A new energy battery separator mold closing alignment compensation control method and system

Abstract

The present application relates to the technical field of automatic compression molding control, and discloses a new energy battery separator mold closing alignment compensation control method and system, comprising: collecting the position increment and torque current feedback of the servo shaft when entering the compression interval, and calculating the dynamic impedance derivative; when the dynamic impedance derivative is greater than the trigger threshold and the position increment is less than the dead zone reference value, the discrete time domain integral of the torque current feedback is calculated to generate the energy flow accumulation parameter; the elastic pre-saturation feedforward instruction and the basic pose compensation instruction are calculated and the timing output is controlled accordingly; the present application eliminates the transmission chain elastic deformation dead zone by early feedforward injection, eliminates the phase lag of the control signal conducted to the mold surface, effectively avoids the position instruction accumulation in the precision compression interval and the system self-excited oscillation, and improves the parallel alignment accuracy and long-term stability of the multi-axis closed-loop system.

Description

Technical Field

[0001] This invention relates to a method and system for mold closing and alignment compensation control of a new energy battery separator mold, belonging to the field of automated injection molding control technology. Background Technology

[0002] In the current hot-pressing process of battery separators, the control system uses a multi-axis servo drive structure to regulate mold closure. It works in conjunction with a dynamic impedance mechanism to synchronously collect physical feedback data from each drive axis to calculate the cross-coupling alignment compensation amount. This is used to correct the spatial pose deviation generated during the multi-axis mold closing action. Under continuous mass production conditions, the mold is subjected to high-frequency alternating heat load and high-stress extrusion. Its structure undergoes non-uniform thermal expansion and structural deformation. When the mold closing mechanism approaches the micron-level precision alignment range, the position impedance balance between the multi-axis drifts and induces asymmetric off-center load internal forces. This drives the control system to frequently issue micro-adjustment commands to counteract the off-center load. The inherent screw backlash, reducer return clearance, and elastic yield resistance of the transmission shaft in the multi-axis transmission chain together constitute the transmission deformation zone. Before the kinetic energy carried by the micro-control signal is transmitted to the mold force surface, it is preferentially converted into nonlinear elastic loss inside the transmission components. This causes the physical motion response of the mold surface to have a phase lag relative to the adjustment commands issued by the control loop.

[0003] Conventional linear improvement approaches, such as replacing the servo motor with a high-power one or increasing the rigidity of transmission components to forcibly eliminate backlash, increase the rotational inertia of the overall mechanism. This can easily induce high-frequency resonance in the servo control loop, increasing the processing and manufacturing costs of the equipment. Relying solely on hardware rigidity enhancement faces numerous limitations. Attempts to address the inherent defects of physical dissipation through conventional software control methods are difficult to overcome. For example, Chinese invention patent application CN121173144A discloses a dual-inertia elastic position servo system and control method. This system uses the motor speed as the input to the second-order linear extended state observer in the active disturbance rejection controller, introducing speed observation errors to calculate elastic disturbances including those caused by transmission mechanism deformation in real time. While existing technologies possess the theoretical basis for suppressing macroscopic mechanical resonance, their underlying logic is highly dependent on the objective physical mechanism of error-driven feedback observation. Under the micron-level pressing condition of the battery separator in an extremely narrow range, the material stress and mechanical stiffness exhibit a sharp nonlinear change. The method of relying on the observer to capture speed deviations and push out linear disturbances is inherently lagging behind the actual deformation on the time axis. There is a fundamental mismatch between the control mechanism and the ultra-high dynamic response micro-pressing environment. Not only is it impossible to pre-neutralize the elastic dead zone of the transmission chain before the spatial position command is issued, but it is also very easy for the high-frequency accumulation of position commands in the micro-range of mold closing and the system self-excited oscillation to be aggravated by the superposition of the lag in digital calculation of state observation and the lag in the mechanical native elasticity.

[0004] Therefore, how to synchronously collect multi-axis position and torque feedback data to quantify the degree of elastic deformation of transmission components, and reconstruct the command output timing before the alignment and correction action so that the transmission chain reaches the rigid critical state, becomes the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for mold closing and alignment compensation control of a new energy battery separator mold, comprising the following steps:

[0006] Step S101: Obtain the real-time current and position increment of each independent drive shaft in the controlled drive system, and calculate the real-time torque change and dynamic impedance derivative accordingly. Step S102: Construct a sliding time window containing 5 consecutive control action cycles, calculate the baseline drift slope based on the steady-state impedance bias baseline at the end of each action within the sliding time window, and translate and correct the dynamic impedance reference envelope before starting the next control action cycle. Step S103: When the real-time torque change exceeds the lower limit of torque fluctuation and the position increment is less than the preset dead zone reference value, calculate the discrete time domain integral value of the real-time current within the time window when the current position increment is lower than the preset dead zone reference value, and generate the energy flow accumulation parameter. Step S104: Calculate the basic pose compensation command based on the difference between the dynamic impedance derivative and the corrected dynamic impedance reference envelope, and calculate the elastic presaturation feedforward command based on the energy flow accumulation parameter and the system stiffness mapping matrix. In step S105, the elastic pre-saturation feedforward command is configured to be output before the basic pose compensation command in terms of timing. The elastic pre-saturation feedforward command is input first to neutralize the dead zone of elastic deformation of the transmission chain. In response to each independent drive shaft reaching the rigid critical state, the basic pose compensation command is triggered to complete the parallel alignment.

[0007] Preferably, the calculation of the real-time torque change of each independent drive shaft in step S101 includes: scaling the real-time current to obtain the real-time torque value, calculating the difference between the real-time torque values ​​between adjacent control cycles as the real-time torque change; obtaining the dynamic impedance derivative by calculating the rate of change of the real-time torque relative to the position change; and obtaining the lower limit of torque fluctuation by acquiring the fixed torque fluctuation corresponding to the inherent cogging torque and static friction force of the system, and reconstructing the amplitude of the fixed torque fluctuation.

[0008] Preferably, step S105, which uses a decoupling compensation matrix to process the asymmetric torsional vector and generate a two-level timing command pulse containing a pre-saturation compensation pulse and a position compensation pulse, includes the following sub-steps: Step S1051, inputting the asymmetric torsional vector generated by the difference mapping into the decoupling matrix model mapping to eliminate the planar tilting eccentric load internal force caused by multi-axis mutual coupling; Step S1052, extracting the transient displacement compensation amount of each independent drive shaft after decoupling matrix model mapping, and superimposing the transient displacement compensation amount of each independent drive shaft as a feedforward correction amount onto the controlled drive system. In the position loop; in step S1053, the feedforward correction is split into two time-separated control signals. The first-level control signal is a pre-saturation compensation pulse, which is used to output within the micro-range before approaching. Its pulse amplitude has a specific proportional relationship with the deformation stiffness of the transmission chain, so as to use the controlled current to make the transmission components produce pre-elastic deformation. The second-level control signal is a position compensation pulse, which is triggered to output in response to the step of the drive shaft torque slope and reaching the rigid critical state, so as to drive the controlled surface to fine-tune the parallelism, so that the dynamic impedance derivatives of each independent drive shaft are forced to converge.

[0009] Preferably, step S105, which triggers the output of the basic pose compensation command in response to each independent drive shaft reaching the rigidity critical state, includes the following sub-steps: Step S1054, during the output of the pre-saturation compensation pulse generated according to the elastic pre-saturation feedforward command, the feedback torque change rate of each independent drive shaft is monitored in real time; Step S1055, the feedback torque change rate is compared with the preset rigidity judgment threshold in real time. When the feedback torque change rate is greater than or equal to the preset rigidity judgment threshold for three consecutive control action cycles, it is determined that the elastic deformation dead zone of the transmission chain has been completely absorbed, and a rigidity critical state confirmation signal is output; Step S1056, the control loop stops outputting the pre-saturation compensation pulse and simultaneously releases the position compensation pulse as the basic pose compensation command, thereby cutting off the attitude compensation phase hysteresis caused by mechanical elastic physical dissipation.

[0010] Preferably, the calculation of the basic pose compensation command based on the difference between the derivative of the dynamic impedance and the corrected dynamic impedance reference envelope includes the following sub-steps: Step S1041, comparing the real-time torque change with the lower limit of torque fluctuation; Step S1042, when the real-time torque change is lower than the lower limit of torque fluctuation, determining that the current fluctuation is non-real deformation noise caused by mechanical static friction, the control loop maintains the command output of the previous control cycle, and shields the interference to the position loop; Step S1043, when the real-time torque change is greater than or equal to the lower limit of torque fluctuation, using this as the filtering environment to activate the built-in impedance digital low-pass filter, truncating high-frequency impedance abrupt change data higher than the inherent resonant frequency of the mechanical system, and extracting the pure surface dynamic impedance features; Step S1044, defining the difference vector between the pure surface dynamic impedance features and the spatial geometric center of the corrected dynamic impedance reference envelope as an asymmetric torsion vector, and using this as the difference to calculate the basic pose compensation command.

[0011] Preferably, step S102, which extracts the steady-state impedance bias baseline at the end of each action within the sliding time window, includes the following sub-steps: Step S1021, during the voltage holding steady-state phase of each control action cycle, the steady-state voltage and steady-state current of each independent drive shaft are collected; Step S1022, the steady-state impedance observation value of each cycle is calculated based on the steady-state voltage and steady-state current; Step S1023, within the sliding time window, the steady-state impedance observation values ​​of five consecutive cycles are subjected to mean filtering to remove transient impedance fluctuations caused by microscopic non-uniformity of material thickness, thereby obtaining a steady-state impedance bias baseline that can characterize the slow thermal accumulation effect of the long-cycle temperature field.

[0012] Preferably, the controlled drive system includes a main controller that establishes a data mapping association through an industrial Ethernet bus and four independent servo drives; the communication cycle between the main controller and the servo drives is 0.25ms to 1.0ms; the real-time current of each independent drive axis in the controlled drive system in step S101 includes: the main controller synchronously reads the original sampling data of the current loop inside the four independent servo drives through the industrial Ethernet bus at the communication cycle.

[0013] Preferably, processing the asymmetric torsion vector using the decoupling compensation matrix to make the dynamic impedance derivatives of each independent drive shaft converge includes: adjusting the position loop gain of each independent drive shaft to reduce the dynamic impedance derivative of the drive shaft that contacts earlier to increase virtual flexibility, and increasing the approach speed of the drive shaft that contacts later, until the dynamic response feedback loop detects that the maximum deviation between the dynamic impedance derivatives of each independent drive shaft is less than 0.05, thereby completing the closed-loop correction control of multi-axis plane parallelism.

[0014] Preferably, the current amplitude of the pre-saturation compensation pulse is 20% to 40% of the rated current of the servo motor; the step S105 first inputs the elastic pre-saturation feedforward command to neutralize the elastic deformation dead zone of the transmission chain, which includes: in the range of 2.0mm to 5.0mm before approaching the surface micro-contact, inputting the pre-saturation compensation pulse to the current loop of each independent drive shaft, so that the electromagnetic torque of the servo motor and the deformation resistance torque of the transmission chain reach a balanced state, so as to offset the physical transmission error without changing the mechanical topology.

[0015] A mold closing alignment compensation control system for a new energy battery separator mold, which is used to implement a mold closing alignment compensation control method for a new energy battery separator mold, including: The data acquisition module is used to acquire the real-time current and position increment of each independent drive shaft in the controlled drive system, and to calculate the real-time torque change and dynamic impedance derivative accordingly. The baseline correction module is used to construct a sliding time window containing five consecutive control action cycles. It calculates the baseline drift slope based on the steady-state impedance bias baseline at the end of each action within the sliding time window, and translates and corrects the dynamic impedance reference envelope before the start of the next control action cycle to obtain the corrected dynamic impedance reference envelope. The parameter generation module is used to calculate the discrete time-domain integral value of the real-time current within the time window when the real-time torque change exceeds the lower limit of torque fluctuation and the position increment is less than the preset dead zone reference value, and generate energy flow accumulation parameters. The feedforward instruction module is used to calculate the basic pose compensation instruction based on the difference between the dynamic impedance derivative and the corrected dynamic impedance reference envelope, and to calculate the elastic presaturation feedforward instruction based on the energy flow accumulation parameter and the system stiffness mapping matrix. The alignment control module is used to set the elastic pre-saturation feedforward command to be output before the basic pose compensation command in terms of timing. The elastic pre-saturation feedforward command is input first to neutralize the dead zone of elastic deformation of the transmission chain. In response to each independent drive shaft reaching the rigidity critical state, the basic pose compensation command is triggered to complete the parallel alignment.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. In the mold closing and alignment compensation control of new energy battery separator mold, the pressing stagnation state at the mold closing end is identified by the current position increment of the multi-axis mold closing mechanism and the torque current feedback. When the position increment is lower than the set deformation dead zone reference value, the time domain discrete integration is started to quantify and generate energy flow accumulation parameters that characterize the degree of elastic deformation of the transmission components. In terms of timing, the elastic pre-saturation feedforward command generated based on the parameter is configured to be output before the basic position and posture compensation command, so that the transmission components reach the rigid critical state before the spatial correction and alignment action occurs. This eliminates the position response lag caused by the elastic hysteresis of the ball screw and reducer in the conventional position closed-loop control loop, avoids high frequency accumulation of control commands and system self-excitement oscillation at the pressing cutoff boundary, and ensures that the small torque output by the control system is smoothly and timely transmitted to the mold force surface.

[0017] 2. By synchronously acquiring the transient physical parameters of multiple mold-closing axes at high frequency, the ratio of the real-time torque change rate to the position change rate of each axis is calculated to generate dynamic impedance derivatives. When the difference in impedance derivatives of opposite or diagonal drive axes exceeds the set contact tolerance threshold, an asymmetric torsional vector characterizing spatial pose deviation is generated. The transient displacement compensation of each drive axis is calculated using the decoupling compensation matrix and superimposed on the position loop as a feedforward correction. This increases the virtual flexibility of the early contact axis and increases the approach speed of the lagging contact axis, forcibly converging the dynamic impedance derivatives of each independent drive axis. Without changing the mechanical topology of the mold, this solves the problem of eccentric internal force resistance caused by non-uniform thermal expansion of multiple axes, and improves the planar parallelism and trajectory tracking accuracy of the closed-loop system in the micron-level alignment stage.

[0018] 3. By constructing a sliding time window containing multiple consecutive mold closing production cycles, the steady-state impedance bias baseline at the end of each mold closing within the sliding time window is extracted and the baseline drift slope is calculated. Before the start of the next production cycle, the dynamic impedance reference envelope containing the standard no-load impedance and material deformation resistance is translated and corrected along the drift direction. The slow heat accumulation effect accompanying the production process is converted into a predictable feedforward translation amount. This works in conjunction with the transient impedance closed-loop adjustment mechanism in a single mold closing process to form a dual defense control architecture under multiple time scales. This eliminates the degradation of position control accuracy caused by long-term time-varying temperature field of the mold and maintains the long-term stability of the control system under continuous mass production conditions. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating the overall steps of the new energy battery separator mold alignment compensation control method of the present invention. Figure 2 This is a diagram showing the two-stage signal shunting and closure correction of the mold alignment compensation control method for the new energy battery separator mold of the present invention.

[0020] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0022] A method for mold closing alignment compensation control of a new energy battery separator mold includes the following steps: Step S101: Obtain the real-time current and position increment of each independent drive shaft in the controlled drive system, and calculate the real-time torque change and dynamic impedance derivative accordingly. Step S102: Construct a sliding time window containing 5 consecutive control action cycles, calculate the baseline drift slope based on the steady-state impedance bias baseline at the end of each action within the sliding time window, and translate and correct the dynamic impedance reference envelope before starting the next control action cycle. Step S103: When the real-time torque change exceeds the lower limit of torque fluctuation and the position increment is less than the preset dead zone reference value, calculate the discrete time domain integral value of the real-time current within the time window when the current position increment is lower than the preset dead zone reference value, and generate the energy flow accumulation parameter. Step S104: Calculate the basic pose compensation command based on the difference between the dynamic impedance derivative and the corrected dynamic impedance reference envelope, and calculate the elastic presaturation feedforward command based on the energy flow accumulation parameter and the system stiffness mapping matrix. In step S105, the elastic pre-saturation feedforward command is configured to be output before the basic pose compensation command in terms of timing. The elastic pre-saturation feedforward command is input first to neutralize the dead zone of elastic deformation of the transmission chain. In response to each independent drive shaft reaching the rigid critical state, the basic pose compensation command is triggered to complete the parallel alignment.

[0023] Preferably, the calculation of the real-time torque change of each independent drive shaft in step S101 includes: scaling the real-time current to obtain the real-time torque value, calculating the difference between the real-time torque values ​​between adjacent control cycles as the real-time torque change; obtaining the dynamic impedance derivative by calculating the rate of change of the real-time torque relative to the position change; and obtaining the lower limit of torque fluctuation by acquiring the fixed torque fluctuation corresponding to the inherent cogging torque and static friction force of the system, and reconstructing the amplitude of the fixed torque fluctuation.

[0024] Preferably, step S105, which uses a decoupling compensation matrix to process the asymmetric torsional vector and generate a two-level timing command pulse containing a pre-saturation compensation pulse and a position compensation pulse, includes the following sub-steps: Step S1051, inputting the asymmetric torsional vector generated by the difference mapping into the decoupling matrix model mapping to eliminate the planar tilting eccentric load internal force caused by multi-axis mutual coupling; Step S1052, extracting the transient displacement compensation amount of each independent drive shaft after decoupling matrix model mapping, and superimposing the transient displacement compensation amount of each independent drive shaft as a feedforward correction amount onto the controlled drive system. In the position loop; in step S1053, the feedforward correction is split into two time-separated control signals. The first-level control signal is a pre-saturation compensation pulse, which is used to output within the micro-range before approaching. Its pulse amplitude has a specific proportional relationship with the deformation stiffness of the transmission chain, so as to use the controlled current to make the transmission components produce pre-elastic deformation. The second-level control signal is a position compensation pulse, which is triggered to output in response to the step of the drive shaft torque slope and reaching the rigid critical state, so as to drive the controlled surface to fine-tune the parallelism, so that the dynamic impedance derivatives of each independent drive shaft are forced to converge.

[0025] Preferably, step S105, which triggers the output of the basic pose compensation command in response to each independent drive shaft reaching the rigidity critical state, includes the following sub-steps: Step S1054, during the output of the pre-saturation compensation pulse generated according to the elastic pre-saturation feedforward command, the feedback torque change rate of each independent drive shaft is monitored in real time; Step S1055, the feedback torque change rate is compared with the preset rigidity judgment threshold in real time. When the feedback torque change rate is greater than or equal to the preset rigidity judgment threshold for three consecutive control action cycles, it is determined that the elastic deformation dead zone of the transmission chain has been completely absorbed, and a rigidity critical state confirmation signal is output; Step S1056, the control loop stops outputting the pre-saturation compensation pulse and simultaneously releases the position compensation pulse as the basic pose compensation command, thereby cutting off the attitude compensation phase hysteresis caused by mechanical elastic physical dissipation.

[0026] Preferably, the calculation of the basic pose compensation command based on the difference between the derivative of the dynamic impedance and the corrected dynamic impedance reference envelope includes the following sub-steps: Step S1041, comparing the real-time torque change with the lower limit of torque fluctuation; Step S1042, when the real-time torque change is lower than the lower limit of torque fluctuation, determining that the current fluctuation is non-real deformation noise caused by mechanical static friction, the control loop maintains the command output of the previous control cycle, and shields the interference to the position loop; Step S1043, when the real-time torque change is greater than or equal to the lower limit of torque fluctuation, using this as the filtering environment to activate the built-in impedance digital low-pass filter, truncating high-frequency impedance abrupt change data higher than the inherent resonant frequency of the mechanical system, and extracting the pure surface dynamic impedance features; Step S1044, defining the difference vector between the pure surface dynamic impedance features and the spatial geometric center of the corrected dynamic impedance reference envelope as an asymmetric torsion vector, and using this as the difference to calculate the basic pose compensation command.

[0027] Preferably, step S102, which extracts the steady-state impedance bias baseline at the end of each action within the sliding time window, includes the following sub-steps: Step S1021, during the voltage holding steady-state phase of each control action cycle, the steady-state voltage and steady-state current of each independent drive shaft are collected; Step S1022, the steady-state impedance observation value of each cycle is calculated based on the steady-state voltage and steady-state current; Step S1023, within the sliding time window, the steady-state impedance observation values ​​of five consecutive cycles are subjected to mean filtering to remove transient impedance fluctuations caused by microscopic non-uniformity of material thickness, thereby obtaining a steady-state impedance bias baseline that can characterize the slow thermal accumulation effect of the long-cycle temperature field.

[0028] Preferably, the controlled drive system includes a main controller that establishes a data mapping association through an industrial Ethernet bus and four independent servo drives; the communication cycle between the main controller and the servo drives is 0.25ms to 1.0ms; the real-time current of each independent drive axis in the controlled drive system in step S101 includes: the main controller synchronously reads the original sampling data of the current loop inside the four independent servo drives through the industrial Ethernet bus at the communication cycle.

[0029] Preferably, processing the asymmetric torsion vector using the decoupling compensation matrix to make the dynamic impedance derivatives of each independent drive shaft converge includes: adjusting the position loop gain of each independent drive shaft to reduce the dynamic impedance derivative of the drive shaft that contacts earlier to increase virtual flexibility, and increasing the approach speed of the drive shaft that contacts later, until the dynamic response feedback loop detects that the maximum deviation between the dynamic impedance derivatives of each independent drive shaft is less than 0.05, thereby completing the closed-loop correction control of multi-axis plane parallelism.

[0030] Preferably, the current amplitude of the pre-saturation compensation pulse is 20% to 40% of the rated current of the servo motor; the step S105 first inputs the elastic pre-saturation feedforward command to neutralize the elastic deformation dead zone of the transmission chain, which includes: in the range of 2.0mm to 5.0mm before approaching the surface micro-contact, inputting the pre-saturation compensation pulse to the current loop of each independent drive shaft, so that the electromagnetic torque of the servo motor and the deformation resistance torque of the transmission chain reach a balanced state, so as to offset the physical transmission error without changing the mechanical topology.

[0031] A mold closing and alignment compensation control system for a new energy battery separator includes: The data acquisition module is used to acquire the real-time current and position increment of each independent drive shaft in the controlled drive system, and to calculate the real-time torque change and dynamic impedance derivative accordingly. The baseline correction module is used to construct a sliding time window containing five consecutive control action cycles. It calculates the baseline drift slope based on the steady-state impedance bias baseline at the end of each action within the sliding time window, and translates and corrects the dynamic impedance reference envelope before the start of the next control action cycle to obtain the corrected dynamic impedance reference envelope. The parameter generation module is used to calculate the discrete time-domain integral value of the real-time current within the time window when the real-time torque change exceeds the lower limit of torque fluctuation and the position increment is less than the preset dead zone reference value, and generate energy flow accumulation parameters. The feedforward instruction module is used to calculate the basic pose compensation instruction based on the difference between the dynamic impedance derivative and the corrected dynamic impedance reference envelope, and to calculate the elastic presaturation feedforward instruction based on the energy flow accumulation parameter and the system stiffness mapping matrix. The alignment control module is used to set the elastic pre-saturation feedforward command to be output before the basic pose compensation command in terms of timing. The elastic pre-saturation feedforward command is input first to neutralize the dead zone of elastic deformation of the transmission chain. In response to each independent drive shaft reaching the rigidity critical state, the basic pose compensation command is triggered to complete the parallel alignment.

[0032] Example 1: When the system faces the continuous operation of a new battery separator injection production line with high frequency and micron-level closing precision, the four independent servo mold closing execution axes, which are distributed at the four corners of the mold and have established a data mapping association with the main controller through an industrial Ethernet bus, generate non-uniform heat accumulation under long-term operation. This causes the overall system to be asymmetrically tilted and unbalanced in the micro-pinch pressing zone at the end of the mold. As a result, the position adjustment signal sent by the main controller is preferentially converted into nonlinear elastic deformation loss inside the ball screw and reducer before being transmitted to the mold under force. Because the original closed-loop architecture cannot detect this deformation, response lag and high-frequency accumulation of commands occur. This induces mechanical self-excited oscillation between multiple servo drive axes and causes the battery separator thickness consistency to deviate from the set process parameters.

[0033] When the controlled drive system enters the set micro-pinch zone, the main controller synchronously reads the native sampling data of the current loop inside the four independent servo drives at a communication cycle ranging from 0.25ms to 1.0ms to obtain the real-time current and position increment of each independent drive axis, and calculates the real-time torque change and dynamic impedance derivative accordingly. The baseline drift slope is calculated by using the steady-state impedance bias baseline at the end of each action within a sliding time window containing five consecutive control action cycles. This allows for the translation and correction of the dynamic impedance reference envelope before the start of the next control action cycle. When the transient real-time torque change in the operating state exceeds the lower limit of torque fluctuation obtained by reconstructing the inherent cogging torque and static friction of the guide post, and the position increment... When the value is less than the set dead zone reference value, the main controller activates the elastic pre-saturation feedforward loop of the drive chain and calculates the discrete integral value of the real-time current in the time domain when the increment at the current position remains below the set dead zone reference value, in order to generate the energy flow accumulation parameter. The calculation formula is as follows: ,in, For energy flow accumulation parameters, and This represents the torque current feedback value for each sampling period. Based on the underlying sampling period set by the main controller, this invention defines the current integral value as an energy flow accumulation parameter. Its physical meaning lies in characterizing the mechanical power consumption of the servo motor during position locking. Since the linear motion of the drive shaft is halted within this micro-range, the electromagnetic energy output by the motor is reflected through current feedback as the continuous accumulation of motor current along the time axis. According to the linear mapping relationship between electromagnetic torque and current, this accumulation directly corresponds to the strain energy density of the transmission component under fixed physical constraints. To eliminate dimensional differences, this invention maps the charge accumulation value to the elastic strain corresponding to the dimensional conversion coefficient in the system stiffness mapping matrix. Equivalent mechanical force parameters are used to achieve the quantitative conversion from electrical quantities to mechanical elastic potential energy. Based on the basic electromechanical energy conversion principle, within the physical boundary where the displacement increment remains below the dead zone reference value, the spatial linear motion of the transmission components stops, and all the mechanical work output by the servo motor is converted into the elastic potential energy inside the transmission chain components. Based on the fixed linear positive correlation between electromagnetic torque and torque current in the steady-state range, discrete integration of torque current within a specified time window is used to directly output the dimension-reduced equivalent characteristic value representing the accumulated elastic deformation stress inside the compressed transmission components. The main controller extracts the pre-stored system stiffness mapping matrix and accumulates the energy flow parameters. The corresponding coefficients of the stiffness mapping matrix are multiplied to calculate the output elastic presaturation feedforward command, and the dynamic impedance derivative is used as the basis. The difference between the base pose compensation command and the modified dynamic impedance reference envelope is used to calculate the basic pose compensation command. In terms of timing logic, the elastic pre-saturation feedforward command is output to the main circuit of the corresponding servo clamping execution axis before the basic pose compensation command to neutralize the elastic deformation dead zone of the transmission chain.

[0034] In response to the torque change rate of each independent drive shaft being greater than or equal to the set rigidity judgment threshold for three consecutive control action cycles under the elastic pre-saturation feedforward command prior to the pose compensation output, the control loop confirms that the elastic deformation dead zone of the transmission chain has been eliminated and outputs a rigidity critical state confirmation signal. Simultaneously, the basic pose compensation command is released. By adjusting the position loop gain of each independent drive shaft, the dynamic impedance derivative of the drive shaft that contacts earlier is reduced to increase virtual flexibility, and the approach speed of the drive shaft that contacts later is increased. Until the dynamic response feedback loop detects that the maximum deviation between the dynamic impedance derivatives of each independent drive shaft is less than 0.05, the closed-loop correction control of multi-axis planar parallelism is completed. This eliminates the phase lag in the transmission of control signals to the mold surface and, without changing the physical transmission mechanism, blocks the accumulation of position commands in the extremely narrow pressing interval. This allows the multi-axis synchronous correction action to break free from the rigid constraints of the physical transmission mechanism. The entire control loop flows stably after command synchronization and ensures that the thickness consistency of the formed battery separator meets the set target.

[0035] Example 2: When the controlled drive system runs in a multi-axis servo linkage control test platform to test the mold closing alignment accuracy, the control test platform includes four independent corner servo mold closing axes. Each servo mold closing axis is equipped with a position sensor with a measurement range of 0 to 50 mm, a resolution of 0.01 μm, and a sampling frequency of 10 kHz. It is also equipped with a current sensor with a measurement range of -20 A to +20 A, an accuracy of 0.01 A, and a sampling frequency of 20 kHz to collect multi-channel bus feedback data streams. To simulate non-ideal interference conditions under production conditions, Gaussian white noise with a signal-to-noise ratio of 25 dB is superimposed on the main circuit signal source, and a non-uniform temperature distribution of 0.02 °C per second is introduced by an external heating source. This forms a raw physical signal stream containing high-frequency disturbances and time-varying tilt load at the input end. The setting of the communication cycle in the control loop depends on the trade-off between the dynamic response timeliness of the alignment control and the bus bandwidth load. When the dynamic impedance derivative... When the transient rate of change increases, in order to avoid control phase lag under the discrete sampling theorem and block the high-frequency accumulation of position commands, the communication cycle needs to be reduced to improve the refresh rate of the servo loop. When the system faces the condition that the position deviation speed at the clamping cutoff boundary reaches 0.1 μm per millisecond, the communication cycle is determined to be 0.5 ms to establish a balance between the control loop adjustment efficiency and the bus communication load. Under the same environmental interference background, the experiment verifies the control performance by constructing a problem intensity gradient control system. The initial pose offset between the four servo clamping axes is set to a low intensity gradient of 0.2 mm, a medium intensity gradient of 0.5 mm, and a high intensity gradient of 1.2 mm, respectively. Under each gradient... The control group, which only uses basic position closed-loop feedback, the out-of-range control group with the deformation dead zone reference value set to exceed the upper limit of the protection range (5.0 μm), and the experimental group using a complete feedforward compensation control method were run separately. After starting under a high-intensity off-center load gradient of 1.2 mm, the original sensor feedback signal of the control group showed a response hysteresis. Due to the elastic deformation of the ball screw absorbing the initial command, the servo axis generated a position oscillation with an amplitude of 18.4 μm after entering the micro-pinch zone. The high-frequency accumulation of commands caused the drive shaft torque current feedback to exhibit periodic electromagnetic large-amplitude oscillations. Correspondingly, when the experimental group faced the above-mentioned high-intensity off-center load, the main controller, after entering the micro-pinch zone, detected a position increment. When the deformation dead zone is less than the set reference value of 1.5 μm and the real-time torque change exceeds the lower limit of torque fluctuation, the time-domain discrete integration is initiated and the energy flow accumulation parameter is calculated in real time. The calculation formula is as follows: ,in, For energy flow accumulation parameters, and This represents the torque current feedback value for each sampling period. The underlying sampling period is set by the main controller.

[0036] As the integration time domain progresses, the energy flow accumulation parameters calculated by the experimental group... Starting from 0.12, the value rises non-linearly and converges to an integral saturation stability value of 3.45 at the instant the transmission chain reaches the rigidity critical state. Based on this data, the main controller calculates and outputs the corresponding elastic pre-saturation feedforward command to neutralize the elastic deformation dead zone of the transmission chain in advance. The feedback loop releases the basic pose compensation command after detecting that the torque change rate of the four drive shafts is greater than or equal to the rigidity judgment threshold for three consecutive control action cycles. In the test group, the maximum deviation of the mold plane parallelism under low, medium and high strength gradients is stable at 0.01mm, 0.02mm and 0.03mm respectively. The thickness consistency error of the formed battery separator is 1.2μm under the highest off-center load. However, in the out-of-range control group, when the deformation dead zone reference value is set to 5.0μm, the energy flow accumulation parameter is significantly exceeded due to the large deviation from the reasonable elastic deformation range. The calculation of the feedforward compensation amount that leads to excessive extension and output deviation induces a mechanical transient impact stress of 450N at the moment of eliminating the deformation dead zone, resulting in nonlinear rigid impact damage to the mold guide mechanism. This confirms that the reference value range of the deformation dead zone is a working window determined by balancing the deformation compensation of the transmission chain with the impact isolation of the mechanism through in-depth engineering.

[0037] Based on the data evolution patterns of the above-mentioned multiple sets of gradient control experiments, the verification experiment confirms that in industrial field environments containing nonlinear deformation of the transmission chain and high-frequency noise interference, the technical path of reducing position response hysteresis and uneven mechanical oscillation loss in multi-axis synchronous systems by calculating energy flow accumulation parameters in real time and reconstructing the timing of secondary control commands reduces position response hysteresis and uneven mechanical oscillation loss in multi-axis synchronous systems. This data-driven position feedforward and rigid critical state identification mechanism does not rely on changes in the physical structure of the mold. Instead, the main controller performs digital domain calculations on the original parameters of the electrical loop at the component level, converting nonlinear deformation hysteresis into position loop gain and feedforward pulse distribution, thereby correcting the multi-axis planar parallelism deviation within the set position. After command synchronization, the control loop flows stably and the thickness consistency of the formed battery separator reaches the set target.

[0038] Example 3: This example combines Figures 1 to 2 This document describes a method and system for mold closing and alignment compensation control of a new energy battery separator mold, such as... Figure 1 As shown, in step S101, the real-time current and position increment are obtained, and the real-time torque change and dynamic impedance derivative are calculated. Then, the system enters step S102, calculates the baseline drift slope according to the sliding time window, and then translates and corrects the dynamic impedance reference envelope. Next, step S103 is executed to calculate the discrete time domain integral value of the real-time current to generate the energy flow accumulation parameter, and then proceeds to step S104 to calculate the basic pose compensation command and the elastic pre-saturation feedforward command. Finally, in step S105, the operation of first inputting the pre-saturation feedforward command and then outputting the pose compensation command after reaching the rigid critical state is executed.

[0039] like Figure 2As shown, the aforementioned real-time torque change is compared with the lower limit of torque fluctuation, thereby activating the built-in impedance digital low-pass filter to truncate high-frequency impedance abrupt change data, extracting pure surface dynamic impedance features. By calculating the difference between the deviation from the spatial geometric center, the result of the pure surface dynamic impedance features deviating from the center of the corrected dynamic impedance reference envelope is obtained, thus outputting an asymmetric torsion vector. This asymmetric torsion vector is input to the decoupling matrix model for mapping. The decoupling matrix model is generated according to the spatial cross stiffness calibration procedure. In the static state where the mold is unloaded, the main controller sequentially inputs test torque commands with stepped amplitudes to each independent servo mold-closing axis, simultaneously reading the small spatial displacements fed back by four position sensors. The quotient of a single test torque change and the corresponding small displacements at the four measuring points is calculated, and column vectors are extracted and concatenated to construct a 4×4 dimension cross compliance matrix. The main controller uses the Gaussian elimination algorithm to perform the inverse matrix operation on the cross compliance matrix, extracting the system cross stiffness matrix and storing it in the storage unit as the decoupling matrix model. The asymmetric torsion vector and the decoupling matrix model undergo direct matrix multiplication to eliminate... In addition to the planar tilting and eccentric loading internal forces caused by multi-axis coupling, the transient displacement compensation of each independent drive shaft is extracted and superimposed on the position loop as a feedforward correction. This signal is then split into two time-separated control signals. The first-level control signal is a pre-saturation compensation pulse, output within the micro-range before approach. Controlled current induces pre-elastic deformation in the transmission components, offsetting physical transmission errors and neutralizing the dead zone of elastic deformation in the transmission chain. From this branch, a hysteresis contact drive shaft control shunt is derived, adjusting the position loop gain of the corresponding drive shaft to increase its approach speed. The second-level control signal of the other branch is a position compensation pulse, which is triggered to output in response to a step change in the torque slope of the drive shaft and the rigidity critical state. This pulse drives the controlled surface to fine-tune the parallelism. From this branch, the control current of the previously contacted drive shaft is shunted, and the position loop gain of the drive shaft is adjusted to reduce its dynamic impedance derivative to increase virtual flexibility. Finally, combined with the aforementioned current shunting control, closed-loop correction control of multi-axis plane parallelism is executed, so that the dynamic impedance derivatives of each independent drive shaft are forced to converge until the maximum deviation detected by the dynamic response feedback loop is lower than the tolerance limit.

[0040] Example 4: When the system faces continuous operation of a new energy battery separator injection molding production line, the four independent servo mold closing execution axes, distributed at the four corners of the mold and connected to the main controller via an industrial Ethernet bus, generate non-uniform heat accumulation during operation. This causes asymmetric tilting load in the pressing zone at the mold closing end, resulting in the position adjustment signal issued by the main controller being converted into nonlinear elastic deformation loss in the ball screw and reducer before being transmitted to the mold. Due to the lack of means to measure this deformation, the position closed-loop architecture experiences position response lag and position command accumulation. During the mold return and disengagement phase after each set of closed pressing actions, the data processing chip in the device synchronously collects the steady-state impedance data of the four independent drive shafts under no-load conditions for five consecutive control action cycles and stores it in the ring memory. Based on the discrete steady-state impedance data within this time window, the slope of the straight line is calculated to determine the baseline drift slope. Before the start of the next control action cycle, the original dynamic impedance reference envelope is shifted and corrected in the same direction according to the product of the baseline drift slope and the action timing, so that the time-varying drift amount is separated from the measured current signal, and the lower limit of torque fluctuation is kept aligned with the time-varying electrical parameters.

[0041] After the aforementioned envelope surface is shifted and corrected in the same direction, the main controller compares the real-time torque current feedback with the corrected dynamic impedance reference envelope surface and the lower limit of torque fluctuation. When it detects that the transient torque change caused by contact during operation exceeds the lower limit of torque fluctuation after the shift correction and the position increment fed back by the position sensor is exceeded, the main controller will take action. When the data processing chip maintains a value 1.5 μm below the set deformation dead zone reference value, it automatically extracts a time-domain sampling window that meets the trigger constraints and provides feedback on the torque current within that window. Accumulated calculations to solve for energy flow accumulation parameters The calculation formula is as follows: ,in, For energy flow accumulation parameters, and This represents the torque current feedback value for each sampling period. The main controller sets a low-level sampling period to determine the rigid critical state after the transmission chain completely eliminates the deformation dead zone. The main controller utilizes the maximum value of the first derivative of the current fed back when each independent drive shaft pushes the rigid stop block at a constant speed under no-load conditions, and saves 80% of this maximum derivative value as the rigidity threshold for closed-loop determination. When the energy flow accumulation parameter calculated using the method claimed in this invention through the aforementioned discrete sampling and accumulation is applied... When the corresponding servo clamping axis outputs a feedforward current pulse, and the torque change rate monitored and fed back by the subsequent sensors is greater than or equal to the rigidity judgment threshold for three consecutive control action cycles, the control loop confirms that the elastic deformation of the transmission chain of that axis has been eliminated. The main controller then releases a position compensation pulse as the second level of the two-level command pulse as a basic pose compensation command. By adjusting the position loop gain of each axis, the pose deviation is corrected in the clamping plane. The pre-injected pre-saturated energy is used to offset the local transmission strain generated by the ball screw and reducer, so that the basic pose compensation command is converted into a rigid displacement of the mold top surface at the moment of release. This eliminates the accumulation of position commands and oscillation of the multi-axis system caused by position response hysteresis, and keeps the thickness consistency error of the formed battery separator within the set industrial specification range.

[0042] Example 5: When the system faces an initial stiffness mismatch caused by equipment assembly or mold replacement, the hardware transmission gap and sensor zero position generate secondary time delay losses in the initial stage of torque, causing the stiffness matrix stored in the main controller to deviate from the physical mechanism. The issued feedforward pulses cannot offset the deformation of each axis. In the no-load alignment state, the main controller drives the four independent servo mold closing execution axes at a constant speed. Approaching a rigid blocking block, where the velocity The value is set to a range of 1 mm / s to 5 mm / s. The position sensor and current sensor synchronously collect data from each loop. The main controller extracts the torque loop output value and calculates the dynamic impedance damping slope during the deformation stage by combining it with the position deviation. Dynamic impedance damping slope The calculation formula is as follows: ,in, The slope of the dynamic impedance damping, subscript Indicates the slope. The value represents the change in torque current feedback, indicated by the subscript. Indicates the torque axis component. As the position pulse change, the main controller writes the calculated slope value into the system stiffness mapping matrix to update the control gain of the energy flow control loop, so that the output of the position feedforward algorithm matches the elastic strain of the current physical mechanism.

[0043] When the system faces long-term operating cycles leading to component wear and bearing preload decay, the physical frictional resistance of each independent drive shaft exhibits time-varying drift with operating time. Upon receiving a start signal, the main controller regulates the four independent servo mold-closing axes to run one stroke in the mold-disengaged state. Sensors collect the frictional torque-current sequence under no-load conditions to calculate the sliding average value as a function of position, storing this average value in a non-volatile memory area to update the torque fluctuation lower limit. Simultaneously, the difference between the torque fluctuation lower limit of the current operating cycle and the previous operating cycle is calculated, along with the coefficient of variation of this difference. coefficient of variation The calculation formula is as follows: ,in, The coefficient of variation is represented by the subscript. Indicates variability. The standard deviation of the difference. The mean of the differences is the coefficient of variation obtained during calculation. When the torque fluctuation is less than the preset fluctuation threshold, the main controller updates the lower limit of torque fluctuation through linear interpolation to correct the torque comparison benchmark for subsequent sampling periods, so that the consistency error of the molded battery separator thickness is kept within the set specification range.

[0044] Example 6: When the system faces the situation of replacing batches of materials for new energy battery separators, and the initial thickness and compressive modulus of each batch of materials fluctuate, the physical resistance generated by different batches of materials on the mold clamping force surface has nonlinear differences. This causes the position response of each independent drive shaft in the controlled drive system to deviate from the load distribution in a non-monotonic manner. The main controller drives four independent servo mold clamping shafts at a constant speed under no-load conditions. Approaching the mold limit block, where speed Set to 2mm / s, the real-time position changes of each axis are collected at high frequency by the position sensor. The processing chip in the main controller determines the elastic physical dissipation parameters of the transmission chain under the current batch of material conditions by calculating the ratio of the torque current feedback quantity to the corresponding position change quantity. The calculated elastic dissipation parameters of each independent drive shaft are combined into matrix coefficients and written into the pre-stored system stiffness mapping matrix. This allows the output power of the first-stage feedforward current pulse in the two-stage pulse timing to be adaptively adjusted according to the fluctuation of material properties, thereby eliminating the position response hysteresis caused by changes in material thickness.

[0045] When the dead zone reference value at the mold closing boundary is adjusted due to sudden changes in batch pressure, the main controller collects the feed depth and torque axis feedback components transmitted by the position sensor, calculates the first derivative of the torque axis current feedback through discrete differential calculation, and the processing chip uses the first derivative value as an input parameter to the internally stored threshold adjustment function to determine the dead zone correction factor. Dead zone correction factor The calculation formula is as follows: ,in, Dead zone correction factor, subscript Indicates the reference; This refers to the torque current feedback for each axis. Sampling time, Let be the reference constant for the rate of change of torque current, where is the reference constant. The physical characteristic values ​​are obtained by conducting step load tests on each drive shaft on a standard rigid test block. During the step increase of torque current generated by the transmission chain, which is impeded by the rigid test block, the upper limit of the average rate of change of torque current per unit time is measured. This value is pre-calibrated based on the physical inertia of the transmission chain of different batches of molds and is stored in the system storage unit as a critical reference rate for judging whether the system is on the edge of oversaturation. The main controller will then use the dead zone correction factor. Multiplying the energy flow accumulation parameter obtained from the previous cycle calculation, the integral time domain of the subsequent control action cycle is dynamically scaled. During the sharp torque rise, the integral time domain is spontaneously shortened to limit the oversaturation of the feedforward motor. When the feedback loop completes the multi-axis plane parallelism closed correction control action, it detects that the maximum deviation between the dynamic impedance derivatives of the four independent drive shafts converges to 0.03, and the thickness consistency error of the formed battery separator remains within the set process index of 1.1 μm.

[0046] 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.

[0047] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for mold closing and alignment compensation control of a new energy battery separator mold, characterized in that, Includes the following steps: Step S101: Obtain the real-time current and position increment of each independent drive shaft in the controlled drive system, and calculate the real-time torque change and dynamic impedance derivative accordingly. Step S102: Construct a sliding time window containing 5 consecutive control action cycles, calculate the baseline drift slope based on the steady-state impedance bias baseline at the end of each action within the sliding time window, and translate and correct the dynamic impedance reference envelope before starting the next control action cycle. Step S103: When the real-time torque change exceeds the lower limit of torque fluctuation and the position increment is less than the preset dead zone reference value, calculate the discrete time domain integral value of the real-time current within the time window when the current position increment is lower than the preset dead zone reference value, and generate the energy flow accumulation parameter. Step S104: Calculate the basic pose compensation command based on the difference between the dynamic impedance derivative and the corrected dynamic impedance reference envelope, and calculate the elastic presaturation feedforward command based on the energy flow accumulation parameter and the system stiffness mapping matrix. In step S105, the elastic pre-saturation feedforward command is configured to be output before the basic pose compensation command in terms of timing. The elastic pre-saturation feedforward command is input first to neutralize the dead zone of elastic deformation of the transmission chain. In response to each independent drive shaft reaching the rigid critical state, the basic pose compensation command is triggered to complete the parallel alignment.

2. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, The calculation of the real-time torque change of each independent drive shaft in step S101 includes: scaling the real-time current to obtain the real-time torque value, calculating the difference between the real-time torque values ​​between adjacent control cycles as the real-time torque change; the dynamic impedance derivative is obtained by calculating the rate of change of the real-time torque change relative to the position change; the lower limit of torque fluctuation is obtained by acquiring the fixed torque fluctuation corresponding to the inherent cogging torque and static friction force of the system, and reconstructing the amplitude of the fixed torque fluctuation.

3. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, Step S105, which uses a decoupling compensation matrix to process the asymmetric torsional vector and generate a two-level timing command pulse containing a pre-saturation compensation pulse and a position compensation pulse, includes the following sub-steps: Step S1051, inputting the asymmetric torsional vector generated by the difference mapping into the decoupling matrix model mapping to eliminate the planar tilting eccentric load internal force caused by multi-axis mutual coupling; Step S1052, extracting the transient displacement compensation amount of each independent drive shaft after decoupling matrix model mapping, and superimposing the transient displacement compensation amount of each independent drive shaft as a feedforward correction amount onto the position compensation of the controlled drive system. In the loop; in step S1053, the feedforward correction is split into two time-separated control signals. The first-level control signal is a pre-saturation compensation pulse, which is used to output within the micro-range before approaching. Its pulse amplitude has a specific proportional relationship with the deformation stiffness of the transmission chain, so as to use the controlled current to make the transmission components produce pre-elastic deformation. The second-level control signal is a position compensation pulse, which is triggered to output in response to the step of the drive shaft torque slope and reaching the rigid critical state, so as to drive the controlled surface to fine-tune the parallelism, so that the dynamic impedance derivatives of each independent drive shaft are forced to converge.

4. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, Step S105, which triggers the output of the basic pose compensation command in response to each independent drive shaft reaching the rigid critical state, includes the following sub-steps: Step S1054, during the output of the pre-saturation compensation pulse generated according to the elastic pre-saturation feedforward command, the feedback torque change rate of each independent drive shaft is monitored in real time; Step S1055, the feedback torque change rate is compared with the preset rigidity judgment threshold in real time. When the feedback torque change rate is greater than or equal to the preset rigidity judgment threshold for three consecutive control action cycles, it is determined that the elastic deformation dead zone of the transmission chain has been completely absorbed, and a rigidity critical state confirmation signal is output; Step S1056, the control loop stops outputting the pre-saturation compensation pulse and simultaneously releases the position compensation pulse as the basic pose compensation command, thereby cutting off the attitude compensation phase hysteresis caused by mechanical elastic physical dissipation.

5. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, The calculation of the basic pose compensation command based on the difference between the dynamic impedance derivative and the corrected dynamic impedance reference envelope includes the following sub-steps: Step S1041, compare the real-time torque change with the lower limit of torque fluctuation; Step S1042, when the real-time torque change is lower than the lower limit of torque fluctuation, determine that the current fluctuation is non-real deformation noise caused by mechanical static friction, and the control loop maintains the command output of the previous control cycle, shielding the interference to the position loop; Step S1043, when the real-time torque change is greater than or equal to the lower limit of torque fluctuation, use this as the filtering environment to activate the built-in impedance digital low-pass filter, cut off high-frequency impedance abrupt change data higher than the inherent resonant frequency of the mechanical system, and extract the pure surface dynamic impedance features; Step S1044, define the difference vector between the pure surface dynamic impedance features and the spatial geometric center of the corrected dynamic impedance reference envelope as an asymmetric torsion vector, and use this difference to calculate the basic pose compensation command.

6. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, Step S102 extracts the steady-state impedance bias baseline at the end of each action within the sliding time window, which includes the following sub-steps: Step S1021, during the voltage holding steady-state phase of each control action cycle, the steady-state voltage and steady-state current of each independent drive shaft are collected; Step S1022, the steady-state impedance observation value of each cycle is calculated based on the steady-state voltage and steady-state current; Step S1023, within the sliding time window, the steady-state impedance observation values ​​of five consecutive cycles are subjected to mean filtering to remove transient impedance fluctuations caused by microscopic non-uniformity of material thickness, thereby obtaining a steady-state impedance bias baseline that can characterize the slow thermal accumulation effect of the long-cycle temperature field.

7. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, The controlled drive system includes a main controller that establishes data mapping associations via an industrial Ethernet bus and four independent servo drives; The communication cycle between the main controller and the servo drive is 0.25ms to 1.0ms; the real-time current of each independent drive axis in the controlled drive system in step S101 includes: the main controller synchronously reads the original sampling data of the current loop inside the four independent servo drives through the industrial Ethernet bus at the communication cycle.

8. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, The decoupling compensation matrix is ​​used to process the asymmetric torsional vector to make the dynamic impedance derivatives of each independent drive shaft converge. This includes adjusting the position loop gain of each independent drive shaft to reduce the dynamic impedance derivative of the drive shaft that contacts earlier to increase virtual flexibility, and increasing the approach speed of the drive shaft that contacts later. This process continues until the dynamic response feedback loop detects that the maximum deviation between the dynamic impedance derivatives of each independent drive shaft is less than 0.05, thus completing the closed-loop correction control of multi-axis plane parallelism.

9. The method for mold closing and alignment compensation control of a new energy battery separator mold according to claim 1, characterized in that, The current amplitude of the pre-saturation compensation pulse is 20% to 40% of the rated current of the servo motor; in step S105, the elastic pre-saturation feedforward command is first input to neutralize the elastic deformation dead zone of the transmission chain, including: in the range of 2.0mm to 5.0mm before approaching the surface micro-contact, the pre-saturation compensation pulse is input to the current loop of each independent drive shaft, so that the electromagnetic torque of the servo motor and the deformation resistance torque of the transmission chain reach a balanced state, so as to offset the physical transmission error without changing the mechanical topology.

10. A mold closing alignment compensation control system for a new energy battery separator mold, used to implement the mold closing alignment compensation control method for a new energy battery separator mold as described in claim 1, characterized in that, include: The data acquisition module is used to acquire the real-time current and position increment of each independent drive shaft in the controlled drive system, and to calculate the real-time torque change and dynamic impedance derivative accordingly. The baseline correction module is used to construct a sliding time window containing five consecutive control action cycles. It calculates the baseline drift slope based on the steady-state impedance bias baseline at the end of each action within the sliding time window, and translates and corrects the dynamic impedance reference envelope before the start of the next control action cycle to obtain the corrected dynamic impedance reference envelope. The parameter generation module is used to calculate the discrete time-domain integral value of the real-time current within the time window when the real-time torque change exceeds the lower limit of torque fluctuation and the position increment is less than the preset dead zone reference value, and generate energy flow accumulation parameters. The feedforward instruction module is used to calculate the basic pose compensation instruction based on the difference between the dynamic impedance derivative and the corrected dynamic impedance reference envelope, and to calculate the elastic presaturation feedforward instruction based on the energy flow accumulation parameter and the system stiffness mapping matrix. The alignment control module is used to set the elastic pre-saturation feedforward command to be output before the basic pose compensation command in terms of timing. The elastic pre-saturation feedforward command is input first to neutralize the dead zone of elastic deformation of the transmission chain. In response to each independent drive shaft reaching the rigidity critical state, the basic pose compensation command is triggered to complete the parallel alignment.

Images