A self-adaptive cooperative control system for process parameters of a manufacturing line multi-station process

Abstract

The application relates to the technical field of intelligent manufacturing equipment, and discloses a self-adaptive cooperative control system for manufacturing line multi-station process parameters, which comprises a first control module, a displacement sensing module, a cooperative operation module and a second control module. The first control module acquires material initial process deviation parameters and splits the parameters into dynamic evolution components and static fixed components. The cooperative operation module calculates residual deviation characteristics of the material arriving at a subsequent station by using characteristic attenuation coefficients according to displacement increment pulses provided by the displacement sensing module. The second control module generates compensation instructions for actuators according to reconstructed actual input vectors. The application eliminates state dissipation deviation in the material flow conversion process by constructing a cooperative topology with synchronous evolution of data flow and material flow, ensures that the adjusting action is consistent with the instantaneous physical state of the material, and enhances the suppression capability of the production line to process fluctuations.

Description

Technical Field

[0001] This invention belongs to the field of intelligent manufacturing equipment technology, and in particular relates to an adaptive collaborative control system for multi-station process parameters in a manufacturing production line. Background Technology

[0002] The current intelligent manufacturing equipment industry involves high-precision continuous manufacturing production lines, where materials flow continuously between multiple serial execution nodes; distributed industrial control networks are used to adjust and synchronize process parameters at multiple workstations; traditional control schemes are mostly based on independent closed-loop regulation at a single workstation, that is, each control unit only outputs drive signals for the process reference of its own workstation, and completes the entire line processing by sequential triggering from upstream workstation to downstream workstation.

[0003] In actual production conditions, due to equipment wear, environmental fluctuations, and differences in material properties, the actual output parameters of upstream stations experience slight drift. When materials with this drift flow into downstream stations, if the downstream control unit still outputs execution signals according to the preset parameters of the ideal reference, the upstream deviation intrudes into the downstream as an unplanned disturbance, leading to a nonlinear decay in processing accuracy. Transient parameters generated by upstream processing, such as processing thermal stress and mechanical residual vibration, will spontaneously generate energy dissipation and state relaxation during the physical flow of materials to the next station with the conveyor mechanism. In addition to mechanical limitations, existing control logic also has bottlenecks in multi-station dynamic deviation compensation. For example, Chinese invention patent application CN121277112A discloses a multi-station collaborative control method and system based on event triggering and look-ahead synchronization. It captures physical events, calculates and compensates for accumulated time deviations, and uses optimization algorithms to adjust motion curves to align each station on the time scale.

[0004] Therefore, the technical problem to be solved by this invention is how to establish a data mapping mechanism that simulates the physical process of material state evolution between workstations, reconstruct the compensation basis according to the energy dissipation logic within the transmission cycle, and thereby improve the accuracy of multi-workstation collaborative control by utilizing the inherent driving data of the system without adding external detection hardware. Summary of the Invention

[0005] This invention provides an adaptive collaborative control system for multi-station process parameters in a manufacturing production line, comprising: The first control module is configured to acquire the initial process deviation parameter generated by the material at the first work station, and according to the component mapping dimension in the preset state decay model, decompose the initial process deviation parameter into a dynamic evolution component excited by the material temperature field and a static fixed component defined by the work station clamping reference. The displacement sensing module is configured to output displacement increment pulses that change with the displacement transmitted along the production line. The collaborative computing module is connected to the first control module and the displacement sensing module respectively, and is configured to acquire dynamic evolution components, use displacement increment pulses to perform weight mapping on the dynamic evolution components through characteristic attenuation coefficients, and calculate the residual deviation characteristics of the material at the moment of physical arrival at the second work station node. The second control module is connected to the collaborative computing module and is configured to reconstruct the residual deviation features and static fixed components to recover the actual input vector, and generate compensation instructions for the second station actuator based on the actual input vector to offset the spatiotemporal physical lag deviation of the initial process deviation parameter caused by the material displacement.

[0006] Preferably, when performing weighted mapping, the collaborative computing module is configured to use the pulse count output by the displacement sensing module as the independent variable to perform integral calculation of the material flow time, and input the integral result into the state decay model to determine the weighting coefficient of the dynamic evolution component; the characteristic decay coefficient decreases monotonically with the increase of the pulse count, so that the residual deviation characteristic exhibits a nonlinear convergence trend with the increase of the material physical displacement, in order to offset the physical state dissipation of the material during the transmission.

[0007] Preferably, the first control module constructs the state decay model in the following manner: extracting the temperature evolution dimension that spontaneously evolves with the transmission time from the initial process deviation parameter, and defining it as a dynamic evolution component with a constant initial value; extracting the geometric position deviation from the initial process deviation parameter that is not affected by the transmission environment, and defining it as a static fixed component whose value remains constant during the transmission process and is independent of the weight mapping operation.

[0008] Preferably, it also includes: an end-feedback module, used to collect the quality feature vector of the finished product at the end of the production line, map the quality feature vector to a preset state transition matrix, and use the residual value of the quality feature vector to adaptively calibrate the feature attenuation coefficient to offset the physical coupling characteristic shift caused by wear of production equipment or environmental temperature drift.

[0009] Preferably, the second control module is equipped with a compensation threshold determination logic; the compensation threshold determination logic is used to determine whether the actual input vector exceeds the physical compensation limit of the second station actuator; when the determination result is yes, the second control module cuts off the feedforward loop and locks it to the preset reference process parameters to prevent abnormal compensation commands from causing mechanical overload.

[0010] Preferably, the collaborative computing module is also connected to a data buffer module; the data buffer module is used to store the static fixed components with time stamp information collected by the first control module, and according to the cumulative pulse number output by the displacement sensing module, to achieve phase alignment of the static fixed components and residual deviation characteristics at the input end of the second control module for data reconstruction.

[0011] Preferably, the initial process deviation parameters include material surface temperature distribution data, surface coating thickness deviation data, and weld point residual stress data; the state decay model is configured with heterogeneous characteristic decay coefficients for different initial process deviation parameters to characterize the state relaxation characteristics of different physical parameters during material flow.

[0012] Preferably, the first control module and the second control module construct a distributed collaborative topology through an industrial network; when the first control module sends data to the collaborative computing module, it encapsulates the initial count value of the displacement increment pulse to provide a globally consistent displacement reference for cross-node deviation reconstruction.

[0013] Preferably, the resolution of the displacement sensing module is 1024 p / r; the collaborative calculation module dynamically updates the step compensation value of the characteristic attenuation coefficient by counting the displacement increment pulses in real time, so that the calculation accuracy of the residual deviation characteristics is adjusted synchronously with the change of the instantaneous movement speed of the conveying mechanism.

[0014] Compared with existing technologies, the adaptive collaborative control system for multi-station process parameters in the manufacturing production line of this invention has the following advantages: 1. In adaptive cooperative control, by decomposing the process state vector into rigid holding components and transient dissipation components, the physical evolution of the material flow between workstations is matched. Since the transient stress or aftershock generated during processing decays spontaneously over time, by introducing a time-domain decay operator driven by the conveyor belt encoder pulse, the compensation basis obtained by the downstream workstation is transformed from static historical values ​​to dynamic residual characteristics. This mechanism eliminates the feedforward overcompensation phenomenon caused by the change of state during material transmission, and ensures that the adjustment action of the downstream execution unit is consistent with the physical state of the material at the moment of arrival.

[0015] 2. The system combines the timestamp collected by the first control node with the physical integral time obtained by the second control node to construct a collaborative topology in which the data flow evolves synchronously with the material flow. This deep coupling of time-series alignment logic and state transition matrix operation enables the system to adjust the buffer period and attenuation intensity in real time according to the change of transmission speed. The logical interaction of multiple nodes integrates the originally isolated workstation closed loop into a globally collaborative control chain, improving the production line's ability to suppress nonlinear cascading fluctuations of process parameters.

[0016] 3. By using the quality feature vector of the finished product at the end of the production line to iteratively update the state transition matrix, a slow calibration mechanism covering the entire process logic is established. This mechanism, combined with the basic control logic, offsets the physical coupling characteristic shift caused by equipment wear or environmental temperature drift. By converting the end quality residual into correction instructions for the weights of the intermediate node transition matrix, the adaptive accuracy maintenance of the control system is achieved throughout the entire life cycle of the equipment, reducing the dependence on high-frequency manual calibration. Attached Figure Description

[0017] Figure 1 This is a flowchart of the closed-loop process for multi-station data processing and collaborative control in the manufacturing production line of this invention; Figure 2 This is a schematic diagram of the adaptive evaluation and optimization logic for multi-station process parameters in this invention. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0019] It should be noted that all directional and positional terms used in this invention, such as: up, down, left, right, front, back, vertical, horizontal, inner, outer, top, bottom, transverse, longitudinal, center, etc., are only used to explain the relative positional relationship and connection between components in a specific state (as shown in the accompanying drawings). They are only for the convenience of describing this invention and do not require that this invention be constructed and operated in a specific orientation. Therefore, they should not be construed as limiting this invention. In addition, the descriptions of "first," "second," etc., in this invention are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated.

[0020] In the description of this invention, unless otherwise explicitly specified and limited, the terms installation, connection, and linking should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; they can refer to mechanical connections; they can refer to direct connections or indirect connections through an intermediate medium; they can refer to the internal connection of two components. For those skilled in the art, the specific meaning of the above terms in this invention can be understood according to the specific circumstances.

[0021] In the description of this specification, references to the terms "an embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example, and the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0022] An adaptive collaborative control system for multi-station process parameters in a manufacturing production line includes: The first control module is configured to acquire the initial process deviation parameter generated by the material at the first work station, and according to the component mapping dimension in the preset state decay model, decompose the initial process deviation parameter into a dynamic evolution component excited by the material temperature field and a static fixed component defined by the work station clamping reference. The displacement sensing module is configured to output displacement increment pulses that change with the displacement transmitted along the production line. The collaborative computing module is connected to the first control module and the displacement sensing module respectively, and is configured to acquire dynamic evolution components, use displacement increment pulses to perform weight mapping on the dynamic evolution components through characteristic attenuation coefficients, and calculate the residual deviation characteristics of the material at the moment of physical arrival at the second work station node. The second control module is connected to the collaborative computing module and is configured to reconstruct the residual deviation features and static fixed components to recover the actual input vector, and generate compensation instructions for the second station actuator based on the actual input vector to offset the spatiotemporal physical lag deviation of the initial process deviation parameter caused by the material displacement.

[0023] Preferably, when performing weighted mapping, the collaborative computing module is configured to use the pulse count output by the displacement sensing module as the independent variable to perform integral calculation of the material flow time, and input the integral result into the state decay model to determine the weighting coefficient of the dynamic evolution component; the characteristic decay coefficient decreases monotonically with the increase of the pulse count, so that the residual deviation characteristic exhibits a nonlinear convergence trend with the increase of the material physical displacement, in order to offset the physical state dissipation of the material during the transmission.

[0024] Preferably, the logical relationship followed by the collaborative computing module in calculating the residual deviation characteristics is as follows: ,in, Residual deviation characteristics; For dynamic evolution components; β is the preset characteristic decay coefficient; This refers to the cumulative number of displacement increment pulses output by the displacement sensing module during the material's transfer from the first workstation to the second workstation.

[0025] Preferably, the first control module constructs the state decay model in the following manner: extracting the temperature evolution dimension that spontaneously evolves with the transmission time from the initial process deviation parameter, and defining it as a dynamic evolution component with a constant initial value; extracting the geometric position deviation from the initial process deviation parameter that is not affected by the transmission environment, and defining it as a static fixed component whose value remains constant during the transmission process and is independent of the weight mapping operation.

[0026] Preferably, it also includes: an end-feedback module, used to collect the quality feature vector of the finished product at the end of the production line, map the quality feature vector to a preset state transition matrix, and use the residual value of the quality feature vector to adaptively calibrate the feature attenuation coefficient to offset the physical coupling characteristic shift caused by wear of production equipment or environmental temperature drift.

[0027] Preferably, the second control module is equipped with a compensation threshold determination logic; the compensation threshold determination logic is used to determine whether the actual input vector exceeds the physical compensation limit of the second station actuator; when the determination result is yes, the second control module cuts off the feedforward loop and locks it to the preset reference process parameters to prevent abnormal compensation commands from causing mechanical overload.

[0028] Preferably, the collaborative computing module is also connected to a data buffer module; the data buffer module is used to store the static fixed components with time stamp information collected by the first control module, and according to the cumulative pulse number output by the displacement sensing module, to achieve phase alignment of the static fixed components and residual deviation characteristics at the input end of the second control module for data reconstruction.

[0029] Preferably, the initial process deviation parameters include material surface temperature distribution data, surface coating thickness deviation data, and weld point residual stress data; the state decay model is configured with heterogeneous characteristic decay coefficients for different initial process deviation parameters to characterize the state relaxation characteristics of different physical parameters during material flow.

[0030] Preferably, the first control module and the second control module construct a distributed collaborative topology through an industrial network; when the first control module sends data to the collaborative computing module, it encapsulates the initial count value of the displacement increment pulse to provide a globally consistent displacement reference for cross-node deviation reconstruction.

[0031] Preferably, the resolution of the displacement sensing module is 1024 p / r; the collaborative calculation module dynamically updates the step compensation value of the characteristic attenuation coefficient by counting the displacement increment pulses in real time, so that the calculation accuracy of the residual deviation characteristics is adjusted synchronously with the change of the instantaneous movement speed of the conveying mechanism.

[0032] Example 1: In high-precision continuous series manufacturing processes, materials flow from the first station (outputting high-temperature welding actions) to the second station downstream. Affected by environmental thermal fluctuations and machining stress, the material exhibits an initial process deviation parameter upon leaving the first station, including transient thermal stress and rigid geometric deformation. Traditional collaborative mechanisms treat the physical displacement process as a static time delay. When the downstream node extracts the initial peak deviation, which has not undergone physical dissipation evolution, as a feedforward compensation benchmark, the downstream control unit outputs an equal amount of reverse offset command for the thermal stress component that has been dissipated to the environment during transmission. This causes overcompensation in the mechanical actuator, affecting machining accuracy. The nonlinear decay and adaptive cooperative control system are controlled to start and run under this working condition. In response to the technical problem of the intertwining of physical dissipation attributes and rigid persistence attributes of deviation parameters in cross-workstation flow, the first control module obtains the initial process deviation parameters generated by the material at the first workstation, and according to the component mapping dimension in the preset state decay model, the initial process deviation parameters are decomposed into dynamic evolution components excited by the material temperature field and static fixed components defined by the workstation clamping reference. This decomposition logic uses the material flow evolution law to reconstruct the data, so that the compensation basis of the control system is transformed from the static value initially collected upstream into a state variable with physical dissipation evolution characteristics.

[0033] During the continuous physical flow of materials along the production line, the displacement sensing module outputs displacement increment pulses that change synchronously with the conveying displacement in real time. The dynamic evolution components extracted by the first control module and the displacement increment pulses are input into the collaborative calculation module to form a data mapping closed loop. The collaborative calculation module uses the pulse count output by the displacement sensing module as the independent variable to accumulate the physical integral of the material flow time, and uses the characteristic attenuation coefficient to apply a decreasing scalar multiplication operation to the dynamic evolution components based on the state constraint relationship. Solve for the residual deviation characteristics of the material at the moment it physically arrives at the second workstation node, where, This is a characteristic of residual deviation. For dynamic evolution components, β is a preset characteristic decay coefficient. The number of displacement increment pulses cumulatively output by the displacement sensing module during the material flow from the first station to the second station is defined as follows: The hardware pulses fed back by the displacement sensing module serve as the trigger source to initiate the feature weighting calculation within the collaborative computing module. The dynamic evolution component exhibits a nonlinear convergence trend with the increase of physical displacement, ensuring that the weighting rate of the control information flow and the energy dissipation rate of the material flow remain synchronized. When the cumulative displacement increment pulses reach the preset threshold representing the physical position of the second station node, the second control module reconstructs the data of the residual deviation feature with nonlinear exponential convergence properties and the static fixed component that maintains the initial value to restore the actual input vector. The second control module generates a compensation command for the second station actuator based on the actual input vector to offset the spatiotemporal physical lag deviation of the initial process deviation parameter caused by the material displacement. This command eliminates redundant compensation parameters for the dissipated transient thermal stress and suppresses downstream feedforward overcompensation oscillations caused by static delay alignment. The actual driving displacement output by the second station actuator is physically aligned with the true residual deformation state of the material at the moment of arrival at the second station.

[0034] Example 2: In a multi-station physical testing platform involving continuous welding and high-precision assembly, the first station includes a heating and pressurizing electromechanical unit that outputs the initial welding heat source, and the second station includes an assembly robotic arm. A conveyor belt unidirectionally transports the metal substrate between the first and second stations. This multi-station physical testing platform integrates a laser displacement sensor with a sampling rate of 10kHz and a measurement resolution of 0.1μm to acquire surface deformation data of the metal substrate. Simultaneously, it is equipped with an infrared thermal imager with a temperature measurement accuracy of 0.1℃ to collect the transient temperature field of the metal substrate and transmit it to the laser displacement sensor. Gaussian white noise with a signal-to-noise ratio of 20dB is actively injected into the raw position signal output by the sensor, and a 50Hz power frequency interference harmonic is superimposed to simulate a multi-station workflow engineering benchmark environment with physical dissipation and measurement disturbances. The characteristic attenuation coefficient β set internally by the first control module determines the technical trade-off between compensation response sensitivity and high-frequency disturbance amplification effect. The value of this characteristic attenuation coefficient β depends on the mapping relationship between the heat dissipation rate of the workshop environment and the thermal diffusivity of the metal substrate. When the heat dissipation rate of the environment increases, the characteristic attenuation coefficient β tends to the upper limit of the value range to suppress excessive thermal stress. The mechanical servo oscillation caused by compensation is determined to have a working range of 0.005 to 0.030 for the characteristic attenuation coefficient β based on the thermodynamic constant of the metal substrate. A control group using static delay compensation logic is set up, and a multi-group experimental system with parameter gradients is constructed. The experimental group includes three independent data processing channels with characteristic attenuation coefficient β values ​​of 0.005 (lower limit), 0.012 (median), and 0.030 (upper limit), respectively, to quantify the nonlinear boundary effect of the dynamic evolution component weighting rate on the final assembly accuracy. The upper and lower bounds of this working range are based on the natural pair of standard materials. Based on the cooling experiments, the system uses 20 steel with low thermal conductivity and 6061 aluminum alloy with high heat dissipation as calibration references. Combining the rated running speed of the conveyor belt and the clock sampling rate of the displacement sensor, the thermal diffusion response time of the substrate obtained from the test is scaled and converted. The minimum deweighting requirement critical coefficient for 20 steel is found to be 0.005, and the maximum heat dissipation deweighting requirement critical coefficient for aluminum alloy is found to be 0.030. This calculation process establishes a physical optimization safety boundary at the objective hardware level to prevent the system from undercompensating or overcompensating.

[0035] The first control module acquires the initial process deviation parameter excited by high-temperature welding at the first station, and extracts the original dynamic evolution component mixed with Gaussian white noise from this initial process deviation parameter. The collaborative computing module uses a moving average filtering algorithm to suppress 20dB Gaussian white noise and 50Hz power frequency interference harmonics, and extracts the effective dynamic evolution component with an initial peak value of 145.6μm. During the transfer of the metal substrate to the second station, the collaborative computing module substitutes the number of displacement increment pulses output by the displacement sensing module into the state constraint relationship. The residual error characteristics are solved continuously, where... This is a characteristic of residual deviation. For the effective dynamic evolution component, β is the characteristic decay coefficient. The number of displacement increment pulses; measured data from the testing instrument shows that the control group issued an equal compensation command of 145.6 μm to the second station. Because the physical heat dissipation during the 18.4s transmission period was not taken into account, this resulted in an overcompensated reverse deformation error of 42.3 μm at the assembly interface. In the experimental group, when the characteristic attenuation coefficient β was at the lower limit of 0.005, the residual deviation characteristic value output by the system was too large, resulting in a positive deformation error of 18.7 μm. When the characteristic attenuation coefficient β was at the median of 0.012, the weighting rate conformed to the thermal stress relaxation curve of the metal substrate, and the final assembly error converged to 3.2 μm. When the characteristic attenuation coefficient β reaches its upper limit of 0.030, the compensation command excessively attenuates and approaches the zero limit, resulting in a compensation deficiency and a hysteresis error of 25.4μm, exhibiting a performance degradation inflection point under the exponential attenuation law. The gradient verification data for the characteristic attenuation coefficient β provides numerical boundary support for the nonlinear exponential weighting model applied to the cross-station deviation compensation loop. Based on the characteristic attenuation coefficient, residual deviation features are extracted to eliminate the compensation blind zone of the static queue in the physical state dissipation process of the metal substrate, so that the servo drive compensation displacement output by the control node is aligned in situ with the residual state of the metal substrate at the moment of physical arrival.

[0036] Example 3: When a continuous manufacturing line is in a processing condition with multi-source thermal field interference and frequent switching of metal substrate material, the first control module acquires initial process deviation parameters including position coordinates and deformation depth. The first control module extracts the spatial deformation point cloud of the metal substrate at the moment it leaves the first station to construct a discrete deviation matrix. The first control module inputs this discrete deviation matrix into the singular value decomposition operator inside the state decay model. The singular value decomposition operator solves the covariance matrix and extracts the first principal component eigenvector representing the temperature field gradient and the second principal component eigenvector representing the mechanical clamping residual stress. In the discrete deviation matrix of the spatial deformation point cloud, the material thermal expansion deformation excited by the ambient temperature gradient is manifested as a low-frequency and globally covered gentle curved surface peak, while the deformation caused by the pressure applied by the mechanical gripper is... The residual stress deformation is manifested as a sudden change in high-frequency local peaks concentrated at the contact boundary. The singular value decomposition operator does not simply rely on variance processing without physical meaning, but utilizes the difference in spectral distribution of the above two effects in the point cloud spatial topology to directly reconstruct the low-frequency smooth band with large singular values ​​into temperature thermal field features, and decouples and extracts the high-frequency divergent band with small singular values ​​into mechanical clamping features, thereby achieving cross-scale precise stripping of physical field properties. The first control module removes weak feature values ​​less than the elastic deformation threshold based on the yield strength limit of the metal substrate. The first control module reconstructs the denoised dynamic evolution component and static fixed component, and transforms the composite physical deviation into mutually independent orthogonal basis parameters with clear physical dimensions, outputting the numerical input benchmark of the collaborative compensation algorithm.

[0037] In the multi-station physical flow process, distributed array thermocouples configured on the production line collect transient ambient temperature data in real time along the transmission path between the first and second stations. The collaborative computing module extracts the average fluctuation value from this transient ambient temperature data. The collaborative computing module reads the specific heat capacity and thermal conductivity physical properties of the current batch of metal substrates pre-stored in the underlying database, and solves for the characteristic attenuation coefficient β based on the thermodynamic heat transfer balance equation. The constraint formula for the characteristic attenuation coefficient β is as follows: Where k is thermal conductivity, This is the equivalent heat dissipation surface area of ​​the metal substrate. For specific heat capacity, The bulk density of the metal substrate. This represents the average fluctuation value in transient ambient temperature data. Using a preset standard reference temperature, the physical quantities derived from the above formula characterize the transient convective heat transfer scalar of the metallic substrate. Since the results from direct calculations contain complex thermodynamic dimensions, the collaborative computing module introduces a pre-loaded device spatiotemporal conversion coefficient during the underlying calculation. This coefficient is configured with the reciprocal of the thermodynamic dimensions. By performing a normalized multiplication operation with the result on the right side of the equation, the characteristic attenuation coefficient is ultimately substituted into the subsequent model as a dimensionless pure number satisfying the exponential operation rules. Material properties and environmental monitoring parameters are transformed into dimensionless characteristic attenuation coefficients that determine the slope of deviation attenuation, forming the physical heat dissipation mechanism and control. The mathematical correlation of reallocation; the collaborative operation module substitutes the characteristic attenuation coefficient β and displacement increment pulse obtained in real time into the aforementioned state constraint relationship to output the residual deviation feature. The second control module reconstructs the actual input vector by combining the residual deviation feature with the static fixed component output by the singular value decomposition operator. The second control module generates a displacement compensation command to drive the servo motor of the second station based on the actual input vector. The displacement compensation command offsets the nonlinear state evolution drift of the initial process deviation parameter during the flow of the thermal field environment. The mechanical position compensation amount output by the servo motor of the second station achieves physical alignment with the residual deformation state of the metal substrate at the moment of physical arrival at the second station.

[0038] Example 4: When the system faces the initial physical commissioning of the manufacturing line, the first control module initiates offline baseline calibration of the state decay model. Without a metal substrate supporting it, the first control module drives the heating and pressurizing electromechanical unit of the first station to output a test heat source with progressively increasing output. An infrared thermal imager and distributed array thermocouples located on the transmission path simultaneously collect the spatial thermal radiation flux sequence under no-load conditions. The collaborative computing module receives this spatial thermal radiation flux sequence and extracts the steady-state heat dissipation asymptote generated by the physical displacement of the conveyor belt to calibrate the standard reference temperature. This collaborative computing module compares the thermal conductivity k of the metal substrate with the equivalent heat dissipation surface area. In the imported boundary calculation logic, a physical boundary matrix is ​​generated to constrain the adjustment range of the characteristic attenuation coefficient β, thus completing the initial numerical filling of the core parameters of the state attenuation model with respect to the thermodynamic properties of the production workshop environment.

[0039] After the initial values ​​of the state decay model are filled in, the displacement sensing module outputs test displacement increment pulses to measure the data transmission delay of the underlying industrial bus. The collaborative computing module introduces time lead compensation into the exponential weighted model containing the characteristic decay coefficient β based on the data transmission delay. The second control module sends a reference position movement command to the assembly robot arm of the second station and receives the static mechanical dead zone displacement returned by the servo motor encoder. The first control module writes the static mechanical dead zone displacement as the zero-point offset base of the aforementioned static fixed component into the underlying storage medium. The end-to-end data flow nodes complete the displacement mapping alignment for the current physical environment. The adaptive collaborative control system of the multi-station process parameters of the manufacturing line anchors the computing flow architecture within the objectively measured engineering state boundary before carrying the metal substrate.

[0040] During the baseline calibration phase before system commissioning, the first control module extracts historical base noise deformation data under the idling state of the conveyor belt to construct a base noise matrix. The singular value decomposition operator inside the input state attenuation model solves the baseline eigenvalue vector. The maximum component is multiplied by the physical deformation amplification factor under rated load to calculate the threshold for distinguishing between rigid geometric deformation and transient thermal stress elastic deformation. Based on the elastic deformation threshold, the initial process deviation parameters collected under the operating state are truncated and separated into dynamic evolution components and static fixed components. The second control module reads the rated maximum output torque and mechanical transmission ratio parameters of the servo motor of the second station through the industrial bus. The maximum allowable physical compensation displacement limit of the servo motor is calculated based on the product. The maximum physical compensation displacement limit is used as the fixed comparison reference value for the compensation threshold judgment logic and written into the system's underlying memory. When the absolute value of the actual input vector exceeds the fixed comparison reference value, a feedforward loop cutoff command is triggered.

[0041] Example 5: When the manufacturing line operates continuously for a long time, inducing zero-point drift of the sensor components, the first control module extracts the background mechanical vibration displacement data of the conveyor belt during the idle phase to construct a base noise matrix. This base noise matrix is ​​then input into the aforementioned singular value decomposition operator to solve for the baseline eigenvalue vector. The first control module multiplies the maximum component in this baseline eigenvalue vector by a safety factor to calculate the elastic deformation threshold, which is used to eliminate weak eigenvalues ​​with amplitudes below this threshold. This safety factor, based on the tensile yield test data of the metal substrate, falls within the range of 1.5 to 2.0. This calibration process filters out signal contamination from environmental base vibration in the dynamic evolution component extraction step. During the transfer of the metal substrate to the second station, if the displacement sensing module experiences pulse loss due to electromagnetic interference, causing a measurement deviation in the number of displacement increment pulses substituted into the state decay model, the collaborative calculation module reads the standard physical distance between the first and second stations. and the real-time drive linear speed of the conveyor belt Based on the formula Solve for the theoretical total number of pulses, where, This represents the theoretical total number of pulses. Standard physical spacing, For real-time driving linear velocity, For the calibration time equivalent of a single displacement increment pulse, when the absolute value of the difference between the actual number of pulses accumulated by the displacement sensing module and the theoretical total number of pulses exceeds the preset pulse tolerance threshold, the collaborative calculation module uses the theoretical total number of pulses. Replace the actual accumulated number of displacement increment pulses and substitute them into the state constraint relationship. Solve for the residual bias characteristics, where, This is a characteristic of residual deviation. For the dynamic evolution component, β is the characteristic attenuation coefficient. The second control module outputs compensation instructions to the second station execution mechanism based on the reconstructed actual input vector, so that the issued mechanical position compensation amount and the residual deformation state of the metal substrate reaching the physical coordinates maintain physical coordination.

[0042] To address the issues of environmental temperature drift and equipment wear caused by prolonged operation, the end-point feedback module uses a visual inspection instrument to acquire the actual physical dimensions of the finished product assembly interface at the production line terminal. This data is then compared with the preset standard design reference dimensions to obtain the quality residual value. This residual value is then substituted into a preset closed-loop correction algorithm to periodically update the feature attenuation coefficient. The update logic satisfies a linear constraint relationship. ,in, This represents the feature decay coefficient after the current update iteration. The historical characteristic decay coefficient of the previous cycle is given, and α is a correction step size constant with a fixed value between 0.001 and 0.005. To obtain the quality residual value, the collaborative computing module receives the update instruction and then calls... The original parameters are replaced in the weighted calculation of the residual deviation characteristic index. Based on the closed-loop iteration mechanism, the nonlinear drift of the system caused by the accumulation of physical and mechanical wear is suppressed. In order to solve the dimensional discontinuity between the overall mechanical physical size and the surface heat dissipation properties, the correction step size constant α in the formula is preset as a physical bridging operator with both scale compression and dimensional inversion functions. Its dimension is configured as an inverse dimensional structure of assembly error length unit. When the mass residual value with micron-level dimension is multiplied by the correction step size constant, the overall displacement variable is stripped of physical length attribute in the low-level calculation and completely mapped to a pure numerical equivalent penalty factor that takes into account wear degradation. Thus, the physical coupling offset of the mechanical system is implicitly compensated by a physically self-consistent numerical adjustment method.

[0043] The embodiments of this application have been described above with reference to the accompanying drawings. Unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other. This application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit of this application and the scope of protection of this invention, and all of these forms are within the protection scope of this application.

Claims

1. An adaptive collaborative control system for multi-station process parameters in a manufacturing production line, characterized in that, include: The first control module is configured to acquire the initial process deviation parameter generated by the material at the first work station, and according to the component mapping dimension in the preset state decay model, decompose the initial process deviation parameter into a dynamic evolution component excited by the material temperature field and a static fixed component defined by the work station clamping reference. The displacement sensing module is configured to output displacement increment pulses that change with the displacement transmitted along the production line. The collaborative computing module is connected to the first control module and the displacement sensing module respectively, and is configured to acquire dynamic evolution components, use displacement increment pulses to perform weight mapping on the dynamic evolution components through characteristic attenuation coefficients, and calculate the residual deviation characteristics of the material at the moment of physical arrival at the second work station node. The second control module is connected to the collaborative computing module and is configured to reconstruct the residual deviation features and static fixed components to recover the actual input vector, and generate compensation instructions for the second station actuator based on the actual input vector to offset the spatiotemporal physical lag deviation of the initial process deviation parameter caused by the material displacement.

2. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, When performing weighted mapping, the collaborative computing module is configured to use the pulse count output by the displacement sensing module as the independent variable to perform integral calculation of the material flow time, and input the integral result into the state decay model to determine the weighting coefficient of the dynamic evolution component. The characteristic attenuation coefficient decreases monotonically with the increase of pulse count, causing the residual deviation characteristic to exhibit a nonlinear convergence trend with the increase of material physical displacement, thus offsetting the physical state dissipation of the material during transmission.

3. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, The first control module constructs the state decay model in the following way: it extracts the temperature evolution dimension that spontaneously evolves with the transmission time from the initial process deviation parameter and defines it as a dynamic evolution component with a constant initial value; it extracts the geometric position deviation from the initial process deviation parameter that is not affected by the transmission environment and defines it as a static fixed component whose value remains constant during the transmission process and is independent of the weight mapping operation.

4. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, Also includes: The end-feedback module is used to collect the quality feature vector of the finished product at the end of the production line and map the quality feature vector to a preset state transition matrix. The residual value of the quality feature vector is used to adaptively calibrate the feature attenuation coefficient to offset the physical coupling characteristic shift caused by wear of production equipment or environmental temperature drift.

5. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, The second control module is equipped with compensation threshold determination logic; the compensation threshold determination logic is used to determine whether the actual input vector exceeds the physical compensation limit of the second station actuator; When the determination result is yes, the second control module cuts off the feedforward loop and locks it to the preset reference process parameters to prevent abnormal compensation commands from causing mechanical overload.

6. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, The collaborative computing module is also connected to a data buffer module; the data buffer module is used to store the static fixed components with time stamp information collected by the first control module, and according to the cumulative pulse number output by the displacement sensing module, to achieve phase alignment of the static fixed components and residual deviation characteristics at the input of the second control module for data reconstruction.

7. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, The initial process deviation parameters include material surface temperature distribution data, surface coating thickness deviation data, and weld point residual stress data; the state decay model is configured with heterogeneous characteristic decay coefficients for different initial process deviation parameters to characterize the state relaxation characteristics of different physical parameters during material flow.

8. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, The first control module and the second control module construct a distributed collaborative topology through the industrial network. When the first control module sends data to the collaborative computing module, it encapsulates the initial count value of the displacement increment pulse to provide a globally consistent displacement reference for cross-node deviation reconstruction.

9. The adaptive collaborative control system for multi-station process parameters in a manufacturing production line according to claim 1, characterized in that, The displacement sensing module has a resolution of 1024 p / r; the collaborative computing module dynamically updates the step compensation value of the characteristic attenuation coefficient by counting the displacement increment pulses in real time, so that the calculation accuracy of the residual deviation characteristics is adjusted synchronously with the change of the instantaneous movement speed of the conveying mechanism.

Images