Self-adaptive cooperative control method for mattress continuous manufacturing production line

Abstract

The invention relates to the technical field of industrial automatic control, and discloses a self-adaptive cooperative control method for a mattress continuous manufacturing production line, and the method comprises the steps: collecting a drive current signal of a drive frequency actuator and speed feedback data, and carrying out the separation of a torque deviation component representing the deformation feedback of a controlled material through a dynamic load observation model; detecting an algebraic symbol direction and a change gradient of the torque deviation component, and converting the algebraic symbol direction and the change gradient into an additional load impedance value of a control loop; according to the method, active hedging of nonlinear deformation stress is achieved through digital load mapping, response lag caused by viscoelasticity of the soft materials is eliminated, the problem of dynamic instability in multi-station conveying is solved, and the method has the advantages of being simple in structure, convenient to operate and high in practicability. And the speed collaboration among the execution units of the production line is ensured to be in a physical deterministic interval.

Description

Technical Field

[0001] This invention relates to an adaptive collaborative control method for a continuous mattress manufacturing line, belonging to the field of industrial automatic control technology. Background Technology

[0002] Current continuous manufacturing lines for mattresses involve multiple physical stations, such as foam and spring assembly and fabric lamination. The synchronization and coordination between the execution units at each station is fundamental to ensuring manufacturing efficiency. Currently, industrial control systems generally adopt proportional control based on position synchronization, which is based on the assumption that the materials are rigidly connected. It achieves the alignment of the cycle time between each process through preset fixed parameters. The soft materials involved in mattress manufacturing have viscoelastic characteristics. During long-distance transportation, elastic potential energy is accumulated and released asynchronously. Due to this characteristic, the acceleration and deceleration actions of the preceding stations are transmitted within the material in the form of pressure waves, resulting in a response that lags behind the control commands. This leads to transient speed differences between the execution units, which in turn generate internal stress within the material, inducing defects such as wrinkles and uneven dimensional rebound in the finished product.

[0003] To address these issues, the industry has attempted to increase sampling frequency or add tension sensors. However, these methods face technical limitations due to the susceptibility of soft materials to pressure damage and slippage. Furthermore, high-frequency feedback signals can induce self-excited oscillations in the control loop under complex material stress fluctuations. The dynamic accumulation of material potential energy alters the equivalent physical impedance of the controlled object, and the fixed parameters of the existing control framework cannot detect or absorb loop characteristic drift caused by changes in material properties. Even with improvements to the material space alignment and correction actuators, the existing control logic focuses on static geometric position repair, neglecting the impact of internal stress evolution during the conveying process on the system. The impact on stability, for example, Chinese invention patent with publication number CN117383327A discloses a belt viscoelastic material conveying correction device and correction method, which achieves centering correction through multi-axis mechanical linkage. The adjustment logic is based on the hysteresis compensation of geometric position deviation. It only executes mechanical action after identifying that the material deviates from the center line. It cannot sense the real-time torque fluctuation caused by acceleration and deceleration inside the soft material. In the dynamic environment of multi-station high-frequency linkage, it is difficult to eliminate the mechanical friction interference of the actuator. It cannot actively offset the accumulation of potential energy and nonlinear response of material conveying. The control accuracy drifts with the increase of running time.

[0004] Therefore, the technical problem to be solved by this invention is how to realize the nonlinear reconstruction of the control gain based on the potential energy evolution law in the material conveying process, and to establish a physical verification mechanism that can remove the influence of mechanical friction, thereby eliminating the influence of elastic hysteresis on system stability. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: An adaptive collaborative control method for a continuous mattress manufacturing line, comprising the following steps: Step S1: Real-time acquisition of drive current signals and real-time speed feedback data of the drive frequency actuator in the continuous mattress manufacturing line; Step S2: Establish a dynamic load observation model for the drive frequency actuator, and based on the dynamic load observation model and real-time speed feedback data, calculate the torque deviation component characterizing the deformation feedback of the controlled material from the drive current signal. Step S2 includes: extracting frequency characteristic reference quantities based on real-time speed feedback data, separating mechanical loss current and deformation feedback current characterizing internal stress changes in materials from the drive current signal according to the frequency characteristic reference quantities, and confirming the deformation feedback current as the torque deviation component. ; Step S3: Real-time detection of torque deviation component The algebraic sign direction and gradient of change, and the torque deviation component. The additional load impedance value is converted into the control loop of the drive frequency actuator to quantitatively characterize the measured coupling strength between the controlled material and the conveying mechanism. Step S4, based on the torque deviation component The algebraic sign direction, gradient change, and additional load impedance value are used to dynamically adjust the speed loop response gain of the drive frequency actuator and inject speed advance compensation to offset the response lag caused by the deformation of the controlled material, ensuring that the conveying speed difference between each drive frequency actuator is stable within the preset physical deterministic range.

[0006] Preferably, the speed loop response gain includes a proportional adjustment coefficient; the dynamic adjustment of the speed loop response gain and the synchronous injection of speed lead compensation in step S4 includes: when a torque deviation component is detected... When the value is positive and the gradient is increasing, the proportional adjustment coefficient is simultaneously reduced while the injection speed advances the compensation amount; when a torque deviation component is detected... When the value is positive and the gradient of change decreases, the proportional adjustment coefficient is gradually restored and the speed advance compensation amount is maintained.

[0007] Preferably, the method further includes a damping adaptive step for deceleration conditions: Step S301, real-time monitoring of the residual value of the speed advance compensation, and calculation of the predicted value of kinetic energy release of the controlled material in combination with the real-time deceleration slope; Step S302, injection of a dynamic damping factor into the speed loop of the drive frequency actuator based on the predicted value of kinetic energy release. This increases the virtual rotational inertia of the drive frequency actuator and counteracts the physical rebound impact generated by the controlled material during the deceleration phase.

[0008] Preferably, in step S302, a dynamic damping factor is injected. The computational logic follows these rules: ,in, For torque deviation component, This represents the real-time deviation value of the speed advance compensation. These are the pre-defined weighting coefficients for the first-order differential term. These are the pre-defined weighting coefficients for the integral term.

[0009] Preferably, the method further includes an online monitoring step for the conveying coupling state: step S501, injecting a small torque disturbance signal with a frequency of 50Hz to 100Hz into the drive frequency actuator; step S502, analyzing the sensing response component of the drive frequency actuator of the downstream adjacent station to the small torque disturbance signal; step S503, when the sensing response component is lower than a preset coupling threshold, outputting an early warning command characterizing material slippage.

[0010] Preferably, the establishment of the dynamic load observation model in step S2 includes: establishing a second-order transfer function model containing mechanical inertia parameters, electromagnetic torque coefficients, and additional load impedance values, and using real-time velocity feedback data to identify the time-varying parameters in the second-order transfer function model online.

[0011] Preferably, in step S4, when the injection speed is ahead of the compensation amount, the material tension decoupling control between the front station and the back station is achieved by calculating the phase correlation matrix between each drive frequency actuator.

[0012] Preferably, the extraction of frequency feature reference quantities in step S2 includes: using a spectrum analysis algorithm to extract the spectrum amplitude corresponding to the mechanical support frequency of the controlled material from the driving current signal, and using the fluctuation of the spectrum amplitude as the control condition for triggering the update of the dynamic load observation model parameters.

[0013] Preferably, the method further includes an initial material characteristic adaptation step: during the pressure contact stage after the production line starts up, the upper and lower limits of the speed loop response gain are automatically set by monitoring the initial change slope of the drive current signal.

[0014] Preferably, the method further includes a zero-point verification step for the control reference: during the no-load operation phase of the production line, the static friction current in the drive current signal is detected, and the compensation reference in the dynamic load observation model is updated based on the drift of the static friction current. The updated compensation reference is then used to adjust the torque deviation component. Perform zero-position offset compensation.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the continuous manufacturing line of mattresses, a deep decoupling between the control loop gain and the physical characteristics of the material is achieved. By identifying the potential energy accumulation characteristic during the material conveying process as the equivalent impedance component of the control loop of the execution unit, the system can reconstruct the speed adjustment gain in real time according to the evolution trend of the internal stress of the material. This enables the dynamic characteristics of the controller to be dynamically matched with the viscoelastic feedback of the soft material, eliminating the drawback of traditional fixed parameter controllers that are prone to resonance with the rebound frequency of the material when performing phase compensation. This ensures that the closed-loop characteristic root of the control system is always in the stable region, and broadens the stability of the system when handling heterogeneous materials.

[0016] 2. A control zero-point self-verification mechanism based on material characteristic frequency anchoring was established. The micro-resistance fluctuations induced by the specific structure of the material during the conveying process are used as logical anchor points. By establishing a spatiotemporal causal model of characteristic frequency signals and potential energy accumulation characteristics, the mechanical friction offset inside the drive chain can be accurately identified. This process effectively separates the mechanical loss current caused by equipment temperature rise or aging from the material stress current, eliminates the risk of overcompensation caused by control reference drift, and ensures constant control accuracy of the production line during long-cycle operation.

[0017] 3. A potential energy release damping suppression system for unsteady working conditions was constructed. By real-time monitoring of the residual value of the phase compensation and prediction of the material release kinetic energy by combining the deceleration gradient, the system superimposes and injects a dynamic damping factor into the velocity loop to increase the virtual rotational inertia of the execution unit. This mechanism uses logical rigidity to counteract the physical rebound impact of the material, eliminating the material wrinkling and stacking problems caused by uncontrolled release of potential energy during sudden stops or decelerations in the production line, and ensuring the physical determinism of the material tension pattern across the entire speed range. Attached Figure Description

[0018] Figure 1 This is a flowchart of the adaptive collaborative control method for mattress production lines based on dynamic load observation, as described in this invention. Figure 2 This is a five-layer functional architecture block diagram of the adaptive collaborative control system for mattress production lines of the present invention. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining the present invention and are not intended to limit the scope of protection of the present invention.

[0020] This invention provides an adaptive collaborative control method for a continuous mattress manufacturing line, which collects the drive current signals of the execution units at each workstation. In addition to the rotational speed feedback signal, the potential energy accumulation characteristic quantity during the material conveying process is calculated. Based on this characteristic quantity, the processor performs impedance mapping and gain reconstruction of the control loop, involving a technical architecture consisting of load characteristic sensing, friction zero-position verification, adaptive gain adjustment, deceleration kinetic energy suppression, and physical coupling monitoring. In continuous mattress manufacturing lines, soft materials such as foam, spring groups, and fabric composites accumulate elastic potential energy during transportation due to their viscoelastic characteristics, resulting in a lag between control commands and the physical deformation response of the materials. To address the transient speed difference between stations caused by this lag, the processor collects the drive current signal of the drive frequency actuator. and instantaneous angular velocity A dynamic load observation model for the drive frequency actuator was established, which includes mechanical inertia parameters and electromagnetic torque coefficients. And a second-order transfer function model of the additional load impedance value, through the driving current signal rate of change The processor performs a differential mapping with the speed deviation to establish the drive current signal. With mechanical loss current The no-load mapping model, during real-time operation, uses the angular velocity recorded in the current real-time speed feedback data. Search for mechanical loss current at the corresponding frequency And perform arithmetic subtraction. Extract deformation feedback current According to the electromagnetic torque coefficient Deformation feedback current This is converted into the torque deviation component characterizing the deformation feedback of the controlled material, i.e., the potential energy accumulation characteristic quantity. This characteristic quantity is used to characterize the dynamic evolution trend of internal stress in materials.

[0021] Because the mechanical friction resistance of the motor drive system drifts slowly with changes in bearing temperature or lubricating grease viscosity, the system needs to calculate the characteristic quantity of potential energy accumulation. At that time, the spectrum analysis algorithm is used to analyze the driving current signal. Extract the characteristic frequency signal corresponding to the mechanical support frequency of the controlled material. This frequency signal is induced by the microscopic resistance fluctuations generated when the material passes through the conveyor rollers, and the processor establishes the characteristic frequency signal. With real-time linear velocity The causal relationship model, when the system detects a characteristic frequency signal Amplitude change and potential energy accumulation characteristic quantity When the evolution trend does not match, a mechanical friction offset is determined to exist. The system detects the static friction current during the no-load operation phase of the production line and updates the compensation benchmark in the dynamic load observation model based on the drift of this current. The updated compensation benchmark is then used to assess the potential energy accumulation characteristic quantity. Zero-position offset compensation is performed; frequency characteristic reference quantities are extracted, and the drive current signal is processed using a recursive sliding window fast Fourier transform algorithm. A Hanning window function is applied within a time-domain window of 1024 sampling points to suppress spectral leakage. A notch filter is used to filter out harmonic components related to the mechanical frequency of the motor rotor. Energy peak points are searched within the 5 Hz to 50 Hz frequency band, and the frequencies corresponding to these peak points are determined as the characteristic frequency signals. To eliminate the contradiction between the second-level latency caused by long-window calculations and the 50-millisecond transient response, the processor establishes a dual-track processing mechanism: in the frequency domain, the mechanical friction offset reference is updated once per second using the aforementioned 1024-point long window; in the time domain, a micro-sliding window with a length of 20 sampling points is simultaneously opened to calculate the absolute value of the current change rate of the drive current signal within a 1-millisecond period in real time. If this change rate exceeds three times the normal operating slope for three consecutive periods, the torque deviation component compensation action is triggered without waiting for the long-window spectrum analysis to end. This ensures that the transient impact of the material entering the pressure roller is pre-compensated within fifty milliseconds. As a characteristic frequency signal, when the energy integral value within the frequency band exceeds the reference noise spectrum by 15 dB, a reference quantity reflecting the characteristic changes in internal stress of the material is obtained. The torque deviation component is generated by separating the mechanical loss current offset caused by the temperature rise of the mechanical transmission chain from the total current. Zero-position offset compensation amount completes the separation of material deformation feedback current.

[0022] The processor accumulates potential energy features. The additional load impedance value identified as the controlled object is used by the processor in the speed control loop based on the potential energy accumulation characteristic. The algebraic sign direction and gradient of change are used to dynamically adjust the velocity loop response gain when the potential energy accumulation characteristic is detected. When the value is positive and the gradient increases, it is determined that internal stress is accumulating in the material, and the processor calculates according to the formula. Injection speed advance compensation amount At the same time, the proportional adjustment coefficient of the speed loop is reduced. ,in For speed advance compensation, The phase compensation intensity coefficient is determined by the system based on the current compressibility modulus gradient of the controlled material, which is within the range of 0.05 to 0.15. When the potential energy accumulation characteristic is detected When a polarity reversal or gradient descent occurs, it is determined to be a potential energy release, and the processor increases the proportional adjustment coefficient of the speed loop. Determine the torque deviation component. The quantization mapping relationship with the additional load impedance value is performed during the system initialization phase. A calibration procedure based on physical characteristic impedance is executed, and the drive frequency actuator is operated under no-load and with a preset compression ratio material. Drive current signals are collected at different target speeds, the frequency response offset of the system's closed-loop transfer function is calculated, and the torque deviation component is obtained by least squares fitting. With equivalent impedance Correspondence, additional load impedance value Satisfy the formula ,in For the additional load impedance value, The impedance transformation factor, For torque deviation component, The value is determined by comparing the ratio of the increase in mechanical power loss of the material under pressure to the amplitude of torque fluctuation, and the physical viscoelastic characteristics of the controlled material are converted into real-time processing impedance parameters in the control loop.

[0023] When the production line enters unsteady conditions such as emergency stop or deceleration, the system monitors the residual value of the phase compensation in real time and calculates the predicted value of the kinetic energy released by the material potential energy in combination with the deceleration gradient. When the predicted value exceeds the preset stability threshold, the processor injects a dynamic damping factor into the speed loop. Dynamic damping factor The calculation formula is as follows: ,in, The characteristic quantity for accumulating potential energy; This represents the real-time deviation value of the phase compensation amount; , which is the weighting coefficient for the first-order differential term, and has a value of 0.15; The weighting coefficient for the integral term is 0.08. The system counteracts the physical rebound impact of materials by increasing the virtual rotational inertia of the drive frequency actuator. To determine the physical connection strength between the material and the production line, the processor periodically injects a high-frequency torque disturbance signal with a frequency between 50Hz and 100Hz and an amplitude less than the material deformation threshold into the drive frequency actuator. The processor monitors the induced response component of the signal from the downstream adjacent workstations and calculates the measured coupling strength coefficient. If the measured coupling strength coefficient is lower than the preset connection threshold, the system determines that material slippage has occurred and automatically reduces the output gain of the phase compensation quantity, while simultaneously outputting a warning command. In addition, by using the total active power during the phase compensation process, the useful power component caused by the material load is eliminated to obtain the accompanying residual power characterizing the transmission loss. When the time series offset of the accompanying residual power reaches the preset mechanical fatigue threshold, the system performs gating restriction on the output range of the phase compensation quantity. A phase correlation matrix is ​​constructed, and in the... The first-stage actuator injects a pulse disturbance signal with an amplitude of 5% of the rated torque to monitor the adjacent first-stage actuator. The speed fluctuation response and phase deviation increment generated by the stage execution unit are calculated using the cross-correlation function to determine the coupling gain between the two stages. ,in This represents the current workstation level. For the first Level and First The coupling gain between the workstations is used to construct a Jacobian matrix that characterizes the material tension transmission characteristics. When the injection speed is ahead of the compensation amount, the mutual interference cancellation amount of each execution unit is calculated based on the inverse matrix of the Jacobian matrix. This ensures that the pressure wave caused by the acceleration action of the preceding workstation is offset by the phase pre-compensation action before it is transmitted to the subsequent workstation through the controlled material, thus maintaining the dynamic synchronization accuracy between the execution units of the production line.

[0024] Example 1: In a continuous manufacturing line for mattresses comprising an independent pocket spring layer and a high-density latex layer, the conveying system introduces two materials with different viscoelasticities into a pressure bonding station. Because the compression ratio of the materials at this station switches from 20% to 50%, the controlled material experiences nonlinear potential energy accumulation along the long conveying path. The processor monitors the drive current signal of the actuator in real time. Within a 50ms time window when the material enters the composite pressure roller, the deformation feedback current component caused by the damping characteristics of the latex material was detected to jump from 0.5A to 1.8A. The potential energy accumulation characteristic quantity, which characterizes the tendency of internal stress accumulation, was calculated. The value is 1.25 N·m, and the characteristic frequency signal is used. Friction zero-point compensation is performed on this value to separate the torque deviation component.

[0025] The system will accumulate the characteristic quantity of this potential energy. Real-time mapping to the additional load impedance value of the control loop, when monitored When the gradient of change exhibits a monotonically increasing trend, the speed loop controller lowers the proportional adjustment coefficient. The value was reduced from the baseline of 1.0 to 0.65, and an 8.2ms speed lead compensation was injected into the drive frequency actuator by lowering the proportional adjustment coefficient. To reduce the risk of the control loop's adaptability to fluctuations in material deformation feedback, when the composite station's operating cycle ends and it enters the deceleration phase, and the residual displacement of the phase compensation is detected to be 3mm, according to the formula... A dynamic damping factor of 2.1 was injected. ,in To accumulate characteristic quantities for potential energy, This represents the real-time deviation value of the phase compensation amount. is the weighting coefficient for the first-order differential term and has a value of 0.15. The weighting coefficient for the integral term is 0.08, which offsets the virtual inertia of the execution unit with the kinetic energy released from the potential energy of the material, keeping the transient tension fluctuation on the material surface within 5%.

[0026] Example 2: In a physical experimental platform for a continuous mattress manufacturing line equipped with a variable frequency drive motor and a high-precision sampling module, the drive current signal of the drive frequency actuator is acquired at a sampling frequency of 1kHz. Furthermore, Gaussian white noise with a signal-to-noise ratio of 25dB and 50Hz power frequency interference harmonics are actively superimposed on this signal source to simulate measurement errors under real industrial electromagnetic environments. The experiment verifies the potential energy accumulation characteristic quantity. The effectiveness of the gain reconstruction logic in suppressing fluctuations in the transport of soft materials, based on the sampling period. The setting balances the real-time performance of signal capture with the computational load of the processor. When the effective bandwidth of the controlled signal is in the range of 100Hz to 200Hz, in order to satisfy the Nyquist sampling theorem and reserve a spectral resolution margin of more than three times, the sampling period is adjusted. The parameter was set to 1ms, providing high-fidelity discretized input data for the dynamic load observation model. The experiment simulated different load conditions by varying the compression ratio of the high-density latex material, and used a spectrum analysis algorithm to analyze the noisy drive current signal. Extracting characteristic frequency signals This signal is used to identify and separate the 0.15A mechanical friction offset caused by the temperature rise of the transmission chain, ensuring the potential energy accumulation characteristic quantity. The calculation results only point to the feedback torque generated by material deformation. See Table 1, which records the evolution trajectory of intermediate variables inside the system and the final speed deviation rate index as the latex sponge compression ratio increases in a gradient.

[0027] Table 1: Comparison of Control Parameters and Performance Indicators under Different Load Conditions

[0028] Data analysis shows that as the material compression ratio increases from 15% to 45%, the potential energy accumulation characteristic quantity calculated by the experimental group... The speed loop proportional adjustment coefficient increases linearly from 0.32 N·m to 1.45 N·m. The automatic reconfiguration from 1.15 to 0.70, this adaptive gain adjustment method, kept the system speed deviation rate at 0.68%. In contrast, control group A, using fixed parameters, induced loop oscillation under a high compression ratio of 45% because it could not accommodate the elastic rebound torque of the material, resulting in a speed deviation rate of 3.15%. Control group B, which only had static phase compensation but lacked impedance mapping logic, although it achieved a speed deviation rate of 3.15% through a preset lower impedance mapping logic, achieved a speed deviation rate of 0.68%. While the oscillations were reduced, its deviation rate of 1.95% was still higher than that of the experimental group in this invention. When the material compression ratio was further increased to the out-of-range boundary of 65%, the step amplitude of the deformation feedback torque exceeded the linear compensation capability of the actuator due to the material entering the viscoelastic supersaturation region, and the proportional adjustment coefficient... The performance inflection point occurred after reaching the limit value of 0.35, causing the speed deviation rate to rise to 5.25%. This phenomenon provides a definite physical boundary basis for the adaptive adjustment range defined in this invention. The experimental results prove the potential energy accumulation characteristic extracted by the dynamic load observation model. It can characterize the dynamic evolution of internal stress in materials in real time. The control gain reconstruction procedure executed based on this characteristic effectively offsets the feedback disturbance caused by the viscoelasticity of the material, enabling the continuous manufacturing line to maintain speed synchronization under variable load environments and realizing dynamic matching between control logic and material physical characteristics.

[0029] Example 3: This example combines Figures 1 to 2 An adaptive collaborative control method for a continuous mattress manufacturing line is described, such as... Figure 1 As shown, step S1 involves acquiring the drive current signal and real-time speed feedback data of the drive frequency actuator in the continuous mattress manufacturing line in real time. Step S2 involves establishing a dynamic load observation model for the drive frequency actuator, extracting frequency characteristic reference quantities based on the real-time speed feedback data, separating the mechanical loss current and deformation feedback current from the drive current signal, and confirming the deformation feedback current as the torque deviation component characterizing the deformation feedback of the controlled material. Next, step S3 is executed to detect the torque deviation component in real time. The algebraic sign direction, and the changing gradient and torque deviation components The additional load impedance value is converted into the control loop of the drive frequency actuator to quantitatively characterize the measured coupling strength between the controlled material and the conveying mechanism; finally, step S4 is executed based on the torque deviation component. The algebraic sign direction, gradient change, and additional load impedance value are used to dynamically adjust the speed loop response gain and inject speed advance compensation to offset the response lag caused by the deformation of the controlled material, ensuring that the conveying speed difference between each drive frequency actuator is stable within the preset physical deterministic range.

[0030] like Figure 2As shown, the adaptive collaborative control system for the mattress production line has a five-layer structure. The data perception layer is equipped with functional modules for acquiring drive current signals, obtaining real-time speed feedback, and extracting frequency characteristic reference quantities. The model observation layer contains processing units for establishing load observation models, separating mechanical loss current, and calculating torque deviation components. The logic decision layer is responsible for performing logical operations such as detecting characteristic gradients and directions, mapping additional load impedance, and verifying the friction zero-position reference. The collaborative execution layer is used to realize control actions such as dynamically adjusting the speed loop gain, injecting speed advance compensation, and superimposing deceleration dynamic damping. The safety monitoring layer undertakes the safety protection tasks of injecting small torque disturbances, monitoring coupling strength coefficients, and outputting slip warning commands.

[0031] Example 4: In a manufacturing environment involving the production of mattresses with a mixture of multiple elastic moduli of foam, the system faces the risk of control loop instability caused by abrupt changes in material hardness. To determine the dynamic damping factor... Weighting coefficients and During the no-load initialization phase of the production line, the drive frequency actuator executes a single-step displacement command and records the drive current signal. The transient response curve was obtained, and the inherent mechanical damping ratio of the system was extracted by performing a least-squares fitting on the response curve. and natural frequency The processor will add weighting coefficients. The linear proportional parameter for the acceleration feedback gain is determined, and its value is set to ensure that the overshoot of the velocity loop is less than 2% at the instant of impact. This is applied to the weighting coefficient of the integral term. The processor injects torque disturbances and adjusts them under constant low-speed operating conditions. The residual deviation of the phase compensation quantity is brought to below the preset noise level within three sampling periods. The combination of physical parameters determined by this calibration procedure is stored in memory as the initial configuration; for the potential energy accumulation characteristic quantity... The characteristic frequency signal on which the calculation process depends The processor uses a recursive sliding window fast Fourier transform algorithm to process the drive current signal. Spectral analysis is performed. Within a time-domain window of 1024 sampling points, the processor applies the Hanning window function to suppress spectral leakage and calculates the power spectral density distribution. To separate mechanical losses from material deformation feedback, the processor searches for peaks in the 5Hz to 50Hz frequency band and uses a centroid frequency algorithm to calculate the characteristic frequency signal. When the energy integral value of this frequency band in the power spectrum exceeds the reference noise spectrum by 15dB, it is determined that a valid characteristic reference quantity has been obtained. The processor removes harmonic components related to the mechanical frequency of the motor rotor in the frequency domain, thereby obtaining the deformation feedback current characterizing the internal stress change of the material and converting it into torque deviation component. .

[0032] During online monitoring of the conveying coupling state, the system determines the logical boundary of the coupling threshold through dynamic load testing. It then drives the frequency actuator to gradually increase the output value of the phase compensation under a known material compression ratio until a nonlinear drop occurs in the downstream inductive response component, corresponding to the critical point of maximum static friction between the material and the conveying mechanism. The processor calculates the coupling strength coefficient corresponding to this critical point and uses it as the coupling threshold. Coupling threshold The calculation formula is as follows: ,in, The coupling threshold, This represents the incremental induced torque detected at the downstream adjacent workstation, in N·m. The amplitude of the disturbance signal injected into the current execution unit, in N·m. The mechanical efficiency constant of the transmission chain is given when the measured coupling strength coefficient drops to the coupling threshold. When the output range of the phase compensation is below 90%, the system performs amplitude limiting processing on the output range of the phase compensation amount, and the displacement synchronization accuracy of the production line is maintained above 99.8% under the operating environment of 20m / min.

[0033] Example 5: In a calibration scenario involving new production line deployment and drive motor replacement, the system performs calibration on the electromagnetic torque coefficient. Compared with the reference noise spectrum The offline calibration process involves the processor driving the motor to perform a frequency sweep motion at a speed range of 0.1 m / s to 1.0 m / s in the initial state of the production line being unloaded and without any conveying materials. The average value of the bus current at each speed point is recorded in real time, and the electromagnetic torque coefficient is calculated by fitting using the least squares method. The measured values, of which The electromagnetic torque coefficient is expressed in N•m / A. Simultaneously, under a constant target speed, a recursive sliding window algorithm is used to continuously acquire the drive current signal at 5000 sampling points. The average distribution of its power spectral density was calculated, and its curve was used as the reference noise spectrum. It is stored in a static storage unit to accumulate characteristic quantities for subsequent potential energy. The solution provides a subtraction hedging benchmark.

[0034] When the system faces mechanical efficiency issues due to wear of the transmission mechanism During fluctuating operating conditions, the initial boundary of the control gain is corrected by executing a closed-loop convergence debugging process. Under the premise of applying a standard load, a step excitation signal with an amplitude of 5% of the rated torque is injected into the speed loop, and the residual deviation of the phase compensation is monitored. proportional adjustment coefficient The damped oscillation characteristics resulting from the reduction are adjusted by the processor based on the deviation convergence time and the proportional coefficient. The mapping relationship of the changing gradient is used to calculate the equivalent transmission impedance under the current physical environment, and the calculation results are fed back to the gain reconstruction logic of the dynamic load observation model to achieve the matching of control parameters and physical entity state. The system enters the operating state with a phase synchronization deviation of less than 0.2ms in the first operating cycle after initial power-on.

[0035] Example 6: In the production line integration and deployment process, the system executes a process targeting the mechanical efficiency constant. The on-site identification procedure enables adaptive matching of the actuator impedance characteristics. The processor controls the drive frequency of the actuator in an initial environment where it is unloaded and not connected to the conveyor belt. The drive motor accelerates from zero speed to 50% of the rated speed and maintains a constant speed. The processor synchronously reads the voltage vector component and bus current signal output by the inverter. The active power is calculated and subtracted from the electromagnetic torque coefficient. The determined stator copper loss and no-load rotational loss are used to obtain the measured mechanical efficiency constant, which reflects the energy conversion efficiency of the transmission chain. ,in Let be the mechanical efficiency constant. For driving current signal, The electromagnetic torque coefficient is expressed in N·m / A. The processor compares this measured value with the theoretical mapping value in memory, and recalibrates the torque mapping weight coefficient in the dynamic load observation model under the trigger condition that the deviation exceeds 2%, thereby completing the matching of the control logic with the current physical transmission chain impedance state.

[0036] When the system is applied to batches of materials with different compression modulus gradients, the processor initiates the initial boundary debugging process for the gain adaptive adjustment procedure. Under the condition of attaching a standard reference load, the system injects a low-frequency fluctuation signal with a frequency of 10Hz and an amplitude of 5% of the rated torque into the speed regulation loop of the actuator. The processor captures the real-time deviation value of the phase compensation in real time. proportional adjustment coefficient The response characteristics of gradient changes are studied, and the proportional adjustment coefficient is determined by identifying the critical point where the derivative of the rate of change of phase difference approaches zero and the system's eigenvalues ​​are in the stable region. The lower limit of the working scope, of which This represents the real-time deviation value of the phase compensation amount. As a proportional adjustment coefficient, this lower limit is written into the configuration register of the drive frequency actuator as a gating limit to prevent instability, thereby enabling the system to maintain a response convergence time of less than 50ms under disturbance conditions where the material compression ratio changes abruptly.

[0037] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0038] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An adaptive collaborative control method for a continuous mattress manufacturing line, characterized in that, Includes the following steps: Step S1: Real-time acquisition of drive current signals and real-time speed feedback data of the drive frequency actuator in the continuous mattress manufacturing line; Step S2: Establish a dynamic load observation model for the drive frequency actuator, and based on the dynamic load observation model and real-time speed feedback data, calculate the torque deviation component characterizing the deformation feedback of the controlled material from the drive current signal. Step S2 includes: extracting frequency characteristic reference quantities based on real-time speed feedback data, separating mechanical loss current and deformation feedback current characterizing internal stress changes in materials from the drive current signal according to the frequency characteristic reference quantities, and confirming the deformation feedback current as the torque deviation component. ; Step S3: Real-time detection of torque deviation component The algebraic sign direction and gradient of change, and the torque deviation component. The additional load impedance value is converted into the control loop of the drive frequency actuator to quantitatively characterize the measured coupling strength between the controlled material and the conveying mechanism. Step S4, based on the torque deviation component The algebraic sign direction, gradient change, and additional load impedance value are used to dynamically adjust the speed loop response gain of the drive frequency actuator and inject speed advance compensation to offset the response lag caused by the deformation of the controlled material, ensuring that the conveying speed difference between each drive frequency actuator is stable within the preset physical deterministic range.

2. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, The speed loop response gain includes a proportional adjustment coefficient; the dynamic adjustment of the speed loop response gain and the synchronous injection of speed lead compensation in step S4 include: when a torque deviation component is detected... When the value is positive and the gradient is increasing, the proportional adjustment coefficient is simultaneously reduced while the injection speed advances the compensation amount; when a torque deviation component is detected... When the value is positive and the gradient of change decreases, the proportional adjustment coefficient is gradually restored and the speed advance compensation amount is maintained.

3. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, The method also includes a damping adaptive step for deceleration conditions: Step S301, real-time monitoring of the residual value of the speed advance compensation, and calculation of the predicted value of kinetic energy release of the controlled material in combination with the real-time deceleration slope; Step S302, based on the predicted value of kinetic energy release, superimposing and injecting a dynamic damping factor into the speed loop of the drive frequency actuator. This increases the virtual rotational inertia of the drive frequency actuator and counteracts the physical rebound impact generated by the controlled material during the deceleration phase.

4. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 3, characterized in that, In step S302, a dynamic damping factor is injected. The computational logic follows these rules: ,in, For torque deviation component, This represents the real-time deviation value of the speed advance compensation. These are the pre-defined weighting coefficients for the first-order differential term. These are the pre-defined weighting coefficients for the integral term.

5. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, The method also includes an online monitoring step for the conveying coupling state: step S501, injecting a small torque disturbance signal with a frequency of 50Hz to 100Hz into the drive frequency actuator; step S502, analyzing the sensing response component of the drive frequency actuator of the downstream adjacent station to the small torque disturbance signal; step S503, when the sensing response component is lower than the preset coupling threshold, outputting an early warning command characterizing material slippage.

6. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, Step S2 involves establishing a dynamic load observation model, which includes establishing a second-order transfer function model containing mechanical inertia parameters, electromagnetic torque coefficients, and additional load impedance values, and using real-time velocity feedback data to identify the time-varying parameters in the second-order transfer function model online.

7. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, When the injection speed advance compensation amount is calculated in step S4, the material tension decoupling control between the front station and the back station is realized by calculating the phase correlation matrix between each drive frequency actuator.

8. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, Step S2 involves extracting frequency feature reference values ​​by using a spectrum analysis algorithm to extract the spectrum amplitude corresponding to the mechanical support frequency of the controlled material from the driving current signal, and using the fluctuation of the spectrum amplitude as the control condition for triggering the update of the dynamic load observation model parameters.

9. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, The method also includes an initial material characteristic adaptation step: during the pressure contact stage after the production line starts up, the upper and lower limits of the speed loop response gain are automatically set by monitoring the initial change slope of the drive current signal.

10. The adaptive collaborative control method for a continuous mattress manufacturing line according to claim 1, characterized in that, The method also includes a zero-point verification step for the control reference: during the no-load operation phase of the production line, the static friction current in the drive current signal is detected, and the compensation reference in the dynamic load observation model is updated based on the drift of the static friction current. The updated compensation reference is then used to adjust the torque deviation component. Perform zero-position offset compensation.

Images