An integral thickness coordination control system and method for an expansion joint

Abstract

The application provides a kind of expansion joint integral forming thickness coordination control system and method, it is related to industrial interconnection and adaptive control system technical field, the method is executed by the processor of coordination control system, obtains pipe blank real-time operation sequence and instruction time stamp from industrial data gateway and extracts hardware response hysteresis evaluation value and cumulative work mapping index;Hardware hysteresis evaluation value is injected into industrial forming deduction model, and dynamic mapping scaling is executed to built-in work safety constraint boundary to generate adaptive work constraint boundary;Work mapping index is input into adaptive boundary to execute out-of-limit evaluation to extract forming feed correction parameter;Adaptive control signaling is generated based on the parameter and issued to servo queue.Effectively solve the problem of closed-loop control failure caused by static model misplacement under complex disturbance and long-term service wear of equipment, realize adaptive coordinated control.

Description

Technical Field

[0001] This invention relates to the field of industrial interconnection and adaptive control system technology, specifically to a control system and method for the integral forming thickness coordination of expansion joints. Background Technology

[0002] In the modern intelligent manufacturing network architecture, the Manufacturing Execution System (MES), as the global data scheduling hub, is gradually extending to edge devices in order to realize the online adaptive evolution of the mechanism model based on microsecond-level industrial data, thereby dealing with the complex nonlinear rheological disturbances in the forming of metal pipes (such as expansion joints).

[0003] In the existing field of thickness forming control, Chinese patent document CN122411018A discloses an adaptive compensation control method for wall thickness fluctuations in aluminum profile extrusion processing. This existing solution utilizes a high-frame-rate line scan camera and a multi-point ultrasonic array to simultaneously acquire profile cross-sectional data, thereby constructing a dynamic map of the wall thickness-sensitive area. Combined with a lightweight temporal memory network, it achieves online prediction and functional area compensation, guiding the extrusion equipment to correct die clearance or extrusion speed, thus forming a closed-loop wall thickness adjustment at the digital level. However, facing the unsteady disturbances of complex industrial environments, the aforementioned existing technology still faces certain limitations under long-term continuous service conditions. In industrial manufacturing environments, servo actuators and mechanical transmission devices inevitably experience mechanical wear after long-term operation, leading to sluggish action response. Existing solutions often rely on the spatial dimensions of external sensors for static feedforward or post-compensation, failing to decouple the physical degradation characteristics of the equipment over time from the control link. This static, fixed parameter control architecture prevents the pre-defined industrial mechanism model within the control system from sensing the dynamic transmission resistance of the hardware, resulting in a severe long-term mapping misalignment between the safety physical boundaries set in the digital space and the actual physical execution capabilities. When encountering complex conditions such as localized material hardening or transient pressure spikes in fluids, traditional fixed models struggle to achieve dynamic decoupling of the multi-dimensional physical variables of "displacement-pressure." This often leads to closed-loop control failure due to servo actions lagging behind algorithm commands, ultimately resulting in uneven stress on the billet and irreversible localized material tearing defects. A technical solution capable of addressing these problems is urgently needed in this field. Summary of the Invention

[0004] The purpose of this invention is to provide a control system and method for the overall forming thickness coordination of expansion joints, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for collaboratively controlling the thickness of an expansion joint during integral forming, executed by the processor of a collaborative control system, includes the following steps:

[0007] The real-time hydraulic status sequence and displacement feed sequence of the current expansion joint tube blank are obtained from the industrial control data gateway of the manufacturing execution system, and the instruction response timestamp sequence of feed control command issuance timestamp and status feedback confirmation timestamp is obtained by parsing the forming control log.

[0008] The communication response delay pattern is extracted periodically from the instruction response timestamp sequence to extract the hardware response hysteresis evaluation value that characterizes the current hardware transmission resistance. At the same time, the dynamic mechanical power dissipation feature is extracted from the real-time hydraulic state sequence and displacement feed sequence to generate a cumulative work mapping index that characterizes the amount of mechanical deformation dissipation work that the current expansion joint tube blank has undergone.

[0009] The hardware response hysteresis evaluation value is injected into the preset industrial forming simulation model, and its built-in preset forming work safety constraint boundary is dynamically mapped and scaled to generate an adaptive work constraint boundary that is dynamically adjusted based on the equipment aging state; the cumulative work mapping index is input into the adaptive work constraint boundary to perform over-limit state evaluation, and forming feed correction parameters used to compensate for work over-limit deviation are extracted.

[0010] Based on the forming feed correction parameters, an adaptive control signaling containing a dynamic damping compensation strategy is generated in the processor's embedded control logic center, and the adaptive control signaling is directed to the corresponding servo drive logic queue to update the mapping relationship of the feed and pressure dual closed-loop control parameters of the servo actuator at the control logic layer.

[0011] Preferably, the communication response delay pattern is extracted periodically from the instruction response timestamp sequence, and a hardware response hysteresis evaluation value characterizing the current hardware transmission hindrance is extracted, including:

[0012] Extract the feed control command issuance timestamp and the corresponding status feedback confirmation timestamp from the command response timestamp sequence, and calculate the single execution response time difference set;

[0013] The timing deviation tracking algorithm is applied to the set of single execution response time differences to remove discrete interference components that characterize random communication jitter data, and a smooth hysteresis baseline feature that characterizes the wear and decay trend of the device over time is fitted and output. The current timing value of the smooth hysteresis baseline feature is used as the hardware response hysteresis evaluation value.

[0014] Preferably, dynamic mechanical work dissipation feature extraction is performed on the real-time hydraulic state sequence and displacement feed sequence to generate a cumulative work mapping index characterizing the amount of mechanical deformation dissipation work that the current expansion joint billet has undergone, including:

[0015] The real-time hydraulic state sequence and displacement feed sequence are aligned with the spatiotemporal dimension features based on the system global clock synchronization tag of the industrial control bus, and the transient compressive stress feature scalar and single-step displacement increment parameter mapped under the unified time reference are extracted respectively.

[0016] Based on the preset material mechanics mapping rules of the current tube blank, the effective cross-sectional area of ​​the servo hydraulic cylinder and the cross-dimensional mapping constant are introduced. The transient compressive stress characteristic scalar and the single-step displacement increment parameter are subjected to cross-dimensional incremental product operation with dimension alignment as the benchmark. Then, the results of the incremental product operation are rolled and accumulated along the time axis in the current entire processing cycle. The accumulated total value is introduced with an elastic loss factor representing the preset tube blank elastic deformation redundancy for system-level multiplication filtering and stripping. Finally, the accumulated work mapping index is output.

[0017] Preferably, the hardware response hysteresis evaluation value is injected into a pre-set industrial forming simulation model, and its built-in preset forming work safety constraint boundary is dynamically mapped and scaled to generate an adaptive work constraint boundary that is dynamically adjusted based on the equipment aging state, including:

[0018] The hardware response hysteresis evaluation value is input into a pre-configured transmission efficiency loss mapping dictionary. Through feature matching query, the loss coefficient characterizing the current tube blank deformation work transmission efficiency is output.

[0019] Using the loss coefficient of the current billet deformation work transfer efficiency as a multiplier factor, the multidimensional tolerance boundary of the preset forming work safety constraint boundary is proportionally reduced and mapped to generate the dynamically shrunken adaptive work constraint boundary.

[0020] Preferably, the cumulative work mapping index is input into the adaptive work constraint boundary to perform an over-limit state evaluation, and the forming feed correction parameters used to compensate for the work over-limit deviation are extracted, including:

[0021] Extract the transient work change rate parameter of the cumulative work mapping index in the current control cycle in real time;

[0022] The cumulative work mapping index and its transient work change rate parameter are compared with the lower limit boundary characteristics of the adaptive work constraint boundary by performing logical deviation quantification.

[0023] If the logic deviation quantification result indicates that the cumulative work mapping index breaks through the adaptive work constraint boundary, then the over-limit deviation amplitude of the cumulative work mapping index exceeding the lower limit boundary characteristic is calculated, and the current transient fluid pressure stress reference value is obtained simultaneously. Combining the preset flow-displacement conversion characteristics and dimension equivalence constant, the spatial dimension inverse mapping solution is performed on the over-limit deviation amplitude, and the physical yield equivalent is extracted as the forming feed correction parameter.

[0024] Preferably, based on the forming feed correction parameters, an adaptive control signaling incorporating a dynamic damping compensation strategy is generated within the processor's embedded control logic center, including:

[0025] Based on the forming feed correction parameters, an evaluation index is constructed in the feature space to characterize the current billet instability state, and the target intervention mode is locked as flexible deceleration compensation mode or constant pressure safety interception mode according to the evaluation index.

[0026] If it is the constant pressure safety interception mode, then based on the basic scheduling strategy of the forming controller, the default position closed-loop control logic is reconstructed into constant pressure relief protection logic, and an adaptive control signaling containing safety deceleration command and explosion-proof pressure relief threshold parameters is organized and constructed.

[0027] Preferably, the method further includes a dynamic closed-loop update step:

[0028] Establish a mapping relationship between the hardware response hysteresis evaluation value and the preset dynamic filtering configuration library inside the industrial control data gateway;

[0029] Based on the hardware aging ladder value range of the currently obtained hardware response hysteresis evaluation value, the corresponding range identifier is extracted and converted into a query mapping identifier. Based on the query mapping identifier, the dynamic filter configuration library is adaptively retrieved to extract the matching target low-pass filter frequency parameter.

[0030] The filtering and interception rules for hydraulic state sequences in the industrial control data gateway are updated in real time using the target low-pass filter frequency parameters.

[0031] Preferably, the logical deviation quantification calculation is performed between the cumulative work mapping index and its transient work change rate parameter and the lower limit boundary feature of the adaptive work constraint boundary, including:

[0032] Based on the transient work change rate parameter, within a preset control simulation time window, the cumulative work mapping index is subjected to a numerical extension calculation based on the prediction period to generate a predictive work evolution trajectory that characterizes the future work absorption trend of the current billet.

[0033] Perform a numerical intersection comparison operation between the predictive work evolution trajectory and the lower limit boundary features of the adaptive work constraint boundary, and extract the trajectory intersection point timestamps in the corresponding trajectory intersection states;

[0034] If the timestamp of the trajectory intersection falls within the control simulation time window, it is determined that the current forming trend exceeds the adaptive work constraint boundary, and the observation node is locked at the end section of the control simulation time window. The absolute work difference between the extended trajectory prediction scalar and the adaptive lower limit boundary feature under the end section is extracted as the expected work overflow value and output as the logic deviation quantization calculation result.

[0035] Preferably, based on the forming feed correction parameters, an evaluation index is constructed in the feature space to characterize the current billet instability state, and the target intervention mode is locked as a flexible deceleration compensation mode or a constant pressure safety interception mode according to the evaluation index, including:

[0036] The extracted forming feed correction parameters and the current transient compressive stress characteristic scalar in the real-time hydraulic state sequence are subjected to multivariate logical weighted aggregation operation to calculate and output the local stress instability assessment value used to characterize the degree of local stress distortion of the current billet.

[0037] If the local stress instability assessment value does not reach the pre-configured safety damping threshold, the target intervention mode will be locked as the flexible deceleration compensation mode.

[0038] In response to the stress distortion exceeding the limit state where the local stress instability assessment value is greater than or equal to the safety damping threshold, it is determined that the current billet has entered a dangerous state of rapid material stress concentration. The target intervention mode is directly switched to the constant pressure safety interception mode in the embedded control logic center to block the continuous injection of excessive mechanical work.

[0039] An expansion joint integral forming thickness coordination control system, the system being used to execute the expansion joint integral forming thickness coordination control method, comprising:

[0040] The data perception and analysis module is configured to obtain the real-time hydraulic status sequence and displacement feed sequence of the current expansion joint billet from the industrial control data gateway of the manufacturing execution system, and to parse the forming control log to obtain the instruction response timestamp sequence of the feed control instruction issuance timestamp and status feedback confirmation timestamp.

[0041] The dual-track feature calculation module is configured to perform periodic communication response delay pattern extraction on the instruction response timestamp sequence, and extract a hardware response hysteresis evaluation value that characterizes the current hardware transmission resistance level; at the same time, it performs dynamic mechanical power dissipation feature extraction on the real-time hydraulic state sequence and displacement feed sequence, and generates a cumulative work mapping index that characterizes the amount of mechanical deformation dissipation work that the current expansion joint tube blank has borne.

[0042] The constraint boundary dynamic scaling and over-limit assessment module is configured to inject the hardware response hysteresis assessment value into a preset industrial forming simulation model, dynamically map and scale its built-in preset forming work safety constraint boundary to generate an adaptive work constraint boundary dynamically adjusted based on the equipment aging state; and input the cumulative work mapping index into the adaptive work constraint boundary to perform over-limit state assessment, and extract forming feed correction parameters used to compensate for work over-limit deviation.

[0043] The signaling reconstruction and dual-loop update module is configured to generate adaptive control signaling containing a dynamic damping compensation strategy within the embedded control logic center of the processor based on the forming feed correction parameters, and to direct the adaptive control signaling to the corresponding servo drive logic queue, so as to update the mapping relationship of the feed and pressure dual-loop control parameters of the servo actuator at the control logic layer.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] This invention synchronously acquires the running sequence and command response timestamps from an industrial control data gateway, extracting response hysteresis evaluation values ​​characterizing hardware transmission resistance and cumulative work mapping indices characterizing deformation dissipation, respectively. Utilizing these hardware hysteresis characteristics, the invention dynamically maps and scales the safety work constraint boundaries in a pre-set industrial forming simulation model, enabling the control safety envelope in the digital space to adaptively adjust in real-time to follow the mechanical wear and degradation state of the hardware. Based on the limit-crossing evaluation results, this mechanism generates adaptive control signaling incorporating a dynamic damping compensation strategy, updates the dual closed-loop parameter mapping at the control logic layer, and ultimately achieves coordinated uniform forming and tear resistance of the expansion joint thickness under long-term service conditions.

[0046] Furthermore, this invention introduces a timing deviation tracking algorithm and a dynamic closed-loop update mechanism on the basis of the above, eliminating discrete random communication jitter, and establishing a mapping relationship between hardware hysteresis evaluation values ​​and the dynamic filter configuration library within the gateway. This system can adaptively retrieve and distribute target low-pass filter frequency parameters according to the hardware aging ladder of the equipment, updating the filtering and interception rules in real time from the data intake source, effectively blocking the penetration and contamination of low-frequency resonant noise excited by late-stage mechanical runout into the core solution network, and improving the determinism of the data acquisition link.

[0047] Preferably, the present invention introduces a cross-dimensional mapping constant to perform cross-dimensional incremental product operations and injects an elastic loss factor in the rolling accumulation of the time axis, so as to accurately reduce the dimension of discrete and heterogeneous pressure and displacement into deformation mechanical work indicators stripped of non-plastic rebound redundancy; at the same time, it combines the flow-displacement conversion characteristics to perform inverse mapping calculation on the over-limit deviation amplitude, so as to safely reconstruct the energy domain deviation determined by digital space into a three-dimensional spatial physical yield equivalent that can be directly executed by the servo motor.

[0048] Furthermore, by employing a multivariate logical weighted aggregation mechanism of feed correction parameters and transient pressure to construct a local stress instability assessment value, this invention establishes a tiered defense gating system. In response to the assessment value exceeding the safety damping threshold, the intervention mode is directly switched to a constant pressure safety interception mode, maintaining the overflow valve in normal operation by utilizing a signaling mechanism that includes an explosion-proof pressure relief threshold. This strategy, while maintaining the physical baseline of the static anti-fall static pressure benchmark for heavy molds, blocks transient stress mutations induced by latent hardening nodules in the material, avoiding secondary problems caused by abrupt displacement compensation. Attached Figure Description

[0049] Figure 1 This is a closed-loop control flow diagram illustrating the core physical application scenario and technical route of the present invention.

[0050] Figure 2 This is a schematic diagram of the technical process of the present invention. Detailed Implementation

[0051] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0052] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0053] Example 1:

[0054] Please see Figure 1 and Figure 2 The present invention provides a technical solution:

[0055] A method for coordinated control of the integral forming thickness of an expansion joint, executed by the processor of a coordinated control system, includes the following steps:

[0056] The real-time hydraulic status sequence and displacement feed sequence of the current expansion joint tube blank are obtained from the industrial control data gateway of the manufacturing execution system, and the instruction response timestamp sequence of feed control command issuance timestamp and status feedback confirmation timestamp is obtained by parsing the forming control log.

[0057] The communication response delay pattern is extracted periodically from the instruction response timestamp sequence to extract a hardware response hysteresis evaluation value that characterizes the current hardware transmission resistance. At the same time, the dynamic mechanical power dissipation feature is extracted from the real-time hydraulic state sequence and displacement feed sequence to generate a cumulative work mapping index that characterizes the amount of mechanical deformation dissipation work that the current expansion joint tube blank has borne.

[0058] The hardware response hysteresis evaluation value is injected into the preset industrial forming simulation model, and its built-in preset forming work safety constraint boundary is dynamically mapped and scaled to generate an adaptive work constraint boundary that is dynamically adjusted based on the equipment aging state; the cumulative work mapping index is input into the adaptive work constraint boundary to perform over-limit state evaluation, and forming feed correction parameters used to compensate for work over-limit deviation are extracted.

[0059] Based on the forming feed correction parameters, an adaptive control signaling containing a dynamic damping compensation strategy is generated in the processor's embedded control logic center, and the adaptive control signaling is directed to the corresponding servo drive logic queue to update the mapping relationship of the feed and pressure dual closed-loop control parameters of the servo actuator at the control logic layer.

[0060] This embodiment operates within a collaborative control system, which includes at least: a core processor for performing logic calculations and strategy distribution; a servo drive logic queue for high-frequency electrical signal interaction with the multi-axis hydraulic forming equipment; and a Manufacturing Execution System (MES) industrial control data gateway for cross-system integration of production data. The core processor maintains microsecond-level synchronous communication with each sensing and execution node via a high-speed industrial control network (such as an EtherCAT communication link). All subsequent multi-dimensional feature calculations, adaptive constraint contraction, and edge control signaling transitions in this embodiment are performed by the core processor based on an embedded industrial mechanism model after acquiring industrial data, ultimately sending the generated control signals to the physical servo nodes to address the technical problem of thickness instability in expansion joints under complex material rheology. The collaborative control system internally instantiates logical processing execution entities mirroring the above processing stages, specifically including: a data sensing and parsing module, a dual-track feature calculation module, a constraint boundary dynamic scaling and over-limit evaluation module, and a signaling reconstruction and dual-loop update module. Under the unified time scheduling of the processor, the aforementioned logical entities complete the cross-logic module transmission and interaction of multi-dimensional state parameters through the high-speed communication link within the system, collaboratively constructing a hardware-software integrated closed-loop defense line from heterogeneous state perception to top-level adaptive control interception. The specific execution logic includes the following steps:

[0061] The first execution step: This method is executed by the processor of the collaborative control system as the core logic control center. The processor of the collaborative control system first performs an offline calibration stage for the control mechanism. This is achieved by constructing a standardized data sensing source for extracting mechanical benchmarks. The processor of the collaborative control system batch-extracts full-volume forming samples of the current expansion joint material billet within the historical processing cycle from the historical production database of the external manufacturing execution system. The processor extracts the historical full-dimensional sensor time-series stream bound to each forming action as the input feature matrix for this sample set, and simultaneously analyzes the actual wall thickness reduction 3D distribution map uploaded after quality inspection by the 3D optical scanner as the target supervision label. The processor injects the above input feature matrix (its data dimension is preferably configured as M×K, where M represents the total number of historical sample collection batches of the material billet, K represents the total number of feature channels of the system's multi-dimensional sensors, and based on the current sensing hardware topology of the hydraulic forming equipment, its conventional point value configuration is 12 to 24 dimensional channels) into the initial state of the industrial forming simulation model. After performing spatiotemporal alignment of the feature dimensions based on a unified microsecond-level timestamp benchmark, the model calculates and outputs the predicted geometric deformation distribution vector through the internal feedforward network simulation structure. The processor extracts the loss evaluation parameter L based on the absolute geometric deviation computation framework, and its feature mapping expression is as follows: Where N is the total number of discrete detection nodes acquired by the 3D optical scanner during the current evaluation period; and These represent the predicted wall thickness reduction parameter derived by the model at the i-th spatial node and the physical wall thickness reduction parameter obtained from the actual 3D scan, respectively. Both are measured in millimeters (mm), the absolute unit of length. This transforms the geometric deviation of material rheology into a definite feedback measure, thereby activating the error backpropagation algorithm in the internal computational graph. By continuously and adaptively adjusting the link weights between network nodes, which characterize the extreme yield value of the metal lattice, in multiple iterations, until this loss evaluation parameter is consistently lower than the preset convergence tolerance threshold (characterizing the geometric deviation between the model derivation and the actual 3D optical scan result, preferably within the range of 0.01 mm to 0.05 mm) in multiple consecutive test batches (preferably 5 to 10 consecutive evaluation cycles), the model is deemed to have reached a globally stable state.

[0062] In this stable state, the offline state envelope fitting logic is activated, specifically performing a dimensionality reduction transformation from the geometric space to the physical work domain. The processor retrieves the abnormal dataset from historical forming samples that were identified as local material tearing defects by physical quality inspection, and extracts the extreme combination of the ultimate compressive stress scalar and transient deformation slope within the control time window preceding each actual material tearing. A discrete feature scatter plot representing the failure state is constructed in the two-dimensional control feature coordinate system, and a density-based envelope fitting rule (e.g., an Alpha-shape edge extraction algorithm or a contour line fitting algorithm based on kernel density estimation) is invoked. This ensures that the interface does not contain any tearing failure sample points and that the nearest failure sample point reaches a preset safety distance (e.g., Euclidean distance) as a constraint. The optimal physical geometric interface that isolates all historical safe forming samples from the aforementioned tearing failure samples is calculated. The processor extracts the inner boundary topological feature parameters of this optimal physical geometric interface, specifically the polynomial fitting coefficients of the discrete vertex coordinate set or boundary curve constituting the two-dimensional envelope surface. These parameters are statically fixed in the control logic core, and the preset forming work safety constraint boundary is extracted using this method, denoted as... The forming work safety constraint boundary encompasses a multi-dimensional state envelope encompassing both absolute compressive stress and transient deformation slope dimensions. It characterizes the maximum non-destructive mechanical work limit that the billet can withstand without causing irreversible tearing of the material under the current production line's basic mechanical properties. In this embodiment, the upper limit of the absolute compressive stress within this multi-dimensional state envelope is convergently constrained within a safe stress range of 20 to 28 MPa. In engineering measurements for standard stainless steel corrugated pipe forming, a preferred calibration point value of 25 MPa is selected. To prevent stress concentration due to excessively rapid material deformation evolution, the upper limit of the transient deformation slope within this envelope is locked within a dynamic domain of 0.3 to 0.6 mm / ms, with an extraction point value of 0.5 mm / ms under the current baseline state.

[0063] Further defining the processor execution data acquisition steps of the collaborative control system: Communication interaction data within the current control cycle is retrieved in real-time from the industrial control data gateway of the manufacturing execution system via the standard OPCUA industrial communication gateway data interface to establish the first data sensing link. The processor synchronously acquires and incorporates the real-time hydraulic state sequence using industrial-grade pressure-sensitive sensors deployed in the servo hydraulic pipeline network or mold cavity, denoted as... It characterizes the transient compressive stress fluctuation state experienced by the expansion joint tube blank within the mold cavity. In a preferred embodiment, the data sampling frequency of this sequence is set to a high-frequency range of 1000Hz. The processor, along the aforementioned first data sensing link, reads the displacement state register data stored after high-speed hardware counting and digital quantization of the feedback pulse stream from the servo driver to the standard photoelectric encoder or linear grating ruler, and synchronously acquires and introduces the displacement feed sequence, denoted as... It characterizes the absolute advance distance and feed rate state vector of the servo forming actuator in three-dimensional physical space.

[0064] The processor of the collaborative control system parses the forming control log of the current forming cycle by calling the system's operation log reading interface. Through structured parsing of the log strings, it extracts the communication protocol transaction ID (or global instruction sequence number) recorded in the log. Using this transaction ID as the primary key, it performs context matching to extract a sequence of instruction response timestamps containing the issuance time of each feed instruction and the corresponding status feedback confirmation timestamp, denoted as... It represents the closed-loop state of total communication and mechanical time consumption in an industrial control network, from the issuance of commands to the feedback response of physical servo nodes.

[0065] The second execution step involves the processor of the collaborative control system acquiring the output of the first execution step and performing parallel time-domain analysis of the hardware degradation state and energy dissipation calculation of the billet deformation state. Specifically addressing the aging of physical equipment and communication delays in industrial control networks, the processor of the collaborative control system extracts the periodic communication response delay pattern from the currently resident instruction response timestamp sequence. The processor iterates through the instruction response timestamp sequence and extracts the timestamp of each feed control instruction issuance and its corresponding status feedback confirmation timestamp, calculating the time-domain deviation between the two. This allows the processor to extract the set of single-execution response time differences at the logic layer, denoted as... It represents the discrete time delay state cluster of the current servo system from receiving the command to the actual action of the mechanical valve body and the feedback.

[0066] The processor of the collaborative control system applies a timing deviation tracking algorithm (e.g., Kalman filter state observation logic) to the set of single-execution response time differences. Specifically, to address the high-frequency random jitter of the industrial control network, this data processing mechanism manifests as discrete recursive estimation logic deployed within the control domain. Before entering the filter update of the current k-th control cycle, the processor performs a linear extrapolation of the time axis based on the posterior estimated state fixed in the previous control cycle to generate the expected time delay parameters for the current cycle. The calculation model is fixed through prior state extrapolation as follows: In this algebraic mapping equation The expected time delay parameter (ms) for the kth control cycle is derived from the simulation. The a posteriori hardware hysteresis time node (ms) is extracted and fixed by the system in the previous control cycle (i.e., the k-1th cycle). To characterize the evolution law of the system's hysteresis state, the discrete state transition constant matrix is ​​based on a first-order evolution model of oil viscosity deterioration. The main diagonal elements of this matrix are preferably fixed between 0.95 and 0.99 in engineering configuration. The incremental control input parameter (ms) is injected for compensation. Discrete state transition constant matrix. The control input incremental parameters are determined offline by the processor of the collaborative control system based on the viscosity degradation trend curve of the actuator hydraulic oil; The opening slope of the control valve port in the main oil circuit is determined by the servo controller based on the current command. This is achieved through real-time linear multiplication using a preset valve port response slope-time hysteresis mapping coefficient. In engineering practice for medium-sized proportional servo valves, the preferred reference value for this mapping coefficient is calibrated to 0.25. The processor calculates the logical information deviation between the expected time delay derived from the internal mechanism model and the observed time difference. Its computational relationship is mapped as Based on the logical information deviation characteristic, an adaptive weighting operation is performed on the filter update gain matrix, and the specific state mapping rule is solidified as follows: .in This represents the actual observation time difference obtained through gateway parsing in the current k-th control cycle; The filter update gain matrix represents the current control cycle, and its dimensionless weighting coefficients characterize the system confidence level. The prediction error covariance parameter in the system state space is preferably configured to be 10 at the initial diagonal covariance point value when the system is cold-started at time zero, and it will adaptively converge in subsequent recursions. The basic measurement is the noise covariance parameter; and This is a nonlinear penalty adjustment function constructed based on the statistical distribution patterns of historical electromagnetic interference in industrial networks under heavy load and concurrency. As a specific application example, the prediction error covariance parameter... Covariance update equation from Kalman filter Iterative calculation determines, where The process excitation noise covariance matrix characterizes the fluctuation amplitude of random losses within the mechanical system. Its preferred initial static point value matrix has diagonal elements configured to 0.05; the basic measurement noise covariance parameter... The noise floor variance of the industrial control network communication is determined by static loading. The system's embedded adjustment mapping equation is set as follows: In this mapping logic The millisecond-level logical information deviation generated by the preceding calculation; The preset high-frequency interference threshold is used to shield the background noise of conventional network communication. Its quantization physical unit is milliseconds (ms), and the preferred fixed value in this Ethernet communication architecture is 50 milliseconds. The exponential amplification gain coefficient, calibrated based on channel characteristics, is preferably 0.15 in this embodiment. The processor makes a determination based on the aforementioned nonlinear mapping action: when the observation deviation... High frequency interference threshold Within the tolerance range, the nonlinear penalty adjustment function outputs a stationary constant of 1, maintaining the basic filtering; once... Once the high-frequency interference threshold is exceeded, the output value of the nonlinear penalty adjustment function experiences a steep exponential jump.

[0067] When the calculated current response time difference exceeds three times the standard deviation of the historical smoothed baseline, or the absolute sudden increase in deviation is greater than 50 milliseconds and exhibits a high-frequency isolated spike, it is characterized as random communication jitter data triggered by electromagnetic interference or a surge in transient load on the communication channel. These discrete interference components are then stripped from the feature space by reducing the corresponding gain weight. When the deviation data exhibits a continuous, monotonically increasing trend over time (e.g., the time deviation slope remains positive and the cumulative delay increment exceeds 15 milliseconds over 10 to 20 consecutive control cycles), it is determined to be a physical hindrance evolution caused by hydraulic fluid viscosity degradation or guide rail interference. Through this interactive process of observation and recursive correction to eliminate false jitter contamination, a smooth hysteresis baseline characteristic representing the irreversible decay trend of mechanical equipment over time is fitted and output.

[0068] For the set of single-execution response time differences, the preferred detection lower limit is 0 milliseconds, and the upper limit is limited by the timeout packet loss constant of the current industrial Ethernet protocol (e.g., set to 500 milliseconds). The processor extracts the transient value of this smooth hysteresis baseline feature at the end of the current control cycle and performs parameter fixing. This transient value is the hardware response hysteresis evaluation value, denoted as... It characterizes the systematic execution lag caused by the combined effects of hydraulic fluid viscosity degradation and mechanical guideway interference wear. In the engineering calibration based on the conventional hydraulic station of a small and medium-sized forming unit, the effective range of the hardware response lag evaluation value is mapped to between 10 milliseconds and 150 milliseconds (e.g., the point value under the current decay state is 45 milliseconds).

[0069] In another parallel data processing link, to address the issue that simple spatial displacement monitoring cannot predict material tearing, the processor of the collaborative control system performs dynamic mechanical power dissipation feature extraction on the real-time hydraulic state sequence and displacement feed sequence. Based on the system global clock synchronization tag embedded in the industrial control communication protocol, the processor performs spatiotemporal dimension feature alignment on the hydraulic state sequence (pressure dimension) and the displacement feed sequence (length dimension). Considering the often asynchronous difference between the high-frequency sampling rate of pressure-sensitive elements (e.g., 1000Hz) and the physical position feedback frequency of servo actuator encoders in industrial control environments, the processor activates resampling compensation logic during feature alignment: using a uniform 1-millisecond clock beat as the absolute alignment reference axis, for the missing intermediate timestamps in the displacement feed sequence due to the relatively low sampling rate, the zero-order hold operator is called to extract the value of the most recent valid displacement sampling point and perform translation filling. This resampling mapping operation eliminates asynchronous misalignment pollution caused by inconsistent network sampling paces of multi-source sensors, ensuring that in any subsequent 1-millisecond calculation slice, there are pressure-displacement parameter pairs from the same source and at the same time.

[0070] After completing the spatiotemporal alignment of the high-frequency clock, the processor invokes the preset material mechanics mapping rules for the current billet to perform cross-dimensional logical mapping on the acquired multidimensional features to eliminate dimensional differences. The processor executes a structured logical mapping process in the logic operation unit. Specifically, based on the alignment reference, it extracts the transient compressive stress feature scalar at each synchronous sampling time node t. And simultaneously acquire the corresponding single-step displacement increment parameters. The processor performs a numerical product operation on the aforementioned heterogeneous features. This reduces the dimension of discrete fluid pressure and mechanical displacement signals and reconstructs them into intermediate features representing transient mechanical work quantities characterizing the energy injection intensity. The processor then processes these transient work features along the current forming process evolution time T. current Continuous rolling integral accumulation is performed along the time axis, and an elastic loss factor is introduced into the final accumulation result to remove the redundant potential energy of non-plastic deformation that can rebound autonomously after decompression, thereby smoothly outputting a mapping index representing the final cumulative deformation dissipation in the control center.

[0071] This embodiment introduces hardware topology constants to achieve dimension alignment. Specifically, the following feature mapping equation is used to output the final work mapping index: ;in The effective cross-sectional area of ​​the hydraulic servo cylinder, used as a static configuration constant, has a physical dimension of millimeters (mm²). Considering the physical specifications of a medium-sized pipe forming unit, the optimal value for this cross-sectional area parameter corresponds to the area of ​​a standard servo master cylinder with an inner diameter of 200mm. The specific value is adjusted according to the actual cylinder size. A cross-dimensional mapping constant is introduced into the product term. Using this transdimensional mapping constant, its value is 10. -3 A differential adjustment is performed on the intermediate cumulative products to offset carry-over misalignment between units. This allows for the coherent accumulation of the product of each microsecond-level slice in digital space along the time axis. A dimensionless elastic loss factor is introduced. Perform multiplication filtering. The final output is... The cumulative work done index (joules, J) represents the total effective mechanical energy absorbed and dissipated by the metal lattice during the nonlinear plastic deformation of the current expansion joint billet. Under actual operating conditions based on a medium-to-large servo hydraulic pump station, the time-series evolution of this dynamic cumulative value spans from 0 to 5000 joules; for example, the transient value extracted at the end of the current processing cycle is 3450.5 joules.

[0072] In the above accumulation logic, the synchronous sampling time node t represents the discrete data slice base point under the global clock synchronization protocol of the industrial control network. The step span of this time base point is constrained to 1 millisecond (e.g., the current control polling cursor is at the 450th millisecond); at the same time, the integration upper limit parameter is established as the current forming process evolution duration, denoted as... It represents the cumulative physical time span experienced by the current billet from the moment it first contacts the mold and undergoes stress deformation until the current controlled simulation section. Based on the single forming cycle of the expansion joint, the dynamic range of this upper limit of the time domain is mapped between 0 and 3000 milliseconds (e.g., the point value obtained from the current simulation section is 1500 milliseconds).

[0073] In the standard stainless steel tube blank forming process, the characteristic scalar of transient compressive stress The effective sensing boundary is limited to the range of 0 to 30 MPa. Single-step displacement increment parameters participate simultaneously in cross-dimensional calculations. Limited by the mechanical decoding resolution of the servo hydraulic cylinder, its value range is between 0 and 0.5 mm.

[0074] For elastic loss factor The underlying mechanism is that in the initial stage of die extrusion, the internal lattice stretching of the billet metal does not exceed the physical yield limit. At this time, the mechanical work injected by the servo actuator into the billet is merely converted into elastic potential energy that can fully rebound autonomously after the pressure is removed. During the offline configuration stage, based on the specific billet material characteristics imported externally (such as elastic modulus and initial yield strength), a fixed loss baseline coefficient is pre-assigned to this batch of tasks at the control logic layer. In continuous forming sequences for commonly used materials such as 304 austenitic stainless steel, this recovery component is typically rigidly mapped and configured within an absolute loss threshold range of 3% to 5% of the theoretical total deformation energy; for example, the preferred value calibrated under the current stable process batch is 4.2%.

[0075] The third execution step is further defined: the processor of the collaborative control system will evaluate the hardware response hysteresis value output in the previous stage. The system is injected into a pre-defined industrial forming simulation model, and dynamic scaling of the safety boundaries in the digital space is performed. Specifically, it actively addresses and retrieves the pre-configured transmission efficiency loss mapping dictionary residing in the memory unit. This static dictionary underwent an offline engineering calibration process: high-frequency bench test logs of hydraulic pump stations of the same model as the current actuator were collected throughout their entire lifecycle. Discrete time points reflecting the degree of valve core interference were extracted. For each discrete time point of hysteresis, the theoretical total mechanical work output of the pump station was extracted based on the pump station's rated operating parameters. Its physical calculation model was then solidified as follows: .in The theoretical total mechanical work output of the pumping station (J); Rated output power (kW) of the servo actuator hydraulic pump; The actual duration (ms) of the valve core action recorded at the current discrete node; This is a dimensional alignment constant. The rated output power of the servo actuator hydraulic pump. The static load setting on the servo pump station's nameplate indicates its preferred configuration is 15kW; the actual valve core operating time... The data is determined by the high-frequency clock sampling node of the control system, with a preferred configuration of 45ms; dimensional alignment constant. The system is statically preset to compensate for carry-over errors from kilowatts to milliseconds to joules, with a preferred configuration of 1. After extracting the theoretical work done, the bench test system synchronously acquires the actual work absorbed by the billet (effective plastic deformation work) by calling the aforementioned dynamic mechanical work dissipation feature extraction logic (using real-time compressive stress and displacement increment for rolling integration), and records the physical difference between the theoretical output total mechanical work of the pump station and the actual work absorbed by the billet at each node. The offline processor performs feature alignment and proportional conversion, dividing the extracted hysteresis time span (10 milliseconds to 150 milliseconds) into multiple discrete hysteresis grade pointers according to a preset equidistant step size (preferably set to 2 milliseconds to 5 milliseconds, for example, set to be divided every 5 milliseconds). At the same time, the actual work absorbed by the billet is divided by the theoretical output total mechanical work of the pump station to calculate the billet deformation work transfer conversion rate (the percentage coefficient of total work to effective work) corresponding to each pointer. Using the hysteresis level pointer as the primary key and the corresponding energy conversion rate coefficient as the mapping value, a structured binding is performed in the relational database to generate a read-only feature mapping set that reflects the binding relationship between mechanical waste and energy loss.

[0076] The hardware response hysteresis evaluation value obtained when executing the adaptive compensation query logic. As the primary key for retrieval, it is injected into the transmission efficiency loss mapping dictionary for addressing. Further smoothing mapping logic is set: if the current hardware response hysteresis evaluation value... If the value falls on a static scale point preset in the transmission efficiency loss mapping dictionary, then the corresponding loss coefficient of work transmission by tube deformation is directly read. If the hardware response hysteresis evaluation value Scattered within the dictionary are two adjacent hierarchical guide cursors and Within the defined interval, the processor extracts the corresponding upper and lower boundary values ​​and calls the following one-dimensional linear interpolation dimensionality reduction algorithm to perform the calculation: .in and Representing the mapping dictionary and and The bound lower and upper limit loss coefficients are both dimensionless constants falling between 0 and 1. The loss coefficient represents the conversion rate of the total mechanical energy output by the servo pump station to the billet and converted into effective plastic deformation under the current hysteretic wear state of the physical equipment. The theoretically calculated value of this loss coefficient fluctuates between 0 and 1; the safe range of this loss coefficient allowed in the actual control logic is limited to between 0.85 and 1. The over-limit circuit breaker monitoring logic runs concurrently during the execution of the above dictionary mapping. Once the processor determines the currently extracted loss coefficient through mapping... A value falling below 0.85 indicates a dangerous underlying anomaly in the system: suggesting physical slippage in the mechanical transmission chain or a large-scale leak in the hydraulic lines. In response to this hardware degradation, the processor immediately abandons subsequent adaptive fine-tuning links, broadcasts a basic protection anomaly interruption to the global control network, urgently stops the servo oil pump, and locks the pressure-holding valve. This is followed by verification of the loss coefficient. Once the value is within the valid range, the processor will apply the loss factor. As a dimensionality reduction penalty multiplier factor, based on Expansion operation; where This is the pre-obtained forming work safety constraint boundary; using this multiplier, the compressive stress threshold and deformation slope limit within the aforementioned offline-calibrated multidimensional tolerance boundary are simultaneously subjected to proportional numerical reduction mapping. Finally, the adaptive work constraint boundary after dynamic shrinkage is calculated and output, denoted as... .

[0077] Furthermore, the processor of the cooperative control system maps the previously generated cumulative work performance index. The transient work rate parameter obtained by its time-domain difference is combined to generate joint observation features. Specifically, in the multi-dimensional feature coordinate system, adaptive work constraint boundaries are extracted. The coordinates of the most stringent lower boundary feature are represented in the middle. The absolute numerical distance between the current cursor point of the joint observation feature and the topological surface of this lower boundary feature is calculated and used as an intermediate observation feature to buffer the current safety margin (the length of the intercept line segment representing the energy spillover in physical space). This serves as the original metric for generating compensation intervention commands in the next stage.

[0078] Real-time extraction of cumulative work mapping indicators The transient work rate parameter during the current control period is denoted as... It characterizes the instantaneous slope of mechanical energy injected into the lattice of the tube blank, reflecting the severity of material deformation after the metal enters the plastic yielding stage. Under the thermodynamic state monitoring based on the forming of stainless steel tube blanks, the effective quantitative monitoring range of this rate of change parameter is mapped within a continuous dynamic domain of 0 to 50 joules per millisecond.

[0079] Load the preset control simulation time window, denoted as Its characterization is the forward-looking perspective period for the system to observe trends and calculate trajectories towards the future time axis. In this embodiment, considering the scheduling frequency and data processing capacity limits of small and medium-sized industrial controllers, the preferred time length of this window is configured to a time domain range of 50 milliseconds to 200 milliseconds (e.g., a setpoint value of 120 milliseconds for complex surface forming). The processor of the collaborative control system performs dynamic trend deduction. The transient work rate parameter... As a time-domain slope traction factor, within a preset control simulation time window Within the span, the current cumulative work mapping index Perform numerical extension calculations based on the prediction period. Its computational logic is as follows: based on the current cumulative work done, a mapping index is generated. Base point, along the transient work rate parameter The indicated energy injection slope is directly plotted as a forward extension vector in the spatiotemporal characteristic coordinate system, thereby generating a predictive work evolution trajectory that characterizes the current billet's short-term work absorption evolution trend, denoted as... .

[0080] Further extract adaptive work constraint boundaries The lower boundary characteristics of the safety red line are represented in the middle, and they are compared with the aforementioned generated predictive work evolution trajectory. Both are placed in a two-dimensional spatiotemporal collision computation domain, and numerical intersection comparison operations are performed. By solving the spatial intersection zero points of the extrapolated trajectory's oblique line features and the lower boundary constraint topological surface, the absolute coordinates of the extrapolated trajectory when it physically penetrates the lower boundary red line are extracted. The processor removes the time dimension attribute of the absolute coordinates and represents them as trajectory intersection timestamps, denoted as . The predicted time when the tube blank's energy absorption trend will precisely break the safety envelope's predicted limit in the future is indicated. The predicted offset of the trajectory intersection timestamp relative to the current system time falls within 10 to 200 milliseconds.

[0081] This embodiment places the predictive work evolution trajectory and the lower boundary features of the adaptive work constraint boundary into a two-dimensional spatiotemporal computation domain, and performs numerical intersection solving of the algebraic equation system. Specifically, it determines the transient work change rate parameter. Assuming the values ​​are greater than zero, use a linear extrapolation model to calculate the timestamps of the trajectory intersection points in the corresponding trajectory intersection states. Its algebraic mapping equation is . The absolute time point of clock synchronization that represents the current system's execution logic deduction is expressed in milliseconds (ms). The threshold value representing the most stringent lower limit of work safety feature extracted within the adaptive work constraint boundary is expressed in joules (J). The current cumulative work mapping index (J) generated from the preceding calculation; This is a parameter representing the transient rate of change of work, reflecting the energy injection slope within the current control cycle, with dimensions in joules per millisecond (J / ms). If If the result is zero, it is determined that there is no trend of evolution beyond the limit, and the intersection solution step is skipped; by performing the algebraic root of the above intersection equation in the internal arithmetic unit, the absolute coordinate point attribute when the evolution trajectory physically penetrates the lower limit red line is extracted.

[0082] The processor then performs a conditional gating decision: if the trajectory intersection timestamp Falling within the control simulation time window Internally, this conditional judgment result sends a rigid physical danger signal to the system, indicating that the current forming trend will exceed the adaptive work constraint boundary within milliseconds. In this warning state, to transform the work warning in virtual space into executable physical intervention instructions, the processor directly relies on the feature coordinate system, since the evolution trajectory intersects at the timestamps... The algebraic difference at a given point theoretically approaches zero. To quantify the absolute deviation of the energy exceeding the limit, the observation node is extended and locked at the end section of a preset extrapolation time window (the absolute time axis) in the spatiotemporal characteristic coordinate system. (Location). The processor extracts the predicted scalar of the extended trajectory under the end section, and performs absolute work difference calculation with the adaptive lower limit boundary features. The calculation equation is as follows: The absolute intercept segment length in physical space is extracted by executing this formula. , representing the expected work overflow value. This overflow value characterizes the overload dissipated energy (unit: joules) that the equipment will illegally inject into the billet at the end of the simulation window if the control architecture does not perform any intervention.

[0083] This embodiment further retrieves the flow-displacement conversion ratio coefficient of the current servo pump station. And use this proportionality coefficient to represent the expected work overflow value expressed by the over-limit deviation amplitude. Perform a reverse mapping solution to extract the physical setback. The algebraic mapping relationship for this reverse derivation is as follows: . The transient fluid compressive stress reference value, measured in megapascals (MPa), is sampled at the absolute time point during the current system simulation. As an application example, the transient fluid compressive stress reference value... The current sampling frame of the pressure-sensitive sensing element is determined by the processor of the collaborative control system polling it in real time. The aforementioned flow-to-displacement conversion ratio, characterizing the servo cylinder extension stroke triggered by a unit volume of hydraulic oil pumped in, has a quantified value range of 0.05 to 0.2 mm / mL. In this embodiment, the preferred value is 0.12 mm / mL. As a specific application example, the flow-to-displacement conversion ratio... The effective cross-sectional area of ​​the hydraulic cylinder of the servo actuator is determined by offline calibration through physical geometric conversion. This is an embedded fixed-dimensional alignment constant, with a theoretical value of 1, used to statically cancel and confirm the zero-level difference conversion between macroscopic units such as Joules, megapascals, and milliliters. By introducing this precise inverse physical mapping formula, the control system reduces and reconstructs the energy domain deviation that is determined to overflow in the virtual digital space into a three-dimensional spatial displacement reduction equivalent that the servo motor can directly execute, ensuring that the issued correction intervention signal has underlying electromechanical executability. This calculation action, at the logic layer, reduces and reconstructs the expected excess mechanical dissipation energy that exceeds the limit into a three-dimensional spatial feed reduction equivalent that the servo can directly execute. This spatial physical deviation equivalent is the over-limit deviation amplitude and is registered as a forming feed correction parameter. Its physical meaning represents the absolute displacement extreme value (e.g., a calculated compensation point value of 0.15 mm) that the servo actuator must retract in the next control cycle in order to pull the rheological state of the metal that is about to become unstable back into a safe physical envelope.

[0084] A fourth execution step is further implemented, where the processor of the collaborative control system constructs a tiered adaptive shunt control logic within its embedded control logic hub. Specifically, a preset basic tolerance displacement limit is retrieved from the global configuration library. This limit characterizes the maximum physical extreme value of the servo cylinder's corrected feed allowed within a single control cycle without triggering a sudden stress-induced fracture in the mold cavity. For the mechanical calibration of medium-sized tubular components, this basic tolerance displacement limit is locked within an operating range of 0.1 mm to 0.3 mm (e.g., a safety point of 0.2 mm is set under the current mold operating conditions). During the judgment process, the forming feed correction parameters extracted in real time are... A bidirectional numerical comparison is performed with the aforementioned basic tolerance displacement limit. If the comparison shows that the forming feed correction parameter is less than the basic tolerance displacement limit, it indicates that the current physical yield is still within the plastic absorption range of the mold, and the system logic gating initially locks the target intervention mode to the flexible deceleration compensation mode. However, if the comparison result triggers the over-limit flag (the required compensation distance is greater than the basic tolerance displacement limit), it is determined that blindly executing large displacement compensation leads to secondary overload collisions, thus directly blocking the execution of conventional control and switching the signal reconstruction target to the constant pressure safety interception mode.

[0085] After performing the initial gating comparison based on spatial displacement limits, the processor of the collaborative control system further activates the cross-dimensional joint decision logic branch. Specifically, for the initially locked intervention mode, it starts from the currently resident real-time hydraulic state sequence. The current transient pressure extreme value is extracted simultaneously.

[0086] Perform dimensionless dimensionality reduction transformation based on physical extremum calibration. Specifically, extract the forming feed correction parameters. And the highest transient pressure value in the real-time hydraulic state sequence at the current time point. Based on preset feature alignment and normalization scales, a cross-dimensional multivariate logical weighted aggregation operation is performed. Its core mapping logic is as follows: .in The absolute ultimate tensile strength scalar value (such as 505 MPa, which is preferred for conventional 304 stainless steel) extracted by pre-existing destructive mechanical tensile testing on this batch of tube blanks is used as the denominator for dimensionality reduction. The extreme value of the basic tolerance displacement limit that can be characterized within a single control cycle of the mold cavity (its preferred value is 0.2 mm). and A pre-defined attention allocation mask, serving as a dimensionless static feature weight constant, is bound to the pressure and displacement dimensions, respectively. In this embodiment, based on discrete fracture attribution statistics, these values ​​are preferably fixed to 0.65 and 0.35, respectively. The feature fusion mechanism specifically works as follows: relying on the feature alignment matrix, the heterogeneous "spatial intervention deviation vector" and "transient hydraulic stress scalar" are mapped to a unified benchmark measurement scale. Then, the pre-defined attention allocation mask is concatenated to perform nonlinear tensor dimension fusion, ultimately calculating and outputting a local stress instability assessment value, denoted as... This assessment value characterizes the severity of local stress distortion induced by the sudden obstruction of metal rheology caused by the sudden encounter with material hardening nodules under the strong constraint of the mold cavity in a local area of ​​the billet. The effective measurement range of this local stress instability assessment value is limited to a dimensionless continuous value between 0 and 1.

[0087] In this embodiment, the pre-defined attention allocation mask is presented in a data structure as a one-dimensional static weighted array containing two feature channels. The specific weight parameters within this mask are derived by the offline diagnostic component of the control system based on historical failure logs. By batch traversing all historical processing cycle data marked as fracture failures in the database, and for the feature stream 100 milliseconds before each failure, a causal attribution calculation based on state covariance eigenvalue decomposition is performed. The specific logic execution process is as follows: the diagnostic component extracts all transient pressure fluctuation data and spatial displacement fluctuation data within the 100-millisecond prediction window, constructs a two-dimensional state covariance matrix in the solution space; by performing standardized eigenvalue decomposition on this matrix, the absolute values ​​of the eigenvalues ​​mapped to the pressure dimension and displacement dimension are extracted respectively. Finally, the eigenvalue values ​​of the pressure dimension are divided by the sum of the total eigenvalue values ​​of the two dimensions to calculate the percentage contribution of pressure distortion to the variance of the final fracture result. In all material bursting accidents caused by localized cold hardening nodules, the contribution of cracking variance caused by transient pressure distortion remained stable at around 65%, while the contribution of cracking caused solely by microscopic spatial displacement accounted for only about 35%. Based on this failure attribution law, the processor internally asymmetrically solidifies this mask parameter: the feature weight of the spatial intervention dimension is anchored at 0.35, while the feature weight of the transient pressure dimension is inversely increased to 0.65.

[0088] In this embodiment, the embedded control logic center synchronously reads the safety damping threshold from the system's global configuration table, denoted as... .

[0089] This safety damping threshold This characterizes the dimensionless probabilistic danger boundary that the control architecture of the metal lattice can tolerate before tearing occurs. The absolute ultimate tensile strength scalar (with a specific megapascal physical dimension; for example, for standard 304 austenitic stainless steel billets, the preferred physical test point value for this absolute ultimate tensile strength scalar at the time of physical fracture is collected using a destructive tensile mechanics platform for this batch of billets.) is then established using a normalized ratio mapping channel with the real-time collected deformation stress characteristics as the numerator and the aforementioned destructive absolute ultimate scalar as the denominator. Through this division-based dimensionality reduction mechanism, the specific megapascal values ​​originally discretely distributed across various models are uniformly compressed into a severity coefficient spanning 0 to 1. After calibration and fitting of the pre-fracture data, the feature mapping range approaching the lattice fracture critical point is constantly anchored within this dimensionless high-risk interval of 0.75 to 0.85 (the gate trigger point value for thick-walled 304 steel billets is 0.78).

[0090] The processor executes conditional interception and shunt control logic based on the above parameters: if the logic comparison result shows a local stress instability assessment value... Less than the pre-configured safety damping threshold If the current rheology of the billet is still within the controllable plastic range, the target intervention mode is locked to the conventional flexible deceleration compensation mode, which only superimposes a smooth deceleration slope on the original control curve (for example, issuing a dynamic negative feedback compensation command to the servo controller to reduce the current feed rate by 10% to 25%, or setting a deceleration parameter of -0.05 mm / m²).

[0091] Response to local stress instability assessment value Greater than or equal to the safety damping threshold When the stress distortion exceeds the limit, the processor determines that the current billet has entered a high-risk state of rapid stress concentration. At this time, the embedded control logic center directly triggers the asymmetric defense mechanism, switching the target intervention mode to the constant pressure safety interception mode. When the above-mentioned edge condition interception mechanism is triggered, the control architecture takes over. The processor reconstructs its core scheduling module to focus on "constant pressure relief protection logic" for fluid discharge. According to the underlying standard of industrial communication protocol, a dual-track control word is assembled and generated in the control logic: its first track is overwritten with a safety deceleration instruction with an absolute value of zero.

[0092] Its second orbit is overwritten with the explosion-proof pressure relief threshold parameter, and its characteristic solution equation is configured as follows: .in The absolute ultimate tensile strength scalar value of the tube blank batch extracted through the aforementioned mechanical platform (e.g., 505 MPa). The comprehensive conversion mapping constant from the pre-configured mold cavity material stress to the cavity fluid pressure is obtained by static calibration of the mold geometric curvature and rheological contact area. Based on the fluid contact characteristics of the standard stainless steel bellows mold, the value range of this conversion mapping constant is constrained to between 0.015 and 0.025 (preferably 0.025 in this embodiment). The fixed derating safety factor (preferred value locked at 0.95); This refers to the static anti-fall hydrostatic pressure reference (MPa) generated by the weight of the heavy upper slide block and servo cylinder in the fluid. In this embodiment, the static anti-fall hydrostatic pressure reference is constructed based on the laws of physical and mechanical equilibrium, and its static solution model is fixed as follows: .in The physical mass (kg) of the heavy-duty upper slide of the mold; denoted as the physical mass (kg) of the servo hydraulic cylinder body; g is the standard gravitational acceleration constant (9.8 m / s²). This represents the effective cross-sectional area of ​​the previously loaded servo hydraulic cylinder. As an application example, the static anti-fall hydrostatic pressure reference... The parameter configuration is determined offline and written into the control system parameter library based on static physical calculations using mold drawings and gravity parameters before the equipment leaves the factory. The preferred configuration is 3.5 MPa. After the processor performs the above algebraic extraction, it will calculate and output... As a parameter for explosion-proof pressure relief threshold.

[0093] The effective quantization command range for the explosion-proof pressure relief threshold parameter is within a constant low-pressure safety range of 5 to 10 MPa (for example, the calculated value for the current model is 8.5 MPa).

[0094] The control data combination containing the above dual-track control word is mapped to adaptive control signaling, denoted as... The current diversion control architecture, consisting of the flexible speed reduction compensation mode and the constant pressure safety interception mode, is the dynamic damping compensation strategy executed by the system.

[0095] Through the above adaptive control signaling In the analysis and execution, the servo drive uses the safety deceleration command of the first track to overwrite the original displacement increment parameters to lock the physical propulsion, and simultaneously uses the explosion-proof pressure relief threshold parameter of the second track to reset the opening reference control signal of the proportional relief valve. In this way, the substantial reconstruction and safety update of the mapping relationship between the feed and pressure dual closed-loop control parameters of the servo actuator is completed at the control logic layer, so as to block the continuous injection of excessive mechanical work.

[0096] Furthermore, in the solid forming test of tube blanks made of thick-walled 304 austenitic stainless steel with latent cold work hardening defects, the rheological resistance of the metal lattice inside the mold cavity did not follow a linear evolution law. A pressure-sensitive sensor network and a high-resolution servo grating ruler captured the heterogeneous multi-source physical abrupt changes during the hydraulic cylinder propulsion process at high frequencies under continuous millisecond-level clock ticks. Quantitative calculations and asymmetric control interception targeting hidden fracture risks were performed, as detailed in Table 1 below.

[0097] Table 1: Evolution of Local Stress Instability Assessment and Asymmetric Interception Execution under Cold-hardened Nodule Intrusion Condition

[0098] Control time slice t / ms Transient fluid compressive stress reference value Expected work overflow value Forming feed correction parameters Local stress instability assessment value Targeted intervention model Explosion-proof pressure relief threshold parameters 100 22.5 9.37 0.05 0.37 Flexible speed reduction compensation mode / 110 24.8 24.8 0.12 0.52 Flexible speed reduction compensation mode / 120 27.1 40.65 0.18 0.66 Flexible speed reduction compensation mode / 130 29.5 61.45 0.25 0.81 Constant pressure safety interception mode 8.5

[0099] As shown in Table 1, In the initial stage of the process, the billet material is subjected to uniform stress, and the transient fluid compressive stress reference value is at a normal level of 22.5 MPa. The processor is based on the dimensional constant. With preset flow-displacement conversion ratio (Set point value is 0.12 mm / mL), perform inverse spatial mapping calculation on the energy overflow value of 9.37 J to obtain the forming feed correction parameters. The value is 0.05 mm. Further, the ultimate tensile strength scalar value of this batch of steel was retrieved. 505 MPa and the extreme value of the basic tolerance displacement limit 0.2mm, combined with fixed weight and Perform cross-dimensional normalized weighted operations to obtain It is only 0.37. This value is less than the preset safety damping threshold of 0.78. The original flexible deceleration compensation mode was maintained, and conventional negative feedback deceleration commands were sent to the servo queue. As the time axis smoothly advanced to 130ms, the mold slider reached the hardened nodule region inside the tube blank, and the rheological resistance surged. The sensor captured that the transient fluid pressure stress reference value soared to 29.5MPa within 10 milliseconds. At this time, the reverse-extracted yield feed extreme value was obtained. Reaching 0.25mm, this physically surpasses the 0.2mm limit for single mold movement. The processor, using the same algebraic weighted logic, calculates the current... The value climbed to 0.81. Since this value directly exceeded the safety damping threshold of 0.78, it was determined that the metallic material had entered a critical fracture state of rapid stress concentration. The control center cut off the basic speed regulation network, instantly switching the target intervention mode to constant pressure safety interception mode.

[0100] The processor of the cooperative control system also includes a dynamic closed-loop update step that performs a pre-processor algorithm reconstruction on the input link. A pre-embedded dynamic filter configuration library is introduced, denoted as... The configuration library is constructed as a static digital index table, which pre-establishes a mapping relationship between various tiers of hardware response hysteresis and high-frequency parasitic noise shielding rules for different frequency bands. This static digital index table undergoes spectrum observation calibration. Calibration personnel pre-collect transient hydraulic impact waveforms of the valve core at various wear stages, from brand new to severely jammed, using high-frequency vibration sensors during destructive testing of the hydraulic pump station's entire life cycle. Subsequently, Fast Fourier Transform (FFT) is performed on the waveforms collected at each stage to extract the main frequency band drift pattern of spurious resonant noise induced by increased mechanical play. A mapping rule for synchronously pressing down the hysteresis extrema and cutoff frequency in opposite directions is constructed at the logic configuration layer, and this structured key-value pair array is injected in batches into the gateway's local configuration library, thereby completing the loading of the dynamic filtering configuration library and enabling it to be directly applied to online addressing for anti-oscillation filtering.

[0101] This embodiment verifies the interference of physical wear degradation of the servo actuator on the data acquisition link under long-term continuous operation, as well as the effectiveness of the system's adaptive noise immunity. Utilizing a servo hydraulic pump station with a rated output power of 15kW and its supporting OPCUA industrial communication data gateway, mechanical response data at different wear stages are collected during a continuous reciprocating forming stamping test cycle. The processor of the collaborative control system continuously captures the hardware state evolution characteristics representing the aging of the physical equipment by periodically extracting the timestamps of feed control command issuance and corresponding status feedback confirmation timestamps, and associates the hardware hysteresis scalar with the execution of front-end filtering and interception rules.

[0102] Table 2: Comparison Table of Hardware Response Hysteresis Characteristics and Adaptive Update of Dynamic Filtering Rules Across Service Cycles

[0103] Hardware response hysteresis evaluation value / ms Logical information deviation / ms Hardware aging step numerical range Transmission efficiency loss coefficient γ Target low-pass filter frequency parameter / Hz System response status 15 2.1 Excellent normal range (10-40) 0.99 500 Maintain routine monitoring across the entire frequency band 55 18.4 Mild interference range (41-90) 0.94 300 Intercepting mid-frequency parasitic oscillations 125 42.6 Severe hysteresis range (91-150) 0.86 150 Forced down-cutoff frequency to filter out low-frequency resonance 155 68.3 Circuit breaker trigger range (>150) 0.84 / Triggering an abnormal emergency stop and locking the pressure-holding valve

[0104] As shown in Table 2, during the initial stamping cycle after equipment commissioning, the hydraulic cylinder seals and guide rails were in a brand new, undamaged state, and the obtained hardware response hysteresis evaluation values ​​were... The process takes only 15ms. The processor injects this value into a pre-configured discrete logic comparison array, locks it into the "optimal normal range," and extracts a transmission efficiency loss coefficient of 0.99 from a pre-configured transmission efficiency loss mapping dictionary. The processor then configures the dynamic filtering library. The query retrieves the matching 500Hz wideband target low-pass filter frequency parameters. This ensures the lossless penetration of high-frequency pressure transient signals. With the continuous accumulation of stamping cycles, metal debris inevitably accumulates in the servo main hydraulic circuit, accompanied by oil emulsification. When the hardware response hysteresis evaluation value... When the scalar value climbed to 125ms, the valve core opening process was already accompanied by low-frequency mechanical resonance. This scalar was determined to fall into the "severe hysteresis region," and the loss coefficient was calculated using a one-dimensional linear interpolation dimensionality reduction algorithm. It has decayed to 0.86. To prevent false low-frequency oscillations from contaminating the work integral network due to increased mechanical play, the control center extracts the interval query mapping identifier from the dynamic filter configuration library. Extract the 150Hz downgrade target low-pass filter frequency parameters. This data is then directly sent to the industrial control data gateway. Upon receiving this value, the gateway immediately overwrites the low-pass filter interception rules for the real-time hydraulic state sequence, cutting off parasitic frequency noise during periods of severe wear. Mechanical wear breaches the 150ms tolerance threshold, and the corresponding acquired loss coefficient is detected. When the value dropped to 0.84, which was below the preset safety threshold of 0.85, the processor immediately cut off all closed-loop filtering and adaptive fine-tuning link calculations, and directly broadcast a hardware physical slippage anomaly alarm to the global gateway. The hydraulic pump station was then brought to an emergency stop to prevent a serious accident.

[0105] Furthermore, the control center pre-sets a stepped range of hardware aging values ​​(10-40 milliseconds is defined as the optimal normal range, 41-90 milliseconds as the mild interference range, and 91-150 milliseconds as the severe hysteresis range). When the hardware response hysteresis assessment value... When a scalar value flows in, the processor performs a logical comparison to confirm that the instantaneous value falls precisely within the hardware aging ladder value range. The processor extracts the identifier of this locked range and converts it into a mapping query identifier, which is then used in the dynamic filter configuration library. Adaptive comparison addressing is performed in the static mapping layer. Based on this discretized pointer mapping, the target low-pass filter frequency parameters that match the configuration library can be stably extracted. This effectively prevents configuration disorder caused by high-frequency parameter jitter. Under dynamic scheduling logic, the target low-pass filter frequency parameter and hysteresis characteristics show an inverse mapping trend: under normal wear-free conditions, its quantized value is configured as a wideband 500Hz; while when the extracted hardware response hysteresis evaluation value... When the wear and tear on the equipment intensifies and the mechanical play increases, the target low-pass filter frequency parameter will be adaptively suppressed and shifted down to a frequency range of 150Hz or even lower (e.g., for mid-to-late stage wear, the extraction point value is 120Hz). This extracted target low-pass filter frequency parameter is then sent to the gateway layer. The numerical values ​​are overwritten and dynamically updated in real time on the hardware acquisition side, and the low-pass filtering interception rules for the real-time hydraulic state sequence are blocked from entering the core integral calculation network due to false high-frequency jitter caused by mechanical aging at the source of data intake.

[0106] Figure 1 The entity is a servo hydraulic forming equipment. Its pipeline pressure sensor and guide rail linear grating ruler correspond to the data acquisition action of "obtaining the real-time hydraulic status sequence and displacement feed sequence of the current expansion joint billet from the industrial control data gateway of the manufacturing execution system, and parsing the command response timestamp sequence" (corresponding to multi-source state perception and timestamp extraction). In the data panel floating above the cabinet on the right side of the screen, the inward contraction of the safety constraint envelope and the soaring work trajectory are presented, intuitively mapping the "extraction of hardware response hysteresis evaluation value from the timestamp sequence; extraction of cumulative work mapping index from the hydraulic and displacement sequences" (corresponding to hardware hysteresis and cumulative work calculation), and the calculation logic of "injecting the hardware hysteresis evaluation value into the industrial forming simulation model, dynamically scaling its built-in safety constraint boundary, and inputting the cumulative work index into the adaptive work constraint boundary to perform over-limit state evaluation" (corresponding to constraint boundary adaptive scaling and evaluation). The dashed line extending from the cabinet to the bottom of the open proportional overflow valve corresponds to the terminal physical intervention action of "extracting forming feed correction parameters, generating adaptive control signaling containing dynamic damping compensation strategy, sending it to the servo drive logic queue, and executing explosion-proof leakage to block over-limit work and retain anti-fall static pressure" (corresponding to damping signaling reconstruction and explosion-proof interception).

[0107] An expansion joint integral forming thickness coordination control system, the system being used to execute the expansion joint integral forming thickness coordination control method, comprising:

[0108] The data perception and analysis module is configured to obtain the real-time hydraulic status sequence and displacement feed sequence of the current expansion joint billet from the industrial control data gateway of the manufacturing execution system, and to parse the forming control log to obtain the instruction response timestamp sequence of the feed control instruction issuance timestamp and status feedback confirmation timestamp.

[0109] The dual-track feature calculation module is configured to perform periodic communication response delay pattern extraction on the instruction response timestamp sequence, and extract a hardware response hysteresis evaluation value that characterizes the current hardware transmission resistance level; at the same time, it performs dynamic mechanical power dissipation feature extraction on the real-time hydraulic state sequence and displacement feed sequence, and generates a cumulative work mapping index that characterizes the amount of mechanical deformation dissipation work that the current expansion joint tube blank has borne.

[0110] The constraint boundary dynamic scaling and over-limit assessment module is configured to inject the hardware response hysteresis assessment value into a preset industrial forming simulation model, dynamically map and scale its built-in preset forming work safety constraint boundary to generate an adaptive work constraint boundary dynamically adjusted based on the equipment aging state; and input the cumulative work mapping index into the adaptive work constraint boundary to perform over-limit state assessment, and extract forming feed correction parameters used to compensate for work over-limit deviation.

[0111] The signaling reconstruction and dual-loop update module is configured to generate adaptive control signaling containing a dynamic damping compensation strategy within the embedded control logic center of the processor based on the forming feed correction parameters, and to direct the adaptive control signaling to the corresponding servo drive logic queue, so as to update the mapping relationship of the feed and pressure dual-loop control parameters of the servo actuator at the control logic layer.

[0112] It should be noted that, in order to decouple the core algorithm of this invention from the specific application strategy, and to ensure the configurability and ease of debugging of the technical solution, in the specific implementation path of this invention, all configurable operating parameters are read through a standardized "configuration interface". The data source of this configuration interface is a "data storage module" (e.g., a non-transitory computer-readable storage medium, such as a configuration file, database entry, or cloud configuration service), which is configured to store configuration data in key-value pair format.

[0113] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for coordinated control of the thickness of an expansion joint during integral forming, characterized in that, Executed by the processor of the collaborative control system, the specific steps include: The real-time hydraulic status sequence and displacement feed sequence of the current expansion joint tube blank are obtained from the industrial control data gateway of the manufacturing execution system, and the instruction response timestamp sequence of feed control command issuance timestamp and status feedback confirmation timestamp is obtained by parsing the forming control log. The communication response delay pattern is extracted periodically from the instruction response timestamp sequence to extract the hardware response hysteresis evaluation value that characterizes the current hardware transmission resistance. At the same time, the dynamic mechanical power dissipation feature is extracted from the real-time hydraulic state sequence and displacement feed sequence to generate a cumulative work mapping index that characterizes the amount of mechanical deformation dissipation work that the current expansion joint tube blank has undergone. The hardware response hysteresis evaluation value is injected into the preset industrial forming simulation model, and its built-in preset forming work safety constraint boundary is dynamically mapped and scaled to generate an adaptive work constraint boundary that is dynamically adjusted based on the equipment aging state; the cumulative work mapping index is input into the adaptive work constraint boundary to perform over-limit state evaluation, and forming feed correction parameters used to compensate for work over-limit deviation are extracted. Based on the forming feed correction parameters, an adaptive control signaling containing a dynamic damping compensation strategy is generated in the processor's embedded control logic center, and the adaptive control signaling is directed to the corresponding servo drive logic queue to update the mapping relationship of the feed and pressure dual closed-loop control parameters of the servo actuator at the control logic layer.

2. The method for coordinated control of the integral forming thickness of an expansion joint according to claim 1, characterized in that: Periodic communication response delay patterns are extracted from the instruction response timestamp sequence, and hardware response hysteresis evaluation values ​​characterizing the current degree of hardware transmission hindrance are extracted, including: Extract the feed control command issuance timestamp and the corresponding status feedback confirmation timestamp from the command response timestamp sequence, and calculate the single execution response time difference set; The timing deviation tracking algorithm is applied to the set of single execution response time differences to remove discrete interference components that characterize random communication jitter data, and a smooth hysteresis baseline feature that characterizes the wear and decay trend of the device over time is fitted and output. The current timing value of the smooth hysteresis baseline feature is used as the hardware response hysteresis evaluation value.

3. The method for coordinated control of the integral forming thickness of an expansion joint according to claim 1, characterized in that: Dynamic mechanical power dissipation feature extraction is performed on the real-time hydraulic state sequence and displacement feed sequence to generate a cumulative work mapping index characterizing the amount of work dissipated by the mechanical deformation of the current expansion joint billet, including: The real-time hydraulic state sequence and displacement feed sequence are aligned with the spatiotemporal dimension features based on the system global clock synchronization tag of the industrial control bus, and the transient compressive stress feature scalar and single-step displacement increment parameter mapped under the unified time reference are extracted respectively. Based on the preset material mechanics mapping rules of the current tube blank, the effective cross-sectional area of ​​the servo hydraulic cylinder and the cross-dimensional mapping constant are introduced. The transient compressive stress characteristic scalar and the single-step displacement increment parameter are subjected to cross-dimensional incremental product operation with dimension alignment as the benchmark. Then, the results of the incremental product operation are rolled and accumulated along the time axis in the current entire processing cycle. The accumulated total value is introduced with an elastic loss factor representing the preset tube blank elastic deformation redundancy for system-level multiplication filtering and stripping. Finally, the accumulated work mapping index is output.

4. The method for coordinated control of the integral forming thickness of an expansion joint according to claim 1, characterized in that: The hardware response hysteresis evaluation value is injected into a pre-set industrial forming simulation model, and its built-in preset forming work safety constraint boundary is dynamically mapped and scaled to generate an adaptive work constraint boundary that is dynamically adjusted based on the equipment aging state, including: The hardware response hysteresis evaluation value is input into a pre-configured transmission efficiency loss mapping dictionary. Through feature matching query, the loss coefficient characterizing the current tube blank deformation work transmission efficiency is output. Using the loss coefficient of the current billet deformation work transfer efficiency as a multiplier factor, the multidimensional tolerance boundary of the preset forming work safety constraint boundary is proportionally reduced and mapped to generate the dynamically shrunken adaptive work constraint boundary.

5. The method for coordinated control of the integral forming thickness of an expansion joint according to claim 4, characterized in that: The cumulative work mapping index is input into the adaptive work constraint boundary to perform an over-limit state evaluation, and the forming feed correction parameters used to compensate for the work over-limit deviation are extracted, including: Extract the transient work change rate parameter of the cumulative work mapping index in the current control cycle in real time; The cumulative work mapping index and its transient work change rate parameter are compared with the lower limit boundary characteristics of the adaptive work constraint boundary by performing logical deviation quantification. If the logic deviation quantification result indicates that the cumulative work mapping index breaks through the adaptive work constraint boundary, then the over-limit deviation amplitude of the cumulative work mapping index exceeding the lower limit boundary characteristic is calculated, and the current transient fluid pressure stress reference value is obtained simultaneously. Combining the preset flow-displacement conversion characteristics and dimension equivalence constant, the spatial dimension inverse mapping solution is performed on the over-limit deviation amplitude, and the physical yield equivalent is extracted as the forming feed correction parameter.

6. The method for coordinated control of the integral forming thickness of an expansion joint according to claim 1, characterized in that: Based on the forming feed correction parameters, an adaptive control signaling system incorporating a dynamic damping compensation strategy is generated within the processor's embedded control logic core, including: Based on the forming feed correction parameters, an evaluation index is constructed in the feature space to characterize the current billet instability state, and the target intervention mode is locked as flexible deceleration compensation mode or constant pressure safety interception mode according to the evaluation index. If it is the constant pressure safety interception mode, then based on the basic scheduling strategy of the forming controller, the default position closed-loop control logic is reconstructed into constant pressure relief protection logic, and an adaptive control signaling containing safety deceleration command and explosion-proof pressure relief threshold parameters is organized and constructed.

7. The method for coordinated control of the integral forming thickness of an expansion joint according to claims 1 to 6, characterized in that: The method also includes a dynamic closed-loop update step: Establish a mapping relationship between the hardware response hysteresis evaluation value and the preset dynamic filtering configuration library inside the industrial control data gateway; Based on the hardware aging ladder value range of the currently obtained hardware response hysteresis evaluation value, the corresponding range identifier is extracted and converted into a query mapping identifier. Based on the query mapping identifier, the dynamic filter configuration library is adaptively retrieved to extract the matching target low-pass filter frequency parameter. The filtering and interception rules for hydraulic state sequences in the industrial control data gateway are updated in real time using the target low-pass filter frequency parameters.

8. The method for coordinated control of the integral forming thickness of an expansion joint according to claim 5, characterized in that: The logical deviation quantification calculation is performed between the cumulative work mapping index and its transient work change rate parameter and the lower boundary feature of the adaptive work constraint boundary, including: Based on the transient work change rate parameter, within a preset control simulation time window, the cumulative work mapping index is subjected to a numerical extension calculation based on the prediction period to generate a predictive work evolution trajectory that characterizes the future work absorption trend of the current billet. Perform a numerical intersection comparison operation between the predictive work evolution trajectory and the lower limit boundary features of the adaptive work constraint boundary, and extract the trajectory intersection point timestamps in the corresponding trajectory intersection states; If the timestamp of the trajectory intersection falls within the control simulation time window, it is determined that the current forming trend exceeds the adaptive work constraint boundary, and the observation node is locked at the end section of the control simulation time window. The absolute work difference between the extended trajectory prediction scalar and the adaptive lower limit boundary feature under the end section is extracted as the expected work overflow value and output as the logic deviation quantization calculation result.

9. The method for coordinated control of the integral forming thickness of an expansion joint according to claim 6, characterized in that: Based on the forming feed correction parameters, an evaluation index is constructed in the feature space to characterize the current billet instability state. According to the evaluation index, the target intervention mode is locked as either a flexible deceleration compensation mode or a constant pressure safety interception mode, including: The extracted forming feed correction parameters and the current transient compressive stress characteristic scalar in the real-time hydraulic state sequence are subjected to multivariate logical weighted aggregation operation to calculate and output the local stress instability assessment value used to characterize the degree of local stress distortion of the current billet. If the local stress instability assessment value does not reach the pre-configured safety damping threshold, the target intervention mode will be locked as the flexible deceleration compensation mode. In response to the stress distortion exceeding the limit state where the local stress instability assessment value is greater than or equal to the safety damping threshold, it is determined that the current billet has entered a dangerous state of rapid material stress concentration. The target intervention mode is directly switched to the constant pressure safety interception mode in the embedded control logic center to block the continuous injection of excessive mechanical work.

10. A thickness coordination control system for integral forming of expansion joints, characterized in that: The system is used to execute the expansion joint integral forming thickness collaborative control method according to any one of claims 1-9, including: The data perception and analysis module is configured to obtain the real-time hydraulic status sequence and displacement feed sequence of the current expansion joint billet from the industrial control data gateway of the manufacturing execution system, and to parse the forming control log to obtain the instruction response timestamp sequence of the feed control instruction issuance timestamp and status feedback confirmation timestamp. The dual-track feature calculation module is configured to perform periodic communication response delay pattern extraction on the instruction response timestamp sequence, and extract a hardware response hysteresis evaluation value that characterizes the current hardware transmission resistance level; at the same time, it performs dynamic mechanical power dissipation feature extraction on the real-time hydraulic state sequence and displacement feed sequence, and generates a cumulative work mapping index that characterizes the amount of mechanical deformation dissipation work that the current expansion joint tube blank has borne. The constraint boundary dynamic scaling and over-limit assessment module is configured to inject the hardware response hysteresis assessment value into a preset industrial forming simulation model, dynamically map and scale its built-in preset forming work safety constraint boundary to generate an adaptive work constraint boundary dynamically adjusted based on the equipment aging state; and input the cumulative work mapping index into the adaptive work constraint boundary to perform over-limit state assessment, and extract forming feed correction parameters used to compensate for work over-limit deviation. The signaling reconstruction and dual-loop update module is configured to generate adaptive control signaling containing a dynamic damping compensation strategy within the embedded control logic center of the processor based on the forming feed correction parameters, and to direct the adaptive control signaling to the corresponding servo drive logic queue, so as to update the mapping relationship of the feed and pressure dual-loop control parameters of the servo actuator at the control logic layer.

Images