A thin film capacitor winding process

Abstract

The present application relates to the technical field of thin film capacitor, especially to a thin film capacitor winding process, comprising the following steps: step one, interface tension control in upstream area; step two, in-situ modification treatment of the edge before the cutting point; step three, negative pressure lamination and winding; step four, in-situ online detection of multi-physical state; and step five, nonlinear dynamics closed-loop compensation across procedures. The present application applies the first physical field to the metalized film during the transportation to the winding forming path through the master control system, to control the viscoelasticity and inhibit microslip, to modify the evaporation cutting edge area online to eliminate the edge electric field concentration, to meet the nonlinear decoupling balance condition in the negative pressure lamination and winding to discharge the interlayer air gap, to form the real-time state characteristic sequence of the multi-physical field through the in-situ online detection, and to calculate the maximum Lyapunov index according to the four-stage control cycle of prediction, pre-adjustment, verification and correction to cooperatively adjust the multi-procedure parameters, so as to realize the multi-procedure space-time dynamic cooperative intervention and in-situ closed-loop control in the thin film capacitor winding process.

Description

Technical Field

[0001] This invention relates to the field of thin-film capacitor technology, and more particularly to a thin-film capacitor winding process. Background Technology

[0002] As a core passive component of power electronic systems, film capacitors are typically made from metallized films through multiple processes such as slitting, conveying, winding, and laminating.

[0003] Existing thin-film capacitor winding processes generally adopt a discrete production mode with independent control between processes. That is, the film conveying tension, edge treatment, winding and forming and quality inspection are executed by independent control systems according to preset static parameters. There is a lack of real-time sharing and collaborative control mechanism for the dynamic evolution characteristics of multi-physical fields between processes.

[0004] During the winding process, multiple physical factors, such as the viscoelastic behavior of the metallized thin film, the electric field distribution of the vapor deposition edge cutting area, and the state of the interlayer air gap, are coupled with each other and evolve nonlinearly over time. Existing technologies can only monitor the local state of a single process offline or with lag, and cannot detect and correct defects such as micro-slippage, interlayer bubbles, and edge electric field concentration caused by cross-process coupling in situ. As a result, it is difficult to guarantee the capacitance consistency, equivalent series resistance stability, and withstand voltage reliability of the capacitor core.

[0005] Therefore, how to achieve multi-process spatiotemporal dynamic coordinated intervention and in-situ closed-loop control during the winding of thin-film capacitors is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by providing a thin-film capacitor winding process. This process involves a main control system applying a first physical field along the metallized thin film's transport path to the winding stage to regulate viscoelasticity and suppress micro-slip; online in-situ modification of the vapor-deposited edge-cutting area to eliminate edge electric field concentration; satisfying nonlinear decoupling equilibrium conditions during negative pressure lamination and co-winding to eliminate interlayer air gaps; and in-situ online detection to form a real-time multi-physics field state characteristic sequence. Furthermore, a four-stage control cycle of prediction, pre-adjustment, verification, and correction is used to calculate the maximum Lyapunov exponent and coordinately adjust multi-process parameters. This achieves multi-process spatiotemporal dynamic coordinated intervention and in-situ closed-loop control during the thin-film capacitor winding process. This solves the problem in existing thin-film capacitor winding processes where independent control of each process prevents in-situ sensing and forward-looking correction of defects such as micro-slip, interlayer bubbles, and edge electric field concentration caused by cross-process coupling.

[0007] To achieve the above objectives, the present invention provides a thin-film capacitor winding process, wherein the thin-film capacitor winding process, under the control of a main control system, involves multi-process spatiotemporal dynamic collaborative intervention and in-situ closed-loop control along the path from the metallized thin film to the winding formation, and includes the following steps:

[0008] Step 1: Upstream region interface conveying tension regulation. A first physical field is applied to at least two metallized films during conveying to regulate the viscoelasticity of the metallized film surface, and the microscopic relative slip rate of the metallized film during the conveying process is controlled within a set slip threshold, and a metallized film with predetermined surface energy morphology characteristics is output. Step 2: In-situ modification of the front edge of the cutting point. Received from Step 1, the main control system performs online in-situ modification of the vapor-deposited cutting edge area of ​​the metallized film according to the current winding parameters to improve the local surface resistivity of the vapor-deposited cutting edge area. Step 3: Negative pressure lamination and co-winding at the tangent point. Radial lamination force is applied to the capacitor core formed by overlapping winding in a negative pressure environment, so that the metallized film modified in Step 2 is overlapped and wound into shape; wherein, the interlayer suction force generated by the negative pressure environment and the radial lamination force satisfy the nonlinear decoupling equilibrium condition of interlayer gas extrusion resistance. Step 4: In-situ online detection of multiple physical quantities. During the winding process, the dynamic electrical parameters and interlayer air gap state of the currently wound part of the capacitor core are detected in-situ online to form a real-time state feature sequence for characterizing the dynamic evolution of multiple physical fields inside the core. Step 5: Cross-process nonlinear dynamic closed-loop compensation. Based on the real-time state characteristic sequence collected in Step 4, calculate the maximum Lyapunov exponent of the dynamic capacitance change rate time series, and adjust the first physical field in Step 1, the current winding parameters in Step 2, and the radial lamination external force in Step 3 in a four-stage control cycle of prediction, pre-adjustment, verification, and correction.

[0009] The beneficial effects of this invention are as follows: By applying a first physical field to the metallized thin film on the winding path through the main control system to regulate viscoelasticity and suppress micro-slippage, online in-situ modification of the evaporation edge trimming area to eliminate edge electric field concentration, satisfying nonlinear decoupling equilibrium conditions to discharge interlayer air gaps during negative pressure lamination and co-winding, forming a real-time state characteristic sequence of multi-physical fields through in-situ online detection, and calculating the maximum Lyapunov exponent according to the four-stage control cycle of prediction-pre-adjustment-verification-correction to coordinate and adjust the parameters of multiple processes, the spatiotemporal dynamic coordinated intervention and in-situ closed-loop control of the thin film capacitor winding process are realized. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating the steps of the present invention. Detailed Implementation

[0011] The present invention will now be described in detail with reference to the accompanying drawings.

[0012] like Figure 1As shown, the present invention discloses a thin-film capacitor winding process. Under the control of the main control system, the thin-film capacitor winding process performs multi-process spatiotemporal dynamic collaborative intervention and in-situ closed-loop control on the path from the metallized film to the winding. Through the main control system, the process implements unified spatiotemporal dynamic collaborative scheduling and in-situ closed-loop control of each process on the entire path from the metallized film to the winding, thereby breaking the discrete production mode of independent control of each process in the prior art. It establishes a real-time sharing and collaborative control mechanism for the dynamic evolution characteristics of multi-physical fields across processes, and overcomes the shortcomings of the prior art in that it cannot perform in-situ perception and forward-looking correction of cross-process coupling defects.

[0013] The thin-film capacitor winding process includes the following steps: Step 1: Upstream region interface conveying tension regulation. Apply a first physical field to at least two metallized films during conveying to regulate the viscoelasticity of the metallized film surface, and control the microscopic relative slip rate of the metallized film during the conveying process within a set slip threshold, and output a metallized film with predetermined surface energy morphology characteristics. By applying a first physical field to at least two metallized films during transport, and utilizing the interaction between the first physical field and the polymer chain segments on the surface of the metallized films to regulate the viscoelasticity of the metallized film surface, the frictional characteristics between the interface between the metallized film and the guide component are changed. The micro-relative slip rate is controlled within a set slip threshold, thereby suppressing the micro-slip caused by viscoelastic mismatch during high-speed transport of the metallized film, eliminating the wrinkles and tension fluctuations on the film surface caused by the accumulation of micro-slip, and outputting a metallized film with controllable surface energy morphology characteristics. This solves the problem in the prior art that it is impossible to detect and proactively correct micro-slip defects caused by cross-process coupling in situ.

[0014] Step 2: In-situ modification of the front edge of the cutting point. Receive the metallized film output from Step 1. The main control system performs online in-situ modification of the vapor deposition cutting edge area of ​​the metallized film according to the current winding parameters to improve the local surface resistivity of the vapor deposition cutting edge area. By adjusting the process conditions of online in-situ modification in real time according to the current winding parameters through the main control system, the evaporation trimming area of ​​the metallized thin film is modified, thereby changing the microstructure of the evaporation trimming area and increasing the local surface resistivity. This eliminates the edge electric field concentration effect in the evaporation trimming area, prevents edge discharge breakdown, and solves the problem that the electric field distribution of the evaporation trimming area evolves nonlinearly over time in the existing technology, which can only be monitored offline or with lag.

[0015] Step 3: Negative pressure lamination and co-winding at the tangent point. Radial lamination force is applied to the capacitor core formed by overlapping winding in a negative pressure environment, so that the metallized film modified in Step 2 is overlapped and wound into shape. Among them, the interlayer suction force generated by the negative pressure environment and the radial lamination force satisfy the nonlinear decoupling equilibrium condition of the interlayer gas extrusion resistance. By applying radial lamination force to the overlapping wound capacitor core in a negative pressure environment, the interlayer suction force generated by the negative pressure environment and the radial lamination force satisfy the nonlinear decoupling equilibrium condition of the interlayer gas extrusion resistance. Even if the interlayer suction force is not lower than the interlayer gas extrusion resistance caused by lamination, the synergistic effect of negative pressure suction and radial lamination is realized to effectively discharge the interlayer micro-gap from the inside of the wound core, avoiding partial discharge and capacitor decay caused by the residue of interlayer gas gaps. This solves the problem that the interlayer gas gap state evolves nonlinearly over time in the existing technology and the existing technology cannot perform in-situ sensing and look-ahead correction.

[0016] Step 4: In-situ online detection of multiple physical quantities. During the winding process, the dynamic electrical parameters and interlayer air gap state of the currently wound part of the capacitor core are detected in situ online to form a real-time state feature sequence for characterizing the dynamic evolution of multiple physical fields inside the core. By performing in-situ online detection of the dynamic electrical parameters and interlayer air gap state of the currently wound portion of the capacitor core during the winding process, the dispersed multi-physical quantity measurement data are integrated into a real-time state feature sequence to characterize the dynamic evolution of the multi-physical field inside the core. This achieves an upgrade from local offline monitoring of a single process to in-situ online real-time sensing of coupled multi-physical quantities, providing continuous and multi-dimensional state information input for cross-process closed-loop control, and solving the problem of lack of real-time sharing of the dynamic evolution characteristics of the multi-physical field between processes in the existing technology.

[0017] Step 5: Cross-process nonlinear dynamic closed-loop compensation. Based on the real-time state characteristic sequence collected in Step 4, calculate the maximum Lyapunov exponent of the dynamic capacitance change rate time series, and adjust the first physical field of Step 1, the current winding parameters of Step 2, and the radial lamination external force of Step 3 in a four-stage control cycle of prediction, pre-adjustment, verification, and correction.

[0018] By calculating the maximum Lyapunov exponent based on the real-time state characteristic sequence to determine the nonlinear dynamic stability of the winding process, and by coordinating and adjusting the process parameters of steps one to three according to the four-stage control cycle of prediction, pre-adjustment, verification and correction, the independent static control of each process is upgraded to cross-process spatiotemporal dynamic collaborative intervention and in-situ closed-loop control. When divergent instability trends occur in the winding process, timely intervention is carried out through forward correction to prevent micro-defects from cascading and amplifying into macro-defects. This solves the problem in the existing technology that it is impossible to forward correct defects caused by cross-process coupling.

[0019] During operation, the metallized film, after being output from the unwinding rack, enters the upstream interface conveying tension control process. The main control system applies a first physical field to at least two metallized films during conveying. This first physical field regulates the viscoelasticity of the metallized film surface, controlling the microscopic relative slip rate of the metallized film during conveying within a set slip threshold, resulting in a metallized film with predetermined surface energy morphology characteristics. Subsequently, the metallized film enters the in-situ modification process before the cutting point. The main control system performs online in-situ modification of the vapor-deposited cutting edge area of ​​the metallized film according to the current winding parameters, increasing the local surface resistivity of the vapor-deposited cutting edge area. The modified metallized film then enters the negative pressure lamination and co-winding process at the cutting point. In a negative pressure environment, a radial lamination force is applied to the overlapping wound capacitor core. The interlayer suction force generated by the negative pressure environment and the radial lamination force satisfy the nonlinear decoupling equilibrium condition of the interlayer gas extrusion resistance. During the winding process, the dynamic electrical parameters of the capacitor core and the interlayer air gap state are detected online in situ to form a real-time state characteristic sequence. The main control system calculates the maximum Lyapunov exponent based on the real-time state characteristic sequence, and performs coordinated adjustment of the first physical field, current winding parameters and radial lamination external force according to the four-stage control cycle of prediction, pre-adjustment, verification and correction.

[0020] The first physical field can be an ultrasonic energy field, a near-infrared radiation heating field, a superposition of ultrasonic energy field and near-infrared thermal field, or a mid-frequency electromagnetic induction heating field.

[0021] In one specific embodiment of the present invention, the first physical field for tension control at the interface in the upstream region of step one is an ultrasonic energy field. Taking BOPP (biaxially oriented polypropylene) metallized film as an example, the aluminum vapor deposition layer thickness is 30 nm, the total film thickness is 4.5 μm, the winding operating temperature is 45℃~55℃, and the winding linear speed is 2~5 m / s. The main control system embeds a piezoelectric ultrasonic transducer array inside the guide roller in the upstream region of the conveying path, and sets a set of piezoelectric ultrasonic transducers every 200~500 mm along the conveying direction of the metallized film. The α relaxation characteristic time τ of the BOPP film at the winding temperature is 4~8 μs, corresponding to the α relaxation characteristic frequency. The frequency range is 20–40 kHz. The main control system sets the ultrasonic wave transmission frequency to a value that matches the α relaxation characteristic frequency, i.e. This process induces resonant coupling between ultrasonic energy and the relaxation motion of polymer segments on the metallized film surface. Through this resonant coupling, the polymer segments on the metallized film surface achieve excited motion with maximum amplitude, increasing the segment motion amplitude by 3–5 times. This significantly alters the viscoelasticity of the metallized film surface, increasing the elastic response component, decreasing the viscous hysteresis component, and improving the equivalent loss factor. Reduced by 30% to 50%.

[0022] The microscopic relative slip ratio between the BOPP film and the stainless steel guide roller interface in the absence of an ultrasonic energy field. The value is 0.05–0.10; after applying an ultrasonic energy field, the microscopic relative slip rate decreases according to a negative exponential law, i.e. Where α is the acoustic-friction coupling coefficient, and for the interface between BOPP film and stainless steel guide roller, α is 0.3–0.8 s·cm² / W. IUS is the ultrasonic power density, set to 5–15 W / cm². When IUS = 10 W / cm² and α = 0.5 s·cm² / W, the microscopic relative slip ratio ε is approximately 0.007, meaning the microscopic relative slip ratio drops to about 7% of that without an ultrasonic energy field. The main control system sets the slip threshold to 0.01 and monitors the microscopic relative slip ratio in real time. When the microscopic relative slip ratio exceeds the set slip threshold, the main control system automatically increases the ultrasonic power density.

[0023] After ultrasonic resonance treatment, the segment orientation order parameter S on the surface of the metallized film increased from approximately 0.2 to over 0.6. The polymer segments were oriented and orderly arranged along the conveying direction of the metallized film. The surface roughness Ra of the metallized film decreased from approximately 15 nm to below 8 nm, and the surface energy of the metallized film was adjusted from approximately 28 mN / m to the range of 32–35 mN / m. Furthermore, a micro-thermocouple array was embedded on the surface of the guide roller to measure the temperature distribution of the metallized film along the conveying direction in real time; the α-relaxation characteristic time τ followed the temperature... The α-relaxation activation energy Ea of the BOPP film is approximately 1.0 eV. When the temperature rises from 45℃ to 55℃, the α-relaxation characteristic time τ shortens from approximately 8μs to approximately 3μs, corresponding to an increase in the resonant frequency from approximately 20kHz to approximately 53kHz. The main control system adjusts the frequency of the ultrasonic transmitters at each guide roller in real time based on the measured local temperature from the micro thermocouple array, ensuring that the resonance matching condition is maintained continuously despite the temperature drift throughout the metallized film transport process. When the temperature gradient |dT / dx| between adjacent guide rollers exceeds 1 K / mm, the ultrasonic frequency switching between adjacent guide rollers uses linear interpolation to avoid abrupt changes in the acoustic field standing wave mode caused by frequency steps.

[0024] In another specific embodiment of the present invention, the first physical field in step one is a near-infrared radiation heating field. Taking a PET (polyester) metallized film as an example, the zinc vapor deposition layer thickness is 40 nm, the total film thickness is 6 μm, and the winding working temperature is 35℃~45℃. The main control system sets two sets of near-infrared radiation heating modules in the upstream region of the conveying path. The near-infrared radiation wavelength is 1.0~2.0 μm, the power density is 0.5~3 W / cm², and the distance between the two sets of near-infrared radiation heating modules along the conveying direction of the metallized film is 300 mm. At 35℃~45℃, the PET film is in the secondary relaxation region below the glass transition temperature, and the movement of polymer chain segments is significantly affected by temperature. Near-infrared radiation heating rapidly raises the surface temperature of the metallized film from the ambient temperature of 25℃ to the temperature range of 40℃~50℃, with a heating rate of 5~10℃ / s. By increasing the temperature, the mobility of the polymer chains on the surface of the metallized film is enhanced, and the viscoelasticity of the metallized film surface changes from a rigid state to a flexible state: the elastic modulus decreases from about 3.5 GPa to about 2.0 GPa, and the loss factor tanδ increases from about 0.015 to about 0.035. The change in the viscoelasticity of the metallized film surface transforms the interface between the metallized film and the guide roller from a rigid, high-friction contact to a flexible, low-friction bonding, with the coefficient of friction decreasing from about 0.35 to about 0.20.

[0025] The microscopic relative slip ratio between the PET film and the stainless steel guide roller interface when there is no near-infrared radiation heating field. The coefficient of friction is 0.08–0.15; after applying a near-infrared radiation heating field, the friction coefficient decreases, causing the microscopic relative slip ratio to drop to approximately 0.02–0.04. The main control system sets the slip threshold to 0.05 and adjusts the near-infrared radiation power density based on feedback from the online slip ratio sensor. Near-infrared radiation heating gives the molecular chains on the surface of the PET metallized film thermal energy. The thermal softening effect of the micro-protrusions on the surface of the metallized film reduces the surface roughness Ra of the metallized film from approximately 20 nm to approximately 12 nm, and increases the surface energy of the metallized film from approximately 38 mN / m to approximately 42–45 mN / m. The main control system strictly controls the surface temperature of the metallized film in the near-infrared radiation heating zone to not exceed 50°C to prevent irreversible deformation of the PET film when it approaches the glass transition temperature. The main control system dynamically adjusts the near-infrared radiation power according to the metallized film conveying speed. When the metallized film conveying speed increases, the main control system increases the near-infrared radiation power to maintain the same temperature rise; when the metallized film conveying speed decreases, the main control system decreases the near-infrared radiation power.

[0026] In step two of this embodiment, the specific steps of the online in-situ modification process include: A pulsed laser beam, linked in a closed loop with the winding shaft, is dynamically triggered based on the real-time acquired winding radius and the preset inner tab spacing. This allows for localized laser melting of periodic sites on the surface of the metallized thin film, resulting in localized thickening of the metal layer. This enables the in-situ construction of distributed, integrated inner tabs within the capacitor core. By employing a pulsed laser beam linked in a closed loop with the winding shaft, and dynamically triggering the laser pulse based on the real-time acquired winding radius and the preset inner tab spacing, localized laser melting of periodic sites on the surface of the metallized thin film results in localized thickening of the metal layer. This achieves the in-situ construction of distributed, integrated inner tabs within the capacitor core, replacing the traditional post-welded tab process and eliminating the risk of damage to the thin film dielectric from the welding heat-affected zone.

[0027] The distribution density of the distributed integrated inner electrode decreases as the winding radius increases, resulting in a higher electrode distribution density in the inner ring of the capacitor core compared to the outer ring. By making the distribution density of the distributed integrated inner electrode decrease as the winding radius increases, i.e., the inner ring distribution density is higher than the outer ring, the current density concentration effect caused by the small radius of the inner ring is compensated, and the Joule heat distribution of the inner and outer rings tends to be uniform.

[0028] The pulse width of the laser pulse satisfies the following thermal diffusion constraints: During the laser pulse, the thermal diffusion depth does not exceed the thickness of the metal layer, thus confining the heat-affected zone (HAZ) within the metal layer. By limiting the laser pulse width to a range where the thermal diffusion depth does not exceed the thickness of the metal layer, and utilizing the finite time window of thermal diffusion to confine the HAZ within the metal layer, thermal damage to the thin film substrate beneath the metal layer due to laser thermal effects is avoided, ensuring that the dielectric properties of the metallized thin film remain unaffected.

[0029] Specifically, the distributed integrated inner tabs constructed on the surface of the metallized thin film by pulsed laser beams appear as a stepped array of protrusions arranged periodically along the film transport direction in macroscopic topology. By making the distributed integrated inner tabs appear as a stepped array of protrusions arranged periodically along the film transport direction in macroscopic topology, a uniformly distributed current collection path is provided by the periodic stepped protrusion structure, avoiding local current congestion caused by isolated tab points.

[0030] At the microscopic level, a single integrated inner tab contains, in the thickness direction of the metal layer, a fused thickening layer, a grain refinement transition layer, and an unmelted conductive base layer that extend sequentially from the outside to the inside. By forming an integrated transition extension structure of the fused thickening layer, the grain refinement transition layer, and the unmelted conductive base layer in the thickness direction of the metal layer, no interface abruptness between the three layers is achieved, avoiding the interface contact resistance of traditional welded tabs.

[0031] The thickness of the fused thickening layer is 1.5 to 3 times the initial thickness of the unmelted conductive substrate, and the average grain size of the grain refinement transition layer is smaller than the average grain size of the unmelted conductive substrate. By setting the thickness of the fused thickening layer to 1.5 to 3 times the initial thickness of the unmelted conductive substrate to increase the current-carrying cross-sectional area, and at the same time making the average grain size of the grain refinement transition layer smaller than the unmelted conductive substrate to improve the local conductivity, the equivalent resistance of the tab site is reduced through the dual effects of thickening and grain refinement.

[0032] The distributed integrated inner tabs are interlocked and bonded together by the radial lamination force in step three, creating an axially continuous low-impedance current-collecting channel in situ inside the capacitor core. By utilizing the radial lamination force in step three to interlock and bond the distributed integrated inner tabs on the upper and lower metallized films, an axially continuous low-impedance current-collecting channel is created in situ inside the capacitor core. This achieves the goal of connecting discrete tab sites into an axially continuous current-collecting channel through interlocking bonding, significantly reducing the equivalent series resistance.

[0033] Specifically, when performing localized laser melting on periodic sites, the specific steps include: In the spatial confinement step, before the pulsed laser beam acts, an annular heat sink cooling zone is constructed around the periodic sites using an airflow jet field or mechanical clamping structure, thus limiting the heat conduction boundary generated by the pulsed laser beam to a diameter not exceeding 500 mm. Within a local micro-region; by constructing an annular heat sink cooling zone around the periodic sites through an airflow jet field or mechanical clamping structure before the pulsed laser beam acts, the heat conduction boundary is limited to a local micro-region with a diameter of no more than 500 μm, thus achieving precise control of the thermal influence range of laser melting and preventing heat from diffusing to surrounding non-target areas, which would cause thermal damage to the thin film substrate.

[0034] The phase modulation step involves simultaneously injecting an inert protective gas stream to periodic sites during the pulsed laser beam's action. By controlling the energy density waveform of the pulsed laser beam, the metal layer undergoes a solidification process characterized by rapid melting and accumulation followed by segmented, stepped cooling. By simultaneously injecting the inert protective gas stream during the pulsed laser beam's action to prevent oxidation of the molten metal, and by controlling the energy density waveform of the pulsed laser beam to achieve a solidification process characterized by rapid melting and accumulation followed by segmented, stepped cooling, the microstructure of the metal layer is optimized through speed-controlled solidification. This reduces solidification shrinkage cavities and hot cracks, and improves the conductivity and mechanical reliability of the inner tab.

[0035] Specifically, as the radius of the winding shaft increases, the main control system dynamically adjusts the single-pulse energy of the pulsed laser beam to ensure that the axial width W of a single integrated inner tab and the contact length L in the direction of metallization film transport satisfy the following limiting control relationship: 2.5 Where W is the axial width and L is the contact length.

[0036] By dynamically adjusting the single-pulse energy of the pulsed laser beam according to the radius of the winding shaft through the main control system, the ratio of the axial width W to the contact length L of a single integrated inner tab is controlled between 1.2 and 2.5. This ensures that the inner tab has sufficient current-carrying width in the axial direction and appropriate contact length in the conveying direction, avoiding insufficient axial current-carrying capacity due to W / L being too small or excessively large resulting in too small spacing between adjacent tabs in the conveying direction.

[0037] In a typical embodiment of the present invention, in-situ modification is carried out using the production of high-voltage, high-current film capacitor cores as an example. In this embodiment, the metallization film used is a biaxially oriented polypropylene (BOPP) film substrate, and its thickness is preferably 3.0 μm (in practical applications, it can also be selected within the aforementioned range or a wide range of 2.0 μm to 8.0 μm according to the voltage withstand requirements); a vapor-deposited layer is deposited on the surface of the metallization film, specifically a zinc-aluminum composite metal layer (Zn-Al) in this embodiment, and the initial thickness of the metal layer is, for example, 30 nm.

[0038] In the process flow of this invention, the online in-situ modification system in step two employs a high-precision pulsed fiber laser, with an output wavelength preferably of 1064 nm and a pulse width set to [value missing]. = 15 ns. The main control system maintains real-time closed-loop linkage with the servo motor and photoelectric encoder of the winding shaft, thereby enabling in-situ, online acquisition of the current winding radius R0. Specifically, when the winding is in its initial stage and the initial winding radius R1 = 20 mm (i.e., the inner ring region of the capacitor core) is collected in real time, the main control system dynamically calculates and triggers laser pulses based on the preset axial spacing of the inner tabs, controlling the center distance between adjacent inner tabs in the metallization film conveying direction to be d1 = 5 mm. As the winding continues, when the winding axis radius increases to the outer ring region and the real-time collected winding radius R2 = 60 mm, the main control system adaptively reduces the pulse triggering frequency through an algorithm, thereby increasing the center distance between adjacent inner tabs to d2 = 15 mm. Thus, this invention, through the dynamic frequency intervention of the main control system, makes the distribution density of the distributed integrated inner tabs exhibit a non-linear decreasing trend with the increase of the winding radius. This results in a significantly higher tab distribution density in the inner ring of the capacitor core compared to the outer ring, fundamentally compensating for the current density concentration effect caused by the small radius of the inner ring, and making the Joule heat distribution of the inner and outer rings of the final capacitor core more uniform during service.

[0039] During laser in-situ melting control, the pulse width of the laser pulse must strictly meet specific thermal diffusion constraints. The overall thermal diffusivity of the Zn-Al metal layer material in this embodiment is known. According to the governing equation for thermal diffusion depth: Theoretical calculations were performed, and the specific pulse width of this embodiment was substituted. Then, its theoretical thermal diffusion depth can be obtained. ≈42.4nm. In actual process control, by utilizing the instantaneous energy decay characteristics of ultrashort pulse lasers at the microsecond level, this thermal diffusion depth is precisely locked within a safe boundary that does not exceed the initial thickness of the metal layer. By confining the thermal diffusion depth during laser pulse action to the interior of the metal layer, the finite time window of thermal diffusion is used to block the conduction of heat to the depth of the thin film substrate. This perfectly avoids thermal damage such as microscopic thermal melting, thermal depression, or perforation caused by laser thermal effects on the polypropylene thin film substrate below the metal layer, ensuring that the initial dielectric strength of the metallized thin film substrate (e.g., maintained at >650μm) remains unaffected.

[0040] Furthermore, during the local laser melting of periodic sites, the following spatial confinement and phase control steps were specifically implemented in the process. In the spatial confinement step, 0.5 ms before the pulsed laser beam focuses, an annular airflow jetting mechanism synchronously jets a cold argon gas with a preferred pressure of 0.4 MPa around the target periodic site (in a modified embodiment, a precision mechanical clamping structure can also be used to block heat transfer), thereby constructing an annular dynamic heat sink cooling zone around the periodic site. In-situ measurement using infrared thermal microscopy showed that this heat sink cooling zone successfully confined the heat conduction boundary generated by the pulsed laser beam to a local micro-region with a diameter not exceeding 500 μm (the actual measured value in this embodiment is 350 μm), achieving precise mechanical locking of the laser melting thermal influence range and preventing heat diffusion to surrounding non-target areas. In the phase control step, during the pulsed laser beam action and the subsequent solidification period, an inert protective gas flow is continuously injected into the periodic sites to prevent oxidation of the molten metal. Simultaneously, the main control system uses digital signals to control the laser power modulator, nonlinearly modulating the energy density waveform of the pulsed laser beam to create a stepped energy characteristic with a "steep initial and gradual subsequent" pattern: the energy density rapidly spikes to a peak of 2.5 J / cm² within the first 5 ns. 2 This drives the metal layer to completely melt locally and instantaneously, generating microfluids that accumulate and thicken towards the center; in the subsequent 10 ns, the energy density decreases stepwise and remains at 0.8 J / cm². 2This process guides the metal layer through a segmented, stepped cooling solidification control process. This rate-controlled solidification optimizes the microcrystalline structure of the metal layer, significantly reduces solidification shrinkage cavities and hot cracks caused by rapid quenching and crystallization, and substantially improves the mechanical reliability and electrical continuity of the subsequently formed inner tabs.

[0041] Through the combined intervention of spatial confinement and phase modulation steps described above, the distributed integrated inner tabs constructed on the surface of the metallized thin film by pulsed laser beams exhibit a macroscopic topological appearance as a stepped array of protrusions periodically arranged along the film transport direction. Microscopic morphology characterization using focused ion beam (FIB) sectioning combined with scanning electron microscopy (SEM) reveals that a single integrated inner tab comprises, from the outside to the inside, a fused thickening layer, a grain refinement transition layer, and an unmelted conductive substrate layer, extending in an integrated manner from the outside to the inside. The three layers exhibit a continuous atomic-level lattice transition, without any physical interfaces or macroscopic heterogeneous interfaces, thus eliminating the unavoidable interfacial contact resistance inherent in traditional tab welding processes. In this embodiment, the thickness of the fused thickened layer was found to be 75 nm, which is exactly 2.5 times the initial thickness of the unmelted conductive substrate (30 nm) (strictly falling within the preferred range of 1.5 to 3 times), thus significantly increasing the local current-carrying cross-sectional area geometrically. At the same time, due to the aforementioned controlled-rate condensation, the average grain size of the grain refinement transition layer is 12 nm, which is significantly smaller than the average grain size of the unmelted conductive substrate (45 nm). By utilizing the conductivity enhancement effect of ultrafine grains, the equivalent resistance of the tab sites is greatly reduced through the dual physical effects of thickening and grain refinement.

[0042] As the radius of the winding shaft continues to increase, in order to ensure the standardization and adaptive adaptation of the geometric dimensions of each loop of the electrode, the main control system dynamically adjusts the single-pulse energy of the pulsed laser beam through an algorithm based on the current instantaneous linear velocity and radius increment (for example, dynamically compensating and fine-tuning the single-pulse energy between 15μJ and 22μJ). This ensures that the axial width W of a single integrated inner electrode and the contact length L in the direction of metallization film transport strictly meet the defined control relationship and maintain their ratio. The ratio is between 1.2 and 2.5 (in this embodiment, the ratio is constantly controlled around 1.8). This control method perfectly ensures that the inner tab has a sufficiently wide current-carrying cross-section in the axial direction, while maintaining an appropriate geometric span in the film conveying direction. This completely avoids the technical defects of insufficient axial current-carrying capacity due to an excessively small ratio, or accidental overlapping and misalignment during winding due to an excessively large ratio causing the spacing between adjacent tabs in the conveying direction to be too small.

[0043] Finally, when at least two metallized films that have undergone the aforementioned in-situ online modification are delivered and converge at the winding cut point, step three is performed to apply a radial lamination force of 0.3 MPa to the capacitor core being overlapped and wound. Under the mechanical drive of this radial mechanical stress, the spatially corresponding distributed integrated inner tabs (i.e., the stepped protrusion microarray) on the two or more metallized films undergo a high-density micro-mechanical interlocking at the cut point. This bonding action connects and reconstructs the tab sites, which were originally discretely distributed on a single film, in situ within the capacitor core into one or more axially connected, low-impedance current-collecting channels, replacing the traditional post-welding tab process. This fundamentally eliminates the risk of further damage to the film dielectric from the welding heat-affected zone and significantly reduces the equivalent series resistance (ESR) of the finished film capacitor.

[0044] In step two, the specific implementation method of online in-situ modification is as follows: In this embodiment, the metallized thin film is specifically a polypropylene thin film with an aluminum metal layer deposited on it. The initial thickness of the aluminum metal layer is 30 nanometers. The pulsed laser beam is specifically an infrared nanosecond pulsed fiber laser with a wavelength of 1064 nanometers.

[0045] The main control system employs an infrared nanosecond pulsed fiber laser that is linked in a closed loop with the winding shaft. Based on real-time acquisition of the winding radius and a preset inner tab spacing, the laser pulse is dynamically triggered to locally melt periodic sites on the polypropylene film surface, causing localized thickening of the aluminum metal layer. This results in the in-situ construction of distributed, integrated inner tabs within the capacitor core. The pulse width of the infrared nanosecond pulsed fiber laser is set to 15 nanoseconds. This 15-nanosecond pulse width ensures that the thermal diffusion depth during the laser pulse is 20 nanometers, not exceeding the initial 30-nanometer thickness of the aluminum metal layer.

[0046] The distribution density of the distributed integrated inner tabs decreases as the winding radius increases. When the winding radius is 30 mm, the distribution density of the distributed integrated inner tabs is set to 50 per meter; when the winding radius increases to 60 mm, the distribution density of the distributed integrated inner tabs is set to 20 per meter, making the tab distribution density of the inner ring of the capacitor core higher than that of the outer ring. The distributed integrated inner tabs constructed on the surface of the polypropylene film by an infrared nanosecond pulsed fiber laser appear in a macroscopic topology as a stepped array of protruding microarrays periodically arranged along the film transport direction.

[0047] At the microscopic level, a single integrated inner tab comprises, from the outside to the inside, an integrally transitioning fused thickening layer, a grain refinement transition layer, and an unmelted conductive substrate layer, extending sequentially from the outside to the inside of the aluminum metal layer. The thickness of the fused thickening layer is set to 60 nanometers. Since the initial thickness of the aluminum metal layer is 30 nanometers, after removing the unmelted portion, the initial thickness of the unmelted conductive substrate layer is 10 nanometers. The thickness of the fused thickening layer, 60 nanometers, is twice the initial thickness of the unmelted conductive substrate layer, 10 nanometers. This 2 times is between 1.5 and 3 times, meeting the process requirements. The average grain size of the grain refinement transition layer is 50 nanometers, smaller than the average grain size of the unmelted conductive substrate layer, 200 nanometers. The distributed integrated inner tabs are interlocked and bonded together by the radial lamination force in step three, constructing an axially continuous low-impedance current collection channel in situ inside the capacitor core.

[0048] When performing localized laser melting on periodic sites, the specific steps include spatial confinement and phase modulation steps: In the spatial confinement step, before the infrared nanosecond pulsed fiber laser is applied, a ring-shaped heat sink cold zone is constructed around the periodic site using a gas jet field. Specifically, the gas jet field uses compressed nitrogen gas at -10 degrees Celsius to confine the heat conduction boundary generated by the infrared nanosecond pulsed fiber laser within a local micro-region with a diameter of 400 micrometers.

[0049] In the phase modulation step, during the operation of the infrared nanosecond pulsed fiber laser, an inert protective gas flow is simultaneously injected into the periodic sites. In this embodiment, the inert protective gas flow is specifically selected as argon gas with a purity of 99.99%. The main control system controls the energy density waveform of the infrared nanosecond pulsed fiber laser to cause the aluminum metal layer to undergo a solidification process of rapid melting and accumulation followed by segmented step-cooling.

[0050] As the winding shaft radius increases, the main control system dynamically adjusts the single-pulse energy of the infrared nanosecond pulsed fiber laser. When the winding shaft radius is 45 mm, the main control system adjusts the axial width W of the single integrated inner tab to 1.5 mm and the contact length L in the metallization film conveying direction to 1.0 mm. At this time, the ratio of axial width W to contact length L is 1.5. 1.5 falls between 1.2 and 2.5, satisfying the ratio constraint control relationship.

[0051] In step two, another specific implementation method for online in-situ modification is as follows: In this embodiment, the metallized thin film is specifically a polyethylene terephthalate film with a zinc-aluminum alloy metal layer deposited on it. The initial thickness of the zinc-aluminum alloy metal layer is 45 nanometers. The pulsed laser beam is specifically a 355-nanometer ultraviolet picosecond pulsed laser.

[0052] The main control system employs a UV picosecond pulsed laser linked in a closed loop with the winding shaft. Based on real-time data acquisition of the winding radius and a preset inner tab spacing, the laser pulses are dynamically triggered to locally melt periodic sites on the surface of the polyethylene terephthalate (PET) film, causing localized thickening of the zinc-aluminum alloy metal layer. This results in the in-situ construction of distributed, integrated inner tabs within the capacitor core. The pulse width of the UV picosecond laser is set to 10 picoseconds. This 10 picosecond pulse width ensures that the thermal diffusion depth during the laser pulse is 15 nanometers, not exceeding the initial 45-nanometer thickness of the zinc-aluminum alloy metal layer.

[0053] The distribution density of the distributed integrated inner tabs decreases as the winding radius increases. When the winding radius is 40 mm, the distribution density of the distributed integrated inner tabs is set to 40 per meter; when the winding radius increases to 80 mm, the distribution density of the distributed integrated inner tabs is set to 15 per meter, making the tab distribution density of the inner ring of the capacitor core higher than that of the outer ring. The distributed integrated inner tabs constructed on the surface of polyethylene terephthalate film by ultraviolet picosecond pulsed lasers appear in macroscopic topology as a stepped array of protruding microarrays periodically arranged along the film transport direction.

[0054] At the microscopic level, a single integrated inner tab comprises, from the outside to the inside, an integrally transitioning fused thickening layer, a grain refinement transition layer, and an unmelted conductive substrate layer, extending sequentially from the outside to the inside of the zinc-aluminum alloy metal layer. The thickness of the fused thickening layer is set to 75 nanometers. Since the initial thickness of the zinc-aluminum alloy metal layer is 45 nanometers, after removing the unmelted portion, the initial thickness of the unmelted conductive substrate layer is 25 nanometers. The thickness of the fused thickening layer, 75 nanometers, is three times the initial thickness of the unmelted conductive substrate layer, 25 nanometers. This three-fold difference falls between 1.5 and 3 times, meeting the process requirements. The average grain size of the grain refinement transition layer is 30 nanometers, smaller than the average grain size of the unmelted conductive substrate layer, 150 nanometers. The distributed integrated inner tab is interlocked and bonded to the two layers through radial lamination force in step three, creating an axially continuous low-impedance current collection channel in situ inside the capacitor core.

[0055] When performing localized laser melting on periodic sites, the specific steps include spatial confinement and phase modulation steps: In the spatial confinement step, before the ultraviolet picosecond pulsed laser is applied, a ring-shaped heat sink cooling zone is constructed around the periodic sites using a mechanical clamping structure. Specifically, the mechanical clamping structure employs a ring-shaped copper micro-pressure head, which is internally circulated with cooling water to confine the heat conduction boundary generated by the ultraviolet picosecond pulsed laser within a local micro-region with a diameter of 350 micrometers.

[0056] In the phase modulation step, during the operation of the ultraviolet picosecond pulsed laser, an inert protective gas flow is simultaneously injected into the periodic sites. In this embodiment, the inert protective gas flow is specifically a helium gas flow with a purity of 99.99%. The main control system controls the energy density waveform of the ultraviolet picosecond pulsed laser to cause the zinc-aluminum alloy metal layer to undergo a solidification process of rapid melting and accumulation followed by segmented step-cooling.

[0057] As the winding shaft radius increases, the main control system dynamically adjusts the single-pulse energy of the ultraviolet picosecond laser. When the winding shaft radius is 70 mm, the main control system adjusts the axial width W of the single integrated inner tab to 2.2 mm and the contact length L in the metallization film conveying direction to 1.0 mm. At this time, the ratio of axial width W to contact length L is 2.2. 2.2 is between 1.2 and 2.5, satisfying the ratio constraint control relationship.

[0058] In step three of this embodiment, the negative pressure environment is constructed by a local negative pressure field covering the winding cut point, and the interlayer micro-air gaps are simultaneously extracted along the axial direction to both ends of the winding shaft through the micro-perforated air extraction channel opened circumferentially inside the winding shaft. By constructing a negative pressure environment by covering the winding cut point with a local negative pressure field, and simultaneously extracting the interlayer micro-air gaps along the axial direction to both ends through the micro-perforated air extraction channel opened circumferentially inside the winding shaft, the negative pressure effect is precisely positioned in the winding cut point area. Air is directly extracted from the interlayer through the micro-perforated channel inside the shaft, which significantly shortens the air gap discharge path and improves exhaust efficiency compared to external air extraction methods.

[0059] The strength of the negative pressure field and the radial lamination force satisfy the following equilibrium condition: the interlayer suction force generated by the negative pressure field is not less than the interlayer gas extrusion resistance caused by lamination. By setting the interlayer suction force generated by the negative pressure field to be not less than the interlayer gas extrusion resistance caused by lamination, the suction force and extrusion resistance are balanced, ensuring that the negative pressure suction capacity is sufficient to overcome the resistance of gas being squeezed into the interlayer during lamination, and avoiding the formation of micro-gap defects due to insufficient suction force causing gas to remain in the interlayer.

[0060] In step three, the negative pressure environment is constructed by a local negative pressure field covering the winding cut point, and the interlayer micro-air gap removal process is performed. The specific implementation method of this embodiment is as follows: In this embodiment, the local negative pressure field is specifically designed by using a semi-enclosed arc-shaped negative pressure suction hood to cover the winding cut point. The semi-enclosed arc-shaped negative pressure suction hood is connected to an external high-vacuum rotary vane pump to control the negative pressure field intensity at the winding cut point to an absolute pressure of 100 Pascals.

[0061] The microporous air extraction channel circumferentially opened inside the winding shaft is specifically made of a circumferential microporous array processed on the surface of the winding shaft substrate. The micropore diameter in the circumferential microporous array is set to 50 micrometers. The circumferential microporous array is connected to the vacuum exhaust unit through the hollow air extraction channel in the mandrel inside the winding shaft.

[0062] During the winding process, a semi-enclosed arc-shaped negative pressure suction hood and a circumferential micropore array work together to simultaneously remove interlayer micro-air gaps along the axial direction of the winding shaft to both ends. The main control system ensures that the intensity of the negative pressure field and the radial lamination force meet the equilibrium condition.

[0063] In this embodiment, the radial lamination force is set to 40 Newtons. Based on this radial lamination force of 40 Newtons, fluid dynamics modeling calculations were performed, and the interlaminar gas extrusion resistance caused by lamination was calculated to be 0.015 Newtons per meter. To ensure that the interlaminar suction force generated by the negative pressure field is not less than the interlaminar gas extrusion resistance caused by lamination, the main control system fine-tunes the pumping speed of the high-vacuum rotary vane pump, controlling the interlaminar suction force corresponding to an absolute pressure of 100 Pascals to 0.018 Newtons per meter. This interlaminar suction force of 0.018 Newtons per meter is higher than the interlaminar gas extrusion resistance of 0.015 Newtons per meter, thus ensuring that the interlaminar suction force is not less than the interlaminar gas extrusion resistance.

[0064] In step three, the negative pressure environment is constructed by a local negative pressure field covering the winding cut point, and the interlayer micro-air gap removal process is performed. Another specific implementation of this embodiment is as follows: In this embodiment, the local negative pressure field is specifically designed by using an integrally enclosed sealed negative pressure chamber enclosed at the winding cut point. The integrally enclosed sealed negative pressure chamber is connected to an external Roots vacuum pump unit to control the negative pressure field intensity at the winding cut point to an absolute pressure of 50 Pascals.

[0065] The microporous air extraction channel circumferentially opened inside the winding shaft is specifically made of a porous metal sintered layer covering the outer layer of the winding shaft substrate. The average filtration accuracy of the porous metal sintered layer is set to a micropore diameter of 15 micrometers. The porous metal sintered layer is connected to the vacuum exhaust unit through the partitioned air collection pipe inside the winding shaft.

[0066] During the winding process, the overall closed-loop sealed negative pressure chamber and the porous metal sintered layer work together to simultaneously remove interlayer micro-air gaps along the axial direction of the winding shaft to both ends. The main control system ensures that the intensity of the negative pressure field and the radial lamination force meet the equilibrium condition.

[0067] In this embodiment, the radial lamination force is set to 80 Newtons. Based on this radial lamination force of 80 Newtons, fluid dynamics modeling calculations were performed, and the interlaminar gas extrusion resistance caused by lamination was calculated to be 0.035 Newtons per meter. To ensure that the interlaminar suction force generated by the negative pressure field is not less than the interlaminar gas extrusion resistance caused by lamination, the main control system fine-tunes the frequency converter of the Roots vacuum pump unit, controlling the interlaminar suction force corresponding to an absolute pressure of 50 Pascals to 0.040 Newtons per meter. This interlaminar suction force of 0.040 Newtons per meter is higher than the interlaminar gas extrusion resistance of 0.035 Newtons per meter, thus ensuring that the interlaminar suction force is not lower than the interlaminar gas extrusion resistance.

[0068] In this embodiment, the radial lamination external force adopts a nonlinear decreasing control process based on real-time inversion of the interlayer stress field, specifically including: Based on the measured dynamic electrical parameters and interlayer insulation resistance from step four, the main control system uses a capacitance-stress coupling model to invert the current interlayer stress distribution and determines the baseline value of the radial lamination force for the next round based on the radial integral value of the inverted stress. This ensures that the radial lamination force decreases non-linearly with the accumulation of residual stress in the wound portion. By using the capacitance-stress coupling model to invert the current interlayer stress distribution based on the measured dynamic electrical parameters and interlayer insulation resistance, and determining the baseline value of the radial lamination force for the next round based on the radial integral value of the inverted stress, the radial lamination force decreases non-linearly with the accumulation of residual stress. This achieves dynamic adjustment of the lamination pressure according to the actual stress state inside the core, avoiding excessive lamination of the outer layer that leads to excessive stress accumulation in the inner layer.

[0069] When the inverted interlayer stress distribution shows localized stress concentrations and the maximum stress exceeds the critical stress threshold, localized lamination force compensation is applied to the localized stress concentration areas, rather than globally decreasing the radial lamination force baseline value. By adjusting the lamination force only to the localized stress concentration areas without changing the global baseline value when localized stress concentrations are detected during inversion and the maximum stress exceeds the critical stress threshold, precise compensation for localized abnormal stresses is achieved while maintaining the overall nonlinear decreasing trend, avoiding overpressure or underpressure in other areas due to global parameter adjustments.

[0070] The specific implementation method of the nonlinear decreasing control process based on real-time inversion of the interlayer stress field in the radial lamination external force control process is as follows: Based on the measured dynamic electrical parameters and interlayer insulation resistance from step four, the main control system uses a capacitance-stress coupling model to invert the current interlayer stress distribution. In this embodiment, the dynamic electrical parameters specifically use dynamic complex capacitance values. The measured interlayer insulation resistance is 5000 megohms. Using the capacitance-stress coupling model, the main control system calculates the current interlayer stress distribution based on the dynamic complex capacitance value and the 5000 megohm interlayer insulation resistance.

[0071] The main control system determines the radial lamination force reference value for the next turn based on the radial integral value of the inverted stress, so that the radial lamination force decreases non-linearly with the accumulation of residual stress in the wound portion. In this embodiment, when the number of wound turns is 100, the radial integral value of the inverted stress is 150 kN / m, and the main control system determines the radial lamination force reference value for the 101st turn to be 100 N; when the number of wound turns increases to 200, due to the accumulation of residual stress in the wound portion, the radial integral value of the inverted stress rises to 320 kN / m, and the main control system determines the radial lamination force reference value for the 201st turn to decrease non-linearly to 65 N.

[0072] In this embodiment, the critical stress threshold is set to 25 MPa. When the process reaches the 250th cycle, the current interlaminar stress distribution inverted by the master control system shows local stress concentration in the inner edge region of the core, and the maximum stress in the inner edge region reaches 28 MPa. Since the maximum stress of 28 MPa exceeds the critical stress threshold of 25 MPa, the master control system performs local lamination force compensation adjustment in the inner edge region, rather than globally reducing the radial lamination force reference value.

[0073] Specifically, the lamination actuator in this embodiment adopts a segmented electromagnetic pressure roller. The main control system keeps the global radial lamination force reference value of 50 Newtons unchanged in the 251st rotation, and only controls the third fine-tuning electromagnetic coil in the segmented electromagnetic pressure roller corresponding to the inner edge area to reduce the control current, thereby individually compensating and reducing the local lamination force in the inner edge area from 50 Newtons to 35 Newtons, thus accurately eliminating the local stress concentration in the inner edge area.

[0074] Another specific implementation method of the nonlinear decreasing control process based on real-time inversion of interlayer stress field in radial lamination external force control is as follows: Based on the measured dynamic electrical parameters and interlayer insulation resistance from step four, the main control system uses a capacitance-stress coupling model to invert the current interlayer stress distribution. In this embodiment, the dynamic high-frequency dissipation factor is specifically selected as the dynamic electrical parameter. The measured interlayer insulation resistance is 8000 megohms. Using the capacitance-stress coupling model, the main control system calculates the current interlayer stress distribution based on the dynamic high-frequency dissipation factor and the 8000 megohm interlayer insulation resistance.

[0075] The main control system determines the radial lamination force reference value for the next turn based on the radial integral value of the inverted stress, so that the radial lamination force decreases non-linearly with the accumulation of residual stress in the wound portion. In this embodiment, when the number of wound turns is 50, the radial integral value of the inverted stress is 80 kN / m, and the main control system determines the radial lamination force reference value for the 51st turn to be 120 N; when the number of wound turns increases to 150, due to the accumulation of residual stress in the wound portion, the radial integral value of the inverted stress rises to 240 kN / m, and the main control system determines the radial lamination force reference value for the 151st turn to decrease non-linearly to 80 N.

[0076] In this embodiment, the critical stress threshold is set to 35 MPa. When the process reaches the 180th cycle, the current interlayer stress distribution inverted by the main control system shows local stress concentration in the axial central region of the electrode lead-out end, and the maximum stress in the axial central region reaches 39 MPa. Since the maximum stress of 39 MPa exceeds the critical stress threshold of 35 MPa, the main control system performs local lamination force compensation adjustment in the axial central region, rather than globally reducing the radial lamination force reference value.

[0077] Specifically, the lamination actuator in this embodiment adopts a multi-segment pneumatic lamination roller. The main control system keeps the global radial lamination force reference value of 70 Newtons unchanged in the 181st rotation, and only controls the fifth precision pneumatic proportional valve in the corresponding axial central region of the multi-segment pneumatic lamination roller to reduce the air pressure output, thereby separately compensating and reducing the local lamination force in the axial central region from 70 Newtons to 45 Newtons, thus accurately eliminating the local stress concentration in the axial central region.

[0078] In this embodiment, before the two metallized films overlap and intersect, the film capacitor winding process further includes a step of introducing a high-voltage electrostatic polarization field to perform microscopic pre-bonding and defect pre-breakdown self-healing of the metallized films. Specifically, it includes: By applying opposite charges to two metallized thin films using a high-voltage electrostatic polarization field, microscopic pre-bonding of the two metallized thin films is achieved due to Coulomb attraction before they enter the winding tangent point. This method enables the physical bonding of the two films to be established before winding, reducing the bonding displacement deviation at the winding tangent point.

[0079] The field strength of the high-voltage electrostatic polarization field is set between one-third and one-half of the dielectric breakdown field strength of the metallized thin film. The field strength is used to drive the defects in the vapor-deposited layer of the metallized thin film to undergo local pre-breakdown self-healing. By setting the field strength of the high-voltage electrostatic polarization field between one-third and one-half of the dielectric breakdown field strength, and using this field strength to drive the defects in the vapor-deposited layer to undergo local pre-breakdown self-healing, the random self-healing process after leaving the factory is transformed into a controllable pre-self-healing process before winding, eliminating the potential threat of defects to the long-term reliability of the capacitor.

[0080] In step two, the surface resistivity of the vapor-deposited edge-cutting region after online in-situ modification is higher than the critical surface resistivity value for charge leakage. By making the surface resistivity of the vapor-deposited edge-cutting region after online in-situ modification in step two higher than the critical surface resistivity value for charge leakage, the path for polarized charge to leak to the outside through the vapor-deposited edge-cutting region is blocked, ensuring that the polarized charge is concentrated in the effective area of ​​the thin film to generate pre-bonded Coulomb attraction.

[0081] The main control system adjusts the field strength of the high-voltage electrostatic polarization field and the film transport tension in step one, so that the pre-bonding positive pressure generated by Coulomb attraction and the film transport tension work together to construct a microscopic wrinkle-free stress transition zone before the winding cut point. By adjusting the field strength of the high-voltage electrostatic polarization field and the film transport tension in step one, the main control system enables the pre-bonding positive pressure generated by Coulomb attraction and the film transport tension to work together to construct a microscopic wrinkle-free stress transition zone, thus eliminating film wrinkles caused by the mismatch between the pre-bonding positive pressure and the transport tension, and ensuring that the two films are flatly bonded before the winding cut point.

[0082] Preferably, before the metallized thin films are overlapped and bonded, the process further includes an online casting coating of a solid polymer electrolyte medium and an in-situ crosslinking and curing under ultraviolet radiation to form a solid composite medium layer. The main control system regulates the field strength of the high-voltage electrostatic polarization field and the cross-linking and solidification process of the solid composite dielectric layer, so that when the high-voltage electrostatic polarization field and the rheological state of the solid composite dielectric layer enter the winding tangent point, they satisfy the following electro-rheological nonlinear dynamic coupling constraint formula: , in, The surface charge density, Rheological loss factor This is the equivalent stress.

[0083] Specifically, surface charge density The corresponding high-voltage electrostatic polarization field drives the two metallized films to generate microscopic pre-bonding Coulomb attraction before entering the winding cutting point. The main control system regulates the output field strength of the electrostatic generator and the conveying speed of the metallized film online. Rheological loss factor The ratio of the viscous modulus to the elastic modulus of the solid composite dielectric layer characterizes the rheological hardening state of the solid composite dielectric layer at the winding cut point. It is controlled by the main control system through online adjustment of the output power of the ultraviolet radiation field and the time interval from the casting coating to the winding cut point. Equivalent stress The mechanical compressive stress generated at the winding tangent point by the radial lamination external force is driven by the main control system to perform real-time adjustment of the lamination actuator based on the current interlayer stress distribution inverted from the capacitance-stress coupling model.

[0084] A solid composite dielectric layer is formed by online casting and coating of a solid polymer electrolyte medium before overlapping and bonding of metallized films, followed by in-situ crosslinking and curing under ultraviolet radiation. The main control system regulates the high-voltage electrostatic polarization field strength and the crosslinking and curing process to achieve the desired charge surface density. Rheological loss factor tanδ and equivalent stress Satisfying 0.1≤ ·tanδ / The coupling constraint of ≤0.5 enables the quantitative coupling constraint of the electrical effect of the polarization field with the rheological state of the dielectric layer, ensuring that the dielectric layer simultaneously meets the electrical pre-bonding requirements and the mechanical damping dissipation requirements at the winding tangent point.

[0085] The surface charge density is controlled by adjusting the output field strength of the electrostatic generator and the delivery speed of the metallized thin film online through the main control system. This allows the surface charge density to be incorporated as a controllable variable into closed-loop regulation, ensuring that the pre-bonding effect generated by Coulomb attraction is dynamically adapted to the process conditions.

[0086] By adjusting the output power of the ultraviolet radiation field and the time interval from the casting coating to the winding cut point online through the main control system, the rheological loss factor tanδ is controlled, thus incorporating the rheological hardening state of the dielectric layer as a controllable variable into the closed-loop regulation, ensuring that the dielectric layer reaches a suitable viscoelastic ratio at the winding cut point.

[0087] The main control system drives the lamination actuator to adjust the equivalent stress in real time based on the current interlayer stress distribution derived from the capacitance-stress coupling model. This enables dynamic correlation between equivalent stress and actual interlayer stress state, avoiding overpressure or underpressure caused by fixed interlayer pressure.

[0088] Specifically, the main control system limits the ratio of the electro-rheological nonlinear dynamic coupling constraint relationship to between 0.1 and 0.5, and executes online compensation control for the following two dynamic boundary scenarios: In the low-limit stress protection scenario, when the ratio is below 0.1, it characterizes the equivalent stress. If the radial lamination force is too large or the interlayer damping dissipation capacity is insufficient, the main control system reduces the rheological loss factor by decreasing the radial lamination force or increasing the output power of the ultraviolet radiation field. To prevent the distributed integrated inner electrode from being mechanically damaged; through when ·tanδ / When the ratio is less than 0.1, the main control system reduces the radial lamination force or increases the ultraviolet radiation power to reduce tanδ, thus enabling timely load reduction when the equivalent stress is too large or the damping dissipation is insufficient, preventing mechanical damage caused by excessive lamination of the distributed integrated inner tab.

[0089] In the high-limit air gap elimination scenario, when the ratio is higher than 0.5, it characterizes the equivalent stress. If the solid composite medium layer has insufficient or excessive viscous flow resistance, the main control system increases the radial lamination force to improve flow. Alternatively, the transport speed of the metallized film can be reduced to prolong the cross-linking and curing time, thereby eliminating interlayer micro-gaps and maintaining an interlayer gap thickness of less than 0.1 mm. Through when ·tanδ / When the ratio is higher than 0.5, the main control system increases the radial lamination force to improve performance. Alternatively, the conveying speed can be reduced to extend the curing time, enabling timely pressurization or extended curing when the equivalent stress is insufficient or the viscous flow resistance of the medium layer is too high, thus eliminating micro-air gaps between layers.

[0090] Before the two metallized films overlap and are bonded together, the thin-film capacitor winding process also includes a step of introducing a high-voltage electrostatic polarization field to perform microscopic pre-bonding and defect pre-breakdown self-healing of the metallized films. The specific implementation method of this embodiment is as follows: In this embodiment, the metallized thin film is specifically a polypropylene film with an aluminum metal layer deposited on it. The dielectric breakdown field strength of the polypropylene film is 400 volts per micrometer. The field strength of the high-voltage electrostatic polarization field is set to 160 volts per micrometer. The field strength of 160 volts per micrometer is between 133.3 volts per micrometer and 200 volts per micrometer, that is, the field strength of the high-voltage electrostatic polarization field is set between one-third and one-half of the dielectric breakdown field strength of the polypropylene film.

[0091] Two metallized thin films are applied with opposite charges using a high-voltage electrostatic polarization field, causing them to microscopically pre-bond due to Coulomb attraction before entering the winding cut point. A field strength of 160 volts per micrometer is used to drive local pre-breakdown self-healing of defects in the aluminum vapor-deposited layer of the polypropylene film. In step two, the surface resistivity of the vapor-deposited cut edge region after online in-situ modification is specifically 1013 ohms per square, and the critical surface resistivity value for charge leakage is set to 1011 ohms per square. The surface resistivity of the vapor-deposited cut edge region after online in-situ modification is higher than the critical surface resistivity value for charge leakage.

[0092] The main control system adjusts the field strength of the high-voltage electrostatic polarization field and the film transport tension in step one, so that the pre-bonding positive pressure generated by Coulomb attraction and the film transport tension work together to construct a microscopic wrinkle-free stress transition zone before the winding cutting point. In this embodiment, the film transport tension is set to 15 Newtons.

[0093] Before the metallization films are overlapped and laminated, the process also includes an online casting coating of a solid polymer electrolyte medium and an in-situ crosslinking and curing process under ultraviolet radiation to form a solid composite dielectric layer. Specifically, the solid polymer electrolyte medium is a mixture of polyethylene oxide and lithium bis(fluorosulfonyl)imide, and the ultraviolet radiation field is provided by an ultraviolet light-emitting diode curing light source with a center wavelength of 365 nm.

[0094] The main control system regulates the field strength of the high-voltage electrostatic polarization field and the cross-linking and solidification process of the solid composite dielectric layer, so that the rheological state of the high-voltage electrostatic polarization field and the solid composite dielectric layer satisfies the electro-rheological nonlinear dynamic coupling constraint formula when entering the winding tangent point.

[0095] In the normal processing stage of this embodiment, the main control system adjusts the charge surface density to 0.02 coulombs per square meter by online adjustment of the output field strength of the electrostatic generator and the conveying speed of the metallized film. The main control system also adjusts the ratio of the viscous modulus to the elastic modulus of the solid composite dielectric layer (i.e., the rheological loss factor) to 0.6 by online adjustment of the output power of the ultraviolet radiation field and the time interval from casting to the winding cut point. Based on the current interlayer stress distribution derived from the capacitance-stress coupling model, the main control system drives the lamination actuator to make real-time adjustments, controlling the mechanical compressive stress (equivalent stress) generated by the radial lamination force at the winding cut point to 0.08 MPa. At this point, the ratio of the electro-rheological nonlinear dynamic coupling constraint relationship is calculated by multiplying the charge surface density by the rheological loss factor and then dividing by the equivalent stress, resulting in a value of 0.15. 0.15 is between 0.1 and 0.5, meeting the preset process requirements.

[0096] When the process operates under low-limit stress protection conditions, mechanical fluctuations cause an abnormal increase in equivalent stress to 0.15 MPa, resulting in a decrease in the ratio of charge surface density multiplied by the rheological loss factor and then divided by the equivalent stress to 0.08. When the main control system detects that the ratio is below 0.1, indicating excessive equivalent stress, it executes online compensation control, reducing the radial lamination force from 80 N to 40 N by driving the lamination actuator, restoring the equivalent stress to 0.08 MPa, thereby bringing the ratio back to 0.15 and preventing mechanical damage to the distributed integrated inner tab.

[0097] Before the two metallized films overlap and are bonded together, the thin-film capacitor winding process also includes a step of introducing a high-voltage electrostatic polarization field to perform microscopic pre-bonding and defect pre-breakdown self-healing of the metallized films. Another specific implementation of this embodiment is as follows: In this embodiment, the metallized thin film is specifically selected as a polyethylene terephthalate (PET) film with a zinc-aluminum alloy metal layer deposited on it. The dielectric breakdown field strength of the PET film is 500 volts per micrometer. The field strength of the high-voltage electrostatic polarization field is set to 220 volts per micrometer. The field strength of 220 volts per micrometer is between 166.7 volts per micrometer and 250 volts per micrometer, that is, the field strength of the high-voltage electrostatic polarization field is set between one-third and one-half of the dielectric breakdown field strength of the PET film.

[0098] Two metallized thin films are applied with opposite charges using a high-voltage electrostatic polarization field, causing them to microscopically pre-bond due to Coulomb attraction before entering the winding cut point. A field strength of 220 volts per micrometer is used to drive local pre-breakdown self-healing of defects in the zinc-aluminum alloy vapor-deposited layer of the polyethylene terephthalate film. In step two, the surface resistivity of the vapor-deposited cut-edge region after online in-situ modification is specifically 10¹⁴ ohms per square, and the critical surface resistivity for charge leakage is set to 10¹¹ ohms per square. The surface resistivity of the vapor-deposited cut-edge region after online in-situ modification is higher than the critical surface resistivity for charge leakage.

[0099] The main control system adjusts the field strength of the high-voltage electrostatic polarization field and the film transport tension in step one, so that the pre-bonding positive pressure generated by Coulomb attraction and the film transport tension work together to construct a microscopic wrinkle-free stress transition zone before the winding cutting point. In this embodiment, the film transport tension is set to 30 Newtons.

[0100] Before the metallized films are overlapped and laminated, the process also includes an online casting coating of a solid polymer electrolyte medium and an in-situ crosslinking and curing process under ultraviolet radiation to form a solid composite dielectric layer. Specifically, the solid polymer electrolyte medium is a mixture of polymethyl methacrylate and lithium perchlorate, and the ultraviolet radiation field is provided by a high-pressure mercury lamp curing light source.

[0101] The main control system regulates the field strength of the high-voltage electrostatic polarization field and the cross-linking and solidification process of the solid composite dielectric layer, so that the rheological state of the high-voltage electrostatic polarization field and the solid composite dielectric layer satisfies the electro-rheological nonlinear dynamic coupling constraint formula when entering the winding tangent point.

[0102] In the normal processing stage of this embodiment, the main control system adjusts the charge surface density to 0.03 coulombs per square meter by online adjustment of the output field strength of the electrostatic generator and the conveying speed of the metallized film. The main control system also adjusts the ratio of the viscous modulus to the elastic modulus of the solid composite dielectric layer (i.e., the rheological loss factor) to 0.8 by online adjustment of the output power of the ultraviolet radiation field and the time interval from casting to the winding cut point. Based on the current interlayer stress distribution derived from the capacitance-stress coupling model, the main control system drives the lamination actuator to make real-time adjustments, controlling the mechanical compressive stress (equivalent stress) generated by the radial lamination force at the winding cut point to 0.06 MPa. At this point, the ratio of the electro-rheological nonlinear dynamic coupling constraint relationship is calculated by multiplying the charge surface density by the rheological loss factor and then dividing by the equivalent stress, resulting in a value of 0.4. This value is between 0.1 and 0.5, meeting the preset process requirements.

[0103] When the process reaches the high-limit air gap elimination scenario, the solid composite dielectric layer is not fully cured due to the attenuation of the ultraviolet radiation field, resulting in an abnormal increase in the rheological loss factor to 1.2. This causes the ratio of charge surface density multiplied by the rheological loss factor and then divided by the equivalent stress to rise to 0.6. When the main control system detects a ratio higher than 0.5, it indicates excessive viscous flow resistance of the solid composite dielectric layer. The main control system executes online compensation control, reducing the metallization film conveying speed from 50 meters per minute to 30 meters per minute to prolong the crosslinking and curing time. This allows the rheological loss factor to decrease and recover to 0.8, thereby bringing the ratio back down to 0.4, eliminating interlayer micro-air gaps and maintaining an interlayer air gap thickness of less than 0.1 micrometers.

[0104] In this embodiment, before the metallization film overlap and bonding, the thin-film capacitor winding process further includes a step of introducing an interlayer dielectric flow damping reconstruction process, specifically including: A solid polymer electrolyte dielectric layer is online cast onto the surface of a metallized thin film and then in-situ crosslinked and cured by ultraviolet radiation to form a solid composite dielectric layer between the capacitor core layers. By online casting of a solid polymer electrolyte dielectric layer onto the surface of a metallized thin film and in-situ crosslinking and curing by ultraviolet radiation to form a solid composite dielectric layer between the capacitor core layers, the traditional air gap can be replaced as the interlayer dielectric, eliminating the risk of partial discharge caused by interlayer air gaps.

[0105] The time interval from the start of casting to the completion of cross-linking and curing is no greater than the time required for the metallized film to travel from the casting station to the winding cut point. This ensures that the solid composite dielectric layer is in a non-viscous rheologically hardened state when it enters the winding cut point. The anisotropic viscoelastic deformation of the solid composite dielectric layer dissipates the radial lamination external force in step three and the mechanical stress caused by the inner tab protrusions constructed in step two.

[0106] By limiting the time interval from casting to crosslinking curing to no more than the time required for the film to travel from the casting station to the winding cut point, it is possible to ensure that the solid composite medium layer is cured before entering the winding cut point, thus avoiding uncontrollable rheological changes in the uncured medium under lamination external force, which would lead to uneven distribution of the interlayer medium.

[0107] By placing the solid composite dielectric layer in a non-viscous rheologically hardened state when it enters the winding tangent point, the anisotropic deformation of the solid composite dielectric layer dissipates the radial lamination external force and the mechanical stress brought by the inner tab protrusion. This transforms the stress concentration in traditional rigid lamination into anisotropic dissipation of the viscoelastic medium, reducing the local stress peak at the inner tab protrusion.

[0108] Before the metallization film is overlapped and bonded, the thin-film capacitor winding process also includes a step of introducing an interlayer dielectric flow damping reconstruction process. The specific implementation of this embodiment is as follows: A solid polymer electrolyte medium is online cast onto the surface of the metallized film. In this embodiment, the solid polymer electrolyte medium is specifically a mixture of polyethylene oxide and lithium bis(fluorosulfonyl)imide. Subsequently, in-situ crosslinking and curing are performed under ultraviolet radiation to form a solid composite dielectric layer between the capacitor core layers. In this embodiment, the ultraviolet radiation field is specifically provided by an ultraviolet light-emitting diode curing light source with a center wavelength of 365 nm.

[0109] In this embodiment, the running time required for the metallized film to travel from the casting station to the winding cut point is 0.8 seconds. The time interval from the start of casting coating to the completion of crosslinking and curing is set to 0.6 seconds. Since the 0.6-second time interval is no greater than the 0.8-second running time, the solid composite dielectric layer has already completed in-situ crosslinking and curing when it enters the winding cut point, thus being in a non-viscous rheologically hardened state. In this embodiment, the non-viscous rheologically hardened state is specifically a highly elastic state.

[0110] The main control system utilizes the viscoelastic deformation anisotropy of the highly elastic solid composite dielectric layer to dissipate the radial lamination force in step three and the mechanical stress caused by the inner tab protrusions constructed in step two. In this embodiment, the radial lamination force in step three is set to 45 Newtons, and the peak local mechanical stress caused by the inner tab protrusions is dissipated by the anisotropy of the highly elastic polyoxyethylene body, reducing it from 32 MPa before the introduction of the interlayer medium flow damping reconstruction process to 12 MPa.

[0111] Before the metallization film is overlapped and bonded, the thin-film capacitor winding process also includes a step of introducing an interlayer dielectric flow damping reconstruction process. Another specific implementation of this embodiment is as follows: A solid polymer electrolyte medium is online cast onto the surface of the metallized film. In this embodiment, the solid polymer electrolyte medium is specifically a mixture of polymethyl methacrylate and lithium perchlorate. Subsequently, in-situ crosslinking and curing are performed under ultraviolet radiation to form a solid composite dielectric layer between the capacitor core layers. In this embodiment, the ultraviolet radiation field is specifically provided by a high-pressure mercury lamp curing light source.

[0112] In this embodiment, the running time required for the metallized film to travel from the casting station to the winding cut point is 1.5 seconds. The time interval from the start of casting coating to the completion of crosslinking and curing is set to 1.1 seconds. Since the 1.1-second time interval is no greater than the 1.5-second running time, the solid composite dielectric layer has already completed in-situ crosslinking and curing when it enters the winding cut point, thus being in a non-viscous rheologically hardened state. In this embodiment, the non-viscous rheologically hardened state is specifically a glassy state.

[0113] The main control system utilizes the viscoelastic deformation anisotropy of the glassy solid composite dielectric layer to dissipate the radial lamination force in step three and the mechanical stress caused by the inner tab protrusions constructed in step two. In this embodiment, the radial lamination force in step three is set to 80 Newtons, and the peak local mechanical stress caused by the inner tab protrusions is dissipated by the anisotropy of the glassy polymethyl methacrylate matrix, reducing it from 45 MPa before the interlayer medium flow damping reconstruction process to 15 MPa.

[0114] The four-stage control loop real-time compensation process of prediction, pre-adjustment, verification, and correction in step five of this embodiment includes: In the prediction phase, the main control system predicts the expected values ​​of electrical parameters and air gap thickness for the next number of turns based on the current number of turns already wound. By establishing a prediction model based on the current number of turns already wound, the main control system predicts the expected values ​​of electrical parameters and air gap thickness for the next number of turns, thus realizing a control upgrade from passive lag response to active look-ahead prediction.

[0115] In the pre-adjustment stage, based on the expected values ​​of electrical parameters and air gap thickness, the film conveying tension, radial lamination force, and output power of the first physical field in step one are pre-adjusted for the next number of winding turns. By pre-adjusting the film conveying tension, radial lamination force, and output power of the first physical field for the next number of winding turns based on the expected values ​​of electrical parameters and air gap thickness obtained in the prediction stage, parameter pre-adjustment is achieved before deviations occur, upgrading post-correction to pre-prevention.

[0116] During the verification phase, after the next number of turns are wound, the electrical parameters and interlayer air gap thickness of the current turn are measured, and the deviation between the measured values ​​and the expected values ​​of the electrical parameters and air gap thickness is calculated. By measuring the electrical parameters and interlayer air gap thickness after the next number of turns are wound and calculating the deviation between the measured values ​​and the expected values, the accuracy of the prediction and the actual deviation are quantified, providing an accurate error feedback signal for the correction phase.

[0117] During the correction phase, when the deviation value exceeds the preset deviation threshold, the compensation parameter for the next number of winding turns is corrected according to the deviation direction; when the deviation value for at least 3 consecutive turns is lower than the preset deviation threshold, the sampling frequency of the in-situ online detection is reduced. By correcting the compensation parameter according to the deviation direction when the deviation value exceeds the preset deviation threshold, and reducing the sampling frequency when the deviation value for at least 3 consecutive turns is lower than the preset deviation threshold, timely correction is achieved when the deviation exceeds the limit, and the sampling frequency is reduced after entering steady state to reduce the computational load, thus balancing control accuracy and computational efficiency.

[0118] The convergence criteria for the four-stage control cycle are: the rate of change of capacitance is less than 0.5% for three consecutive cycles and the interlayer air gap thickness is less than 0.1. By setting the capacitance change rate deviation to be less than 0.5% for three consecutive cycles and the interlayer air gap thickness to be less than 0.1 μm as convergence criteria, quantifiable dual indicators are used as objective criteria for closed-loop convergence, avoiding premature or delayed exit from the compensation state due to subjective judgment.

[0119] In step five, the specific implementation method of the four-stage control cycle real-time compensation process of prediction, pre-adjustment, verification, and correction is as follows: During the prediction phase, the main control system predicts the expected electrical parameters and air gap thickness for the next winding turn (the 51st turn) based on the current number of winding turns (50 turns). In this embodiment, the expected electrical parameters are specifically the expected equivalent series resistance of the 51st turn, which is set to 1.2 milliohms, and the expected air gap thickness is set to 0.3 micrometers.

[0120] During the pre-adjustment phase, based on the expected equivalent series resistance of 1.2 milliohms and the expected air gap thickness of 0.3 micrometers, the film transport tension, radial lamination force, and output power of the first physical field in step one are pre-adjusted for the 51st cycle. In this embodiment, the first physical field specifically adopts a high-frequency infrared radiation field. The main control system pre-adjusts the film transport tension for the 51st cycle from 20 Newtons to 22 Newtons, the radial lamination force from 50 Newtons to 55 Newtons, and the output power of the high-frequency infrared radiation field from 500 watts to 520 watts.

[0121] During the verification phase, after the 51st turn is wound, the electrical parameters and interlayer air gap thickness of the 51st turn are measured. In this embodiment, the measured electrical parameters of the current turn are specifically the measured equivalent series resistance value of the 51st turn. The measured equivalent series resistance value is 1.5 milliohms, and the measured interlayer air gap thickness is 0.5 micrometers. The main control system calculates that the resistance deviation between the measured equivalent series resistance value and the expected equivalent series resistance value is 0.3 milliohms, and the thickness deviation between the measured interlayer air gap thickness and the expected air gap thickness value is 0.2 micrometers.

[0122] During the correction phase, the preset deviation thresholds in this embodiment include a preset resistance deviation threshold of 0.1 milliohms and a preset thickness deviation threshold of 0.05 micrometers. Since the resistance deviation value of 0.3 milliohms exceeds the preset resistance deviation threshold of 0.1 milliohms and the thickness deviation value of 0.2 micrometers exceeds the preset thickness deviation threshold of 0.05 micrometers, the main control system corrects the compensation parameters for the next winding turn, i.e., the 52nd turn, according to the deviation direction, further compensating and increasing the film conveying tension of the 52nd turn to 24 Newtons.

[0123] The winding process continues until the process reaches a stable state. When the resistance and thickness deviation values ​​of the 100th, 101st, and 102nd consecutive turns are all below the preset deviation threshold, the main control system reduces the sampling frequency of the in-situ online detection from 100 times per second to 20 times per second.

[0124] Throughout the control process, the convergence criteria for the four-stage control loop are: the capacitance change rate deviation for three consecutive cycles is less than 0.5% and the interlayer air gap thickness is less than 0.1 micrometers. When the capacitance change rate deviation for three consecutive cycles is detected to be 0.3% and the interlayer air gap thickness is 0.08 micrometers, the main control system determines that the four-stage control loop has reached the convergence criteria and exits the compensation state.

[0125] In step five, another specific implementation of the four-stage control cycle real-time compensation process of prediction, pre-adjustment, verification, and correction is as follows: During the prediction phase, the main control system predicts the expected values ​​of electrical parameters and air gap thickness for the next winding turn, i.e., the 201st turn, based on the current number of winding turns of 200. In this embodiment, the expected values ​​of electrical parameters are specifically the expected value of the dielectric loss factor for the 201st turn, which is set to 0.0002, and the expected value of the air gap thickness is set to 0.2 micrometers.

[0126] During the pre-adjustment phase, based on the expected dielectric loss factor of 0.0002 and the expected air gap thickness of 0.2 micrometers, the film conveying tension, radial lamination force, and output power of the first physical field in step one are pre-adjusted for the 201st cycle. In this embodiment, the first physical field specifically adopts an ultraviolet curing physical field. The main control system pre-adjusts the film conveying tension for the 201st cycle from 40 Newtons to 38 Newtons, the radial lamination force from 120 Newtons to 110 Newtons, and the output power of the ultraviolet curing physical field from 800 watts to 750 watts.

[0127] During the verification phase, after the 201st turn is wound, the electrical parameters and interlayer air gap thickness of the 201st turn are measured. In this embodiment, the measured electrical parameters of the current turn are specifically the measured dielectric loss factor of the 201st turn. The measured dielectric loss factor is 0.0004, and the measured interlayer air gap thickness is 0.35 micrometers. The main control system calculates that the loss deviation between the measured dielectric loss factor and the expected value is 0.0002, and the thickness deviation between the measured interlayer air gap thickness and the expected value is 0.15 micrometers.

[0128] During the correction phase, the preset deviation thresholds in this embodiment include a preset loss deviation threshold of 0.0001 and a preset thickness deviation threshold of 0.08 micrometers. Since the loss deviation value of 0.0002 exceeds the preset loss deviation threshold of 0.0001 and the thickness deviation value of 0.15 micrometers exceeds the preset thickness deviation threshold of 0.08 micrometers, the main control system corrects the compensation parameters for the next winding turn, i.e., the 202nd turn, according to the deviation direction, further compensating and increasing the radial lamination force of the 202nd turn to 125 Newtons.

[0129] The winding process continues until the process reaches a stable state. When the loss deviation and thickness deviation values ​​of the four consecutive turns (350, 351, 352, and 353) are all below the preset deviation threshold, the main control system reduces the sampling frequency of the in-situ online detection from 200 times per second to 50 times per second.

[0130] Throughout the control process, the convergence criteria for the four-stage control loop are: the capacitance change rate deviation for three consecutive cycles is less than 0.5% and the interlayer air gap thickness is less than 0.1 micrometers. When the capacitance change rate deviation for three consecutive cycles is detected to be 0.2% and the interlayer air gap thickness is 0.05 micrometers, the main control system determines that the four-stage control loop has reached the convergence criteria and exits the compensation state.

[0131] In the four-stage control loop real-time compensation process of this embodiment, a step of nonlinear dynamic stability monitoring of the dynamic capacitance change rate time series acquired in real time in step four is also included, specifically including: The main control system uses a sliding window algorithm to calculate the maximum Lyapunov exponent of the dynamic capacitance change rate time series. The length of the sliding window algorithm is set to at least 20 recent winding data turns. By using the sliding window algorithm to calculate the maximum Lyapunov exponent of the dynamic capacitance change rate time series, and setting the sliding window length to at least 20 recent winding data turns, the nonlinear dynamic stability of the winding process can be determined by the sign of the maximum Lyapunov exponent, upgrading the traditional threshold judgment to dynamic stability judgment.

[0132] When the maximum Lyapunov exponent is negative, the current closed-loop feedback gain is maintained. By maintaining the current closed-loop feedback gain when the maximum Lyapunov exponent is negative, it is possible to confirm that the winding process is in a stable convergence state without additional intervention, thus avoiding unnecessary gain adjustments that could introduce oscillations.

[0133] When the maximum Lyapunov exponent is positive, the closed-loop feedback gain coefficient is increased to 1.5 to 2 times the current feedback gain value. Simultaneously, based on the magnitude of the positive value of the maximum Lyapunov exponent, the output power of the first physical field in step one is linearly and proportionally increased. This utilizes the enhanced first physical field to reconstruct the viscoelasticity of the thin film interface in real time, suppressing chaotic slip. By increasing the closed-loop feedback gain coefficient to 1.5 to 2 times the current feedback gain value when the maximum Lyapunov exponent is positive, and simultaneously linearly and proportionally increasing the output power of the first physical field based on the magnitude of the positive value of the maximum Lyapunov exponent, the enhanced first physical field is used to reconstruct the viscoelasticity of the thin film interface and suppress chaotic slip. This achieves cascaded amplification to suppress chaotic slip through dual intervention of gain amplification and physical field enhancement when the winding process enters a divergent instability trend.

[0134] For example, during process operation, when the main control system determines that the maximum Lyapunov exponent is positive through nonlinear dynamic stability monitoring (indicating that the winding system is abruptly changing from a laminar stable state to a chaotic unstable state), the main control system will adaptively linearly quantize and adjust the output power of the first physical field (specifically the ultrasonic energy field) in step one.

[0135] Specifically, the adjusted output power of the first physical field is calculated using the following power-dynamic linear adaptive compensation control equation:

[0136] The definitions and industrial implementation constraints of each physical quantity and parameter in the formula are as follows: The real-time output power of the adjusted first physical field (ultrasonic energy field) is expressed in watts (W).

[0137] To adjust the reference output power for the first physical field operation, in this embodiment, the reference output power is determined based on the initial tension of the thin film substrate. Set within the range of 200 W to 500 W.

[0138] This is the positive (dimensionless) value of the maximum Lyapunov exponent, calculated in real time using the sliding window algorithm. Its typical positive value range usually falls between 0.01 and 0.15 when microslippage or wrinkling occurs during the chaotic evolution of thin film winding.

[0139] This is a preset power-dynamic nonlinear adaptive compensation gain coefficient, measured in watts (W). This coefficient characterizes the linear proportional mapping relationship between the power boost magnitude and the instability rate. In this embodiment, to ensure that the ultrasonic energy can instantaneously suppress the abnormal slippage of polymer chain segments, The value range is strictly limited to 1500 W to 3000 W.

[0140] Example of industrial compensation implementation scenario: If the ultrasonic field reference operating power in step one is currently... = 300 W. When the real-time state feature sequence in step four becomes abnormal, the main control system calculates through the sliding window of the most recent 20 cycles that the current system's maximum Lyapunov exponent has a sudden change, and is positive. = 0.08.

[0141] At this point, the main control system adaptively matches the power-dynamic gain coefficient (e.g., taking...). = 2500 W), automatically calculate the power amplitude to be compensated ΔP = = 2500 0.08 = 200W.

[0142] The main control system then issued a control command, instantly and precisely increasing the output power of the ultrasonic energy field from 300 W to [unclear]. = 500 W. By instantaneously modulating the viscoelasticity of the thin film interface in the far field region through a sudden surge in sound intensity, the microscopic relative slip rate is forcibly pushed back to within a set slip threshold, thereby blocking the expansion of chaotic instability.

[0143] After the winding and forming is completed in step five of this embodiment, the process also includes a dual-mode stress relief step for the core tube, specifically including: The central core tube of the winding shaft is constructed using shape memory polymer material, with a glass transition temperature of 45℃~55℃. By using shape memory polymer material with a glass transition temperature of 45℃~55℃ to construct the central core tube of the winding shaft, it is possible to maintain rigid support when the central core tube is wound at room temperature and generate controllable radial centripetal contraction when thermally triggered.

[0144] After winding, a heat flow of 50℃~60℃ is introduced into the core tube to trigger radial centripetal contraction. The radial centripetal contraction rate of the core tube is controlled at 1%~3%, which is the ratio of the difference between the initial outer diameter and the contracted outer diameter of the core tube to the initial outer diameter. By introducing a heat flow of 50℃~60℃ into the core tube after winding to trigger radial centripetal contraction and controlling the contraction rate at 1%~3%, the inner ring winding stress is released from the inside to the outside through the gentle contraction of the core tube, avoiding residual inner ring stress caused by forced lamination of the outer layer.

[0145] During the shrinkage process of the central core tube, the main control system executes dual-mode adaptive control, specifically including: In stress relief mode, the central core tube is controlled to shrink slowly, while the in-situ online detection in step four is used to continuously monitor the interlayer air gap status. By controlling the central core tube to shrink slowly in stress relief mode and using the in-situ online detection in step four to continuously monitor the interlayer air gap status, it is possible to monitor the air gap changes in real time while releasing the inner ring stress, ensuring that no new air gap defects are introduced during the stress relief process.

[0146] In the air gap preservation mode, when the interlayer air gap thickness is detected to be lower than the preset lower threshold, the central core tube is controlled to decelerate or pause contraction. By controlling the central core tube to decelerate or pause contraction when the interlayer air gap thickness is lower than the preset lower threshold in the air gap preservation mode, it is possible to prevent excessive contraction of the central core tube from causing excessive compaction of the interlayer and resulting in local electric field distortion.

[0147] Throughout the entire shrinkage stress relief process, the capacitance change rate deviation of the capacitor core was maintained within 1%. By maintaining the capacitance change rate deviation within 1% throughout the entire shrinkage stress relief process, the capacitance change rate deviation was used as a quality constraint indicator for the stress relief process, ensuring that the stress relief effect was achieved within an acceptable range of electrical performance.

[0148] After the winding and forming is completed in step five, the capacitor core process also includes a dual-mode stress relief step for the core tube. The specific implementation method of this embodiment is as follows: In this embodiment, a shape memory polymer material is used to construct the central core tube of the winding shaft. Specifically, shape memory polyurethane material is selected as the shape memory polymer material, and the glass transition temperature of the shape memory polyurethane material is 48°C.

[0149] After winding, a 52°C hot air stream is introduced into the central core tube to trigger radial centripetal contraction. In this embodiment, the hot air stream is specifically a gaseous circulating hot air. The initial outer diameter of the central core tube is 100 mm, and the outer diameter after contraction is 98.5 mm. The difference between the initial outer diameter and the contracted outer diameter accounts for 1.5% of the initial outer diameter, meaning the radial centripetal contraction rate of the central core tube is controlled at 1.5%.

[0150] During the shrinkage of the central core tube, the main control system executes dual-mode adaptive control. In stress-relief mode, the main control system controls the airflow speed of the circulating hot air to 3 meters per second, causing the central core tube to shrink slowly. Simultaneously, the in-situ online detection from step four continuously monitors the interlayer air gap status. In this embodiment, the in-situ online detection specifically employs an ultrasonic reflection method online detection system, with the preset lower limit threshold for the interlayer air gap thickness set at 5 micrometers.

[0151] In the air gap preservation mode, when the ultrasonic reflection online detection system detects an interlayer air gap thickness of 4.2 micrometers, which is below the preset lower threshold of 5 micrometers, the main control system controls the central core tube to slow down its contraction by reducing the input power of the circulating hot air. Throughout the entire contraction and stress relief process, the capacitance change rate deviation of the capacitor core is maintained within 0.6% through the dynamic adjustments of the ultrasonic reflection online detection system and the main control system.

[0152] After the winding and forming is completed in step five, the capacitor core process also includes a dual-mode stress relief step for the core tube. Another specific implementation of this embodiment is as follows: In this embodiment, a shape memory polymer material is used to construct the central core tube of the winding shaft. Specifically, the shape memory polymer material is a shape memory epoxy resin material with a glass transition temperature of 53°C.

[0153] After winding, a 58°C hot flow is introduced into the central core tube to trigger radial centripetal contraction. In this embodiment, the hot flow is specifically constant-temperature, low-pressure heat transfer oil. The initial outer diameter of the central core tube is 150 mm, and the outer diameter after contraction is 146.25 mm. The difference between the initial and contracted outer diameters accounts for 2.5% of the initial outer diameter, meaning the radial centripetal contraction rate of the central core tube is controlled at 2.5%.

[0154] During the shrinkage of the central core tube, the main control system executes dual-mode adaptive control. In stress-relief mode, the main control system controls the circulation flow rate of the heat transfer oil to 10 liters per minute, causing the central core tube to shrink slowly. Simultaneously, the in-situ online detection from step four continuously monitors the interlayer air gap status. In this embodiment, the in-situ online detection specifically employs an X-ray computed tomography online detection system, with the preset lower limit threshold for the interlayer air gap thickness set at 8 micrometers.

[0155] In the air gap preservation mode, when the X-ray computed tomography (CT) online detection system detects that the interlayer air gap thickness is 7.5 micrometers, which is below the preset lower threshold of 8 micrometers, the main control system temporarily cuts off the circulation of the heat transfer oil to control the central core tube to pause its contraction. Throughout the entire contraction and stress relief process, the capacitance change rate deviation of the capacitor core is maintained within 0.8% through dynamic adjustments by the X-ray CT online detection system and the main control system.

[0156] In this embodiment, during the thermal shrinkage stress relief process of the central core tube, at least three grating sensing points embedded equidistantly inside the central core tube along the axial direction are used to monitor the radial shrinkage strain at each axial position in real time, and to perform axial uniformity control accordingly. Specifically, this includes: When the strain difference measured at any two grating sensing points exceeds the allowable threshold for axial uniformity, the main control system locally increases the flow rate of heat flow or prolongs the residence time of heat flow in the axial region with smaller strain, and locally decreases the flow rate of heat flow or shortens the residence time of heat flow in the axial region with larger strain, until the strain difference measured at each grating sensing point does not exceed the allowable threshold for axial uniformity. The allowable threshold for axial uniformity is set at 10% to 15% of the maximum shrinkage rate; the wavelength resolution used for monitoring is no greater than 1 pm, and the corresponding strain resolution is no greater than 1 microstrain.

[0157] By using at least three grating sensing points embedded at equal intervals along the axial direction inside the central core tube to monitor the radial shrinkage strain at each axial position in real time, the shrinkage strain distribution at each position along the axial direction of the central core tube was obtained, and the axial shrinkage non-uniformity was discovered.

[0158] By increasing the hot air velocity or extending the residence time in areas with smaller strain when the strain difference measured at any two grating sensing points exceeds the allowable threshold for axial uniformity, and decreasing the hot air velocity or shortening the residence time in areas with larger strain until the strain difference converges, closed-loop control of axial shrinkage uniformity is achieved through zoned heat flow regulation, thus avoiding uneven axial deformation of the core caused by insufficient or excessive local shrinkage.

[0159] By setting the allowable threshold for axial uniformity to 10% to 15% of the maximum shrinkage rate, the allowable threshold for axial uniformity is set based on the maximum shrinkage rate, so that the uniformity requirement is adaptively matched with the actual shrinkage range.

[0160] By limiting the wavelength resolution of the grating sensing point to no more than 1 pm and the strain resolution to no more than 1 microstrain, it is possible to capture minute differences in axial contraction with high-resolution strain detection, ensuring that axial uniformity control has sufficient sensing accuracy.

[0161] The specific implementation method of this embodiment during the thermal shrinkage stress relief process of the central core tube is as follows: In this embodiment, the central core tube is made of polyvinylidene fluoride (PVDF), and its maximum shrinkage rate is set to 12%. Three grating sensing points are embedded axially at equal intervals inside the central core tube, specifically using fiber Bragg grating sensors with different center wavelengths. The wavelength resolution for monitoring is set to 1 picometer, and the corresponding strain resolution is set to 1 microstrain. The axial uniformity tolerance threshold is then set to 10% of the maximum shrinkage rate, i.e., 1.2%.

[0162] During the processing, the main control system monitors the radial shrinkage strain at each axial position in real time using fiber Bragg grating sensors. When the radial shrinkage strain measured at the first grating sensing point is 11.5% and the radial shrinkage strain measured at the second grating sensing point is 10.0%, the strain difference between the first and second grating sensing points is 1.5%. This 1.5% strain difference exceeds the allowable threshold of 1.2% for axial uniformity, and the axial region corresponding to the second grating sensing point is a region with relatively low strain, while the axial region corresponding to the first grating sensing point is a region with relatively high strain.

[0163] At this point, the main control system performs axial uniformity control. In this embodiment, high-pressure hot air is used for heat flow. The main control system controls the inverter of the high-pressure blower to increase the output frequency, locally increasing the flow velocity of the high-pressure hot air in the axial region where the strain of the second grating sensor point is small, from 5 meters per second to 8 meters per second. Simultaneously, the main control system controls the proportional regulating valve to decrease the valve opening, locally decreasing the flow velocity of the high-pressure hot air in the axial region where the strain of the first grating sensor point is large, from 5 meters per second to 3 meters per second. The main control system maintains closed-loop control until the strain difference measured by each fiber Bragg grating sensor decreases to 1.0%, that is, the strain difference measured by each fiber Bragg grating sensor does not exceed the axial uniformity allowable threshold of 1.2%, at which point the main control system controls the high-pressure blower to resume normal heating mode.

[0164] Another specific implementation method of this embodiment is as follows during the heat shrinkage stress relief process of the central core tube: In this embodiment, the central core tube adopts a composite tube structure with a stainless steel lining and an outer cross-linked polyethylene coating. The maximum shrinkage rate of the central core tube is set to 8%. Five grating sensing points, specifically fiber Bragg grating arrays, are embedded axially at equal intervals inside the central core tube. The wavelength resolution for monitoring is set to 0.5 picometers, corresponding to a strain resolution of 0.6 microstrain. The axial uniformity tolerance threshold is then set to 15% of the maximum shrinkage rate, i.e., 1.2%.

[0165] During the manufacturing process, the main control system monitors the radial shrinkage strain at each axial position in real time using a fiber Bragg grating array. When the radial shrinkage strain measured at the third grating sensing point in the middle is 7.8%, while the radial shrinkage strain measured at the fifth grating sensing point at the edge is 6.2%, the strain difference between the third and fifth grating sensing points is 1.6%. This strain difference of 1.6% exceeds the allowable threshold of 1.2% for axial uniformity, and the axial region corresponding to the fifth grating sensing point is a region with relatively low strain, while the axial region corresponding to the third grating sensing point is a region with relatively high strain.

[0166] At this point, the main control system performs axial uniformity control. In this embodiment, the heat flow is infrared radiation heat flow, provided by multiple infrared heating lamps arranged axially. The main control system extends the residence time of the infrared heating lamps in the axial region where the strain is relatively small (the fifth grating sensing point) is located, increasing the single infrared irradiation time in this region from 40 seconds to 60 seconds. Simultaneously, the main control system shortens the residence time of the infrared heating lamps in the axial region where the strain is relatively large (the third grating sensing point) is located, reducing the single infrared irradiation time from 40 seconds to 25 seconds. The main control system maintains closed-loop adjustment control until the strain difference measured by each fiber Bragg grating array converges to 0.8%, thereby achieving adaptive uniform matching of contraction strain at each axial position.

[0167] The above description is only a preferred embodiment of the present invention. For those skilled in the art, there will be changes in the specific implementation and application scope based on the ideas of the present invention. The content of this specification should not be construed as a limitation of the present invention.

Claims

1. A thin-film capacitor winding process, characterized in that, The thin-film capacitor winding process, under the control of the main control system, involves multi-stage spatiotemporal dynamic collaborative intervention and in-situ closed-loop regulation along the path from the metallized thin film to the winding formation, including the following steps: Step 1: Upstream region interface conveying tension regulation. A first physical field is applied to at least two metallized films during conveying to regulate the viscoelasticity of the metallized film surface, and the microscopic relative slip rate of the metallized film during the conveying process is controlled within a set slip threshold, and a metallized film with predetermined surface energy morphology characteristics is output. Step 2: In-situ modification of the front edge of the cutting point. Received from Step 1, the main control system performs online in-situ modification of the vapor-deposited cutting edge area of ​​the metallized film according to the current winding parameters to improve the local surface resistivity of the vapor-deposited cutting edge area. Step 3: Negative pressure lamination and co-winding at the tangent point. Radial lamination force is applied to the capacitor core formed by overlapping winding in a negative pressure environment, so that the metallized film modified in Step 2 is overlapped and wound into shape; wherein, the interlayer suction force generated by the negative pressure environment and the radial lamination force satisfy the nonlinear decoupling equilibrium condition of interlayer gas extrusion resistance. Step 4: In-situ online detection of multiple physical quantities. During the winding process, the dynamic electrical parameters and interlayer air gap state of the currently wound part of the capacitor core are detected in-situ online to form a real-time state feature sequence for characterizing the dynamic evolution of multiple physical fields inside the core. Step 5: Cross-process nonlinear dynamic closed-loop compensation. Based on the real-time state characteristic sequence collected in Step 4, calculate the maximum Lyapunov exponent of the dynamic capacitance change rate time series, and adjust the first physical field in Step 1, the current winding parameters in Step 2, and the radial lamination external force in Step 3 in a four-stage control cycle of prediction, pre-adjustment, verification, and correction.

2. The thin-film capacitor winding process according to claim 1, characterized in that, In step two, the specific steps of the online in-situ modification treatment include: A pulsed laser beam, which is linked in a closed loop with the winding shaft, is used to dynamically trigger the laser pulse based on the real-time collected winding radius and the preset inner tab spacing. This laser pulse is used to locally melt the periodic sites on the surface of the metallized thin film, thereby locally thickening the metal layer and constructing a distributed integrated inner tab in situ inside the capacitor core. The distribution density of the distributed integrated inner tabs decreases as the winding radius increases, resulting in a higher tab distribution density in the inner ring of the capacitor core compared to the outer ring. The pulse width of the laser pulse satisfies the following thermal diffusion constraints: The thermal diffusion depth during the laser pulse does not exceed the thickness of the metal layer, so as to confine the heat-affected zone within the metal layer.

3. The thin-film capacitor winding process according to claim 1, characterized in that, In step three, the negative pressure environment is constructed by a local negative pressure field covering the winding cut point, and the interlayer micro air gaps are simultaneously removed from both ends along the axial direction of the winding shaft through the micro-porous air extraction channel opened circumferentially inside the winding shaft. The strength of the negative pressure field and the radial lamination force satisfy the following equilibrium condition: the interlayer suction force generated by the negative pressure field is not less than the interlayer gas extrusion resistance caused by lamination.

4. The thin-film capacitor winding process according to claim 3, characterized in that, The radial lamination external force employs a nonlinear decreasing control process based on real-time inversion of the interlayer stress field, specifically including: The main control system uses the capacitance-stress coupling model to invert the current interlayer stress distribution based on the dynamic electrical parameters and interlayer insulation resistance measured in step four, and determines the radial lamination force reference value for the next turn based on the radial integral value of the inverted stress, so that the radial lamination force decreases nonlinearly with the accumulation of residual stress in the wound part. When the inverted current interlayer stress distribution shows local stress concentration and the maximum stress exceeds the critical stress threshold, the local stress concentration area is adjusted by local lamination external force compensation, rather than globally reducing the radial lamination external force reference value.

5. The thin-film capacitor winding process according to claim 3, characterized in that, Before the two metallized films overlap and are bonded together, the thin-film capacitor winding process further includes a step of introducing a high-voltage electrostatic polarization field to perform microscopic pre-bonding and defect pre-breakdown self-healing on the metallized films, specifically including: The high-voltage electrostatic polarization field is used to apply opposite charges to the two metallized films, causing the two metallized films to undergo microscopic pre-bonding due to Coulomb attraction before entering the winding cut point. The field strength of the high-voltage electrostatic polarization field is set between one-third and one-half of the dielectric breakdown field strength of the metallized thin film. The field strength is used to drive the defects in the vapor deposition layer of the metallized thin film to undergo local pre-breakdown self-healing. The surface resistivity of the vapor-deposited trimmed area after the online in-situ modification treatment in step two is higher than the critical surface resistivity value for charge leakage. The main control system adjusts the field strength of the high-voltage electrostatic polarization field and the film conveying tension in step one, so that the pre-bonding positive pressure generated by the Coulomb attraction and the film conveying tension work together to construct a microscopic wrinkle-free stress transition zone before the winding cut point.

6. The thin-film capacitor winding process according to claim 3, characterized in that, Prior to the metallization film overlap and bonding, the thin-film capacitor winding process further includes a step of introducing an interlayer dielectric flow damping reconstruction process, specifically including: A solid polymer electrolyte dielectric is coated on the surface of the metallized film by online casting, and then cross-linked and cured in situ by ultraviolet radiation field to form a solid composite dielectric layer between the capacitor core layers. The time interval from the start of casting to the completion of crosslinking and curing is no greater than the time required for the metallized film to travel from the casting station to the winding cut point, so that the solid composite dielectric layer is in a non-viscous rheologically hardened state when it enters the winding cut point. The anisotropic viscoelastic deformation of the solid composite dielectric layer dissipates the radial lamination external force in step three and the mechanical stress caused by the inner tab protrusions constructed in step two.

7. The thin-film capacitor winding process according to claim 1, characterized in that, The four-stage control loop real-time compensation process of prediction, pre-adjustment, verification, and correction in step five includes: During the prediction phase, the main control system predicts the expected values ​​of electrical parameters and air gap thickness for the next number of turns based on the current number of turns already wound. In the pre-adjustment stage, based on the expected values ​​of the electrical parameters and the expected value of the air gap thickness, the film conveying tension for the next number of winding turns, the radial lamination force, and the output power of the first physical field in step one are pre-adjusted. During the verification phase, after the next number of turns is completed, the electrical parameters and interlayer air gap thickness of the current turn are measured, and the deviation between the measured values ​​and the expected values ​​of the electrical parameters and the expected values ​​of the air gap thickness are calculated. During the correction phase, when the deviation value exceeds the preset deviation threshold, the compensation parameter for the next number of winding turns is corrected according to the deviation direction; when the deviation value is lower than the preset deviation threshold for at least 3 consecutive turns, the sampling frequency of the in-situ online detection is reduced. The convergence criteria for the four-stage control cycle are: the capacitance change rate deviation for three consecutive cycles is less than 0.5% and the interlayer air gap thickness is less than 0.1 μm.

8. The thin-film capacitor winding process according to claim 7, characterized in that, The four-stage control loop real-time compensation process also includes a step of nonlinear dynamic stability monitoring of the dynamic capacitance change rate time series acquired in step four, specifically including: The main control system uses a sliding window algorithm to calculate the maximum Lyapunov exponent of the dynamic capacitance change rate time series, and the length of the sliding window algorithm is set to the winding data of the most recent 20 turns. When the maximum Lyapunov exponent is negative, maintain the current closed-loop feedback gain. When the maximum Lyapunov exponent is positive, the closed-loop feedback gain coefficient is increased to 1.5 to 2 times the current feedback gain value. At the same time, according to the magnitude of the positive value of the maximum Lyapunov exponent, the output power of the first physical field in step one is linearly and proportionally increased, so as to utilize the enhanced first physical field to reconstruct the viscoelasticity of the thin film interface in real time and suppress chaotic slip.

9. The thin-film capacitor winding process according to claim 1, characterized in that, After winding and forming is completed in step five, the process further includes a dual-mode stress relief step for the core tube, specifically including: The central core tube of the winding shaft is constructed using a shape memory polymer material, and the glass transition temperature of the central core tube is 45℃~55℃; After winding, a hot flow of 50°C to 60°C is introduced into the center core tube to trigger radial centripetal contraction of the center core tube. The radial centripetal contraction rate of the center core tube is controlled at 1% to 3%, and the radial centripetal contraction rate of the center core tube is the ratio of the difference between the initial outer diameter and the outer diameter after contraction to the initial outer diameter. During the shrinkage process of the central core tube, the main control system performs dual-modal adaptive control, specifically including: In stress relief mode, the central core tube is controlled to shrink slowly, while the in-situ online detection described in step four is used to continuously monitor the interlayer air gap status. In the air gap preservation mode, when the interlayer air gap thickness is detected to be lower than the preset lower threshold, the central core tube is controlled to decelerate or pause contraction. Throughout the entire shrinkage stress relief process, the rate of change of capacitance of the capacitor core is maintained within 1%.

10. The thin-film capacitor winding process according to claim 9, characterized in that, During the thermal shrinkage stress relief process of the central core tube, at least three grating sensing points embedded equidistantly within the central core tube along the axial direction are used to monitor the radial shrinkage strain at each axial position in real time, and to perform axial uniformity control accordingly. Specifically, this includes: When the strain difference measured at any two of the grating sensing points exceeds the allowable threshold for axial uniformity, the main control system locally increases the flow rate of the heat flow or prolongs the residence time of the heat flow in the axial region with smaller strain, and locally decreases the flow rate of the heat flow or shortens the residence time of the heat flow in the axial region with larger strain, until the strain difference measured at each of the grating sensing points does not exceed the allowable threshold for axial uniformity. The allowable threshold for axial uniformity is set to 10% to 15% of the maximum shrinkage rate; the wavelength resolution used for monitoring is no greater than 1 pm, and the corresponding strain resolution is no greater than 1 microstrain.

Images