A high-precision thickness tolerance thickness detection system for PET release film

Abstract

This application relates to the field of online detection technology for polymer films and discloses a high-precision thickness tolerance PET release film thickness detection system. This invention solves the technical problem of interference between optical polarization artifacts and physical necking thinning caused by tension stretching in release film production. By introducing the large deformation continuous medium volume conservation theory, optical phase distortion and mechanical deformation are simultaneously stripped from the original optical path difference data, restoring the stress-free true material reference thickness that reflects the pure melt distribution. The collaborative feedback module uses a physical dimension isolation mechanism to separate the quality deviation and deformation deviation, and independently generate opening fine-tuning commands and feedforward torque compensation commands, breaking the control limitations of conventional single closed loops that cause coupled oscillations. This invention achieves complete decoupling of extrusion quality control and mechanical tension suppression, eliminating erroneous adjustment defects caused by false thickness feedback from the physical root, and ensuring the continuous steady-state production of high-precision tolerance films.

Description

Technical Field

[0001] This application relates to the field of online detection technology for polymer films, and in particular to a high-precision thickness tolerance PET release film thickness detection system. Background Technology

[0002] In the high-speed roll-to-roll production process of polyethylene terephthalate (PET) release film, the film winding and traction processes are crucial for ensuring the flatness and roll quality of the finished product. As the production process continues, the actual roll diameter of the winding shaft exhibits a continuous non-linear increasing trend. To maintain the preset tension taper, the programmable logic controller (PLC) at the bottom of the production line must frequently apply fine-tuning commands to the operating torque and speed of the winding and traction motors. Simultaneously, the various levels of transition guide rollers distributed along the entire transmission path inevitably suffer from accumulated assembly errors, mechanical eccentricity, and slight vibrations under high-speed operation. The physical vibrations of the aforementioned mechanical transmission devices and the dynamic adjustments of the electrical control system superimpose each other, causing the mechanical tension experienced by the film during longitudinal transmission and lateral flattening to be unable to remain absolutely constant. Instead, it exhibits complex high-frequency dynamic fluctuations along the time and space axes. These high-frequency tension fluctuations constitute the complex underlying physical operating environment that online film quality monitoring must address.

[0003] Polyethylene terephthalate (PET), as a typical crystalline polymer material, exhibits a significant stress birefringence effect when subjected to external mechanical stress. When the mechanical tension on the production line changes abruptly, the orientation of the polymer chain segments within the film changes instantaneously, directly causing a dynamic drift in the material's intrinsic optical refractive index. Existing online optical thickness measurement equipment generally uses a static mathematical model that divides the measured optical path difference by a preset fixed refractive index constant when calculating the underlying physical data. Furthermore, the thickness measurement control system and the production line tension control system are completely isolated from each other. Because the thickness gauge cannot sense and synchronize the current mechanical strain characteristics of the film in real time, the physical change in refractive index induced by tension fluctuations is directly miscalculated by the static algorithm as a physical increase or decrease in the film's thickness. This pseudo-thickness fluctuation caused by stress distortion not only masks the true thickness tolerance defects but also misleads the upstream extrusion die into performing incorrect closed-loop adjustments. When downstream customers use release films with hidden thickness defects in precision die-cutting or multi-layer coating and lamination processes, undetected thickness deviations can disrupt the stress balance at the interface, leading to serious lamination bubbles or die-cutting accidents. Given current technology, there is an urgent need to establish a data fusion mechanism that bridges the optical detection and mechanical control domains to dynamically compensate for refractive index parameters, thereby completely eliminating the coupling interference of mechanical tension on the optical thickness measurement logic. Summary of the Invention

[0004] This application proposes a high-precision thickness tolerance PET release film thickness detection system to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, this application adopts the following technical solution: a high-precision thickness tolerance PET release film thickness detection system, comprising: a spatiotemporal mapping module, a refractive index calculation module, a pseudo-thickness peeling module, and a collaborative feedback module, wherein; The spatiotemporal mapping module acquires the spindle tension scalar data and the original optical path difference data, establishes a two-dimensional spatial coordinate system, and discretizes the two-dimensional spatial coordinate system to construct a two-dimensional physical grid. The time warping algorithm is used to align the spindle tension scalar data and the original optical path difference data to the grid nodes of the two-dimensional physical grid, and combines them to generate a force-light synchronization mapping feature set. The refractive index calculation module receives the force-optical synchronous mapping feature set generated by the spatiotemporal mapping module, then extracts the principal axis tension scalar data contained in the force-optical synchronous mapping feature set, and calls the preset stress optics tensor polynomial to perform polynomial algebra operations on the principal axis tension scalar data to calculate the dynamic drift of the refractive index. The pseudo-thickness stripping module receives the dynamic refractive index drift calculated by the refractive index calculation module, extracts the original optical path difference data contained in the force-light synchronous mapping feature set generated by the spatiotemporal mapping module, and then inputs the preset intrinsic Poisson's ratio, the original optical path difference data and the dynamic refractive index drift into the preset thickness restoration equation for nonlinear algebraic solution to calculate the stress-free real material reference thickness. The collaborative feedback module receives the stress-free real material reference thickness calculated by the pseudo-thickness stripping module and the force-light synchronous mapping feature set generated by the spatiotemporal mapping module. Then, it performs a difference operation between the stress-free real material reference thickness and the preset target thickness to generate an opening fine-tuning command. It extracts the original optical path difference data contained in the force-light synchronous mapping feature set, divides the original optical path difference data by the preset initial static refractive index to calculate the apparent physical thickness, and performs a difference operation between the apparent physical thickness and the stress-free real material reference thickness to generate a feedforward torque compensation command.

[0006] Furthermore, the spatiotemporal mapping module acquires the spindle tension scalar data and the original optical path difference data, establishes a two-dimensional spatial coordinate system, and discretizes the two-dimensional spatial coordinate system to construct a two-dimensional physical mesh. The specific operations are as follows: Acquire spindle tension scalar data, raw optical path difference data, spindle transmission linear velocity data, lateral position pulse data, and absolute temperature field data; Assign absolute timestamps to the spindle tension scalar data, raw optical path difference data, spindle transmission linear velocity data, lateral position pulse data, and absolute temperature field data; Extract the horizontal position pulse data with absolute timestamps, convert the horizontal position pulse data with absolute timestamps into absolute horizontal coordinates, and set the absolute horizontal coordinates as the horizontal axis; Extract the spindle transmission line velocity data with absolute timestamps, perform integration on the spindle transmission line velocity data with absolute timestamps to calculate the integral displacement value, and set the integral displacement value as the vertical axis; Establish a two-dimensional spatial coordinate system by combining the horizontal and vertical axes; The two-dimensional spatial coordinate system is divided according to a preset discrete resolution to construct a two-dimensional physical grid containing grid nodes.

[0007] Furthermore, the spatiotemporal mapping module uses a time warping algorithm to align the spindle tension scalar data and the original optical path difference data to the grid nodes of the two-dimensional physical grid, combining them to generate a force-light synchronization mapping feature set. The specific operation is as follows: Extract the absolute temperature field data with absolute timestamps, and extract the dynamic Young's modulus corresponding to the absolute temperature field data with absolute timestamps from the preset temperature stiffness mapping table. The mechanical tensile strain term is calculated by dividing the spindle tension scalar data with absolute timestamps by the product of the preset nominal width, preset nominal thickness, and dynamic Young's modulus. The temperature difference variable is calculated by subtracting the preset reference temperature data from the absolute temperature field data with absolute timestamps. The thermal expansion strain term is calculated by multiplying the preset thermal expansion coefficient by the temperature difference variable. The comprehensive strain ratio is calculated by adding the constant, the mechanical tensile strain term, and the thermal expansion strain term. The physical transmission length is calculated by multiplying the spindle transmission linear velocity data with absolute timestamps by the comprehensive strain ratio and integrating the result of the multiplication. The distance penalty is calculated by subtracting the physical transmission length from the preset geometric span. Calculate the differential rate of change of the spindle tension scalar data with absolute timestamps and the differential rate of change of the raw optical path difference data with absolute timestamps; Calculate the Hilbert norm between the differential rate of change of the spindle tension scalar data with absolute timestamps and the differential rate of change of the original optical path difference data with absolute timestamps; The cost function value is calculated by adding the Hilbert norm and the distance penalty. Extract the target timestamp coordinate pairs that minimize the cost function value from the multidimensional time series matrix; The scalar data of spindle tension with absolute timestamps and the original optical path difference data with absolute timestamps are extracted to the grid nodes of the two-dimensional physical grid corresponding to the target timestamp coordinate pair, and combined to generate a force-light synchronization mapping feature set.

[0008] Furthermore, the system receives the force-optical synchronization mapping feature set generated by the spatiotemporal mapping module, and then extracts the spindle tension scalar data contained in the force-optical synchronization mapping feature set. The specific operation is as follows: Extract the principal axis tension scalar data and absolute temperature field data contained in the force-light synchronization mapping feature set; The preset zero-phase-shift low-pass filter algorithm is invoked to perform time-domain smoothing on the spindle tension scalar data, and the steady-state tension scalar data is calculated. Extract the preset nominal width, preset nominal thickness, and preset intrinsic Poisson's ratio from the preset product physical property specification database; Using absolute temperature field data as an index, the dynamic Young's modulus is extracted from the preset underlying memory. The stiffness base is calculated by multiplying the dynamic Young's modulus, the preset nominal width, and the preset nominal thickness. The initial longitudinal tensile strain is calculated by dividing the steady-state tension scalar data by the stiffness base. The shrinkage ratio is calculated by multiplying the initial longitudinal tensile strain by the preset intrinsic Poisson's ratio. The shrinkage factor is calculated by subtracting the shrinkage ratio from the constant. The squared necking factor is calculated by performing a quadratic operation on the shrinkage factor. The nominal physical cross-sectional area is calculated by multiplying the preset nominal width by the preset nominal thickness. Multiply the nominal physical cross-sectional area by the square necking factor to calculate the transient real force cross-sectional area; The transient true tensile stress is calculated by dividing the steady-state tension scalar data by the transient true cross-sectional area.

[0009] Furthermore, the refractive index calculation module calls the preset stress optics tensor polynomial to perform polynomial algebraic operations on the principal axis tension scalar data to calculate the dynamic refractive index drift. The specific operation is as follows: Using absolute temperature field data as an index, the first-order constant, the second-order constant, and the third-order constant of the multiphysics precalibration are extracted from the pre-set rheological matrix library. Calculate the first-order power value, the second-order power value, and the third-order power value of transient true tensile stress. The first-order product value is calculated by multiplying the first-order power value of the transient true tensile stress by the first-order constant of the multiphysics precalibration. The second-order product value is calculated by multiplying the second-order power value of the transient true tensile stress by the second-order constant of the multiphysics precalibration. The third-order product value is calculated by multiplying the third-order power value of the transient true tensile stress by the third-order constant of the multiphysics precalibration. The first, second, and third product values ​​are added together to calculate the initial value of the dynamic refractive index drift. The preset upper limit and lower limit of the elastic deformation zone threshold are extracted. The absolute value of the initial value of the dynamic refractive index drift is compared with the preset upper limit and lower limit of the elastic deformation zone threshold. When the absolute value of the initial value of the dynamic refractive index drift is greater than the upper limit of the preset elastic deformation zone threshold, or when the absolute value of the initial value of the dynamic refractive index drift is less than the lower limit of the preset elastic deformation zone threshold, the value of zero is set as the final dynamic refractive index drift. When the absolute value of the initial value of the dynamic refractive index drift is less than or equal to the upper limit of the preset elastic deformation zone threshold, and the absolute value of the initial value of the dynamic refractive index drift is greater than or equal to the lower limit of the preset elastic deformation zone threshold, the initial value of the dynamic refractive index drift is set as the final dynamic refractive index drift.

[0010] Furthermore, the pseudo-thickness stripping module receives the dynamic refractive index drift calculated by the refractive index calculation module and extracts the original optical path difference data contained in the force-optical synchronization mapping feature set generated by the spatiotemporal mapping module. Specifically, the operation is as follows: Extract the dynamic refractive index drift and the initial longitudinal tensile strain; Extract the raw optical path difference data contained in the force-optical synchronization mapping feature set; Extract the preset initial static refractive index; The real-time absolute refractive index is calculated by adding the preset initial static refractive index and the dynamic refractive index drift. The apparent physical thickness is calculated by dividing the original optical path difference data by the real-time absolute refractive index.

[0011] Furthermore, the preset intrinsic Poisson's ratio, original optical path difference data, and refractive index dynamic shift are input into the preset thickness reduction equation for nonlinear algebraic solution to calculate the stress-free real material reference thickness. The specific operation is as follows: Extract the preset intrinsic Poisson's ratio, add the constant to the initial longitudinal tensile strain, and calculate the absolute ratio of tensile elongation. Using the preset intrinsic Poisson's ratio as the exponent, the absolute ratio of stretching and elongation is exponentially calculated to obtain the volume conservation reverse expansion factor. The apparent physical thickness is multiplied by the volume conservation reverse expansion factor to calculate the stress-free real material reference thickness. Extract the absolute lateral coordinates and integral displacement values ​​corresponding to the grid nodes contained in the two-dimensional physical mesh; The stress-free real material reference thickness is associated with and combined with the absolute lateral coordinate and integral displacement value to output the stress-free real material reference thickness with absolute lateral coordinate and integral displacement value.

[0012] Furthermore, the system receives the stress-free real material reference thickness calculated by the pseudo-thickness stripping module and the force-light synchronous mapping feature set generated by the spatiotemporal mapping module. Then, it performs a difference operation between the stress-free real material reference thickness and the preset target thickness to generate an opening fine-tuning command. The specific operation is as follows: Receive stress-free real material reference thickness and force-optical synchronous mapping feature set; Extract the preset target thickness, subtract the stress-free real material reference thickness from the preset target thickness, and calculate the absolute geometric space error. Extract the preset mold head bolt space mapping ratio coefficient, multiply the absolute geometric space error by the preset mold head bolt space mapping ratio coefficient, calculate the displacement step command, and set the displacement step command as the opening fine adjustment command.

[0013] Furthermore, the original optical path difference data contained in the force-optical synchronization mapping feature set is extracted, and the apparent physical thickness is calculated by dividing the original optical path difference data by the preset initial static refractive index. The specific operation is as follows: The original optical path difference data contained in the force-optical synchronization mapping feature set is extracted, the preset initial static refractive index is extracted, the original optical path difference data is divided by the preset initial static refractive index, and the apparent physical thickness is calculated.

[0014] Furthermore, the collaborative feedback module calculates the difference between the apparent physical thickness and the stress-free real material reference thickness to generate a feedforward torque compensation command. The specific operation is as follows: The thickness difference is calculated by subtracting the stress-free real material reference thickness from the apparent physical thickness, and then the thickness difference is converted into an absolute difference. Extract the preset elastic moment conversion stiffness coefficient, multiply the absolute difference by the preset elastic moment conversion stiffness coefficient, and calculate the macroscopic linear failure traction force. The real-time physical roll diameter of the take-up roll is obtained, and the transient electromagnetic interference torque is calculated by multiplying the macroscopic linear destructive traction force by the real-time physical roll diameter of the take-up roll. Set the transient electromagnetic interference torque as a feedforward torque compensation command.

[0015] The beneficial effects of this invention are as follows: This invention solves the technical problem of interference between optical polarization artifacts and physical necking thinning caused by tension stretching in release film production. By introducing the theory of volume conservation of large deformation continuous medium, optical phase distortion and mechanical deformation are simultaneously separated from the original optical path difference data, restoring the stress-free true material reference thickness that reflects the pure melt distribution. The collaborative feedback module uses a physical dimension isolation mechanism to separate the quality deviation and deformation deviation, and independently generate opening fine-tuning commands and feedforward torque compensation commands, breaking the control limitations of conventional single closed loops that cause coupled oscillations. This invention achieves complete decoupling of extrusion quality control and mechanical tension suppression, eliminating the erroneous adjustment defects caused by false thickness feedback from the physical root, and ensuring the continuous steady-state production of high-precision tolerance films. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort: Figure 1 This is a system framework diagram of the present invention; Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Example like Figure 1 and Figure 2 As shown, this invention discloses a high-precision thickness tolerance PET release film thickness detection system, comprising: a spatiotemporal mapping module, a refractive index calculation module, a pseudo-thickness peeling module, and a collaborative feedback module, wherein; The spatiotemporal mapping module acquires the spindle tension scalar data and the original optical path difference data, establishes a two-dimensional spatial coordinate system, and discretizes the two-dimensional spatial coordinate system to construct a two-dimensional physical grid. The time warping algorithm is used to align the spindle tension scalar data and the original optical path difference data to the grid nodes of the two-dimensional physical grid, and combines them to generate a force-light synchronization mapping feature set.

[0019] Existing technologies typically assume that polyethylene terephthalate (PET) films are absolutely rigid and rely solely on encoder pulse counting for spatial delay calculations. Since PET release films are subjected to both mechanical tensile stress and thermal expansion strain during transmission in a high-temperature oven, a discrepancy inevitably arises between the actual physical deformation of the film and the rigid body theoretical model. This leads to misalignment of data collected by heterogeneous sensors in spatial coordinates. The specific implementation method disclosed in this embodiment solves the problem of spatial coordinate slippage during the transmission of multi-source heterogeneous signals in physical space. Compared with the existing method of using a single static time window for data alignment, this embodiment introduces thermodynamic principles and continuous medium mechanics principles into the computational process, achieving the technical effect of strictly mapping heterogeneous physical signals to the same physical spatial coordinates.

[0020] Specifically, the spatiotemporal mapping module acquires the spindle tension scalar data and the original optical path difference data, establishes a two-dimensional spatial coordinate system, and then discretizes the two-dimensional spatial coordinate system to construct a two-dimensional physical mesh. The specific operations are as follows: The system acquires spindle tension scalar data, raw optical path difference data, spindle transmission linear velocity data, lateral position pulse data, and absolute temperature field data. In this embodiment, the spindle tension scalar data represents the instantaneous mechanical tensile load borne longitudinally by the polyethylene terephthalate release film. The value of the spindle tension scalar data is limited to the range of 50N to 300N. The selection of 50N to 300N is based on the conclusion of the ultimate yield point test analysis of the polymer film anti-breakage process. The spindle tension scalar data plays a role in quantifying physical tension and filtering abnormal electrical signal interference in logical operations. The raw optical path difference data represents the phase difference of interference including stress distortion and physical thickness fluctuation. The value of the raw optical path difference data is set to 10μm to 300μm, based on the hardware optical range reference of the white light interferometer. The raw optical path difference data plays a role in providing undecoupled initial optical characteristics in system operations. The spindle transmission linear velocity data represents the absolute linear velocity of the traction motor driving the polyethylene terephthalate release film forward. The spindle transmission linear velocity data is set from 0.1 m / s to 10.0 m / s, based on the rated mechanical operating parameters of the hot melt coating machine. This data serves as the reference variable for kinematic Riemann integral calculations. The absolute temperature field data represents the physical intensity of convection and radiation from the heat source inside the coating oven. The absolute temperature field data is set from 20°C to 180°C, based on the phase transition temperature range of the polyethylene terephthalate drying and heat setting process. This data drives thermal expansion calculations and provides an index for stiffness softening compensation in the time warping algorithm.

[0021] Absolute timestamps are assigned to the spindle tension scalar data, raw optical path difference data, spindle transmission linear velocity data, lateral position pulse data, and absolute temperature field data. In this embodiment, the spatiotemporal mapping module invokes the distributed clock synchronization protocol of industrial real-time Ethernet. When the field programmable logic controller triggers a low-level hardware interrupt request, it assigns time stamps based on the global network clock to the discrete sampling points of the five heterogeneous data sources, establishing a unified time reference that avoids bus communication delays.

[0022] Extract the lateral position pulse data with absolute timestamps, convert the lateral position pulse data with absolute timestamps into absolute lateral coordinates, and set the absolute lateral coordinates as the horizontal axis. This module multiplies the pulse count value output by the linear motor of the optical scanning frame by a pre-calibrated single-pulse physical equivalent constant, mapping the pulse count value to the actual physical dimension of the lateral movement of the optical probe on the surface of the polyethylene terephthalate release film, thus forming the lateral measurement dimension of the two-dimensional spatial coordinate system.

[0023] The spindle transmission linear velocity data with absolute timestamps is extracted. An integral operation is performed on this data to calculate the integral displacement value, which is then set as the vertical axis. The spatiotemporal mapping module uses the time difference of the absolute timestamps as the integral element to perform algebraic accumulation on the spindle transmission linear velocity data along the time dimension. This calculates the apparent travel length of the polyethylene terephthalate release film in the mechanical transmission direction, forming the vertical measurement dimension of the two-dimensional spatial coordinate system.

[0024] A two-dimensional spatial coordinate system is established by combining the horizontal and vertical axes. The measurement values ​​of the horizontal and vertical axes are extracted and orthogonally combined to establish a planar geometric mapping system that extends in real time with the dynamic operation of the polyethylene terephthalate release film.

[0025] The two-dimensional spatial coordinate system is divided according to a preset discrete resolution to construct a two-dimensional physical mesh containing grid nodes. In this embodiment, the preset discrete resolution represents the granularity of the spatiotemporal mapping module in physically representing the surface of the continuous polyethylene terephthalate release film. The preset discrete resolution is set to 1mm × 1mm. This value is set based on the absolute physical diameter of the laser spot of the thickness probe and the Nyquist space sampling theorem, thereby determining the topological density of the geometric mapping base map and limiting the matrix operation load of the microprocessor.

[0026] The spatiotemporal mapping module uses a time warping algorithm to align the spindle tension scalar data and the original optical path difference data to the grid nodes of the two-dimensional physical grid, and combines them to generate a force-light synchronization mapping feature set. The specific operation is as follows: Absolute temperature field data with absolute timestamps is extracted, and the dynamic Young's modulus corresponding to the absolute temperature field data with absolute timestamps is extracted from a preset temperature stiffness mapping table. In this embodiment, the dynamic Young's modulus represents the physical parameter of the polyethylene terephthalate release film resisting longitudinal tensile deformation under different thermal field conditions. The numerical range of the dynamic Young's modulus is set from 2.0 GPa to 4.5 GPa. The selection of 2.0 GPa to 4.5 GPa is based on the stress-strain decay curves obtained from multiple temperature rise and fall measurements using a polymer multiphysics rheometer. This curve compensates for the softening and drift of the material's mechanical rigidity caused by high-temperature airflow in the formula, preventing distortion of the denominator value in the subsequent mechanical calculation equations.

[0027] The mechanical tensile strain term is calculated by dividing the spindle tension scalar data with absolute timestamps by the product of the preset nominal width, preset nominal thickness, and dynamic Young's modulus. In this embodiment, the preset nominal width represents the transverse physical dimension of the cross-section of the polyethylene terephthalate release film extrusion molding, and the preset nominal width value is set to 1.5m, based on the initial adjustment clearance of the extrusion die lip; the preset nominal thickness represents the longitudinal dimension of the polyethylene terephthalate release film design, and the preset nominal thickness value is set to 50μm, based on the release film product manufacturing specifications. The preset nominal width and preset nominal thickness are multiplied to obtain the stress cross-sectional area. The spatiotemporal mapping module divides the spindle tension scalar data by the product of the stress cross-sectional area and the dynamic Young's modulus to quantify the longitudinal elastic deformation ratio of the polyethylene terephthalate release film due to tensile stress.

[0028] The temperature difference variable is calculated by subtracting a preset reference temperature data from the absolute temperature field data with an absolute timestamp. In this embodiment, the preset reference temperature data represents the absolute zero basis for thermodynamic deviation comparison by the spatiotemporal mapping module. The preset reference temperature data is set to 25°C, which is based on the constant temperature control value of an industrial standard cleanroom. The preset reference temperature data plays a role in the algorithm in calculating the actual thermodynamic increment experienced by the polyethylene terephthalate release film after it enters the oven.

[0029] The thermal expansion strain term is calculated by multiplying the preset coefficient of thermal expansion by the temperature difference variable. In this embodiment, the preset coefficient of thermal expansion represents the linear physical elongation of the polyethylene terephthalate release film per unit temperature increase. The preset coefficient of thermal expansion is set to 6.5 × 10^-51 / °C, and this value is based on the basic engineering property test data of polyethylene terephthalate material. The preset coefficient of thermal expansion serves to convert the thermal field intensity variable into the macroscopic longitudinal physical elongation ratio.

[0030] The constant 1, the mechanical tensile strain term, and the thermal expansion strain term are added together to calculate the comprehensive strain ratio. The spacetime mapping module performs algebraic summation of the constant 1, which represents the ratio of absolute rigid body length, the mechanical tensile strain term, which represents the ratio of elongation under force, and the thermal expansion strain term, which represents the ratio of thermal expansion, to restore the total transient physical deformation of the polyethylene terephthalate release film under the environment of multiple physical fields.

[0031] The physical transmission length is calculated by multiplying the spindle transmission linear velocity data with absolute timestamps by the comprehensive strain ratio and then integrating the result. The spatiotemporal mapping module uses a Riemann integral mathematical model to solve for the area of ​​the product of the spindle transmission linear velocity data and the comprehensive strain ratio in the time dimension. The area value obtained constitutes the true physical transmission length after considering the dual physical effects of thermodynamic expansion deformation and kinematic tensile deformation.

[0032] The distance penalty is calculated by subtracting the physical transmission length from the preset geometric span. In this embodiment, the preset geometric span represents the absolute fixed distance between the tension sensor and the optical thickness probe at the hardware installation level on the production line. The preset geometric span is set to 5.5m, which is based on the three-dimensional laser scanning distance verification value of the as-built drawings of the coating equipment assembly. The preset geometric span serves as a criterion for judging the law of conservation of kinematic integrals in the spatiotemporal mapping module, so as to apply algebraic penalties to false data matching paths that violate the law of conservation of matter.

[0033] Calculate the differential rate of change of the spindle tension scalar data with absolute timestamps and the differential rate of change of the raw optical path difference data with absolute timestamps. The spatiotemporal mapping module extracts the data difference between adjacent absolute timestamps and divides it by the discrete time step to extract the transient change feature slope of the heterogeneous data stream.

[0034] The Hilbert norm is calculated between the differential rate of change of the principal tension scalar data with absolute timestamps and the differential rate of change of the original optical path difference data with absolute timestamps. The spatiotemporal mapping module measures the numerical distance between the mechanical catastrophe eigenvectors and the optical catastrophe eigenvectors in the mapped coordinate space using norm mathematical operators.

[0035] The cost function value is calculated by adding the Hilbert norm and the distance penalty. The joint constraint comprehensive evaluation index is calculated by algebraically summing the Hilbert norm, which characterizes the similarity of data forms, and the distance penalty, which characterizes the constraints of physical kinematic laws.

[0036] The target timestamp coordinate pairs that minimize the cost function value are extracted from the multidimensional time series matrix. The spatiotemporal mapping module extracts the minimum algebraic solution of the cost function value from the time series matrix, locking in the occurrence and arrival time coordinates that conform to the real physical transport laws.

[0037] To objectively verify the technical effectiveness of the spatiotemporal mapping module provided in this embodiment, a comparative example is provided for illustration. The experimental environment was set as follows: the coating oven test length was set to 4.0m, the coating oven operating temperature was set to 150°C, the spindle transmission linear velocity data was stably running at 5.0m / s, and the spindle tension scalar data of the polyethylene terephthalate release film fluctuated sinusoidally within the range of 100N to 150N. The comparative example uses a conventional absolute rigid body model in the prior art (only performing integration calculations based on the spindle transmission linear velocity data, without incorporating mechanical tensile strain terms and thermal expansion strain terms). Experimental data shows that when reaching the thickness probe scanning position, the spatial alignment error between the tension data extracted in the comparative example and the optical data reaches 45mm. After processing the data using the time warping algorithm including a multiphysics coupling compensation model provided in this embodiment, the experimental data shows that the spatial alignment error is reduced to within 1mm. The above experimental data proves that, by introducing mechanical tensile strain terms and thermal expansion strain terms and within a limited parameter range, the method disclosed in this embodiment solves the technical problem of data spatial coordinate slippage.

[0038] The scalar data of principal tension with absolute timestamps and the original optical path difference data with absolute timestamps are extracted to the grid nodes of the two-dimensional physical grid corresponding to the target timestamp coordinate pair, and combined to generate a force-optical synchronization mapping feature set. The spatiotemporal mapping module executes the data tensor splicing command to complete the data combination.

[0039] Through the specific operations performed by the spatiotemporal mapping module, the spatiotemporal mapping module outputs a force-optical synchronous mapping feature set. All data items in the force-optical synchronous mapping feature set share the same spatial physical coordinates and synchronous absolute timestamp, avoiding variable extraction errors caused by asynchronous transmission delays. This provides underlying data matrix support that satisfies physical consistency for subsequent calculation of refractive index dynamic drift using the force-optical synchronous mapping feature set.

[0040] The refractive index calculation module receives the force-optical synchronous mapping feature set generated by the spatiotemporal mapping module, then extracts the principal axis tension scalar data contained in the force-optical synchronous mapping feature set, and calls the preset stress optics tensor polynomial to perform polynomial algebra operations on the principal axis tension scalar data to calculate the dynamic drift of the refractive index.

[0041] Existing technologies for calculating the stress-birefringence effect of polymer films generally use the static nominal cross-sectional area to directly calculate the engineering stress and directly substitute the tension signal collected by the sensor into the equation. This conventional approach has two major physical drawbacks: First, it ignores the transverse and normal Poisson necking deformation that inevitably occurs in polymer materials under tension, resulting in calculated stress values ​​lower than the actual physical pressure inside the material; second, it disregards the high-frequency mechanical vibrations that inevitably accompany the operation of equipment in industrial settings, where minute tension noise is directly fed into high-order sensitive polynomials, leading to an exponential amplification of errors. The specific implementation method disclosed in this embodiment completely solves the technical problem of the intertwined issues of underestimation of engineering stress and mechanical noise pollution. By introducing a zero-phase-shift low-pass filtering mechanism and dynamic compensation calculation based on the real stress cross-sectional area using the Poisson effect, it achieves a high-precision decoupling effect for optical distortion.

[0042] Specifically, the system receives the force-optical synchronization mapping feature set generated by the spatiotemporal mapping module, and then extracts the spindle tension scalar data contained in the force-optical synchronization mapping feature set. The specific operation is as follows: The principal axis tension scalar data and absolute temperature field data contained in the force-light synchronous mapping feature set are extracted. The force-light synchronous mapping feature set output by the spatiotemporal mapping module has completed the alignment compensation of the physical spatiotemporal dimensions. The refractive index calculation module directly extracts the principal axis tension scalar data and absolute temperature field data through the underlying communication link, providing access to the refractive index calculation module's calculation logic for the principal axis tension scalar data and absolute temperature field data.

[0043] The preset zero-phase-shift low-pass filtering algorithm is used to smooth the spindle tension scalar data in the time domain, resulting in steady-state tension scalar data. The refractive index calculation module uses bidirectional digital filtering to remove high-frequency mechanical harmonics from the spindle tension scalar data without changing the physical time phase, thus obtaining pure steady-state tension scalar data for subsequent calculations.

[0044] The preset nominal width, preset nominal thickness, and preset intrinsic Poisson's ratio are extracted from the preset product property specification database. The refractive index calculation module directly retrieves the preset nominal width and preset nominal thickness for subsequent calculations. In this embodiment, the preset intrinsic Poisson's ratio represents the physical constant that quantifies the lateral volume shrinkage rate in the width and thickness directions of the polyethylene terephthalate release film when the longitudinal stretching is converted. The dimensionless value of the preset intrinsic Poisson's ratio is set between 0.38 and 0.42. The basis for setting 0.38 to 0.42 is the experimental measurement based on the assumption of a continuous medium in the material mechanics of polymers. The preset intrinsic Poisson's ratio plays a role in driving the physical calculation of the dynamic reduction of the material's cross-sectional area.

[0045] Using absolute temperature field data as an index, the dynamic Young's modulus is extracted from the preset underlying memory. The refractive index calculation module inputs the absolute temperature field data as an address pointer into the preset underlying memory and directly extracts the matching dynamic Young's modulus, preventing systematic physical omissions caused by the initial longitudinal tensile strain due to thermal softening.

[0046] The stiffness base is calculated by multiplying the dynamic Young's modulus, the preset nominal width, and the preset nominal thickness. In this embodiment, the stiffness base represents the total physical stiffness of the material's overall cross-section against external tensile force. The value of the stiffness base is limited to between 150,000 N and 330,000 N. The range of the stiffness base value is based on the algebraic product of the dynamic Young's modulus, the preset nominal width, and the preset nominal thickness. The stiffness base serves as the denominator in the basic strain calculation.

[0047] The initial longitudinal tensile strain is calculated by dividing the steady-state tension scalar data by the stiffness base. In this embodiment, the initial longitudinal tensile strain represents the unidirectional scalar elongation ratio of the polyethylene terephthalate release film after being subjected to force. The dimensionless value of the initial longitudinal tensile strain is between 0.00015 and 0.002, and the value is based on the ratio of the steady-state tension scalar data to the stiffness base. The initial longitudinal tensile strain serves as the input variable for the Poisson contraction conversion calculation.

[0048] The shrinkage ratio is calculated by multiplying the initial longitudinal tensile strain by a preset intrinsic Poisson's ratio. In this embodiment, the shrinkage ratio represents the physical percentage reduction in the transverse and normal dimensions of the polyethylene terephthalate release film relative to the preset nominal width and preset nominal thickness. The dimensionless value of the shrinkage ratio is between 0.000057 and 0.00084, and the value is derived from the Poisson shrinkage equation in mechanics of materials. The shrinkage ratio serves to quantify the geometric deformation in three-dimensional space.

[0049] The shrinkage factor is calculated by subtracting the shrinkage ratio from the constant. In this embodiment, the shrinkage factor represents the ratio of the actual unidirectional dimension after deformation to the original design dimension. The dimensionless value of the shrinkage factor infinitely approaches one but is always less than one. The value is based on the physical law of conservation of mass of rigid bodies. The shrinkage factor serves to prevent the omission of the attenuation of lateral length and normal length.

[0050] The squared shrinkage factor is calculated by performing a quadratic operation on the shrinkage factor. In this embodiment, the quadratic necking factor represents the multiplier ratio of the overall shrinkage of the two-dimensional physical cross-section of the polyethylene terephthalate release film. The dimensionless value of the quadratic necking factor is strictly less than one, based on the multiplicative shrinkage derivation of the shrinkage of the two-dimensional physical orthogonal area. The quadratic necking factor plays the role of mapping unidirectional linear shrinkage to two-dimensional surface volume shrinkage.

[0051] The nominal physical cross-sectional area is calculated by multiplying the preset nominal width by the preset nominal thickness. In this embodiment, the nominal physical cross-sectional area represents the initial static cross-sectional area when not subjected to physical tension. The value of the nominal physical cross-sectional area is 75 mm^2, calculated based on the product of the length and width dimensions. The nominal physical cross-sectional area serves to provide a reference geometric surface undisturbed by physical forces.

[0052] The transient true cross-sectional area is calculated by multiplying the nominal physical cross-sectional area by the square necking factor. In this embodiment, the transient true cross-sectional area represents the absolute physical area of ​​the polymer cross-section after considering multidimensional Poisson necking deformation. The value of the transient true cross-sectional area is between 74.8 mm^2 and 74.99 mm^2, and is based on the multiplication of the nominal physical cross-sectional area and the square necking factor. The transient true cross-sectional area serves to eliminate the common-sense errors caused by assuming a constant cross-sectional area in engineering stress calculations.

[0053] The transient true tensile stress is calculated by dividing the steady-state tension scalar data by the transient true cross-sectional area under stress. In this embodiment, the transient true tensile stress represents the absolute true physical pressure borne inside the polyethylene terephthalate film after completely deducting the cross-sectional necking deformation. The value of the transient true tensile stress is set in the range of 0 Pa to 150,000,000 Pa, based on the result of the division operation of the steady-state tension scalar data by the transient true cross-sectional area under stress. The transient true tensile stress serves to provide an accurate and dimensionlessly conserved intrinsic mechanical driving variable for nonlinear high-order polynomial operations.

[0054] Furthermore, the refractive index calculation module calls the preset stress optics tensor polynomial to perform polynomial algebraic operations on the principal axis tension scalar data to calculate the dynamic refractive index drift. The specific operation is as follows: Using absolute temperature field data as an index, the first, second, and third order multiphysics pre-calibration constants are extracted from a pre-defined rheological matrix library. In this embodiment, the first order multiphysics pre-calibration constant represents the linear initial response weight of the first-order tensile stress term to the optical polarizability, and its numerical magnitude is set to... The second-order constant in the multiphysics precalibration represents the weight of the second-order deformation polarization nonlinear response corresponding to the quadratic term of tensile stress. The numerical magnitude of the second-order constant in the multiphysics precalibration is set to... The third-order constant in the multiphysics pre-calibration represents the weight of the third-order polymer chain segment limiting orientation response corresponding to the cubic tensile stress term. The numerical magnitude of the third-order constant in the multiphysics pre-calibration is set to 3.0E-281 / Pa^3. The numerical values ​​of the three order constants are extracted based on the joint physical experiment calibration matrix library of polymer rheology and polarization at specific temperatures. The three order constants serve to establish the intrinsic distribution ratio of the transformation of the transient real tensile stress term of different orders to the refractive index drift and to ensure the absolute conservation of physical dimensions in the computational equations.

[0055] The system calculates the first, second, and third exponentiation values ​​of transient true tensile stress. The refractive index calculation module performs one to three consecutive multiplication operations on the transient true tensile stress to establish a higher algebraic order reflecting the Langevin physical orientation of the microscopic polymer chain segments.

[0056] The first-order product value is calculated by multiplying the first-order power value of the transient true tensile stress by the first-order constant of the multiphysics pre-calibration. The second-order product value is calculated by multiplying the second-order power value by the second-order constant of the multiphysics pre-calibration. The third-order product value is calculated by multiplying the third-order power value by the third-order constant of the multiphysics pre-calibration. The refractive index calculation module transforms the single-physical-dimensional mechanical variable into a cross-physical-domain optical intrinsic property interference element through independent multiplication operations.

[0057] The initial value of the refractive index dynamic drift is calculated by summing the values ​​of the first, second, and third products. Preset upper and lower limits of the elastic deformation region threshold are then extracted. The absolute value of the initial refractive index dynamic drift is compared with these two preset limits. In this embodiment, the preset upper limit of the elastic deformation region threshold represents the maximum physical limit of an optical phase transition that can be triggered, and its dimensionless value is fixed at 0.005. The preset lower limit of the elastic deformation region threshold represents the minimum physical limit of an optical phase transition with practical physical compensation significance, and its dimensionless value is fixed at 0.0001. Both values ​​are based on the measured values ​​of the rheological breaking limit of the elastic deformation region of polyethylene terephthalate (PET). In the logic gating, both functions to forcibly intercept illegal mathematical divergence values ​​caused by underlying electromagnetic interference.

[0058] When the absolute value of the initial dynamic refractive index drift is greater than the preset upper limit of the elastic deformation zone threshold, or when the absolute value of the initial dynamic refractive index drift is less than the preset lower limit of the elastic deformation zone threshold, the final dynamic refractive index drift is set to zero. When the absolute value of the initial dynamic refractive index drift is less than or equal to the preset upper limit of the elastic deformation zone threshold, and the absolute value of the initial dynamic refractive index drift is greater than or equal to the preset lower limit of the elastic deformation zone threshold, the initial dynamic refractive index drift is set to the final dynamic refractive index drift. In this embodiment, the final dynamic refractive index drift represents the absolute physical value of mechanical stress converted into optical polarization property disturbance. The dimensionless value of the final dynamic refractive index drift is strictly constrained within the range of 0.0001 to 0.005 or is absolutely zero. The final dynamic refractive index drift is determined based on the logical judgment operation output of the pre-comparator on the safety boundary. The final dynamic refractive index drift serves as the core correction base for optical path difference compensation in thickness reduction calculation. The refractive index calculation module performed a physical extreme value constraint judgment to prevent downtime and eliminated abnormal divergence features.

[0059] To visually demonstrate the technological advancements of this embodiment compared to existing technologies, an experimental comparative example is provided. The experimental environment was set up on an industrial polyethylene terephthalate release film mass production line, where a specific temperature field was present. The high-frequency thermal jump of the winding guide roller, due to the aging of the mechanical bearing, resulted in mechanical resonance noise with an amplitude of 20N. The test traction tension was set to 150N. A comparative example used an existing conventional engineering stress calculation model. Because it did not incorporate a zero-phase-shift low-pass filter algorithm and a transient real-force cross-sectional area necking mechanism, it directly divided the tension containing the 20N noise by the static nominal physical cross-sectional area and fed it into a third-order exponential polynomial for calculation. Experimental results showed that the dynamic refractive index drift calculated by the conventional algorithm diverged to a maximum of 0.85, severely exceeding the limits of real-world optics, leading to a measurement fluctuation error of up to 12μm in the downstream thickness measurement system. After applying the technical solution disclosed in this embodiment, the zero-phase-shift low-pass filter algorithm completely eliminated the 20N resonance noise, and the transient real-force cross-sectional area accurately restored the stress concentration effect. Experimental results showed that the final dynamic refractive index drift was stably locked within the legal physical threshold of 0.0032, and the downstream thickness tolerance decoupling accuracy was improved to 0.5μm. The above experimental data proves that this embodiment has achieved significant technical effects within the limited parameter range and calculation rules.

[0060] Through the specific operations performed by the refractive index calculation module, this embodiment eliminates the optical decoupling divergence defects caused by Poisson necking omissions and high-frequency mechanical noise in polymer materials under high-temperature and high-speed traction conditions. The final dynamic refractive index drift calculated provides a pure and physically consistent benchmark decoupling independent variable for the system to activate the subsequent pseudo-thickness stripping module and completely strip stress optical artifacts from the original interference coherent optical path.

[0061] The pseudo-thickness stripping module receives the dynamic refractive index drift calculated by the refractive index calculation module, extracts the original optical path difference data contained in the force-optical synchronous mapping feature set generated by the spatiotemporal mapping module, and then inputs the preset intrinsic Poisson's ratio, the original optical path difference data, and the dynamic refractive index drift into the preset thickness restoration equation for nonlinear algebraic solution to calculate the stress-free real material reference thickness.

[0062] Existing technologies for high-speed production thickness measurement of polyethylene terephthalate (PET) release films generally employ a single conversion logic of dividing the optical path length by the static refractive index. Even with the introduction of dynamic refractive index compensation in some technologies, existing methods still suffer from serious physical boundary defects: the system only measures the apparent stress thickness of the film under mechanical tension. Since the material inevitably undergoes Poisson necking deformation (i.e., physical thinning) in the transverse and normal directions when subjected to longitudinal tension, directly feeding back the apparent stress thickness, which includes necking thinning errors, to the upstream extrusion die can lead to the die misinterpreting insufficient polymer melt extrusion and incorrectly issuing control commands to increase extrusion. This ultimately results in severe over-thickness defects in the release film after the winding tension is released and the material springs back. The specific implementation method disclosed in this embodiment solves the physical mapping discontinuity problem between apparent thickness and actual extrusion quality. By introducing the theory of volume conservation in large deformation continuous media, a two-stage reverse decoupling is performed at the mathematical level. Compared with existing technologies, this embodiment achieves the technical effect of simultaneously removing optical polarization artifacts and mechanical necking thinning.

[0063] Specifically, the pseudo-thickness stripping module receives the dynamic refractive index drift calculated by the refractive index calculation module and extracts the original optical path difference data contained in the force-optical synchronization mapping feature set generated by the spatiotemporal mapping module. The specific operation is as follows: The dynamic refractive index drift and initial longitudinal tensile strain are extracted. The pseudo-thickness stripping module directly calls the dynamic refractive index drift and initial longitudinal tensile strain already processed by the refractive index calculation module through the underlying internal communication data bus, avoiding repeated deduction and calculation of the same physical driving quantity, ensuring the uniqueness of the system data flow chain and the high efficiency of floating-point computing power.

[0064] The original optical path difference data contained in the force-light synchronization mapping feature set is extracted. The pseudo-thickness stripping module confirms, through a global absolute timestamp verification mechanism, that the acquired original optical path difference data and the refractive index dynamic drift share absolutely aligned physical space node coordinates in the two-dimensional physical mesh.

[0065] Extract the preset initial static refractive index. In this embodiment, the preset initial static refractive index represents the base value of the intrinsic optical properties of the polyethylene terephthalate film under a standard state without any physical stress. The dimensionless value of the preset initial static refractive index is fixed at 1.64. The basis for setting 1.64 is the experimental matrix table of optical property measurement of polymers at room temperature standard state. The preset initial static refractive index serves to provide a reference zero point for dynamic refractive index compensation superposition.

[0066] The real-time absolute refractive index is calculated by adding the preset initial static refractive index to the dynamic refractive index drift. In this embodiment, the real-time absolute refractive index represents the true optical refractive index of the polyethylene terephthalate release film under the current actual tensile polarization state. The dimensionless value of the real-time absolute refractive index is between 1.6401 and 1.6450, and the value is determined based on the algebraic sum of the preset initial static refractive index and the dynamic refractive index drift. The real-time absolute refractive index serves as the precise denominator for reducing the interference optical path difference to the physical length.

[0067] The apparent physical thickness is calculated by dividing the original optical path difference data by the real-time absolute refractive index. In this embodiment, the apparent physical thickness represents the actual geometric thickness after completely eliminating stress birefringence optical artifacts, but still including the Poisson thinning loss due to mechanical stretching. The value of the apparent physical thickness is between 11.5 μm and 198.0 μm, and the value is based on the quotient of the original optical path difference data and the real-time absolute refractive index. The apparent physical thickness serves as the base value of the independent variable for the subsequent inverse thickness compensation in the volume conservation equation.

[0068] Furthermore, the preset intrinsic Poisson's ratio, original optical path difference data, and refractive index dynamic shift are input into the preset thickness reduction equation for nonlinear algebraic solution to calculate the stress-free real material reference thickness. The specific operation is as follows: The preset intrinsic Poisson's ratio is extracted, and the constant is added to the initial longitudinal tensile strain to calculate the absolute elongation ratio. In this embodiment, the absolute elongation ratio represents the absolute physical quotient of the real-time length of the polyethylene terephthalate release film after being stretched under tension and its initial unstressed length. The dimensionless value of the absolute elongation ratio is between 1.00015 and 1.002, and the value is determined by the algebraic summation of the constant and the initial longitudinal tensile strain. The absolute elongation ratio serves as the independent variable for providing the basis for the large deformation volume conservation deduction.

[0069] Using a preset intrinsic Poisson's ratio as an exponent, the absolute ratio of stretching is exponentially calculated to obtain the volume-conserving reverse expansion factor. As a step to overcome physical limitations in this embodiment, the solution logic for the volume-conserving reverse expansion factor abandons linear estimation. In this embodiment, the volume-conserving reverse expansion factor represents a nonlinear multiplier used to mathematically restore the stretched film thickness to the original tension-free extrusion thickness. The dimensionless value of the volume-conserving reverse expansion factor is between 1.00006 and 1.0008, and its value is determined based on the Hench true strain theory of large deformation in continuous media and the power function calculation results of the connected Poisson necking effect. The volume-conserving reverse expansion factor plays a role in positively amplifying the apparent physical thickness to accurately compensate for the normal volume loss caused by longitudinal stretching.

[0070] The stress-free true material reference thickness is calculated by multiplying the apparent physical thickness by the volume conservation reverse expansion factor. In this embodiment, the stress-free true material reference thickness represents the absolute physical thickness of the pure melt extrusion after completely removing optical phase shift artifacts and physical stress-induced necking and thinning distortion. The value of the stress-free true material reference thickness is between 12.0 μm and 200.0 μm, and the value is based on the algebraic product of the apparent physical thickness and the volume conservation reverse expansion factor. The stress-free true material reference thickness serves as an absolute true target for feedback to the extrusion die for closed-loop correction control.

[0071] Extract the absolute lateral coordinates and integral displacement values ​​corresponding to the mesh nodes contained in the 2D physical mesh. The pseudo-thickness stripping module locks the specific physical spatial location features that trigger the calculation.

[0072] The stress-free real material reference thickness is associated and combined with the absolute lateral coordinate and integral displacement value to output the stress-free real material reference thickness with the absolute lateral coordinate and integral displacement value. The pseudo-thickness stripping module encapsulates and binds the pure target thickness value with the system's two-dimensional spatial coordinate system and outputs a multi-dimensional feedback matrix to the global control bus.

[0073] To visually demonstrate the advancements of the reverse thickness decoupling technology in this embodiment compared to existing technologies, a rigorous mass production comparative example is provided. The experimental environment was set up on a polyethylene terephthalate (PET) release film coating line with a nominal production thickness of 50.0 μm. The real-time traction spindle tension was maintained at 200 N, with an initial longitudinal tensile strain calibration of 0.001. The comparative example used a conventional thickness measurement algorithm (only refractive index compensation was performed without introducing a large deformation volume conservation reverse expansion factor). Experimental data showed that the thickness value output by the conventional algorithm was only 49.95 μm (including the apparent stress thickness due to Poisson's necking). Based on the false thinning signal of 49.95 μm, the control system issued an erroneous command to the extrusion die to compensate 0.05 μm. This caused the actual finished product thickness measured by the offline micrometer to become 50.05 μm after the PET release film was finally wound up and the internal tension was released and rebounded, resulting in irreversible out-of-tolerance scrap. After applying the technical solution disclosed in this embodiment, the pseudo-thickness peeling module multiplies the apparent physical thickness of 49.95 μm by the calculated volume conservation reverse expansion factor, and the system outputs a stress-free real material reference thickness of 50.0 μm. The control system maintains stable operation of the extrusion die based on 50.0 μm, and the final winding product thickness is precisely locked at 50.0 μm. Production experimental data proves that this embodiment, by introducing a preset intrinsic Poisson's ratio and power-law expansion compensation mechanism, achieves the technical effect of eliminating pseudo-thickness fluctuation feedback and preventing erroneous compensation defects.

[0074] By using the pseudo-thickness stripping module, this embodiment completely eliminates the interference of the pseudo-stress thinning phenomenon caused by mechanical stretching on the thickness measurement system. By using the calculated stress-free real material reference thickness, it provides the purest physical measurement data for generating accurate feedforward torque commands and mold head fine-tuning commands in the next control stage. The next control stage will be undertaken and executed by the collaborative feedback module.

[0075] The collaborative feedback module receives the stress-free real material reference thickness calculated by the pseudo-thickness stripping module and the force-light synchronous mapping feature set generated by the spatiotemporal mapping module. Then, it performs a difference operation between the stress-free real material reference thickness and the preset target thickness to generate an opening fine-tuning command. It extracts the original optical path difference data contained in the force-light synchronous mapping feature set, divides the original optical path difference data by the preset initial static refractive index to calculate the apparent physical thickness, and performs a difference operation between the apparent physical thickness and the stress-free real material reference thickness to generate a feedforward torque compensation command.

[0076] In the high-speed production of polyethylene terephthalate (PET) release film, existing technologies face the technical problem of interference between the physical properties of the actuators in thickness closed-loop adjustment. Conventional technologies simultaneously feed back mixed thickness measurements, including mechanical deformation errors, to both the extrusion die and the winding motor. This leads to tension fluctuations causing physical thinning of the film, which the extrusion die may incorrectly determine as material deficiency and increase the extrusion volume. Ultimately, after tension is released and the material rebounds, irreversible, excessively thick waste products are produced. The specific implementation method disclosed in this embodiment solves the technical problem of strong coupling of the physical properties of multiple actuators. Through the underlying physical laws of spatial geometric decoupling and elastic deformation separation, the pure extrusion quality deviation and mechanical deformation deviation are dimensionally isolated at the data level. Compared with existing technologies, this embodiment achieves a crosstalk-free independent feedforward closed loop between the extrusion die and the winding motor.

[0077] Specifically, the collaborative feedback module receives the stress-free real material reference thickness calculated by the pseudo-thickness stripping module and the force-light synchronous mapping feature set generated by the spatiotemporal mapping module. Then, it performs a difference operation between the stress-free real material reference thickness and the preset target thickness to generate an opening fine-tuning command. The specific operation is as follows: The system receives the stress-free real material reference thickness and the force-optical synchronous mapping feature set. The collaborative feedback module synchronously calls the stress-free real material reference thickness and the force-optical synchronous mapping feature set through the internal real-time communication bus. The system verifies the absolute timestamp to ensure that the stress-free real material reference thickness and the force-optical synchronous mapping feature set are in an absolutely synchronized and aligned state on the physical time axis and geometric space coordinates, thus establishing a unified data input benchmark for the physical dimensionality reduction and decoupling of the dual links.

[0078] The preset target thickness is extracted, and the absolute geometric space error is calculated by subtracting the stress-free real material reference thickness from the preset target thickness. In this embodiment, the preset target thickness represents the absolute true baseline for evaluating the working stability of the upstream extruder. The preset target thickness is 50.0 μm, and it is determined based on the process drawing specifications of the specific production order, serving to provide an ideal reference zero point for thickness control closed loop. The absolute geometric space error represents the pure and absolute melt extrusion geometric deviation, and its value is between -5.0 μm and 5.0 μm. The absolute geometric space error is determined by performing a pure algebraic subtraction calculation by subtracting the stress-free real material reference thickness from the preset target thickness, serving to completely eliminate the misleading effect of false thinning caused by tension stretching on the extrusion die.

[0079] A preset die head bolt spatial mapping ratio coefficient is extracted. The absolute geometric space error is multiplied by the preset die head bolt spatial mapping ratio coefficient to calculate the displacement step command. The displacement step command is set as the opening fine-tuning command. In this embodiment, the preset die head bolt spatial mapping ratio coefficient represents a conversion multiplier that proportionally maps the film thickness geometric error to the geometric deformation command of the die head steel body. The dimensionless value of the preset die head bolt spatial mapping ratio coefficient is between 0.1 and 1.0, and is determined based on the mechanical deformation calibration test results of the extrusion die lip, thus avoiding the introduction of unmeasurable hydrodynamic parameters. The displacement step command represents the step control amount that drives the opening and closing of the die head lip thermal expansion bolt actuator. The value of the displacement step command is within the mechanical stroke limit of 0μm to 50.0μm. The displacement step command is determined based on the multiplication result of the absolute geometric space error and the preset die head bolt spatial mapping ratio coefficient, thus only responding to the lack or surplus of the intrinsic quality of the material to achieve a robust quality closed loop.

[0080] Furthermore, the original optical path difference data contained in the force-optical synchronization mapping feature set is extracted, and the apparent physical thickness is calculated by dividing the original optical path difference data by the preset initial static refractive index. The specific operation is as follows: The original optical path difference data contained in the force-optical synchronous mapping feature set is extracted, and a preset initial static refractive index is extracted. The original optical path difference data is divided by the preset initial static refractive index to calculate the apparent physical thickness. In order to construct an independent torque feedforward calculation link, the collaborative feedback module uses the geometric inference of Snell's law to perform division dimensionality reduction on the interference signal. In this embodiment, the apparent physical thickness represents a mixed geometric measure captured by the optical interferometer under the physical field of view, including the tensile thinning state under stress and the birefringence effect of residual stress. The value of the apparent physical thickness is between 10.0 μm and 200.0 μm. The apparent physical thickness is determined based on the division quotient of the original optical path difference data and the preset initial static refractive index, which serves to provide the basic geometric input quantity containing the hidden mechanical force characteristics for the subsequent Hooke elasticity physics inverse deduction.

[0081] Furthermore, the collaborative feedback module calculates the difference between the apparent physical thickness and the stress-free real material reference thickness to generate a feedforward torque compensation command. The specific operation is as follows: The thickness difference is calculated by subtracting the stress-free real material reference thickness from the apparent physical thickness, and then converting the thickness difference into an absolute difference. In this embodiment, the physical essence of the absolute difference is the residual evidence of mechanical deformation imprinted on the polyethylene terephthalate film by transient abnormal traction tension. The value of the absolute difference is between 0.0 μm and 2.0 μm. The absolute difference is obtained by taking the absolute value after subtracting the apparent physical thickness from the stress-free real material reference thickness. The absolute difference serves to convert the film body into an ultra-high precision non-contact optical strain gauge.

[0082] A preset elastic moment conversion stiffness coefficient is extracted, and the absolute difference is multiplied by the preset elastic moment conversion stiffness coefficient to calculate the macroscopic linear failure traction force. In this embodiment, the preset elastic moment conversion stiffness coefficient represents the conversion bridge that directly back-calculates the normal thickness physical thinning amount to the longitudinal abnormal tensile force of the accident. The value of the preset elastic moment conversion stiffness coefficient is between 15000 N / μm and 35000 N / μm. The preset elastic moment conversion stiffness coefficient is extracted based on a fixed polymerization constant derived jointly from the material's dynamic Young's modulus and intrinsic Poisson's ratio. The preset elastic moment conversion stiffness coefficient plays a role in performing the rigid inverse deduction of Hooke's law. The macroscopic linear failure traction force represents the abnormal traction force transiently applied to the film surface. The value of the macroscopic linear failure traction force is between 0 N and 50 N. The macroscopic linear failure traction force is calculated based on the multiplication result of the absolute difference and the preset elastic moment conversion stiffness coefficient. The macroscopic linear failure traction force plays a role in converting microscopic deformation geometric errors into macroscopic mechanical physical variables.

[0083] The real-time physical diameter of the take-up roll is obtained, and the transient electromagnetic interference torque is calculated by multiplying the macroscopic linear destructive traction force by the real-time physical diameter of the take-up roll. In this embodiment, the real-time physical diameter of the take-up roll represents the absolute physical lever multiplier that converts the macroscopic abnormal linear tension into the electromagnetic torque of the rotating equipment. The value of the real-time physical diameter of the take-up roll is between 0.1m and 1.0m. The real-time physical diameter of the take-up roll is obtained by dynamically reading from the take-up servo programmable logic controller through the collaborative feedback module. The real-time physical diameter of the take-up roll acts as the base for torque lever amplification. The transient electromagnetic interference torque represents the destructive torque converted to the end of the motor rotating shaft. The value of the transient electromagnetic interference torque is between 0 N·m and 50 N·m. The value is based on the product of the macroscopic linear destructive traction force and the real-time physical diameter of the take-up roll, and serves to quantify the mechanical vibration destructive energy.

[0084] The transient electromagnetic interference torque is set as a feedforward torque compensation command. In this embodiment, the feedforward torque compensation command represents the reverse suppression electromagnetic torque sent to the downstream winding servo driver. The value of the feedforward torque compensation command is between 0 N·m and 50 N·m. The setting of the feedforward torque compensation command is based on the inversion of the direction of the transient electromagnetic interference torque and the encapsulation of the communication protocol. The feedforward torque compensation command plays the role of outputting reverse suppression in advance by utilizing the optical-grade high-sensitivity deformation difference before the traditional mechanical tension sensor produces a hysteresis response.

[0085] To visually demonstrate the substantial advancements of the dual-link independent feedback technology in the collaborative feedback module compared to existing technologies, a real-world mass production comparison is provided. The experimental environment was set up on a polyethylene terephthalate release film coating line with a nominal production thickness of 50.0 μm, where the system encountered a transient tension surge. The comparison used a conventional closed-loop algorithm, where the optical probe measured a mixed thickness of 49.8 μm due to the instantaneous tension of the material. The conventional control system treated 49.8 μm as a material shortage and directly controlled the extrusion die to increase the melt flow. When the traction tension returned to stability, the increased flow caused the actual finished product thickness to become 50.2 μm, resulting in irreversible defects. After applying the technical solution disclosed in this embodiment, the stress-free real material reference thickness received by the collaborative feedback module remained at 50.0 μm. The absolute geometric space error of the opening fine-tuning command calculation link was 0.0 μm, and the extrusion die remained completely stationary, fundamentally preventing erroneous adjustments at the quality level. Simultaneously, the feedforward torque compensation command calculation link utilizes the 0.2μm absolute difference between 50.0μm and 49.8μm, combined with the preset elastic torque conversion stiffness coefficient, to inversely deduce that the tension surged by 10N. The collaborative feedback module immediately generates and issues the corresponding feedforward torque compensation command, and the winding motor instantly performs reverse fine-tuning, rapidly restoring the traction tension to stability. Production test data proves that this invention achieves the technical effect of decoupling extrusion and traction control, and eliminating coupled oscillation waste.

[0086] Through the specific operations performed by the collaborative feedback module, the system completely separates the absolute quality control of materials and mechanical stress intervention into two independent computational links that do not interfere with each other. This avoids the logical chaos that a single PID controller faces when dealing with a multi-input multi-output system, and achieves a precise and stable closed loop in complex production environments. It provides decisive control output support for the high-yield continuous and stable production of high-precision thickness tolerance release films.

[0087] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A thickness detection system for high precision thickness tolerance PET release film, characterized in that, include: The module includes a spacetime mapping module, a refractive index calculation module, a pseudo-thickness stripping module, and a collaborative feedback module; The spatiotemporal mapping module acquires the spindle tension scalar data and the original optical path difference data, establishes a two-dimensional spatial coordinate system, and discretizes the two-dimensional spatial coordinate system to construct a two-dimensional physical grid. The time warping algorithm is used to align the spindle tension scalar data and the original optical path difference data to the grid nodes of the two-dimensional physical grid, and combines them to generate a force-light synchronization mapping feature set. The refractive index calculation module receives the force-optical synchronous mapping feature set generated by the spatiotemporal mapping module, then extracts the principal axis tension scalar data contained in the force-optical synchronous mapping feature set, and calls the preset stress optics tensor polynomial to perform polynomial algebra operations on the principal axis tension scalar data to calculate the dynamic drift of the refractive index. The pseudo-thickness stripping module receives the dynamic refractive index drift calculated by the refractive index calculation module, extracts the original optical path difference data contained in the force-light synchronous mapping feature set generated by the spatiotemporal mapping module, and then inputs the preset intrinsic Poisson's ratio, the original optical path difference data and the dynamic refractive index drift into the preset thickness restoration equation for nonlinear algebraic solution to calculate the stress-free real material reference thickness. The collaborative feedback module receives the stress-free real material reference thickness calculated by the pseudo-thickness stripping module and the force-light synchronous mapping feature set generated by the spatiotemporal mapping module. Then, it performs a difference operation between the stress-free real material reference thickness and the preset target thickness to generate an opening fine-tuning command. It extracts the original optical path difference data contained in the force-light synchronous mapping feature set, divides the original optical path difference data by the preset initial static refractive index to calculate the apparent physical thickness, and performs a difference operation between the apparent physical thickness and the stress-free real material reference thickness to generate a feedforward torque compensation command.

2. The thickness detection system for high-precision thickness tolerance PET release film according to claim 1, characterized in that, The spatiotemporal mapping module acquires the spindle tension scalar data and the original optical path difference data, establishes a two-dimensional spatial coordinate system, and then discretizes the two-dimensional spatial coordinate system to construct a two-dimensional physical mesh. The specific operations are as follows: Acquire spindle tension scalar data, raw optical path difference data, spindle transmission linear velocity data, lateral position pulse data, and absolute temperature field data; Assign absolute timestamps to the spindle tension scalar data, raw optical path difference data, spindle transmission linear velocity data, lateral position pulse data, and absolute temperature field data; Extract the horizontal position pulse data with absolute timestamps, convert the horizontal position pulse data with absolute timestamps into absolute horizontal coordinates, and set the absolute horizontal coordinates as the horizontal axis; Extract the spindle transmission line velocity data with absolute timestamps, perform integration on the spindle transmission line velocity data with absolute timestamps to calculate the integral displacement value, and set the integral displacement value as the vertical axis; Establish a two-dimensional spatial coordinate system by combining the horizontal and vertical axes; The two-dimensional spatial coordinate system is divided according to a preset discrete resolution to construct a two-dimensional physical grid containing grid nodes.

3. The thickness detection system for high-precision thickness tolerance PET release film according to claim 2, characterized in that, The spatiotemporal mapping module uses a time warping algorithm to align the spindle tension scalar data and the original optical path difference data to the grid nodes of the two-dimensional physical grid, and combines them to generate a force-light synchronization mapping feature set. The specific operation is as follows: Extract the absolute temperature field data with absolute timestamps, and extract the dynamic Young's modulus corresponding to the absolute temperature field data with absolute timestamps from the preset temperature stiffness mapping table. The mechanical tensile strain term is calculated by dividing the spindle tension scalar data with absolute timestamps by the product of the preset nominal width, preset nominal thickness, and dynamic Young's modulus. The temperature difference variable is calculated by subtracting the preset reference temperature data from the absolute temperature field data with absolute timestamps. The thermal expansion strain term is calculated by multiplying the preset thermal expansion coefficient by the temperature difference variable. The comprehensive strain ratio is calculated by adding the constant, the mechanical tensile strain term, and the thermal expansion strain term. The physical transmission length is calculated by multiplying the spindle transmission linear velocity data with absolute timestamps by the comprehensive strain ratio and integrating the result of the multiplication. The distance penalty is calculated by subtracting the physical transmission length from the preset geometric span. Calculate the differential rate of change of the spindle tension scalar data with absolute timestamps and the differential rate of change of the raw optical path difference data with absolute timestamps; Calculate the Hilbert norm between the differential rate of change of the spindle tension scalar data with absolute timestamps and the differential rate of change of the original optical path difference data with absolute timestamps; The cost function value is calculated by adding the Hilbert norm and the distance penalty. Extract the target timestamp coordinate pairs that minimize the cost function value from the multidimensional time series matrix; The scalar data of spindle tension with absolute timestamps and the original optical path difference data with absolute timestamps are extracted to the grid nodes of the two-dimensional physical grid corresponding to the target timestamp coordinate pair, and combined to generate a force-light synchronization mapping feature set.

4. The thickness detection system for high-precision thickness tolerance PET release film according to claim 3, characterized in that, The system receives the force-optical synchronization mapping feature set generated by the spatiotemporal mapping module, and then extracts the spindle tension scalar data contained in the force-optical synchronization mapping feature set. The specific operation is as follows: Extract the principal axis tension scalar data and absolute temperature field data contained in the force-light synchronization mapping feature set; The preset zero-phase-shift low-pass filter algorithm is invoked to perform time-domain smoothing on the spindle tension scalar data, and the steady-state tension scalar data is calculated. Extract the preset nominal width, preset nominal thickness, and preset intrinsic Poisson's ratio from the preset product physical property specification database; Using absolute temperature field data as an index, the dynamic Young's modulus is extracted from the preset underlying memory. The stiffness base is calculated by multiplying the dynamic Young's modulus, the preset nominal width, and the preset nominal thickness. The initial longitudinal tensile strain is calculated by dividing the steady-state tension scalar data by the stiffness base. The shrinkage ratio is calculated by multiplying the initial longitudinal tensile strain by the preset intrinsic Poisson's ratio. The shrinkage factor is calculated by subtracting the shrinkage ratio from the constant. The squared necking factor is calculated by performing a quadratic operation on the shrinkage factor. The nominal physical cross-sectional area is calculated by multiplying the preset nominal width by the preset nominal thickness. Multiply the nominal physical cross-sectional area by the square necking factor to calculate the transient real force cross-sectional area; The transient true tensile stress is calculated by dividing the steady-state tension scalar data by the transient true cross-sectional area.

5. The thickness detection system for high-precision thickness tolerance PET release film according to claim 4, characterized in that, The refractive index calculation module calls a preset stress optics tensor polynomial to perform polynomial algebraic operations on the principal axis tension scalar data, calculating the dynamic refractive index drift. The specific operation is as follows: Using absolute temperature field data as an index, the first-order constant, the second-order constant, and the third-order constant of the multiphysics precalibration are extracted from the pre-set rheological matrix library. Calculate the first-order power value, the second-order power value, and the third-order power value of transient true tensile stress. The first-order product value is calculated by multiplying the first-order power value of the transient true tensile stress by the first-order constant of the multiphysics precalibration. The second-order product value is calculated by multiplying the second-order power value of the transient true tensile stress by the second-order constant of the multiphysics precalibration. The third-order product value is calculated by multiplying the third-order power value of the transient true tensile stress by the third-order constant of the multiphysics precalibration. The first, second, and third product values ​​are added together to calculate the initial value of the dynamic refractive index drift. The preset upper limit and lower limit of the elastic deformation zone threshold are extracted. The absolute value of the initial value of the dynamic refractive index drift is compared with the preset upper limit and lower limit of the elastic deformation zone threshold. When the absolute value of the initial value of the dynamic refractive index drift is greater than the upper limit of the preset elastic deformation zone threshold, or when the absolute value of the initial value of the dynamic refractive index drift is less than the lower limit of the preset elastic deformation zone threshold, the value of zero is set as the final dynamic refractive index drift. When the absolute value of the initial value of the dynamic refractive index drift is less than or equal to the upper limit of the preset elastic deformation zone threshold, and the absolute value of the initial value of the dynamic refractive index drift is greater than or equal to the lower limit of the preset elastic deformation zone threshold, the initial value of the dynamic refractive index drift is set as the final dynamic refractive index drift.

6. The thickness detection system for high-precision thickness tolerance PET release film according to claim 5, characterized in that, The pseudo-thickness stripping module receives the dynamic refractive index drift calculated by the refractive index calculation module and extracts the original optical path difference data contained in the force-optical synchronization mapping feature set generated by the spatiotemporal mapping module. The specific operation is as follows: Extract the dynamic refractive index drift and the initial longitudinal tensile strain; Extract the raw optical path difference data contained in the force-optical synchronization mapping feature set; Extract the preset initial static refractive index; The real-time absolute refractive index is calculated by adding the preset initial static refractive index and the dynamic refractive index drift. The apparent physical thickness is calculated by dividing the original optical path difference data by the real-time absolute refractive index.

7. The thickness detection system for high-precision thickness tolerance PET release film according to claim 6, characterized in that, The preset intrinsic Poisson's ratio, original optical path difference data, and refractive index dynamic drift are input into the preset thickness reduction equation for nonlinear algebraic solution to calculate the stress-free real material reference thickness. The specific operation is as follows: Extract the preset intrinsic Poisson's ratio, add the constant to the initial longitudinal tensile strain, and calculate the absolute ratio of tensile elongation. Using the preset intrinsic Poisson's ratio as the exponent, the absolute ratio of stretching and elongation is exponentially calculated to obtain the volume conservation reverse expansion factor. The apparent physical thickness is multiplied by the volume conservation reverse expansion factor to calculate the stress-free real material reference thickness. Extract the absolute lateral coordinates and integral displacement values ​​corresponding to the grid nodes contained in the two-dimensional physical mesh; The stress-free real material reference thickness is associated with and combined with the absolute lateral coordinate and integral displacement value to output the stress-free real material reference thickness with absolute lateral coordinate and integral displacement value.

8. The thickness detection system for high-precision thickness tolerance PET release film according to claim 7, characterized in that, The system receives the stress-free real material reference thickness calculated by the pseudo-thickness stripping module and the force-light synchronous mapping feature set generated by the spatiotemporal mapping module. Then, it performs a difference operation between the stress-free real material reference thickness and the preset target thickness to generate an opening fine-tuning command. The specific operation is as follows: Receives stress-free real material reference thickness and force-optical synchronous mapping feature set; Extract the preset target thickness, subtract the stress-free real material reference thickness from the preset target thickness, and calculate the absolute geometric space error. Extract the preset mold head bolt space mapping ratio coefficient, multiply the absolute geometric space error by the preset mold head bolt space mapping ratio coefficient, calculate the displacement step command, and set the displacement step command as the opening fine adjustment command.

9. A high-precision thickness tolerance PET release film thickness detection system according to claim 8, characterized in that, The original optical path difference data contained in the force-optical synchronization mapping feature set is extracted. The apparent physical thickness is calculated by dividing the original optical path difference data by the preset initial static refractive index. The specific operation is as follows: The original optical path difference data contained in the force-optical synchronization mapping feature set is extracted, the preset initial static refractive index is extracted, the original optical path difference data is divided by the preset initial static refractive index, and the apparent physical thickness is calculated.

10. A high-precision thickness tolerance PET release film thickness detection system according to claim 9, characterized in that, The collaborative feedback module calculates the difference between the apparent physical thickness and the stress-free real material reference thickness to generate a feedforward torque compensation command. The specific operation is as follows: The thickness difference is calculated by subtracting the stress-free real material reference thickness from the apparent physical thickness, and then the thickness difference is converted into an absolute difference. Extract the preset elastic moment conversion stiffness coefficient, multiply the absolute difference by the preset elastic moment conversion stiffness coefficient, and calculate the macroscopic linear failure traction force. The real-time physical roll diameter of the take-up roll is obtained, and the transient electromagnetic interference torque is calculated by multiplying the macroscopic linear destructive traction force by the real-time physical roll diameter of the take-up roll. Set the transient electromagnetic interference torque as a feedforward torque compensation command.

Images