Real-time monitoring, regulating and controlling method for manufacturing process of automobile lightweight component

Abstract

The invention relates to the crossing field of mechanical engineering and intelligent manufacturing, and discloses a real-time monitoring, regulating and controlling method for the manufacturing process of an automobile lightweight component. According to the method, a distributed fiber bragg grating sensing array is embedded in a mold, and a three-dimensional strain field at the flowing front edge of resin is obtained in real time; a multi-scale pore generation kinetic model is combined to perform online inversion on an unwetted area and pore distribution; and the pressure gradient and the mold temperature field of the hot press are adaptively regulated and controlled, and the flow path is dynamically corrected. The system comprises a sensing embedding module, a strain demodulation module, a pore inversion module, a regulation and control instruction generation module and a multi-zone execution module. According to the method, closed-loop precise control over the porosity being lower than 0.5% is achieved, and the consistency and reliability of the mechanical property of the component are remarkably improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of mechanical engineering and intelligent manufacturing, and particularly relates to a real-time monitoring and regulation method for a manufacturing process of a light-weight automobile part. BACKGROUND

[0002] With the increasing requirements of energy saving and emission reduction and endurance performance of the automobile industry, light-weight has become a core strategic direction of vehicle design. Carbon fiber reinforced plastics (CFRP) are widely used in key parts such as vehicle body structural parts and chassis components due to their high specific strength, high specific modulus and excellent corrosion resistance. In the manufacturing process of CFRP parts, resin transfer molding (RTM) or vacuum assisted resin infusion (VARTM) and other laminating molding processes are the mainstream technical routes, the core of which is to realize the full impregnation of the fiber preform by controlling the resin flow, so as to obtain a composite material product with low porosity and high performance. However, this process highly depends on the accurate perception and regulation of micro-scale fluid behavior.

[0003] The dynamic flow front of resin in the complex fiber network directly determines the internal quality of the final product. Due to the difference in capillary effect between carbon fiber bundles, the impregnation rate of resin in different areas is easily unbalanced, resulting in defects such as local dry spots, resin-rich areas or closed bubbles, which further leads to excessive porosity, seriously weakening the mechanical properties and fatigue life. In order to ensure the molding quality, the existing process usually relies on a pre-set pressure-time curve combined with discrete pressure sensors for process monitoring. However, the spatial resolution of traditional pressure sensors is limited, and it is difficult to embed them inside the fiber preform, which cannot capture the transient strain and flow field changes of the resin flow front at the millimeter or even micrometer scale, resulting in regulation lag or even failure.

[0004] Existing research attempts to introduce indirect monitoring methods such as infrared thermal imaging or dielectric sensing, but the former is severely affected by the mold obstruction, and the latter has high sensitivity to resin systems and poor universality. More importantly, the current monitoring system generally lacks closed-loop linkage capability with the actuator, and even if an abnormality is identified, it cannot adjust key parameters such as vacuum degree, injection pressure or mold temperature in real time. SUMMARY

[0005] In order to solve the above technical problems, the present application is proposed. The embodiments of the present application provide a real-time monitoring and regulation method for manufacturing process of automobile lightweight components, which realizes real-time acquisition of dynamic strain field distribution of resin flow front in three-dimensional space by embedding a distributed fiber Bragg grating sensing array inside a carbon fiber reinforced plastic laminated molding die; online inversion calculation of resin uneven infiltration is performed by combining a pre-set multi-scale pore generation dynamics model; based on the inversion result, the pressure gradient distribution and temperature field coordination parameters of the hot press are adjusted adaptively to dynamically correct the resin flow path and inhibit local dry spots and pore aggregation caused by fiber bundle capillary effect, so as to realize closed-loop precise control of the porosity of the molded component.

[0006] The real-time monitoring and regulation method for manufacturing process of automobile lightweight components provided by the present application comprises: After the carbon fiber preform is laid, the distributed fiber Bragg grating sensing array is embedded between the inner cavity surface of the die and the preform in the form of an orthogonal grid; The vacuum assisted resin transfer molding process is started, and during the resin injection stage, the center wavelength shift of each grating node is collected in real time by a wide spectrum light source and a high-speed spectrum demodulator, and the two-dimensional strain gradient field of the resin flow front in the die plane coordinate system at each time step is calculated by decoupling according to the linear coupling relationship between the wavelength shift and the strain and the temperature; The two-dimensional strain gradient field is input into the pre-constructed multi-scale pore generation dynamics model, the model is based on Darcy's law and capillary pressure balance equation, introduces the local permeability tensor and interfacial tension coefficient at the fiber bundle scale, and solves the space-time evolution equation of resin saturation by finite volume method, and outputs the spatial position, area ratio and predicted pore volume fraction of the area where the resin is not completely infiltrated at the current time; According to the deviation of the predicted pore volume fraction and the pre-set threshold value, pressure gradient regulation instructions and temperature field coordination correction instructions are generated; the pressure gradient regulation instructions act on the multi-zone independent hydraulic execution unit of the hot press to adjust the amplitude and rising rate of the pressure applied by each region; the temperature field coordination correction instructions act on the partitioned electric heating sheet built in the die to adjust the set temperature and heating slope of each heating region; The above monitoring, inversion and regulation cycle is continuously performed during the resin flow and curing stage until the resin completely fills the cavity and the pore volume fraction is stable below 0.5%.

[0007] Preferably, the grating period of the sensing array is 532 nanometers, the spatial resolution is 5 millimeters, the strain measurement accuracy is one micro-strain, and the temperature cross-sensitivity is compensated to be lower than 0.5 degrees Celsius per micro-strain.

[0008] Preferably, after the carbon fiber preform is laid, the distributed fiber Bragg grating sensing array is embedded between the surface of the mold cavity and the preform in the form of an orthogonal grid, including: The sensing array composed of 128 grating units inscribed on a single optical fiber is arranged with 64 measuring points in the length direction and the width direction of the mold, and the spacing between adjacent gratings is 10 mm; The sensing array is packaged with a polyimide coating, so that its outer diameter is 125 microns, and the temperature resistance range is -200 degrees Celsius to 300 degrees Celsius; The sensing array is arranged on the interface layer between the surface of the mold cavity and the carbon fiber preform, ensuring that the resin flow front causes local strain changes in the grating area without affecting the resin infiltration path.

[0009] Preferably, the vacuum assisted resin transfer molding process is started, and during the resin injection stage, the center wavelength shift of each grating node is collected in real time by a broadband light source and a high-speed spectrum demodulator. According to the linear coupling relationship between wavelength shift and strain and temperature, the two-dimensional strain gradient field of the resin flow front in the mold plane coordinate system at each time step is decoupled and calculated, including: The broadband light source emits continuous light with a wavelength range of 1520 nm to 1570 nm, which is coupled into the sensing fiber through a circulator. The high-speed spectrum demodulator collects the reflected spectrum sequence at a rate of 2000 frames per second; The center wavelength of each grating is located to the sub-picometer level, and the center wavelength shift is obtained ; Based on the fiber Bragg grating sensing equation: ; is the initial center wavelength, is the photoelastic coefficient, is the thermal expansion coefficient, is the thermo-optic coefficient; is the strain, temperature change; Combined with the temperature reference provided by the reference grating or the environmental temperature sensor, the pure strain component is decoupled and separated ; represents the two-dimensional spatial coordinates on the mold plane, represents time; The strain values of all measuring points are spatially interpolated to construct a two-dimensional strain field covering the entire mold plane, and the strain gradient tensor is calculated to identify the position of the resin flow front.

[0010] Preferably, the two-dimensional strain gradient field is input into a pre-constructed multi-scale pore generation dynamics model, which is based on Darcy's law and capillary pressure balance equation, introduces a local permeability tensor and interfacial tension coefficient at the scale of fiber bundles, discretely solves the space-time evolution equation of resin saturation by finite volume method, and outputs the spatial position, area ratio and predicted pore volume fraction of the current time of the resin incompletely saturated area, including: The mass conservation equation is taken as the control equation, where is the local porosity, is the resin density, is the Darcy velocity; The Darcy velocity is expressed as , where is the local permeability tensor, is the resin viscosity, is the capillary pressure; The capillary pressure is determined by the Young-Laplace equation , where is the resin-air interfacial tension, is the contact angle, is the equivalent radius of the fiber bundle; The two-dimensional strain gradient field is converted into a moving boundary condition, and the resin flow front advancing speed is derived from the strain change rate. The control equation is discretely solved by finite volume method to obtain the resin saturation and the pore volume fraction .

[0011] Preferably, the local permeability tensor K is obtained by offline micro-CT scanning to reconstruct the three-dimensional microstructure of the carbon fiber preform, and the lattice Boltzmann method is used to simulate the resin flow in the structure to calibrate the anisotropic permeability components .

[0012] Preferably, according to the deviation of the predicted pore volume fraction from the preset threshold value, pressure gradient control instructions and temperature field cooperative correction instructions are generated, including: Set the preset pore volume fraction threshold value to 0.5%, and the allowable deviation to 0.1%; When the maximum local pore volume fraction , an inverse optimization problem is constructed, and the objective function is , and the constraint conditions include that the maximum output pressure of the hot press is 30 megapascals, the hydraulic cylinder response delay is not more than 50 milliseconds, and the maximum local pressure gradient limited by the mold structure strength is 5 megapascals per centimeter; Solving the inverse optimization problem, obtaining the target pressure of 16 independent hydraulic execution units with the target temperature of 16 groups of partitioned electric heating sheets , is 1 to 16, is 1 to 16; The optimization result is mapped to the spatial continuous pressure gradient control instruction , for accelerating the resin flow to the low saturation area.

[0013] Preferably, the generation of the temperature field coordination correction instruction includes: Based on the resin viscosity-temperature index relationship , wherein is the initial lipid viscosity, is the activation energy of the resin, is the gas constant, is the absolute temperature, is the resin viscosity when the absolute temperature is ; According to the required viscosity adjustment amount, the local temperature correction amount is calculated To reduce the local temperature to slow down the fast flow caused by the increase in viscosity, or to increase the temperature to reduce the viscosity to promote the resin to fill the dry spot area; The is mapped to the temperature control instruction of 16 groups of partitioned electric heating sheets.

[0014] Preferably, the above monitoring, inversion and control cycle is continuously performed during the resin flow and curing stage until the resin completely fills the cavity and the porosity volume fraction is stable below 0.5%, including: After the resin injection is completed, enter the curing stage, continue to monitor the residual strain caused by the curing shrinkage of each point through the distributed fiber Bragg grating sensing array; Calculate the strain relaxation rate of the whole area, defined as the absolute value of the strain change per unit time; When the strain change rate is lower than 5 micro-strain per minute, it is determined that the curing is completed, and the control cycle is terminated.

[0015] Preferably, the monitoring, inversion and control cycle runs at a frequency of 2 Hz per second, and the control instruction update period is 500 milliseconds.

[0016] Compared with the prior art, the beneficial effects of the present application are: 1、The present application realizes full-field, real-time and non-interfering perception of the micro-fluid dynamics of the resin flow front by integrating a high spatial resolution distributed fiber Bragg grating sensing array inside the mold, overcoming the technical limitations of traditional point-type pressure sensors that cannot capture spatially uneven wetting phenomena; 2、The constructed multi-scale pore generation dynamics model combines macroscopic Darcy flow and microscopic capillary effect, can accurately invert the resin un-infiltrated area and pore generation trend; on this basis, through self-adaptive generation and execution of the coordinated control instructions of pressure gradient and temperature field, the resin flow path is actively corrected, the formation of dry spots and pores is inhibited from the source, the porosity of the final molded part is stably controlled below 0.5%, the mechanical property consistency and structural reliability of the carbon fiber reinforced plastic part are improved, and the strict quality requirements of automobile lightweight on high-performance composite components are met. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 It is the overall technical scheme architecture schematic diagram of the automobile lightweight component manufacturing process real-time monitoring and regulation method provided by the present application; Figure 2 It is the core principle framework schematic diagram of the online inversion of the multi-scale pore generation dynamics in the present application; Figure 3 It is the logic flow framework diagram of the real-time demodulation of the resin flow front dynamic strain field in the present application; Figure 4 It is the logic flow framework diagram of the self-adaptive pressure-temperature coordinated control instruction generation in the present application; Figure 5 It is the multi-level interaction relationship and data flow schematic diagram of the distributed fiber Bragg grating sensing embedding and multi-zone hydraulic-electric heating execution regulation in the present application; Figure 6 It is the closed-loop monitoring and regulation stage division and information feedback framework diagram of the whole process of resin injection to curing in the present application. DETAILED DESCRIPTION

[0018] Please refer to Figures 1 to 6 In the carbon fiber reinforced plastic laminated molding process, the resin flow front is unevenly infiltrated due to the fiber bundle capillary effect, resulting in excessive porosity. The pressure sensor used in the prior art can only obtain the pressure value of a local point, cannot capture the microfluid dynamic characteristics of resin flow in three-dimensional space, and is difficult to effectively warn and intervene the dry spot and pore aggregation. To solve this technical problem, the present embodiment provides an automobile lightweight component manufacturing process real-time monitoring and regulation method, which embeds a distributed fiber Bragg grating sensing array in the mold, senses the strain field change caused by resin flow in real time; combines a preset multi-scale pore generation dynamics model to invert the resin infiltration state online; and based on the inversion result, self-adaptively regulates the pressure gradient distribution of the hot press and the mold temperature field, realizes closed-loop accurate control of the porosity.

[0019] The method comprises the following steps: S1, after the carbon fiber preform laying is completed, the distributed fiber Bragg grating sensing array is embedded between the surface of the mold inner cavity and the preform in the form of orthogonal grid, the grating period of the sensing array is 532 nanometers, the spatial resolution is 5 millimeters, the strain measurement accuracy is one micro-strain, and the temperature cross-sensitivity is less than 0.5 degrees Celsius per micro-strain after packaging compensation.

[0020] Specifically, the distributed fiber Bragg grating sensing array is composed of 128 grating units engraved on a single optical fiber, the interval between adjacent gratings is 10 millimeters, 64 measuring points are arranged along the length direction and the width direction of the mold respectively, and a two-dimensional sensing network covering the entire molding area is formed. The sensing array is packaged with a polyimide coating, has an outer diameter of 125 microns, and can withstand shear stress without breaking or signal attenuation during resin injection. The layout position of the sensing array is the interface layer between the surface of the mold inner cavity and the carbon fiber preform, which ensures that the resin flow front can cause local strain change in the grating area when passing through, and does not affect the structural integrity of the preform and the resin infiltration path.

[0021] The grating period is set to 532 nanometers, so that the center reflection wavelength is near 1550 nanometers, which is within the effective output range of the wide spectrum light source and matches the detection window of the high-speed spectrum demodulator. The spatial resolution of 5 millimeters is determined by the interval between adjacent effective measuring points, which can be further improved to 2.5 millimeters through interpolation algorithm, meeting the identification needs of micro-scale non-uniformity of resin flow front. The strain measurement accuracy of one micro-strain is realized through high signal-to-noise ratio spectrum demodulation and temperature cross-sensitivity compensation algorithm, wherein the temperature cross-sensitivity is suppressed through a double-grating differential structure or the matching of the thermal expansion coefficient of the packaging material, so that the strain decoupling error is less than 0.5 degrees Celsius per micro-strain.

[0022] S2, start the vacuum assisted resin transfer molding process, during the resin injection stage, the center wavelength shift of each grating node is collected in real time by a wide spectrum light source and a high-speed spectrum demodulator, and the two-dimensional strain gradient field of the resin flow front in the mold plane coordinate system at each time step is calculated by decoupling according to the linear coupling relationship between the wavelength shift and the strain and the temperature.

[0023] Specifically, the wide spectrum light source emits continuous light with a wavelength range of 1520 nanometers to 1570 nanometers, which is coupled into the sensing optical fiber through a circulator. Each grating unit reflects a specific wavelength of light signal, and the reflected light returns to the high-speed spectrum demodulator through the same circulator. The high-speed spectrum demodulator collects the reflected spectrum sequence at a sampling rate of 2000 frames per second, and each frame contains the reflection peak information of 128 gratings. Through peak tracking algorithm, the center wavelength of each grating is located at sub-picometer level, and the center wavelength shift is obtained According to the basic sensing equation of fiber Bragg grating, the center wavelength shift and the strain and temperature change satisfy the linear relationship: ; where is the initial center wavelength, is the photoelastic coefficient, is the thermal expansion coefficient, is the thermo-optic coefficient. By pre-calibrating the temperature-strain cross-sensitivity matrix, combined with the temperature reference provided by the reference grating or the environmental temperature sensor, the pure strain component is decoupled and separated out . Then, the strain values of all measurement points are spatially interpolated to construct a two-dimensional strain field covering the entire mold plane. represents the two-dimensional spatial coordinates on the mold plane, represents time. Further calculate the strain gradient tensor , extract the principal strain direction and amplitude, identify the position of the resin flow front - which corresponds to the strain gradient mutation area, that is, the transition zone where the strain rises rapidly from 0 to positive value. The time step is set to 0.5 milliseconds, ensuring non-aliasing sampling of the front dynamic with a resin flow speed of up to 10 millimeters per second.

[0024] S3, input the two-dimensional strain gradient field into the pre-constructed multi-scale pore generation dynamics model, the model is based on Darcy's law and capillary pressure balance equation, introduces the local permeability tensor and interfacial tension coefficient at the scale of fiber bundle, solves the space-time evolution equation of resin saturation by finite volume method, outputs the spatial position, area ratio and predicted pore volume fraction of the area not completely infiltrated by resin at the current time.

[0025] The control equation of the multi-scale pore generation dynamics model is the mass conservation equation: ; where is the local porosity, is the resin density, is the Darcy velocity. The Darcy velocity is given by the generalized Darcy's law: ; is the local permeability tensor, is the resin viscosity, is the capillary pressure. The capillary pressure is determined by the Young-Laplace equation: ; where is the resin-air interfacial tension, is the contact angle, is the fiber bundle equivalent radius. The local permeability tensor K is calibrated by reconstructing the 3D microstructure of the carbon fiber preform using off-line micro-CT scanning and simulating resin flow in the structure using lattice-Boltzmann method , whose values are dynamically adjusted with local fiber volume fraction and orientation. The model domain is consistent with the mold geometry, with a mesh size of 5 mm matching the spatial resolution of the sensor array. The boundary conditions are converted from the strain gradient field output from Step S2: the resin flow front position is taken as a moving boundary, whose advancing speed is derived from the strain rate; the inlet pressure is set by the measured pressure at the resin inlet; the outlet is a vacuum boundary with a pressure of 0. The governing equations are discretized using the finite volume method, with a time step of 0.5 ms synchronized with the sensor sampling. By iterative solving, the resin saturation distribution is obtained, and the area with a resin saturation less than 0.95 is defined as the incompletely impregnated region. The pore volume fraction is determined from the relationship between saturation and local porosity . The model outputs include the spatial coordinates of the incompletely impregnated region, the total area ratio , the total area of the entire study region, and the maximum local pore volume fraction .

[0026] S4, generating pressure gradient control instructions and temperature field cooperative correction instructions according to the deviation of the predicted pore volume fraction from the preset threshold value; the pressure gradient control instructions act on the multi-zone independent hydraulic execution unit of the hot press to adjust the amplitude and rising rate of the pressure applied by each region; the temperature field cooperative correction instructions act on the partitioned electric heating sheet built in the mold to adjust the set temperature and temperature rising slope of each heating region.

[0027] The preset pore volume fraction threshold value is 0.5%, and the allowable deviation is 0.1%. When the model output , the control mechanism is triggered. The control instruction generation is based on the solution of an inverse optimization problem: the objective function is to minimize the squared error integral of the pore volume fraction and the threshold value, that is, ; The constraint conditions include the maximum output pressure of the hot press hydraulic system of 30 MPa, the response delay of each hydraulic cylinder of no more than 50 ms, and the maximum local pressure gradient limited by the mold structure strength of 5 MPa per centimeter. The optimization variables are the target pressure of each hydraulic execution unit and the target temperature of each electric heating sheet, where is 1 to 16, is 1 to 16, corresponding to 16 independent regulation regions. Spatially continuous pressure gradient regulation instruction By optimizing the obtained Mapping to spatially continuous field, used to accelerate resin flow to low saturation area. Temperature correction amount According to the resin viscosity-temperature index relationship: ; is the initial resin viscosity, is the resin activation energy, is the gas constant, is the absolute temperature, is the absolute temperature is Resin viscosity at the required viscosity. Reducing the local temperature can increase the viscosity and slow down the flow; increasing the temperature reduces the viscosity and promotes the resin to fill the dry spot area. Temperature correction amount Corresponding to the required viscosity adjustment amount, obtained by table lookup or real-time calculation. The regulation instruction is updated at a frequency of 2 Hz per second, and is sent to the actuator through industrial Ethernet.

[0028] The table lookup means that the viscosity data of the target resin system at different temperatures is determined by offline experiments in advance, and a "required viscosity adjustment amount-temperature correction amount" corresponding relationship table is established (including the matching results of known parameters such as resin viscosity, activation energy, gas constant, etc.). During actual regulation, according to the current pore inversion result to determine the required viscosity adjustment amount, the corresponding temperature correction amount is directly called from the pre-stored table, without real-time calculation, and the response speed is fast.

[0029] Real-time calculation means that based on the lipid viscosity-temperature index relationship, combined with the current real-time monitoring of the pore volume fraction deviation, the real-time state of the resin (such as the current viscosity, temperature), the required viscosity adjustment amount is dynamically derived, and the corresponding temperature correction amount is reversely calculated. This method does not need to rely on pre-stored data, can flexibly adapt to real-time working condition changes of resin, and has higher precision.

[0030] S5, the above monitoring, inversion and regulation cycle is continuously performed during the resin flow and curing stage, until the resin completely fills the cavity and the pore volume fraction is stable below 0.5%.

[0031] After the resin injection is completed, the curing stage is entered, at which time the resin flow stops, but the chemical cross-linking reaction initiates volume shrinkage, and residual strain is generated. The distributed fiber Bragg grating sensing array continues to monitor the strain changes at each point. The strain relaxation rate is defined as the absolute value of the strain change amount per unit time. When the strain change rate of the entire region is lower than 5 micro-strains per minute, it is determined that the resin curing reaction is basically completed, the molecular chain network structure is stable, and no significant shrinkage is generated. At this time, the regulation cycle is terminated, the hydraulic and electric heating execution units are closed, and the pressure maintaining and cooling stage is entered. During the entire process, the monitoring-inversion-regulation cycle runs at a frequency of 2 hertz per second, ensuring closed-loop control of the entire process of resin flow and curing. The final molded part is detected by ultrasonic C scanning, and the porosity distribution is uniform, with a maximum pore volume fraction of 0.48%, meeting the stringent requirement of automobile structural parts for porosity less than 0.5%.

[0032] Corresponding to the above method, the embodiment also discloses a real-time monitoring and regulation system for a manufacturing process of a lightweight automobile part, which comprises a distributed fiber Bragg grating sensing embedded module, a resin flow front dynamic strain field real-time demodulation module, a multi-scale pore generation kinetics online inversion module, a self-adaptive pressure-temperature cooperative regulation instruction generation module, and a multi-zone hydraulic-electric heating execution regulation module.

[0033] The distributed fiber Bragg grating sensing embedded module is used to arrange the sensing array in the form of an orthogonal grid between the carbon fiber preform and the inner cavity of the mold. The module includes an optical fiber placing mechanical arm, a tension controller and a positioning clamp, which ensures that the optical fiber is accurately attached to the surface of the mold before the preform is laid, without wrinkles or suspension. The sensing array is integrated with 128 gratings in a single optical fiber, with 64 measuring points along the length and width directions of the mold, covering an area of more than 648 square millimeters. The optical fiber is packaged with polyimide, with an outer diameter of 125 microns, and can withstand the shear force and high temperature environment during resin injection.

[0034] The resin flow front dynamic strain field real-time demodulation module comprises a wide-spectrum light source, a circulator, a high-speed spectral demodulator and a data processing unit. The wide-spectrum light source has a power stability better than 0.5% and a wavelength range of 1520 to 1570 nanometers. The high-speed spectral demodulator has a sampling rate of 2000 frames per second and a wavelength resolution of 0.1 picometer. The data processing unit runs a peak tracking algorithm and a temperature-strain decoupling program, and outputs two-dimensional strain field data in real time with a delay of no more than 1 millisecond.

[0035] The multi-scale pore generation dynamics online inversion module is deployed on an industrial computer equipped with a multi-core processor and a graphics acceleration unit. The module loads a pre-constructed numerical model, receives strain field data as boundary conditions, and uses the finite volume method to solve the resin saturation and pore volume fraction. Model parameters include local permeability tensor, interfacial tension, resin viscosity-temperature relationship, etc., which are calibrated through offline experiments. The inversion results are output to the control instruction generation module at a frequency of 2 Hz per second.

[0036] The adaptive pressure-temperature coordinated control instruction generation module runs an inverse optimization algorithm, with the objective function being the minimization of pore volume fraction error and the constraint condition covering device physical limits. The module outputs 16 pressure instructions and 16 temperature instructions, corresponding to the hydraulic cylinders of the hot press and the electric heating sheets of the mold, respectively. The instruction update period is 500 milliseconds, ensuring the timeliness of the control response.

[0037] The multi-zone hydraulic-electric heating execution control module includes 16 independent servo hydraulic cylinders and 16 sets of partitioned electric heating sheets. The hydraulic cylinder has a maximum output pressure of 30 MPa and a position repeatability of ±0.05 mm. The electric heating sheet uses thin film resistance heating technology, with a temperature control accuracy of ±1 degree and a heating rate of up to 10 degrees Celsius per second. After receiving the control instructions, the execution module dynamically adjusts the pressure and temperature of each region to achieve spatial collaborative control.

[0038] The above method and system realize full-process closed-loop monitoring and active control of resin flow and pore generation during the carbon fiber reinforced plastic laminated molding process, solve the technical problems that traditional methods cannot perceive microfluid dynamics and cannot intervene in the formation of local dry spots, and improve the quality consistency and structural reliability of automotive lightweight components.

Claims

1. A method for real-time monitoring and control of the manufacturing process of lightweight automotive components, characterized in that, include: After the carbon fiber preform is laid, a distributed fiber Bragg grating sensor array is embedded in an orthogonal grid between the inner surface of the mold cavity and the preform. The vacuum-assisted resin transfer molding process is initiated. During the resin injection stage, the center wavelength offset of each grating node is collected in real time by a broadband light source and a high-speed spectral demodulator. Based on the linear coupling relationship between wavelength offset and strain and temperature, the two-dimensional strain gradient field of the resin flow front in the mold plane coordinate system at each time step is decoupled and calculated. The two-dimensional strain gradient field is input into a pre-constructed multi-scale pore generation dynamics model. The model is based on Darcy's law and capillary pressure balance equation, and introduces the local permeability tensor and interfacial tension coefficient at the fiber bundle scale. The spatiotemporal evolution equation of resin saturation is solved by discretization using the finite volume method, and the spatial location, area ratio and predicted pore volume fraction of the resin-incompletely wetted area at the current moment are output. Based on the deviation between the predicted pore volume fraction and the preset threshold, a pressure gradient control command and a temperature field collaborative correction command are generated. The pressure gradient control command acts on the multi-zone independent hydraulic actuator of the hot press to adjust the amplitude and rise rate of the pressure applied in each zone. The temperature field collaborative correction command acts on the partitioned electric heating elements built into the mold to adjust the set temperature and heating slope of each heating zone. During the resin flow and curing stages, the above monitoring, inversion and control cycle is continuously performed until the resin completely fills the cavity and the pore volume fraction is stable below 0.5%.

2. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 1, characterized in that, The sensor array has a grating period of 532 nanometers, a spatial resolution of 5 millimeters, a strain measurement accuracy of one microstrain, and a temperature cross-sensitivity of less than 0.5 degrees Celsius per microstrain after encapsulation compensation.

3. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 2, characterized in that, After the carbon fiber preform is laid, a distributed fiber Bragg grating sensor array is embedded in an orthogonal grid between the inner surface of the mold cavity and the preform, including: The sensing array, consisting of 128 grating units etched on a single optical fiber, is arranged with 64 measuring points along the length and width of the mold, with a spacing of 10 mm between adjacent gratings. The sensor array is encapsulated with a polyimide coating, giving it an outer diameter of 125 micrometers and a temperature resistance range of -200 degrees Celsius to 300 degrees Celsius. The sensor array is placed in the interface layer between the inner surface of the mold cavity and the carbon fiber preform to ensure that the resin flow front causes local strain changes in the grating area without affecting the resin penetration path.

4. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 3, characterized in that, The vacuum-assisted resin transfer molding process is initiated. During the resin injection stage, the center wavelength offset of each grating node is acquired in real time using a broadband light source and a high-speed spectrometer. Based on the linear coupling relationship between wavelength offset and strain and temperature, the two-dimensional strain gradient field of the resin flow front in the mold plane coordinate system at each time step is decoupled and calculated, including: Continuous light with a wavelength range of 1520 nm to 1570 nm is emitted through a broadband light source, coupled into the sensing fiber through a circulator, and the reflectance spectrum sequence is acquired by a high-speed spectrometer at a rate of 2000 frames per second. The center wavelength of the reflection peak of each grating is located at the sub-picometer level to obtain the center wavelength offset. ; Based on fiber Bragg grating sensing equation: ; The initial center wavelength, The photoelastic coefficient, The coefficient of thermal expansion is... Thermo-optic coefficient; In response, For temperature change; By combining a temperature reference provided by a reference grating or an ambient temperature sensor, the pure strain component is decoupled and separated. ; Represents the two-dimensional spatial coordinates on the mold plane. Represents time; Spatial interpolation is performed on the strain values ​​at all measuring points to construct a two-dimensional strain field covering the entire mold plane, and the strain gradient tensor is calculated. To identify the location of the resin flow front.

5. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 4, characterized in that, The two-dimensional strain gradient field is input into a pre-constructed multi-scale pore generation kinetic model. This model, based on Darcy's law and the capillary pressure balance equation, introduces the local permeability tensor and interfacial tension coefficient at the fiber bundle scale. The model then uses the finite volume method to discretize and solve the spatiotemporal evolution equation of resin saturation, outputting the spatial location, area percentage, and predicted pore volume fraction of the resin-incompletely wetted region at the current moment. The mass conservation equation As the governing equation, where For local porosity, The density of the resin, For Darcy's speed; Darcy speed is expressed as ,in For local permeability tensor, Resin viscosity, For capillary pressure; Capillary pressure From the Young-Laplace equation Confirmed, among which The resin-air interfacial tension. Contact angle, The equivalent radius of the fiber bundle; The two-dimensional strain gradient field is converted into a moving boundary condition. The resin flow front propulsion velocity is derived using the strain change rate. The governing equations are then discretized and solved using the finite volume method to obtain the resin saturation. and pore volume fraction .

6. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 5, characterized in that, The local permeability tensor K was used to reconstruct the three-dimensional microstructure of the carbon fiber preform using offline micro-CT scanning, and the resin flow in the structure was simulated using the lattice Boltzmann method to calibrate the anisotropic permeability components. .

7. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 6, characterized in that, Based on the deviation between the predicted pore volume fraction and a preset threshold, pressure gradient control commands and temperature field collaborative correction commands are generated, including: Set a preset pore volume fraction threshold The tolerance is 0.5%. It is 0.1%; When the maximum local pore volume fraction Then, construct an inverse optimization problem, with the objective function being... The constraints include a maximum output pressure of 30 MPa for the hot press, a response delay of no more than 50 milliseconds for the hydraulic cylinder, and a maximum local pressure gradient of 5 MPa per centimeter for the strength of the mold structure. Solve the inverse optimization problem to obtain the target pressure of 16 independent hydraulic actuators. The target temperature of the 16-group zoned electric heating element , From 1 to 16, The range is from 1 to 16; The optimized result Mapped to spatial continuous pressure gradient control command It is used to accelerate the flow of resin to the low saturation region.

8. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 7, characterized in that, The generation of the temperature field collaborative correction command includes: Based on the resin viscosity-temperature index relationship ,in The initial grease viscosity, The activation energy of the resin. The gas constant is... Absolute temperature absolute temperature is The resin viscosity at that time; Calculate the local temperature correction amount based on the required viscosity adjustment. To reduce local temperature and increase viscosity to slow down excessively fast flow, or to increase temperature and decrease viscosity to promote resin filling of dry spot areas; Will The mapping is to the temperature control command of 16-group zoned electric heating elements.

9. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 8, characterized in that, During the resin flow and curing stages, the above monitoring, inversion, and control cycle is continuously performed until the resin completely fills the cavity and the pore volume fraction stabilizes below 0.5%, including: After the resin injection is completed, the curing stage begins, and the residual strain caused by curing shrinkage at each point is monitored by a distributed fiber Bragg grating sensor array. The strain relaxation rate of the entire region is calculated and defined as the absolute value of the strain change per unit time. When the strain change rate is less than 5 microstrains per minute, curing is considered complete, and the control cycle is terminated.

10. The method for real-time monitoring and control of the manufacturing process of lightweight automotive components according to claim 9, characterized in that, The monitoring, inversion, and control cycle operates at a frequency of 2 Hz per second, and the control command update cycle is 500 milliseconds.

Images