution characteristics of the pressure in the device contact area. The contact surface curvature refers to the geometric curvature parameter of the device contact surface (unit: 1 / mm), which is obtained by calculating the three-dimensional point cloud data and reflects the influence of the bending degree of the contact surface on the stress distribution. The vector product result refers to the output of the vector product operation of the pressure gradient data and the contact surface curvature tensor, which generates a tangential stress distribution vector (unit: N / mm²) and represents the direction and intensity of the friction and shear forces.

[0097] In the embodiment of the present invention, first, the axial vibration mode characteristics are processed, including performing spectral analysis on the vibration signal in the spindle direction of the device through fast Fourier transform, extracting the spectral amplitude within the characteristic frequency band (usually within the range of ±10% of the natural frequency of the device), and integrating and summing the squares (i.e., energies) of the amplitudes at each frequency point in the frequency domain to obtain the vibration energy value representing the vibration intensity of the device; then, the radial coefficient of thermal expansion characteristics are processed, including multiplying the temperature value at each sampling point of the temperature distribution data collected by the infrared thermal imager by the coefficient of thermal expansion of the material at the corresponding position to obtain the instantaneous thermal expansion amount at each point, and then calculating the difference in thermal expansion amounts between adjacent points through a spatial gradient operator, and finally generating the deformation gradient value reflecting the non-uniform degree of radial deformation of the device; finally, the tangential stress distribution characteristics are processed, including using the pressure gradient data obtained by the pressure sensor array, first reconstructing the three-dimensional curvature characteristics of the device contact surface through a surface fitting algorithm, and then performing a vector product operation on the pressure gradient vector and the curvature normal vector (the product of the pressure gradient magnitude, the radius of curvature, and the sine of the included angle between the two) to obtain the vector product result representing the tangential frictional stress.

[0098] Step 312: Calculate the first deformation contribution rate according to the ratio of the vibration energy value to the maximum allowable deformation threshold; calculate the second deformation contribution rate according to the ratio of the deformation gradient value to the maximum allowable deformation threshold, and dynamically compensate the second deformation contribution rate according to the deviation degree of the ambient temperature from the standard working condition temperature of the material to generate a corrected second deformation contribution rate;

[0099] In this step, the first deformation contribution rate refers to the ratio of the vibration energy value to the maximum allowable deformation threshold, reflecting the contribution degree of vibration to the deformation of the device (dimensionless). The second deformation contribution rate refers to the geometric relationship ratio of the deformation gradient value to the maximum allowable deformation threshold, characterizing the contribution ratio of thermal expansion to deformation (dimensionless). The standard working condition temperature of the material refers to the reference temperature (unit: °C) defined during material property testing or device design, which is used as the reference condition for calibrating the coefficient of thermal expansion. The corrected second deformation contribution rate refers to the result of dynamically adjusting the second deformation contribution rate according to the deviation degree between the ambient temperature and the standard working condition temperature of the material. The adjustment formula is: the second deformation contribution rate × (1 + temperature deviation coefficient × temperature sensitivity coefficient), where the temperature deviation coefficient is obtained by dividing the absolute value of the temperature difference between the current temperature and the standard working condition temperature by the standard temperature value.

[0100] In the embodiment of the present invention, the vibration energy value is divided by the maximum allowable deformation threshold in the process constraint conditions to obtain the dimensionless first deformation contribution rate. Secondly, through the temperature distribution data, the temperature gradient between adjacent measurement points is calculated (obtained by dividing the temperature difference by the spatial distance), and the temperature gradient is multiplied by the coefficient of thermal expansion of the material to generate the deformation gradient value (the amount of deformation per unit length). Then, the gradient value is divided by the maximum allowable deformation threshold and logarithmically mapped to the [0, 1] interval to generate the second deformation contribution rate. For the problem that the ambient temperature deviates from the standard working condition, the temperature deviation coefficient is obtained by dividing the absolute value of the temperature difference between the current temperature distribution data and the standard working condition temperature of the material by the standard temperature value. The coefficient is multiplied by the temperature sensitivity coefficient in the material property parameters to generate a dynamic compensation factor. According to the second deformation contribution rate × (1 + dynamic compensation factor), the corrected second deformation contribution rate is output, realizing the dynamic increase of the contribution rate under high-temperature working conditions or the suppression adjustment under low-temperature working conditions.

[0101] Step 313: According to the cross product result, perform weighted accumulation on the projection component in the tangent direction of the device motion trajectory to calculate the third deformation contribution rate;

[0102] In this step, the projection component refers to the scalar value of the component of the cross product result in the tangent direction of the device motion trajectory, which is obtained through vector dot product calculation and reflects the actual influence of the tangential stress on the motion trajectory. The third deformation contribution rate refers to the weighted accumulation value of the tangential stress projection component. The calculation formula is: Σ (projection component × contact surface friction coefficient) / maximum allowable deformation threshold, where Σ represents the cumulative sum of the products of the projection components and the corresponding friction coefficients of all sampling points on the device contact surface.

[0103] In the embodiment of the present invention, first, the vector product result of the tangential stress distribution characteristics is decomposed into stress vectors of each discrete point on the equipment contact surface, and the unit vector in the tangent direction of the equipment motion trajectory is extracted through the kinematic model; a projection calculation is performed on each stress vector (the dot product of the stress vector and the unit vector in the tangent direction) to obtain the stress projection component of each point in the tangent direction. Subsequently, based on the contact surface friction coefficient distribution (the higher the friction coefficient, the greater the weight), the projection components are weighted: each projection component is multiplied by the friction coefficient weight value at the corresponding position, and then all the weighted projection components are accumulated to obtain the total tangential stress contribution value. Finally, the total tangential stress contribution value is divided by the maximum allowable deformation threshold defined in the process constraint conditions and mapped to the 0-1 interval through a non-linear normalization function (such as the Sigmoid function) to generate the third deformation contribution rate. This process quantifies the dynamic influence degree of the tangential stress on the equipment deformation through the accumulation of stress projections weighted by the friction coefficient.

[0104] Step 314: Calculate the second reference weight according to the corrected second deformation contribution rate and the temperature sensitivity coefficient of the material property parameters in the digital twin.

[0105] In this step, the second reference weight refers to the weight factor generated by the normalized calculation of the corrected second deformation contribution rate and the temperature sensitivity coefficient and is used for multi-source feature fusion. The calculation formula is: corrected second deformation contribution rate × temperature sensitivity coefficient / total contribution rate sum, where the total contribution rate sum is the addition result of the first deformation contribution rate, the corrected second deformation contribution rate, and the third deformation contribution rate.

[0106] In the embodiment of the present invention, the second reference weight = corrected second deformation contribution rate × temperature sensitivity coefficient / total contribution rate sum, and the total contribution rate sum is the addition result of the first deformation contribution rate, the corrected second deformation contribution rate, and the third deformation contribution rate.

[0107] Step 315: Calculate the first reference weight and the third reference weight according to the first deformation contribution rate, the third deformation contribution rate, the vibration energy attenuation coefficient, and the friction coefficient correction factor, and combine the second reference weight, the first quantization value, the second quantization value, and the third quantization value to generate a multi-source fusion feature vector.

[0108] In this step, the first reference weight refers to the weight factor after normalization of the product of the first deformation contribution rate and the vibration energy attenuation coefficient, and the calculation formula is: (the first deformation contribution rate × the vibration energy attenuation coefficient) / the sum of total weights. The third reference weight refers to the functional relationship value between the third deformation contribution rate and the friction coefficient of the contact surface, and the calculation formula is: the third deformation contribution rate × the friction coefficient correction factor / the sum of total weights, where the sum of total weights = (the first deformation contribution rate × the vibration energy attenuation coefficient) + (the corrected second deformation contribution rate × the temperature sensitivity coefficient) + (the third deformation contribution rate × the friction coefficient correction factor), and the friction coefficient correction factor refers to the dynamic adjustment coefficient (dimensionless) generated by fitting historical friction data or real-time working conditions (such as wear and lubrication status).

[0109] In the embodiment of the present invention, the first reference weight is calculated according to (the first deformation contribution rate × the vibration energy attenuation coefficient) / the sum of total weights; the third reference weight is calculated according to the third deformation contribution rate × the friction coefficient correction factor / the sum of total weights. Subsequently, the first reference weight, the second reference weight, and the third reference weight are normalized (each reference weight is divided by the sum of the three), and the normalized weight coefficients are obtained; finally, the first quantization value is multiplied by the normalized first weight coefficient to generate a weighted vibration component, the second quantization value is multiplied by the normalized second weight coefficient to generate a weighted thermal deformation component, and the third quantization value is multiplied by the normalized third weight coefficient to generate a weighted stress component. The three types of weighted components are combined into a multi-source fusion feature vector through vector splicing. This process ensures the precise fusion of the contribution rates and quantization values of the axial vibration, radial thermal expansion, and tangential stress characteristics in the spatio-temporal dimension through a dynamic weight distribution mechanism.

[0110] In the embodiment of the present invention, the normalization evaluation of the deformation influence is realized by combining the contribution rates (the first, second, and third deformation contribution rates) in the three dimensions of axial, radial, and tangential with the maximum allowable deformation threshold, reducing the deformation characterization error; the contribution rate of radial thermal expansion (the second deformation contribution rate) is dynamically corrected according to the deviation degree between the environmental temperature and the standard working condition temperature, and the temperature sensitivity coefficient of the material is introduced to generate the second reference weight, improving the thermal deformation compensation accuracy under high-temperature working conditions; the generated multi-source fusion feature vector has working condition adaptability, and the feature characterization stability is improved in scenarios of sudden load or mechanical wear, providing a high-fidelity input for the subsequent generation of virtual control parameters.

[0111] The present invention provides a specific embodiment. Step 104: Based on the device state fingerprint features, environmental working condition labels, and material flow rate data, generate the actual material flow state, and use a distributed computing architecture to perform a coupling analysis on the target virtual control parameters, the target error compensation amount, and the actual material flow state to generate an optimization strategy sequence for multi-device collaborative operation, specifically including the following steps:

[0112] Step 401: Based on the vibration energy distribution parameters in the device state fingerprint features and the temperature and humidity parameters in the environmental condition label, process the material flow rate data to generate discretized flow rate intervals and the corresponding material bulk density distribution parameters, and combine the vibration energy distribution parameters to calculate the material transfer efficiency attenuation factor between adjacent devices;

[0113] In this step, the vibration energy distribution parameters refer to the distribution characteristics of the vibration energy of the device in different frequency bands or spatial directions during operation, which are the energy density functions obtained by spectral analysis of the device state fingerprint features and are used to quantify the dynamic stability of the device. The temperature and humidity parameters refer to the temperature and humidity measurement data included in the environmental condition label, which are collected in real time by sensors and are used to characterize the thermodynamic state of the environment where the device is located. The discretized flow rate interval refers to the discretized interval formed by segmenting the continuous material flow rate data according to time or the flow rate change rate, and each interval corresponds to a specific flow rate range and time window. The material bulk density distribution parameters refer to the statistical characteristics of the material bulk density at different positions during the device transmission process, which reflect the uniformity of the material flow. The material transfer efficiency attenuation factor refers to the attenuation degree of the material transfer efficiency between adjacent devices, and the larger the value, the lower the transfer efficiency.

[0114] In the embodiments of the present invention, an adaptive binning method is used for the time series of the material flow rate data to discretize the flow rate intervals according to the flow rate change rate, and the interval boundaries are divided according to the change rate threshold obtained by dividing the flow rate standard deviation by the average flow rate; within each flow rate interval, the material bulk density distribution parameter is calculated by combining the temperature and humidity parameters in the environmental condition label. The specific calculation formula is: material bulk density distribution parameter = material weight ÷ effective volume within the interval × temperature and humidity correction factor. The temperature and humidity correction factor is calculated according to the formula: temperature and humidity correction factor = 1 + (measured temperature - standard temperature) × temperature coefficient + (measured humidity - standard humidity) × humidity coefficient. Finally, the material bulk density distribution parameter corresponding to each discretized flow rate interval is obtained; subsequently, the vibration energy distribution parameter between adjacent devices is extracted. The extraction process includes preprocessing the original vibration signal (collected by the acceleration sensor) in the device state fingerprint feature (such as denoising and normalization), and obtaining the vibration signal spectrum analysis result through Fourier transform or power spectral density analysis, and generating the vibration energy distribution parameter based on this; taking the duration ratio of the high vibration energy section calculated according to this parameter (duration ratio = total length of high energy period ÷ monitoring period) as the vibration energy influence coefficient, and at the same time, based on the material bulk density distribution parameter, calculating the material bulk density gradient (density difference between adjacent devices ÷ transmission path length); finally, multiplying the vibration energy influence coefficient by the material bulk density gradient, and superimposing the flow resistance term calculated by the viscosity coefficient model through the temperature and humidity parameters (resistance term = reference viscosity coefficient × (1 + temperature influence coefficient on viscosity × (measured temperature - standard temperature) + humidity influence coefficient on viscosity × (measured humidity - standard humidity)) to generate a material transfer efficiency attenuation factor that comprehensively reflects the influence of mechanical vibration and material characteristics.

[0115] Step 402: Generate a flow state correction coefficient according to the non-linear relationship between the temperature and humidity parameters and the material viscosity coefficient. Using the distributed computing architecture, map the target virtual control parameter to the kinematic constraint boundary of the device node, convert the target error compensation amount into the dynamic tolerance interval of the transmission chain clearance, and embed the flow state correction coefficient as a time-varying weight factor into the matrix dimension to construct a device collaborative analysis matrix;

[0116] In this step, the material viscosity coefficient refers to the physical property parameter of the material, which reflects its flow resistance characteristics and is dynamically adjusted with changes in temperature and humidity. The viscosity coefficient decreases in a high-temperature or high-humidity environment. The dynamic correction factor generated based on the non-linear relationship between the material viscosity coefficient and temperature and humidity parameters is used to adjust the mathematical model parameters of the material flow state. The kinematic constraint boundary refers to the motion limit conditions of the equipment nodes, including the maximum displacement, speed, and acceleration thresholds, which are generated by mapping the target virtual control parameters. The dynamic tolerance interval refers to the real-time error range allowed for the transmission chain clearance, which is dynamically adjusted according to the target error compensation amount and includes the upper tolerance limits of axial displacement and angular offset. The flow state correction coefficient refers to the dynamic adjustment factor generated through a non-linear function based on the material viscosity coefficient and temperature and humidity parameters, which is used to quantify the dynamic change characteristics of the material flow state affected by the environmental working conditions. The time-varying weight factor refers to the dynamic weight parameter embedded in the equipment collaborative analysis matrix, which adjusts the contribution ratio of different physical field characteristics with changes in time or working conditions. The equipment collaborative analysis matrix refers to the multi-dimensional data structure used for multi-equipment collaborative optimization, which integrates the kinematic constraint boundary, dynamic tolerance interval, and time-varying weight factor to support the coupled analysis of energy and materials.

[0117] In the embodiment of the present invention, based on the experimentally calibrated temperature and humidity-viscosity coefficient mapping table, a non-linear calculation model (the product of temperature and humidity parameters and viscosity coefficient divided by the reference working condition value) is generated through polynomial fitting to calculate the flow state correction coefficient; the target virtual control parameters are decomposed into displacement, speed, and acceleration components according to the equipment degrees of freedom by using a distributed computing architecture, and combined with kinematic constraint rules (such as the maximum joint angle and limit speed) to be mapped into the kinematic constraint boundary; the target error compensation amount is decomposed according to the axial and tangential components of the transmission chain clearance, and converted into the axial dynamic tolerance interval (the compensation amount divided by the nominal value of the transmission chain clearance to generate a proportionality coefficient) and the angular tolerance interval (the compensation amount multiplied by the reciprocal of the clearance radius) through linear interpolation and polar coordinate transformation, and combined to form the dynamic tolerance interval; the flow state correction coefficient is sliced according to the time series and embedded in the weight dimension of the equipment collaborative analysis matrix through tensor outer product operation (the coefficient of each time slice is multiplied by the corresponding matrix element); finally, the kinematic constraint boundary (matrix row constraint), dynamic tolerance interval (matrix column relaxation factor), and time-varying weight factor are integrated in the distributed architecture to construct an equipment collaborative analysis matrix with spatio-temporal correlation. This process ensures the efficient fusion of parameters of large-scale equipment nodes through distributed data sharding and parallel computing.

[0118] Step 403: Construct an energy transfer function and a material conservation equation between equipment nodes based on the connection topology in the equipment topology relationship network and the equipment collaborative analysis matrix;

[0119] In this step, the energy transfer function refers to a mathematical model that describes the energy transfer relationship between devices and is constructed based on the difference between the vibration energy distribution parameters and the kinematic constraint boundaries. The material conservation equation represents a mathematical expression for the mass conservation of material flow and is established by discretizing the product relationship between the flow velocity interval and the material accumulation density distribution parameters.

[0120] In the embodiment of the present invention, based on the connection topology defined in the device topology relationship network (clarifying the mechanical coupling relationship between device nodes), the kinematic constraint boundaries are extracted from the device collaborative analysis matrix as the boundary conditions for energy transfer, the dynamic tolerance interval is used as the relaxation variable for material conservation, and the time-varying weight factor is used as the energy distribution coefficient; first, according to the connection topology, determine the energy interaction path between adjacent device nodes (such as a gear transmission chain), compare the vibration energy distribution parameters in the device state fingerprint feature with the kinematic constraint boundaries, and calculate the energy transfer efficiency by multiplying the vibration energy by the time-varying weight factor and then dividing by the equivalent damping coefficient of the transmission chain, establish the energy conservation equation between nodes (the input energy is equal to the sum of the output energy and the loss energy calculated by the equivalent damping coefficient and the square of the vibration energy), and establish the energy transfer relationship between device nodes through the energy conservation equation to construct the energy transfer function between device nodes; subsequently, based on the material flow path in the connection topology (such as a conveyor belt), multiply the discretized flow velocity interval in the preprocessed data packet by the material accumulation density distribution parameters to obtain the real-time material inventory of the node, combine the dynamic tolerance interval to set the allowable inventory fluctuation range, calculate the inventory change amount by multiplying the difference between the inventories of adjacent nodes by the flow velocity time step, establish the material balance equation between nodes (the output of the upstream node is equal to the sum of the input of the downstream node and the inventory change amount), and establish the material flow relationship through the material balance equation to define the material conservation equation.

[0121] Step 404: Generate an initial optimization strategy by iteratively solving the joint solution set of the energy transfer function and the material conservation equation, and optimize the initial optimization strategy by combining the material transfer efficiency decay factor and the device safe operation threshold in the process constraint conditions to obtain a preliminarily optimized optimization strategy;

[0122] In this step, the joint solution set refers to the solution set that simultaneously satisfies the energy transfer function and the material conservation equation and is generated by a numerical iteration method. The initial optimization strategy refers to a preliminary collaborative control scheme generated based on the joint solution set and includes the motion parameter adjustment instructions for each device. The device safe operation threshold refers to the device safe operation limit value defined in the process constraint conditions, including the maximum vibration amplitude, temperature threshold, etc., and is used to screen feasible strategies. The preliminarily optimized optimization strategy refers to the set of optimized strategies screened by the device safe operation threshold and dynamically sorted according to the material transfer efficiency decay factor.

[0123] In the embodiments of the present invention, first, based on the energy transfer function between device nodes and the material conservation equation, an improved Newton iteration method is used to solve the joint solution set. During the iterative solution process, a multi-objective optimization function is constructed through the joint solution set (including the device energy consumption deviation output by the energy transfer function and the material flow deviation calculated by the material conservation equation): the energy consumption deviation in the joint solution set is multiplied by the energy weight coefficient, and the material flow deviation is multiplied by the material weight coefficient, and then weighted summation is performed to form a comprehensive optimization objective; by iteratively adjusting the device control parameters (such as displacement, speed), the objective function is converged to within a preset threshold, and an initial optimization strategy that satisfies the constraints of the joint solution set is output; subsequently, the effectiveness of the initial optimization strategy is evaluated in combination with the material transfer efficiency decay factor (calculated from the vibration energy difference between adjacent devices), and the strategies with a material transfer efficiency lower than the decay factor threshold are eliminated; at the same time, the feasibility of the remaining strategies is verified according to the device safe operation threshold (such as the maximum allowable power, temperature limit), and the strategies exceeding the safety threshold are deleted to generate a preliminarily optimized optimization strategy.

[0124] Step 405: Based on the distributed computing architecture and the device topology relationship network, perform distributed verification on the preliminarily optimized optimization strategy to generate an optimization strategy sequence for multi-device collaborative operation;

[0125] In this step, distributed verification refers to verifying the coordination and executability of the optimization strategy in the distributed computing architecture according to the device topology relationship network, detecting strategy conflicts and dynamically correcting them.

[0126] In the embodiments of the present invention, during the distributed verification process, based on the connection topology of the device topology relationship network, the preliminarily optimized strategy is assigned to the computing units corresponding to each device node. Through conflict detection, strategy adjustment, and sequence generation, an optimization strategy sequence is generated. Specifically, using the kinematic constraint boundary (such as the maximum speed threshold) and the dynamic tolerance interval (such as the allowable deviation of the transmission chain clearance) in the device collaboration analysis matrix, the spatio-temporal consistency of the theoretical motion trajectories after the execution of each node's strategy is verified, and the trajectory overlap degree of adjacent device nodes is calculated (the overlap distance divided by the upper limit of the tolerance interval); when a trajectory conflict (the overlap degree exceeds the preset threshold) is detected, the speed parameter of the conflict node is adjusted according to the time-varying weight factor (dynamically generated by the flow state correction coefficient), and the adjustment amount is the conflict distance multiplied by the weight factor and then divided by the transmission chain stiffness coefficient; through the iterative convergence algorithm, the adjusted strategies of all nodes are globally synchronized. When the motion trajectories of all device nodes satisfy the topological connection constraints and there are no conflicts, the optimization strategy sequence for multi-device collaborative operation is output.

[0127] In the embodiments of the present invention, by constructing a device collaborative analysis matrix and a dynamic coupling analysis mechanism, the problems of inaccurate state modeling and lagging strategy generation in traditional methods for multi-device collaborative optimization are effectively solved; by deeply integrating device state characteristics, operating condition parameters, and material flow data, accurate modeling of energy transfer and material flow between devices is achieved, significantly improving the adaptability and real-time performance of optimization strategies, and overcoming the defect of poor collaborative control effect of existing solutions under complex operating conditions.

[0128] The present invention provides a specific embodiment. In step 402, according to the non-linear relationship between temperature and humidity parameters and the material viscosity coefficient, a flow state correction coefficient is generated. Using a distributed computing architecture, the target virtual control parameters are mapped to the kinematic constraint boundaries of device nodes, the target error compensation amount is converted into the dynamic tolerance interval of the transmission chain clearance, and the flow state correction coefficient is embedded as a time-varying weight factor into the matrix dimension to construct a device collaborative analysis matrix, which specifically includes the following steps:

[0129] Step 411: Based on the non-linear relationship between temperature and humidity parameters and the material viscosity coefficient, a three-dimensional interpolation grid is established at the temperature and humidity sampling points of the environmental condition label to calculate the change gradient of the material viscosity coefficient under continuous operating conditions. Combining with the material flow velocity data, a flow state correction coefficient is generated;

[0130] In this step, the three-dimensional interpolation grid refers to a continuous interpolation model established in the three-dimensional space composed of temperature and humidity parameters (temperature, humidity) and the material viscosity coefficient. The grid nodes are discrete temperature and humidity sampling points (such as temperature sampling points T1 - Tn, humidity sampling points H1 - Hn), and each node corresponds to the material viscosity coefficient value calibrated in the laboratory. The viscosity coefficient under any temperature and humidity combination is calculated through an interpolation algorithm (such as cubic spline interpolation) to generate the change surface of the viscosity coefficient under continuous operating conditions; the change gradient of the material viscosity coefficient refers to the derivative (slope) of the material viscosity coefficient with respect to the change of temperature and humidity, indicating the change amount of the viscosity coefficient caused by the change of unit temperature and humidity. The gradient value is calculated through the three-dimensional interpolation grid to quantify the dynamic impact of environmental condition changes on the material flow resistance;

[0131] In the embodiments of the present invention, first, a three-dimensional interpolation grid (temperature and humidity - viscosity coefficient mapping space) is established at the temperature and humidity sampling points of the environmental condition label, and the cubic spline interpolation algorithm is used to calculate the change gradient of the material viscosity coefficient under continuous operating conditions. Specifically, the discrete temperature and humidity sampling points (such as temperature sampling points T1 - Tn, humidity sampling points H1 - Hn) are used as grid nodes, and a continuous surface is generated through the interpolation function to calculate the viscosity coefficient gradient value under any temperature and humidity combination. Subsequently, the time series of the material flow velocity data is convolved with the viscosity coefficient gradient, and the length of the convolution kernel is adaptively adjusted according to the change frequency of the flow velocity. Finally, the flow state correction coefficient is output, and this coefficient will be used as the time-varying weight factor of the subsequent device collaborative analysis matrix.

[0132] Step 412: Decompose the target virtual control parameters into displacement control quantities, velocity control quantities, and acceleration control quantities of device nodes. Based on the kinematic constraint parameters in the device topology relationship network, use the distributed computing architecture to convert the displacement control quantity into a position boundary threshold, the velocity control quantity into a rate change envelope, and the acceleration control quantity into an inertia compensation quantity, so as to generate the kinematic constraint boundaries of device nodes.

[0133] In this step, the displacement control quantity refers to the device position adjustment instruction parsed from the target virtual control parameters, representing the linear distance or angle that the device node needs to move. For example, the end of the robotic arm needs to move 0.5 meters along the X-axis. The velocity control quantity refers to the velocity adjustment instruction extracted from the target virtual control parameters, representing the motion rate that the device node needs to reach. For example, the conveyor belt speed needs to be increased to 1.2 m / s. The acceleration control quantity refers to the acceleration adjustment instruction parsed from the target virtual control parameters, representing the acceleration value that the device node needs to achieve. For example, the motor needs to accelerate from rest to the rated speed within 0.5 seconds. The position boundary threshold refers to the upper and lower limits of the position range that the device node is allowed to reach, which is generated by superimposing the displacement control quantity on the current actual position and passing through a limiting function. For example, if the current position is 1.5 meters and the displacement control quantity is 0.8 meters, the position boundary threshold is limited to [0, 2 meters]. The rate change envelope refers to the allowable range curve of the device node speed change, which is jointly determined by the velocity control quantity and the maximum acceleration. For example, the initial velocity is 0.5 m / s, the target velocity is 1 m / s, and the maximum acceleration is 0.2 m / s², then the rate change envelope is 0.5 + 0.2t (t is time). The inertia compensation quantity refers to the compensation parameter that offsets the influence of the device motion inertia.

[0134] In an embodiment of the present invention, the target virtual control parameter is generated by the dynamic parameter correction loop of the digital twin, and its decomposition process is as follows: Obtain the displacement control amount of the device node by parsing the position instruction component in the target virtual control parameter, and its numerical range is limited by the maximum stroke of the device defined in the device topology relationship network; Extract the speed instruction component from the target virtual control parameter to generate a speed control amount, and its upper limit is determined by the maximum allowable angular velocity of the device node; Parse the acceleration instruction component from the target virtual control parameter to generate an acceleration control, which is limited by the maximum output torque of the device drive system; Use a linear scaling function to add the displacement control amount to the current actual position of the device, and limit the result within the range of [0, Lmax] through the clamp function. The specific formula is: position boundary threshold = current actual position + displacement control amount. If the result is less than 0, take 0; if it is greater than Lmax, take Lmax. For example, if the maximum stroke of the device is 2 meters, the current actual position is 1.5 meters, and the displacement control amount is 0.8 meters, then the position boundary threshold is limited to 2 meters; Based on the maximum allowable acceleration of the device, generate a rate change envelope. The upper limit of the rate change envelope is the sum of the initial velocity and the integral of the acceleration. The integral direction is determined by the sign difference between the target velocity and the actual velocity. The specific formula is: rate change envelope = initial velocity + maximum acceleration × sign function × time integral of (target velocity - actual velocity). For example, if the initial velocity is 0.5 m / s, the target velocity is 1 m / s, and the maximum acceleration is 0.2 m / s², then the velocity upper limit of this envelope within time t is 0.5 + 0.2t until the target velocity is reached; According to the device inertia matrix (dynamic parameters from the device topology relationship network) and the inertial torque (calculated by the product of acceleration and inertia), convert the acceleration control amount into an inertial compensation amount. The specific formula is: inertial compensation amount = inverse matrix of the inertia matrix × (external torque - inertial torque). The device inertia is defined by the dynamic parameters in the device topology relationship network, and the inertial torque is calculated by the product of acceleration and inertia.

[0135] Step 413: Based on the spatial distribution characteristics of the transmission chain clearance, decompose the target error compensation amount into an axial compensation component and a tangential compensation component, and use a distributed computing architecture to convert the axial compensation component into an axial tolerance interval and the tangential compensation component into an angular tolerance interval. Combine the axial tolerance interval and the angular tolerance interval to generate a dynamic tolerance interval for the transmission chain clearance;

[0136] In this step, the spatial distribution characteristic refers to the distribution law of the transmission chain clearance in the physical space, including the axial linear clearance (along the direction of the equipment main shaft) and the tangential angular clearance (around the rotation direction of the main shaft). For example, the nominal axial clearance of the guide rail is 0.1 mm, and the tangential angular tolerance of the gear is ±0.5°. The axial compensation component refers to the component of the target error compensation amount in the direction of the equipment main shaft, which is used to correct the linear displacement error. For example, among the total compensation amount of 0.5 mm, the axial component is 0.3 mm. The tangential compensation component refers to the component of the target error compensation amount in the tangential direction of the transmission chain, which is used to correct the rotation angle error. For example, the tangential component corresponding to the total compensation amount of 0.5 mm is 0.2 mm (converted to the angular compensation amount Δθ = 0.2 mm / radius). The axial tolerance interval refers to the allowable range of axial displacement error, which is calculated and generated from the axial compensation component and the material stiffness coefficient. The angular tolerance interval refers to the allowable range of rotation angle error, which is generated by converting the tangential compensation component.

[0137] In the embodiment of the present invention, the spatial distribution characteristic of the transmission chain clearance includes the axial linear clearance and the tangential angular clearance. According to the design parameter of the axial linear clearance of the transmission chain (for example, the nominal clearance of the guide rail is 0.1 mm), the target error compensation amount is decomposed into the axial compensation component along the direction of the equipment main shaft; according to the design parameter of the tangential angular clearance of the transmission chain (for example, the meshing angle tolerance of the gear is ±0.5°), the target error compensation amount is decomposed into the tangential compensation component along the tangential direction of the transmission chain; through the linear interpolation method, the axial compensation component is converted into the axial tolerance interval. For example, the nominal axial clearance is 0.1 mm, the stiffness coefficient is 0.8, and the axial compensation component is 0.3 mm, then the axial tolerance interval is [0.1 - 0.8×0.3, 0.1 + 0.8×0.3] = [-0.14 mm, 0.34 mm]. In actual application, the negative value is taken as 0; through the polar coordinate transformation, the tangential compensation component is converted into the angular tolerance interval. For example, the current angle is 30°, and the corresponding tangential compensation component is 0.2°, then the angular tolerance interval is [30° - 0.1°, 30° + 0.1°]; the Cartesian product operation is performed on the axial tolerance interval and the angular tolerance interval to generate the dynamic tolerance interval. For example, after the axial interval [0 mm, 0.34 mm] and the angular interval [29.9°, 30.1°] are combined, it represents the allowable compensation range of the transmission chain clearance in the axial and tangential directions.

[0138] Step 414: Embed the flow state correction coefficient as a time-varying weight factor into the matrix weight, and construct an equipment collaborative analysis matrix in combination with the kinematic constraint boundary, the dynamic tolerance interval, and the connection topology;

[0139] In this step, the time-varying weight factor refers to the dynamic adjustment coefficient embedded in the device collaborative analysis matrix, which is used to quantify the influence weights of multi-source features (such as vibration, temperature, pressure) on the optimization strategy under different working conditions. Its value changes with time and is dynamically updated by the flow state correction coefficient through an exponential decay model.

[0140] In the embodiment of the present invention, the flow state correction coefficient is sliced according to the time series and embedded into the weight dimension of the matrix through the orthogonal projection method. The weight update adopts an exponential decay model. For example, if the decay factor is 0.9 and the current flow state correction coefficient is 0.7, then the new weight = 0.9 × old weight + 0.1 × 0.7, realizing the smooth transition of the weight; the kinematic constraint boundary is encoded as the row constraint condition of the matrix; for example, the position boundary threshold is converted into the inequality "device position ≤ 2 meters", the rate change envelope is converted into "speed ≤ 1 m / s", and the inertia compensation amount is converted into the torque limit condition; the dynamic tolerance interval is converted into the column relaxation factor; for example, the relaxation factor corresponding to the axial tolerance interval of 0.34 mm is 1 / 0.34 ≈ 2.94, and the relaxation factor corresponding to the angular tolerance interval of 0.2° is 1 / 0.2 = 5, allowing the optimization algorithm to adjust parameters within the relaxation range; according to the connection topology (for example, device A is directly connected to device B), the associated index is marked in the matrix. If device i is connected to j, then the matrix element M(i,j) = 1, otherwise it is 0, ensuring that the optimization strategy conforms to the physical connection relationship; based on the results of the above steps, the device collaborative analysis matrix is constructed.

[0141] The embodiment of the present invention constructs a device collaborative analysis matrix based on multi-dimensional parameter fusion, breaking through the technical bottleneck of single constraint conditions and insufficient adaptability in traditional control systems; by dynamically integrating kinematic constraints, error tolerances, and working condition weights, it realizes the comprehensive characterization of the device operating state, provides a more accurate decision-making basis for multi-device collaborative control, and solves the problem of the decline in control accuracy of existing methods under dynamic working conditions.

[0142] The present invention provides a specific embodiment. In step 105, the optimization strategy sequence is converted into the trajectory compensation parameters of the end effector in the device, and combined with the target error compensation amount, data closed-loop transmission is performed, which specifically includes the following steps:

[0143] Step 501: Decompose the optimization strategy sequence into the motion control instruction sets of each end effector. Based on the motion control instruction sets, generate the trajectory reference point coordinates, the interpolation path function between adjacent reference points, and the smoothing coefficient of the trajectory transition section to obtain the initial trajectory compensation parameters.

[0144] In this step, the motion control instruction set refers to the specific control instruction set of each end effector parsed from the optimization strategy sequence, including displacement increment, speed adjustment amount, and acceleration constraint amount. The trajectory reference point coordinates refer to the three-dimensional coordinates (X, Y, Z) of the key nodes of the end effector motion path in the Cartesian coordinate system, which are generated by converting the displacement control amount. For example, in the path of a robotic arm from the starting point A(0, 0, 0) to the ending point B(100, 200, 50), the intermediate reference point is C(30, 60, 15). The interpolation path function between adjacent reference points refers to the continuous motion trajectory function connecting two adjacent trajectory reference points, which is generated using cubic spline interpolation or Bezier curve algorithm. The smoothing coefficient of the trajectory transition segment refers to the parameter that controls the smoothness of the trajectory curve, which is calculated from the acceleration constraint amount and is used to limit the trajectory curvature change rate. For example, the smoothing coefficient K = 0.5 indicates that the curvature change rate of adjacent path segments does not exceed 50%. The initial trajectory compensation parameters refer to the set of original trajectory parameters without error compensation, including trajectory reference point coordinates, interpolation path function, and smoothing coefficient. For example, the initial trajectory compensation parameters are {trajectory reference points A / B / C, cubic spline function, K = 0.5}.

[0145] In the embodiment of the present invention, the optimization strategy sequence is grouped according to the end effector identifiers in the device topology relationship network, and the displacement increment parameters (Δx, Δy, Δz), speed adjustment parameters, and acceleration constraint parameters of each actuator are extracted; based on the kinematic model of the end effector (such as the forward kinematic equation of a SCARA robotic arm), the displacement increment parameters are converted into trajectory reference point coordinates in the Cartesian space;

[0146] For example, for a two-degree-of-freedom robotic arm, the end coordinates corresponding to the joint angles θ1 and θ2 are calculated using geometric relationships: x = L1cosθ1 + L2cos(θ1 + θ2), y = L1sinθ1 + L2sin(θ1 + θ2);

[0147] According to the speed adjustment parameters, a fifth-order polynomial interpolation function is generated between adjacent trajectory reference points to ensure the continuity of speed and acceleration; for example, the path function between two points is: q(t) = a0 + a1t + a2t 2 + a3t 3 + a4t 4 + a5t 5 where a0 - a5 are solved through boundary conditions (starting / ending point positions, speeds, accelerations); based on the acceleration constraint parameters, the Gaussian filtering algorithm is used to smooth the acceleration curve of the trajectory transition segment, and the smoothing coefficient is calculated where is the standard deviation, which determines the window width of the Gaussian filtering, is the time average, representing the time point at the center of the filtering window; is the current time, is the pi, According to dynamically adjust, The larger, the smaller (the higher the filtering intensity).

[0148] Step 502: Based on the spatial distribution characteristics of the transmission chain clearance and the initial trajectory compensation parameters, using a distributed computing architecture, decompose the target error compensation amount into an axial position compensation amount and an angular offset compensation amount, and combine the axial position compensation amount, the angular offset compensation amount, the trajectory reference point coordinates, the interpolation path function, and the smoothing coefficient to generate the fused trajectory compensation parameters;

[0149] In this step, the axial position compensation amount refers to the component of the target error compensation amount along the main axis direction of the device, which is used to correct the linear position deviation of the trajectory reference point. For example, if the axial clearance of the transmission chain causes the reference point B to shift by 0.3 mm, the axial position compensation amount is +0.3 mm. The angular offset compensation amount refers to the component of the target error compensation amount along the tangent direction of the transmission chain, which is used to correct the trajectory angle deviation. For example, if the gear clearance causes the end effector to deflect by 0.2°, the angular offset compensation amount is -0.2°. The fused trajectory compensation parameters refer to the corrected parameters after fusing the initial trajectory parameters and the error compensation amount, including the compensated reference point coordinates (original coordinates + axial compensation amount), the adjusted interpolation path function (superimposed angular compensation amount), and the optimized smoothing coefficient.

[0150] In the embodiment of the present invention, according to the axial linear clearance design parameters of the transmission chain (such as the nominal clearance of the guide rail), decompose the target error compensation amount along the main axis direction of the device to obtain the axial position compensation amount; according to the tangential angular clearance (such as the gear meshing tolerance ±0.5°), decompose the target error compensation amount into an angular offset amount; in the distributed computing architecture, superimpose the axial position compensation amount on the trajectory reference point coordinates through linear interpolation to obtain the compensated trajectory reference point coordinates; for example, the x of the original reference point coordinates (x, y, z) is corrected to: x′ = x + axial position compensation amount × cos where is the angle between the main axis direction of the device and the coordinate axis; fuse the angular offset compensation amount into the tangent direction of the interpolation path function through the homogeneous coordinate transformation matrix to obtain the corrected interpolation path function; perform tensor splicing on the compensated trajectory reference point coordinates, the corrected interpolation path function, and the smoothing coefficient to generate the fused trajectory compensation parameters.

[0151] Step 503: Verify the fused trajectory compensation parameters to obtain the verified trajectory compensation parameters. Use the timestamp alignment mechanism to process the verified trajectory compensation parameters to generate a transmission data packet. Distribute the transmission data packet to each end effector controller through the downlink of the 5G edge computing node, and synchronously collect the residuals between the actual motion trajectory data of each end effector and the predicted trajectory of the digital twin;

[0152] In this step, the verified trajectory compensation parameters refer to the final trajectory parameters that have passed the data integrity verification.

[0153] In the embodiment of the present invention, data integrity verification is performed on the fused trajectory compensation parameters, and the verification and transmission process are as follows: Verify whether the coordinates of the trajectory reference point are within the device workspace. For example, through the bounding box collision detection algorithm, check whether the coordinates (x, y, z) satisfy:

[0154] , , ; Check whether the first derivative (velocity) and second derivative (acceleration) of the interpolation path function are continuous. Use the numerical differentiation method to calculate the derivative difference between adjacent intervals, and set the threshold to ; Compare the matching degree between the smoothing coefficient and the dynamic response frequency band of the transmission chain. If the smoothing coefficient exceeds the preset range (such as 0.8 - 1.2), trigger re - smoothing processing; Package the trajectory compensation parameters that pass the verification into a transmission data packet according to the timestamp alignment mechanism of the 5G edge computing node. Each data packet contains a sequence of trajectory reference point coordinates, interpolation path function coefficients, smoothing coefficients, and timestamp synchronization marks (accuracy ±1μs); Distribute the transmission data packet to the end effector controller through the downlink of the 5G edge computing node. At the same time, use the MEMS sensor to collect the actual motion trajectory data of the actuator in real time, and calculate the residual with the predicted trajectory of the digital twin. The calculation formula is, , where is the actual motion trajectory, is the predicted trajectory of the digital twin.

[0155] Step 504: When the residual exceeds the dynamic tolerance interval of the transmission chain clearance, trigger the dynamic parameter correction loop to recalculate the compensation amount to achieve closed - loop data transmission;

[0156] In the embodiment of the present invention, the closed-loop trigger mechanism is as follows: The residual is compared with the dynamic tolerance interval in real time. For example, if the dynamic tolerance interval is [0mm, 0.5mm], an alarm is triggered when the residual > 0.5mm; the dynamic parameter correction loop of the digital twin is triggered, and the following process is re-executed: Based on the current device state fingerprint features and environmental condition labels, the multi-source fusion feature vector is updated; the sliding window algorithm is used to regenerate the target error compensation amount; the virtual control parameters and trajectory compensation parameters are corrected; the recomputed compensation parameters are transmitted back to the edge computing node through the uplink of the 5G edge computing node to update the optimization strategy sequence, forming a closed-loop control flow.

[0157] In the embodiment of the present invention, by establishing a closed-loop trajectory compensation mechanism, the problems of insufficient trajectory tracking accuracy and error accumulation in the traditional control system are effectively improved. By combining the optimization strategy with real-time error compensation, the dynamic correction of the motion trajectory of the end effector is realized, the compensation ability of the system for mechanical errors such as transmission chain clearance is significantly improved, and the defect of accuracy decay of the existing open-loop control scheme during long-term operation is overcome.

[0158] The present invention provides a specific embodiment. Step 101: Construct a timestamp alignment mechanism in the 5G edge computing node to perform time-domain matching of vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line to generate a preprocessed data packet, which specifically includes the following steps:

[0159] Step 111: Construct a timestamp alignment mechanism in the 5G edge computing node. Based on the pulse signal of the material flow rate data, trigger the timestamp alignment mechanism to perform time-domain matching processing on the vibration spectrum data, the pressure gradient change data, and the temperature distribution data to obtain the vibration spectrum data, pressure gradient change data, and temperature distribution data after time-domain matching;

[0160] In the embodiment of the present invention, the timestamp alignment mechanism is implemented according to the following process: using the material flow rate pulse signal as the global time reference to generate a synchronous trigger signal; when the pulse rising edge arrives, recording the current moment as the reference timestamp T0; setting a dynamic time window for the vibration spectrum data, and the window length is adaptively adjusted according to the pulse interval (window length = 1 / pulse frequency); resampling the vibration signal within the window, and the sampling frequency is the same as the pulse frequency to ensure that the time series of the vibration data is strictly aligned with the material flow rate pulse; using the piecewise cubic Hermite interpolation algorithm for the pressure gradient change data, generating interpolation nodes according to the timestamps of the pulse signal to make the sampling moments of the pressure data match the pulse timestamps; for example, if the pulse arrives at times t1 and t2, then 10 equally spaced data points are interpolated within the interval [t1, t2] for the pressure data; establishing a heat conduction delay model to calculate the physical delay time of the temperature sensor measurement value relative to the material flow (physical delay time = distance from the sensor to the heat source / heat conduction speed), and correcting the timestamps of the original temperature data by advancing the physical delay time to eliminate the influence of heat conduction delay; after the above process, the vibration spectrum data (strictly synchronized with the pulse), the pressure gradient change data (aligned with the pulse after interpolation), and the temperature distribution data (after delay compensation) are output in the time domain matching.

[0161] Step 112: According to the spatial coordinates of the device topology relationship network, decompose the vibration spectrum data after time domain matching into axial vibration components in the direction of the device main axis, convert the pressure gradient change data after time domain matching into a normal pressure distribution matrix on the device contact surface, and map the temperature distribution data after time domain matching into an equivalent heat source intensity parameter;

[0162] In this step, the axial vibration component refers to the vibration energy distribution parameter in the direction of the device main axis (defined by the device topology relationship network), and is obtained by performing spatial projection decomposition on the vibration spectrum data after time domain matching. The normal pressure distribution matrix refers to the quantization matrix of the pressure gradient distribution in the normal direction of the device contact surface (defined by the contact surface geometric parameters), and is generated by projecting the pressure gradient change data after time domain matching onto the normal direction of the contact surface. The equivalent heat source intensity parameter refers to the spatial distribution parameter characterizing the heat source intensity in the device thermodynamic model, and is generated by calculating the temperature distribution data after time domain matching and the material thermal conductivity (from the process constraint conditions).

[0163] In the embodiment of the present invention, the device topology relationship network defines the three-dimensional spatial coordinates and the mechanical spindle direction of each device node. The data decomposition process is as follows: Extract the unit vector of the device spindle direction (defined by the topology relationship network), project the vibration spectrum data after time domain matching onto the spindle direction, and calculate the preliminary axial vibration component (axial vibration component = projection amplitude of the vibration signal in the spindle direction); Extract the spectral characteristics of the preliminary axial vibration component through Fourier transform, and retain the frequency band related to the natural frequency of the device (such as 0 - 500 Hz) to finally obtain the axial vibration component; According to the geometric parameters of the device contact surface (defined by the topology relationship network), establish a local coordinate system of the contact surface; Project the pressure gradient change data after time domain matching onto the local coordinate system of the contact surface along the normal direction of the contact surface to generate a normal pressure distribution matrix, where the matrix element P(i,j) represents the normal pressure value of the grid point at the i-th row and j-th column of the contact surface; Based on the thermodynamic simulation model, convert the temperature distribution data after time domain matching into an equivalent heat source intensity parameter Q, and its calculation formula is: Q = k·ΔT / Δx, where k is the material thermal conductivity (from process constraint conditions), ΔT is the temperature gradient, and Δx is the heat source spacing.

[0164] Step 113: Perform a convolution operation on the time series of the axial vibration component and the time series of the material flow rate data to generate vibration and flow rate correlation features, perform a tensor product operation on the normal pressure distribution matrix and the equivalent heat source intensity parameter to generate pressure and temperature coupling features, and combine the vibration and flow rate correlation features to generate a preprocessing data packet;

[0165] In this step, the vibration and flow rate correlation features refer to the dynamic correlation features in the time domain between the axial vibration component and the material flow rate data, which are generated through convolution operation. The pressure and temperature coupling features refer to the interaction features in the spatial dimension between the normal pressure distribution matrix and the equivalent heat source intensity parameter, which are generated through tensor product operation.

[0166] In the embodiment of the present invention, perform a convolution operation on the time series of the axial vibration component and the time series of the material flow rate data. The length of the convolution kernel is adaptively set according to the flow rate change rate (when the flow rate change rate > 10%, the kernel length is 5, otherwise it is 10); Perform normalization processing on the convolution result to eliminate the dimensional difference and generate vibration and flow rate correlation features; Perform a tensor product operation on the normal pressure distribution matrix and the equivalent heat source intensity parameter to generate a three-dimensional coupling feature tensor, perform dimensionality reduction processing on the three-dimensional coupling feature tensor, and take the maximum value along the depth direction (k-axis) to obtain a two-dimensional pressure and temperature coupling feature; Concatenate the time series of the vibration-flow rate correlation features and the spatial distribution of the pressure and temperature coupling features in the time and space dimensions, create a structured data container in the 5G edge computing node for data encapsulation. The encapsulated data packet includes a timestamp, feature dimension information, and a check code.

[0167] In the embodiments of the present invention, through multi-source data time-domain matching, the problem of insufficient data synchronization accuracy commonly found in the industrial Internet of Things is solved; through an innovative timestamp alignment mechanism and feature fusion technology, high-precision matching of heterogeneous data such as vibration, pressure, and temperature is achieved, providing reliable preprocessed data for subsequent analysis and making up for the deficiencies of traditional methods in data mismatch in high-speed production scenarios.

[0168] Figure 2 The following is a schematic structural diagram of a real-time transmission system for Internet of Things perception data in a digital twin scenario provided by an embodiment of the present invention, as Figure 2 shown. The system includes:

[0169] An acquisition module 21, configured to acquire vibration spectrum data, pressure gradient change data, and temperature distribution data of equipment in an industrial production line, and build a timestamp alignment mechanism in a 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet;

[0170] An analysis module 22, configured to analyze the preprocessing data packet to obtain device status fingerprint features and environmental working condition labels;

[0171] A generation module 23, configured to generate a target virtual control parameter of the device and a target error compensation amount of the clearance of the transmission chain in the device according to the device status fingerprint features and environmental working condition labels, using the dynamic parameter correction loop of the digital twin body, where the digital twin body is constructed based on the device topology relationship network and process constraint conditions corresponding to the device;

[0172] A determination module 24, configured to generate an actual material flow state based on the device status fingerprint features, environmental working condition labels, and material flow rate data, and perform coupled analysis on the target virtual control parameter, target error compensation amount, and actual material flow state using a distributed computing architecture to generate an optimization strategy sequence for multi-device collaborative operation;

[0173] A conversion module 25, configured to convert the optimization strategy sequence into trajectory compensation parameters of the end effector in the device, and perform data closed-loop transmission in combination with the target error compensation amount.

[0174] Figure 2 The described real-time transmission system for Internet of Things perception data in a digital twin scenario can execute Figure 1 The real-time transmission method for Internet of Things perception data in a digital twin scenario described in the embodiments shown. Its implementation principle and technical effects will not be elaborated. For the real-time transmission system for Internet of Things perception data in a digital twin scenario in the above embodiments, the specific ways in which each module and unit perform operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0175] In a possible design, Figure 2 A real-time transmission system based on the Internet of Things perception data in a digital twin scenario in the illustrated embodiment can be implemented as a computing device, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0176] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32.

[0177] The processing component 32 is configured to: collect vibration spectrum data, pressure gradient change data, and temperature distribution data of devices in an industrial production line, construct a timestamp alignment mechanism in a 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet; analyze the preprocessing data packet to obtain device state fingerprint features and environmental working condition labels; according to the device state fingerprint features and environmental working condition labels, use the dynamic parameter correction loop of the digital twin to generate the target virtual control parameters of the device and the target error compensation amount of the clearance of the transmission chain in the device, wherein the digital twin is constructed based on the device topology relationship network and process constraint conditions corresponding to the device; generate an actual material flow state based on the device state fingerprint features, environmental working condition labels, and material flow rate data, use a distributed computing architecture to perform coupled analysis on the target virtual control parameters, target error compensation amount, and actual material flow state, and generate an optimization strategy sequence for multi-device collaborative operation; convert the optimization strategy sequence into trajectory compensation parameters of the end effector in the device, and combine the target error compensation amount to perform data closed-loop transmission.

[0178] Among them, the processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for executing the above method.

[0179] The storage component 31 is configured to store various types of data to support the operation of the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0180] Of course, the computing device may necessarily further include other components, such as an input / output interface, a display component, a communication component, etc.

[0181] The input / output interface provides an interface between the processing component and the peripheral interface module, and the above-mentioned peripheral interface module may be an output device, an input device, etc.

[0182] The communication component is configured to facilitate communication between the computing device and other devices in a wired or wireless manner, etc.

[0183] Among them, the computing device may be a physical device or an elastic computing host provided by a cloud computing platform, etc. At this time, the computing device may refer to a cloud server, and the above-mentioned processing component, storage component, etc. may be basic server resources leased or purchased from a cloud computing platform.

[0184] The embodiment of the present invention also provides a computer storage medium storing a computer program, and when the computer program is executed by a computer, it can implement the above-mentioned Figure 1 real-time transmission method of Internet of Things perception data based on digital twin scenarios shown in the embodiments.

[0185] Those skilled in the art can clearly understand that for the convenience and conciseness of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0186] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative labor.

[0187] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the essence of the above technical solution, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0188] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A real-time transmission method based on Internet of Things perception data in a digital twin scenario, characterized in that, Including: Collect the vibration spectrum data, pressure gradient change data, and temperature distribution data of the equipment in the industrial production line, and construct a timestamp alignment mechanism in the 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet; Analyze the preprocessing data packet to obtain the equipment status fingerprint features and environmental condition labels; According to the equipment status fingerprint features and environmental condition labels, use the dynamic parameter correction loop of the digital twin to generate the target virtual control parameters of the equipment and the target error compensation amount of the transmission chain clearance in the equipment, where the digital twin is constructed based on the equipment topology relationship network and process constraint conditions corresponding to the equipment; Based on the equipment status fingerprint features, environmental condition labels, and material flow rate data, generate the actual material flow state, and use the distributed computing architecture to perform coupled analysis on the target virtual control parameters, target error compensation amount, and actual material flow state to generate an optimization strategy sequence for multi-device collaborative operation; Convert the optimization strategy sequence into the trajectory compensation parameters of the end effector in the equipment, and combine the target error compensation amount to perform data closed-loop transmission.

2. The method according to claim 1, characterized in that, According to the equipment status fingerprint features and environmental condition labels, use the dynamic parameter correction loop of the digital twin to generate the target virtual control parameters of the equipment and the target error compensation amount of the transmission chain clearance in the equipment, where the digital twin is constructed based on the equipment topology relationship network and process constraint conditions corresponding to the equipment, including: Based on the equipment topology relationship network and process constraint conditions, construct a digital twin, and the equipment topology relationship network uses the graph theory modeling method to establish the mechanical connection attributes and signal transmission paths between equipment nodes; Based on the dynamic parameter correction loop in the digital twin, decompose the equipment status fingerprint features into axial vibration mode features, radial thermal expansion coefficient features, and tangential stress distribution features, and combine the maximum allowable deformation threshold in the process constraint conditions and the temperature sensitivity coefficient of the material property parameters in the digital twin to generate a multi-source fusion feature vector; Based on the multi-source fusion feature vector, generate an initial error compensation amount, and use the sliding window algorithm to dynamically adjust the initial error compensation amount to obtain the target error compensation amount, where the window length is adaptively adjusted according to the change rate of the material flow rate data; Based on the working parameter feasible region of each equipment node in the digital twin and the environmental condition label, generate initial virtual control parameters; According to the coupling relationship of the mechanical connection attributes of the equipment and the target error compensation amount, perform dynamic collaborative correction on the initial virtual control parameters to obtain the target virtual control parameters.

3. The method according to claim 2, wherein Combining the maximum allowable deformation threshold in the process constraint conditions and the temperature sensitivity coefficient of the material property parameters in the digital twin to generate a multi-source fusion feature vector, including: Based on the axial vibration mode characteristics, the vibration energy value is obtained through the spectral energy integration of the vibration signal in the direction of the main shaft of the device. Based on the radial thermal expansion coefficient characteristics, the deformation gradient value is calculated through the product relationship between the temperature distribution data and the material thermal expansion coefficient. Based on the tangential stress distribution characteristics, the vector product result is calculated through the pressure gradient data and the contact surface curvature; According to the ratio of the vibration energy value to the maximum allowable deformation threshold, the first deformation contribution rate is calculated; according to the ratio of the deformation gradient value to the maximum allowable deformation threshold, the second deformation contribution rate is calculated. According to the deviation degree of the ambient temperature from the material standard working condition temperature, the second deformation contribution rate is dynamically compensated to generate a corrected second deformation contribution rate; According to the vector product result, the projection components in the tangent direction of the device motion trajectory are weighted and accumulated to calculate the third deformation contribution rate; According to the corrected second deformation contribution rate and the temperature sensitivity coefficient of the material property parameters in the digital twin, the second reference weight is calculated; According to the first deformation contribution rate, the third deformation contribution rate, the vibration energy attenuation coefficient, and the friction coefficient correction factor, the first reference weight and the third reference weight are calculated. Combining the second reference weight, the first quantization value, the second quantization value, and the third quantization value, a multi-source fusion feature vector is generated.

4. The method according to claim 1, wherein Based on the device state fingerprint characteristics, the environmental condition label, and the material flow rate data, the actual material flow state is generated. Using the distributed computing architecture, the target virtual control parameter, the target error compensation amount, and the actual material flow state are coupled and analyzed to generate an optimization strategy sequence for multi-device collaborative operation, including: Based on the vibration energy distribution parameters in the device state fingerprint characteristics and the temperature and humidity parameters in the environmental condition label, the material flow rate data is processed to generate a discretized flow rate interval and the corresponding material accumulation density distribution parameters. Combining the vibration energy distribution parameters, the material transfer efficiency attenuation factor between adjacent devices is calculated; According to the non-linear relationship between the temperature and humidity parameters and the material viscosity coefficient, a flow state correction coefficient is generated. Using the distributed computing architecture, the target virtual control parameter is mapped to the kinematic constraint boundary of the device node, the target error compensation amount is converted into the dynamic tolerance interval of the transmission chain clearance, and the flow state correction coefficient is embedded as a time-varying weight factor into the matrix dimension to construct a device collaborative analysis matrix; Based on the connection topology in the device topology relationship network and the device collaborative analysis matrix, the energy transfer function and the material conservation equation between device nodes are constructed; By iteratively solving the joint solution set of the energy transfer function and the material conservation equation, an initial optimization strategy is generated. Combining the material transfer efficiency attenuation factor and the device safe operation threshold in the process constraint conditions, the initial optimization strategy is optimized to obtain a preliminarily optimized optimization strategy; Based on the distributed computing architecture and the device topology relationship network, the preliminarily optimized optimization strategy is distributedly verified to generate an optimization strategy sequence for multi-device collaborative operation.

5. The method according to claim 4, characterized in that, According to the non - linear relationship between temperature - humidity parameters and material viscosity coefficient, a flow - state correction coefficient is generated. Using a distributed computing architecture, the target virtual control parameters are mapped to the kinematic constraint boundaries of device nodes, the target error compensation amount is converted into the dynamic tolerance interval of the transmission chain clearance, and the flow - state correction coefficient is embedded as a time - varying weight factor into the matrix dimension to construct a device collaborative analysis matrix, including: Based on the non - linear relationship between temperature - humidity parameters and material viscosity coefficient, a three - dimensional interpolation grid is established at the temperature - humidity sampling points of the environmental condition label to calculate the change gradient of the material viscosity coefficient under continuous working conditions. Combining with the material flow velocity data, a flow - state correction coefficient is generated; The target virtual control parameters are decomposed into the displacement control amount, velocity control amount, and acceleration control amount of device nodes. Based on the kinematic constraint parameters in the device topology relationship network, using a distributed computing architecture, the displacement control amount is converted into a position boundary threshold, the velocity control amount is converted into a rate change envelope, and the acceleration control amount is converted into an inertia compensation amount to generate the kinematic constraint boundaries of device nodes; Based on the spatial distribution characteristics of the transmission chain clearance, the target error compensation amount is decomposed into an axial compensation component and a tangential compensation component. Using a distributed computing architecture, the axial compensation component is converted into an axial tolerance interval, the tangential compensation component is converted into an angular tolerance interval, and combining the axial tolerance interval and the angular tolerance interval, a dynamic tolerance interval of the transmission chain clearance is generated; The flow - state correction coefficient is embedded as a time - varying weight factor into the matrix weight. Combining the kinematic constraint boundaries, the dynamic tolerance interval, and the connection topology, a device collaborative analysis matrix is constructed.

6. The method according to claim 1, characterized in that, The optimization strategy sequence is converted into the trajectory compensation parameters of the end - effector in the device. Combining with the target error compensation amount, data closed - loop transmission is carried out, including: The optimization strategy sequence is decomposed into the motion control instruction sets of each end - effector. Based on the motion control instruction sets, the coordinates of the trajectory reference points, the interpolation path function between adjacent reference points, and the smoothing coefficient of the trajectory transition section are generated to obtain the initial trajectory compensation parameters; Based on the spatial distribution characteristics of the transmission chain clearance and the initial trajectory compensation parameters, using a distributed computing architecture, the target error compensation amount is decomposed into an axial position compensation amount and an angular offset compensation amount. Combining the axial position compensation amount, the angular offset compensation amount, the trajectory reference point coordinates, the interpolation path function, and the smoothing coefficient, the fused trajectory compensation parameters are generated; The fused trajectory compensation parameters are verified to obtain the verified trajectory compensation parameters. Using the timestamp alignment mechanism to process the verified trajectory compensation parameters, transmission data packets are generated. The transmission data packets are distributed to each end - effector controller through the downlink of the 5G edge computing node, and the residuals between the actual motion trajectory data of each end - effector and the predicted trajectory of the digital twin are synchronously collected; When the residual exceeds the dynamic tolerance interval of the transmission chain clearance, the dynamic parameter correction loop is triggered to recalculate the compensation amount to achieve data closed - loop transmission.

7. The method according to claim 1, characterized in that, Build a timestamp alignment mechanism in the 5G edge computing node to perform time-domain matching of vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet, including: Build a timestamp alignment mechanism in the 5G edge computing node. Based on the pulse signal of the material flow rate data, trigger the timestamp alignment mechanism to perform time-domain matching processing on the vibration spectrum data, the pressure gradient change data, and the temperature distribution data, and obtain the vibration spectrum data, pressure gradient change data, and temperature distribution data after time-domain matching; According to the spatial coordinates of the device topology relationship network, decompose the vibration spectrum data after time-domain matching into axial vibration components in the main axis direction of the device, convert the pressure gradient change data after time-domain matching into a normal pressure distribution matrix of the device contact surface, and map the temperature distribution data after time-domain matching into an equivalent heat source intensity parameter; Perform a convolution operation on the axial vibration component and the time series of the material flow rate data to generate vibration and flow rate correlation features, perform a tensor product operation on the normal pressure distribution matrix and the equivalent heat source intensity parameter to generate pressure and temperature coupling features, and combine the vibration and flow rate correlation features to generate a preprocessing data packet.

8. A real-time transmission system based on Internet of Things perception data in a digital twin scenario, characterized in that, Including: An acquisition module for acquiring vibration spectrum data, pressure gradient change data, and temperature distribution data of devices in the industrial production line, and building a timestamp alignment mechanism in the 5G edge computing node to perform time-domain matching of the vibration spectrum data, pressure gradient change data, and temperature distribution data with the material flow rate data of the industrial production line, and generate a preprocessing data packet; An analysis module for analyzing the preprocessing data packet to obtain device state fingerprint features and environmental condition labels; A generation module for generating target virtual control parameters of the device and target error compensation amounts for the clearance of the transmission chain in the device according to the device state fingerprint features and environmental condition labels, using the dynamic parameter correction loop of the digital twin, where the digital twin is constructed based on the device topology relationship network corresponding to the device and process constraint conditions; A determination module for generating an actual material flow state based on the device state fingerprint features, environmental condition labels, and material flow rate data, and performing a coupling analysis on the target virtual control parameters, target error compensation amounts, and actual material flow state using a distributed computing architecture to generate an optimization strategy sequence for multi-device collaborative operation; A conversion module for converting the optimization strategy sequence into trajectory compensation parameters of the end effector in the device, and combining the target error compensation amounts to perform data closed-loop transmission.

9. A computing device, characterized in that, Including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a real-time transmission method for Internet of Things perception data in a digital twin scenario as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that, Stores a computer program, and when the computer program is executed by a computer, it implements a real-time transmission method for Internet of Things perception data in a digital twin scenario as described in any one of claims 1 to 7.

Images