A method of testing a resin component of a carbon fibre prepreg

Abstract

The present application relates to the field of on-line nondestructive testing of physical and chemical properties of high-performance composite materials, and discloses a method for testing resin components of carbon fiber prepreg, which comprises: projecting characteristic absorption light and reference light onto the surface of the moving prepreg, collecting diffuse reflection light intensity and surface temperature value, calculating degradation rate according to time sequence fluctuation of reference light intensity, adjusting central emission wavelength when the temperature value deviates from the standard value, and determining a correction operator for uneven fiber distribution according to spatially discrete diffuse reflection intensity, introducing the degradation rate, the adjusted light intensity and the correction operator into an analysis loop to correct differential absorbance, and calculating resin content and curing agent ratio value, wherein the present application uses reference light intensity feedforward to balance light path phase distortion, linkage adjusts wavelength to track functional group absorption peak migration, eliminates mechanical bounce and temperature drift system deviation, and realizes adaptive hedging of multi-source optical path distortion.

Description

Technical Field

[0001] This invention relates to a method for testing the resin composition of carbon fiber prepreg, belonging to the field of online non-destructive testing technology for the physical and chemical properties of high-performance composite materials. Background Technology

[0002] Currently, near-infrared reflectance spectroscopy is commonly used for non-destructive testing of polymer composite materials. This method projects light of a specific wavelength onto the material, utilizing the absorption of this wavelength by the functional groups of the epoxy resin. The intensity of reflected light from the material surface is then collected, and the differential absorbance is calculated. During the detection process, the light signal with the characteristic absorption of the component is typically compared with a reference light signal without the corresponding absorption characteristic. Under static detection conditions, the relative relationship between the baseline reflectance and the absorption peak intensity of the material is relatively stable, thus establishing a correspondence between spectral data and chemical component content. However, in high-performance carbon fiber prepreg... During online monitoring or large-roll re-inspection of continuous production lines, the material is in a state of continuous high-speed transmission and may experience high-frequency mechanical fluctuations and position changes with an amplitude of 2mm to 5mm. At the same time, the diffuse reflection of light on the surface of carbon fiber fabric has obvious directional differences. Within the microsecond-level alternating sampling interval, the dynamic fluctuation of the material will cause instantaneous changes in the position of the incident light spot and the scattered light path. The scattering paths of light with different characteristic wavelengths will also show different degrees of shift. As a result, the originally consistent optical path change relationship between the characteristic absorption channel and the reference channel is disrupted, and large-amplitude and unstable pseudo-component fluctuations are likely to appear in the test signal.

[0003] To address the aforementioned interference, existing methods typically perform digital filtering or multi-point time-domain averaging on the acquired signals. Some methods also employ physical restraint mechanisms to mechanically constrain the surface of the moving material. However, when faced with non-stationary random fluctuations, processing methods such as time-domain averaging are insufficient to reliably eliminate their influence and may even cause subsequent decoupling calculations to become unstable or weaken the response to instantaneous changes. Furthermore, contact-type physical restraint mechanisms are prone to scratching the surface of carbon fiber prepregs during continuous high-speed friction. For example, Chinese invention patent application CN105259135A discloses a pointless temperature-compensated near-infrared measurement method suitable for real-time online applications, attempting to incorporate temperature into the implicit measurement through partial least squares regression. Variable modeling and online recursive update algorithms based on historical sampling bias are used to correct the measurement output. However, this one-dimensional data-level recursive fitting, which relies on implicit variables, implicitly depends on the underlying assumption that the detection environment is generally stable and the interference is relatively simple. When faced with instantaneous and high-frequency mechanical fluctuations under the high-speed transmission line of carbon fiber prepreg and anisotropic scattering caused by unidirectional fabric weight change, the spatial geometric phase of the optical path has undergone severe non-common-mode distortion. At this time, the above-mentioned existing technology relies purely on the back-end historical error recursive calculation with lag, which cannot trace and correct the real-time photon scattering cross-section change that occurs on the front-end transmission path. This can easily lead to baseline drift and decoupling matrix degradation.

[0004] Therefore, how to correct for changes in the optical path caused by material dynamic displacement, ambient temperature changes, and fiber anisotropic scattering during ultra-high frequency spectral acquisition without using external contact mechanical limiting, and how to perform highly sensitive online quantitative testing of resin matrix content and curing agent ratio under high-speed transmission conditions, has become a technical problem that needs to be solved. Summary of the Invention

[0005] To address the problems in the background art, the technical solution of the present invention is as follows: A method for testing the resin composition of carbon fiber prepreg, comprising the following steps:

[0006] Step S1: Multiple monochromatic lights of characteristic wavelengths and reference wavelengths are alternately projected onto the surface of the carbon fiber prepreg in continuous transmission state, diffuse reflection light is captured and the corresponding characteristic light intensity and reference light intensity are output, real-time temperature is collected simultaneously, and discrete diffuse reflection intensity under reference wavelength light is collected through multiple detection channels.

[0007] Step S2: Calculate the transient degradation rate of light intensity based on the temporal fluctuation of the reference light intensity; when the real-time temperature value deviates from the reference temperature and the absolute value of the deviation reaches the preset temperature deviation threshold, generate a wavelength fine-tuning command and adjust the center emission wavelength of multiple characteristic wavelengths according to the preset compensation step size.

[0008] Step S3: Calculate the spatial dispersion range based on the discrete diffuse reflection intensity. When the spatial dispersion range exceeds the stability threshold, determine the weighted correction operator for uneven fiber distribution.

[0009] Step S4: The transient degradation rate of light intensity, the light intensity corresponding to each characteristic wavelength after adjustment, and the weighted correction operator are introduced into the signal analysis loop to correct the differential absorbance of each component and decouple the resin content and curing agent ratio of the current collection area of ​​the carbon fiber prepreg.

[0010] Preferably, in step S1, the surface temperature value of the carbon fiber prepreg is captured in real time by adding an infrared temperature measurement channel on the same side of the transmitter, and is used as the real-time temperature value. The multiple characteristic wavelengths include a first characteristic wavelength for responding to the resin and a second characteristic wavelength for responding to the curing agent. Each characteristic light intensity includes the first characteristic light intensity corresponding to the first characteristic wavelength and the second characteristic light intensity corresponding to the second characteristic wavelength. In step S2, the reference temperature is 25°C, the preset temperature deviation threshold is 0.5°C, and the preset compensation step size is 0.2nm / °C. The optical wavelength adjuster inside the transmitter receives the wavelength fine-tuning command and corrects the emission driving current parameters of the first characteristic wavelength and the second characteristic wavelength accordingly, so as to lock the adjusted center emission wavelength at the intrinsic chemical absorption peak of the molecular structure of the carbon fiber prepreg. In step S4, the differential absorbance of each component includes the first differential absorbance value corresponding to the resin and the second differential absorbance value corresponding to the curing agent. The resin content and curing agent ratio are the resin mass percentage content and the curing agent molar ratio, respectively.

[0011] Preferably, in step S1, the discrete diffuse reflection intensity under the reference wavelength light is simultaneously captured by scattering probes deployed at three discrete geometric spatial positions at 30°, 45° and 60° to the incident optical axis, and used as the discrete diffuse reflection intensity; in step S3, the calculation of the spatial dispersion range and determination of the weighted correction operator based on the discrete diffuse reflection intensity includes the following sub-steps: step S31, calculate the difference between the maximum and minimum values ​​of the discrete diffuse reflection intensity at the three captured discrete geometric spatial positions to obtain the spatial dispersion range; step S32, when the spatial dispersion range exceeds the stability threshold, use the spatial dispersion range as the feature input of the coefficient matrix to calculate the weighted correction operator to correct the differential absorbance of each component.

[0012] Preferably, the method further includes a step performed before step S1: step S401, controlling the projection light source module to alternately emit a first characteristic wavelength, a second characteristic wavelength and a reference wavelength, and performing pulse code modulation on the emitted light intensity to form monochromatic light with coded characteristics, which is input to the surface of the carbon fiber prepreg in step S1.

[0013] Preferably, in step S4, before introducing the light intensity into the signal analysis loop, the following sub-steps are included: Step S41, acquiring the transient power fluctuation data of the transmitter, quantitatively correcting the transient degradation rate of the light intensity by combining the temporal fluctuation of the reference light intensity, and generating a feedforward balancing gain coefficient, which is applied to the captured diffuse reflection light intensity to correct the transient attenuation of the light intensity caused by the continuous operation of the transmitter.

[0014] Preferably, in step S4, the decoupling calculation of the resin mass percentage content and curing agent molar ratio of the current sampling area of ​​the carbon fiber prepreg includes the following sub-steps: Step S42, input the corrected first difference absorbance value and second difference absorbance value into the preset component correlation matrix, and simultaneously calculate the resin mass percentage content and curing agent molar ratio by performing matrix inverse operation under full-rank square matrix constraint on the component correlation matrix.

[0015] Preferably, the method further includes a step performed after step S4: Step S701, after calculating the resin mass percentage content and the curing agent molar ratio, comparing the resin mass percentage content with the preset material composition standard range, and when the resin mass percentage content exceeds the preset material composition standard range, outputting the resin component anomaly index to perform zero-point dynamic calibration of the analysis baseline in the signal analysis loop.

[0016] Preferably, in step S4, the decoupling calculation of the differential absorbance of each component includes the following sub-steps: Step S43, read the first characteristic light intensity, the second characteristic light intensity, and the reference light intensity, perform component differential decoupling calculation, and obtain the first differential absorbance value of the epoxy resin matrix. The second difference absorbance value with the curing agent : ,in, For the test component number, in hour, The first difference absorbance value and The first characteristic light intensity after correction, in hour, The second differential absorbance value and The corrected second characteristic light intensity. This is a real-time sampled value of the reference light intensity.

[0017] Preferably, the method further includes the step of evaluating the uniformity of continuously produced materials along the process: Step S901, continuously record the resin mass percentage content of each collection area, construct a trend sequence of the resin mass percentage content as the winding length evolves, calculate the first derivative of the trend sequence to obtain the component space evolution rate, and when the component space evolution rate continues to exceed the preset fluctuation threshold, output a material evaluation index characterizing the non-uniformity defects along the process inside the carbon fiber prepreg roll.

[0018] Preferably, the method utilizes the velocity tracking unit to match analysis accuracy and includes the following steps:

[0019] Step S1001: Collect the continuous transmission speed of the carbon fiber prepreg. Based on the change in the continuous transmission speed, adjust the transmission frequency of the transmitter proportionally. When the continuous transmission speed is in the range of 15m / min to 30m / min, make the transmission frequency dynamically stable in the frequency domain range corresponding to the current transmission speed, so as to keep the spatial sampling step size of the current collection area constant.

[0020] Compared with the prior art, the beneficial effects of the present invention are:

[0021] 1. In the resin composition testing of carbon fiber prepreg, the measurement deviation caused by material fluctuation and scattering light path changes can be compensated in a timely manner by utilizing the change of reference light intensity. This can reduce the influence of the change of incident light spot position and light path offset on the differential absorbance calculation during high-speed transmission, so that the spectral absorption characteristics and resin matrix content maintain a relatively stable correspondence. Therefore, in the continuous production and large roll re-inspection of carbon fiber prepreg, the accuracy and stability of resin content and curing agent ratio testing can be improved. At the same time, this method does not require the use of mechanical limiting mechanism in contact with the material surface, which can avoid continuous friction causing scratches on the prepreg surface.

[0022] 2. Adjusting the characteristic wavelength according to the temperature change of the carbon fiber prepreg surface can make the detection wavelength change accordingly with the changes in the absorption peak position of the epoxy resin matrix and curing agent. This reduces the influence of red shift or blue shift of the absorption band caused by ambient temperature fluctuations on the test results, thereby reducing the system measurement deviation caused by temperature drift. This ensures that the test baseline of resin content and curing agent ratio remains stable under different temperature conditions, improving the reliability of online component determination.

[0023] 3. By comparing the diffuse reflection intensity of the reference wavelength collected at different spatial locations and correcting the scattering differences caused by uneven distribution of reinforcing fibers, the interference of local basis weight changes of unidirectional carbon fibers on the component differential absorbance can be reduced. This can avoid misjudging the light intensity fluctuations caused by changes in fiber arrangement or density as changes in resin composition, so that the test results can more accurately reflect the actual distribution of resin matrix content and curing agent ratio, and improve the adaptability of this method to different fiber surface states. Attached Figure Description

[0024] Figure 1 This is a flowchart of the adaptive correction process for multi-source optical path distortion of carbon fiber prepreg in this invention.

[0025] Figure 2 This is a structural diagram of the carbon fiber prepreg multi-source optical path distortion adaptive correction system of the present invention.

[0026] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0027] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0028] A method for testing the resin composition of carbon fiber prepreg includes the following steps:

[0029] Step S1: Multiple monochromatic lights of characteristic wavelengths and reference wavelengths are alternately projected onto the surface of the carbon fiber prepreg in continuous transmission state, diffuse reflection light is captured and the corresponding characteristic light intensity and reference light intensity are output, real-time temperature is collected simultaneously, and discrete diffuse reflection intensity under reference wavelength light is collected through multiple detection channels.

[0030] Step S2: Calculate the transient degradation rate of light intensity based on the temporal fluctuation of the reference light intensity; when the real-time temperature value deviates from the reference temperature and the absolute value of the deviation reaches the preset temperature deviation threshold, generate a wavelength fine-tuning command and adjust the center emission wavelength of multiple characteristic wavelengths according to the preset compensation step size.

[0031] Step S3: Calculate the spatial dispersion range based on the discrete diffuse reflection intensity. When the spatial dispersion range exceeds the stability threshold, determine the weighted correction operator for uneven fiber distribution.

[0032] Step S4: The transient degradation rate of light intensity, the light intensity corresponding to each characteristic wavelength after adjustment, and the weighted correction operator are introduced into the signal analysis loop to correct the differential absorbance of each component and decouple the resin content and curing agent ratio of the current collection area of ​​the carbon fiber prepreg.

[0033] Preferably, in step S1, the surface temperature value of the carbon fiber prepreg is captured in real time by adding an infrared temperature measurement channel on the same side of the transmitter, and is used as the real-time temperature value. The multiple characteristic wavelengths include a first characteristic wavelength for responding to the resin and a second characteristic wavelength for responding to the curing agent. Each characteristic light intensity includes the first characteristic light intensity corresponding to the first characteristic wavelength and the second characteristic light intensity corresponding to the second characteristic wavelength. In step S2, the reference temperature is 25°C, the preset temperature deviation threshold is 0.5°C, and the preset compensation step size is 0.2nm / °C. The optical wavelength adjuster inside the transmitter receives the wavelength fine-tuning command and corrects the emission driving current parameters of the first characteristic wavelength and the second characteristic wavelength accordingly, so as to lock the adjusted center emission wavelength at the intrinsic chemical absorption peak of the molecular structure of the carbon fiber prepreg. In step S4, the differential absorbance of each component includes the first differential absorbance value corresponding to the resin and the second differential absorbance value corresponding to the curing agent. The resin content and curing agent ratio are the resin mass percentage content and the curing agent molar ratio, respectively.

[0034] Preferably, in step S1, the discrete diffuse reflection intensity under the reference wavelength light is simultaneously captured by scattering probes deployed at three discrete geometric spatial positions at 30°, 45° and 60° to the incident optical axis, and used as the discrete diffuse reflection intensity; in step S3, the calculation of the spatial dispersion range and determination of the weighted correction operator based on the discrete diffuse reflection intensity includes the following sub-steps: step S31, calculate the difference between the maximum and minimum values ​​of the discrete diffuse reflection intensity at the three captured discrete geometric spatial positions to obtain the spatial dispersion range; step S32, when the spatial dispersion range exceeds the stability threshold, use the spatial dispersion range as the feature input of the coefficient matrix to calculate the weighted correction operator to correct the differential absorbance of each component.

[0035] Preferably, the method further includes a step performed before step S1: step S401, controlling the projection light source module to alternately emit a first characteristic wavelength, a second characteristic wavelength and a reference wavelength, and performing pulse code modulation on the emitted light intensity to form monochromatic light with coded characteristics, which is input to the surface of the carbon fiber prepreg in step S1.

[0036] Preferably, in step S4, before introducing the light intensity into the signal analysis loop, the following sub-steps are included: Step S41, acquiring the transient power fluctuation data of the transmitter, quantitatively correcting the transient degradation rate of the light intensity by combining the temporal fluctuation of the reference light intensity, and generating a feedforward balancing gain coefficient, which is applied to the captured diffuse reflection light intensity to correct the transient attenuation of the light intensity caused by the continuous operation of the transmitter.

[0037] Preferably, in step S4, the decoupling calculation of the resin mass percentage content and curing agent molar ratio of the current sampling area of ​​the carbon fiber prepreg includes the following sub-steps: Step S42, input the corrected first difference absorbance value and second difference absorbance value into the preset component correlation matrix, and simultaneously calculate the resin mass percentage content and curing agent molar ratio by performing matrix inverse operation under full-rank square matrix constraint on the component correlation matrix.

[0038] Preferably, the method further includes a step performed after step S4: Step S701, after calculating the resin mass percentage content and the curing agent molar ratio, comparing the resin mass percentage content with the preset material composition standard range, and when the resin mass percentage content exceeds the preset material composition standard range, outputting the resin component anomaly index to perform zero-point dynamic calibration of the analysis baseline in the signal analysis loop.

[0039] Preferably, in step S4, the decoupling calculation of the differential absorbance of each component includes the following sub-steps: Step S43, read the first characteristic light intensity, the second characteristic light intensity, and the reference light intensity, perform component differential decoupling calculation, and obtain the first differential absorbance value of the epoxy resin matrix. The second difference absorbance value with the curing agent : ,in, For the test component number, in hour, The first difference absorbance value and The first characteristic light intensity after correction, in hour, The second differential absorbance value and The corrected second characteristic light intensity. This is a real-time sampled value of the reference light intensity.

[0040] Preferably, the method further includes the step of evaluating the uniformity of continuously produced materials along the process: Step S901, continuously record the resin mass percentage content of each collection area, construct a trend sequence of the resin mass percentage content as the winding length evolves, calculate the first derivative of the trend sequence to obtain the component space evolution rate, and when the component space evolution rate continues to exceed the preset fluctuation threshold, output a material evaluation index characterizing the non-uniformity defects along the process inside the carbon fiber prepreg roll.

[0041] Preferably, the method utilizes the velocity tracking unit to match analysis accuracy and includes the following steps:

[0042] Step S1001: Collect the continuous transmission speed of the carbon fiber prepreg. Based on the change in the continuous transmission speed, adjust the transmission frequency of the transmitter proportionally. When the continuous transmission speed is in the range of 15m / min to 30m / min, make the transmission frequency dynamically stable in the frequency domain range corresponding to the current transmission speed, so as to keep the spatial sampling step size of the current collection area constant.

[0043] Example 1: This example applies the carbon fiber prepreg resin composition testing method to a continuous coating production line and large roll re-inspection. The carbon fiber prepreg is continuously transported at a speed of 15m / min to 30m / min. Uneven winding tension can cause mechanical fluctuations and displacement changes of 2mm to 5mm. The unidirectional continuous carbon fiber fabric texture on its surface has anisotropic diffuse reflection spatial scattering characteristics, which easily causes the characteristic absorption channel and the reference channel to form different spatial scattering paths during alternating sampling. The speed tracking unit continuously collects the continuous transport speed of the carbon fiber prepreg and adjusts the projected light proportionally with the speed change. The emission frequency of the source module is kept within the frequency domain range corresponding to the current transmission speed to maintain a constant spatial sampling step size in the current acquisition area. Taking the continuous transmission speed of 20 m / min as an example, the scintillation sampling interval of the three monochromatic lights is 50 μs. Within this sampling interval, the mechanical vibration of the material surface will cause the incident light spot and the scattered light path to be geometrically offset, resulting in the spatial scattering path of the characteristic absorption channel and the reference channel to deviate, changing the common-mode optical response of the two and generating non-stationary variation noise in the test signal, which in turn causes the decoupling result of the component differential absorbance to deviate.

[0044] Before projecting monochromatic light, the projection light source module pulse-codes the emitted light intensity corresponding to each wavelength according to the alternating emission sequence of the first characteristic wavelength, the second characteristic wavelength, and the reference wavelength, forming monochromatic light with encoded characteristics. The projection light source module then alternately projects monochromatic light of the first characteristic wavelength of 1430nm, which responds to the absorption of epoxy groups in the epoxy resin matrix, monochromatic light of the second characteristic wavelength of 1510nm, which responds to the absorption of amino groups in the curing agent, and monochromatic light of the reference wavelength of 1300nm, which serves as a physical scattering reference baseline without chemical characteristic absorption, onto the surface of the carbon fiber prepreg. The photodetector located on the same side of the transmitter synchronously captures the diffuse reflection light intensity of the carbon fiber prepreg surface. The signal analysis unit distinguishes the corresponding wavelength channels according to the encoded characteristics of each sampling pulse, forming the first characteristic light intensity sequence, the second characteristic light intensity sequence, and the reference light intensity sequence, respectively.

[0045] The signal analysis unit reads the reference light intensity value of the current sampling period. The reference light intensity value of the previous sampling period The transient degradation rate of light intensity is calculated according to the following formula. : ,in, Indicates the current sampling period number. This represents the reference light intensity value for the current sampling period. This represents the reference light intensity value from the previous sampling period. The signal analysis unit also acquires transient power fluctuation data from the transmitter, which is a dimensionless transient rate of light intensity. This data is then correlated with the temporal fluctuations of the reference light intensity using the same sampling period. This is used to correct the component in the transient rate of light intensity caused by changes in the transmitter power. The corrected transient rate of light intensity is still denoted as... .

[0046] The signal analysis unit will The absolute value is compared with the preset fluctuation deviation threshold. When the absolute value exceeds the fluctuation deviation threshold, As the feedforward balancing gain coefficient, it represents the real-time sampled value of the first characteristic light intensity in the current sampling period. Second characteristic light intensity real-time sampling value The correction is performed to obtain the first characteristic light intensity correction value. Second characteristic light intensity correction value : ,in, This is the real-time sampled value of the first characteristic light intensity in the current sampling period. This is the real-time sampled value of the second characteristic light intensity in the current sampling period. This is the first characteristic light intensity correction value. The second characteristic light intensity correction value is the feedforward balancing gain coefficient applied to the captured diffuse reflection light intensity. It is used to correct the light intensity deviation caused by the change in optical path geometry during alternating sampling and to compensate for the transient light intensity attenuation caused by continuous operation of the transmitter.

[0047] To correct for absorption band migration caused by temperature changes in the workshop, an infrared temperature measurement channel located on the same side of the transmitter collects the real-time temperature value of the carbon fiber prepreg surface. When the real-time temperature value deviates from the standard temperature by 25℃ and the absolute value of the deviation is greater than or equal to 0.5℃, the signal analysis unit generates a wavelength fine-tuning command. The optical wavelength adjuster inside the transmitter receives the command and, based on the pre-calibrated correspondence between the light emission driving current parameters and the center emission wavelength, corrects the light emission driving current parameters of the first characteristic wavelength and the second characteristic wavelength, and adjusts the corresponding center emission wavelength with a compensation step size of 0.2nm / ℃, so that the adjusted center emission wavelength is locked at the intrinsic chemical absorption peak position of the molecular structure of the carbon fiber prepreg.

[0048] The 0.2 nm / ℃ compensation step size used in this embodiment is a macroscopic equivalent linear compensation coefficient determined experimentally for a multiphase crosslinked system of epoxy resin matrix and curing agent with a specific ratio range within a standard workshop temperature range of 25℃ to 35℃. The thermally induced displacements of epoxy groups and amino groups in the microscopic molecular structure are not exactly the same. However, when the ambient temperature undergoes a gradual drift of 0.5℃ to 5℃ in a short period of time, the macroscopic volume expansion of the composite system and the change in the refractive index of the matrix medium produce a common-mode modulation effect on the migration of the main absorption band. Wide-temperature-range environmental simulation chamber experiments show that, under this temperature range and multiphase dispersion medium conditions, adjusting the center emission wavelengths of the first and second characteristic wavelengths synchronously according to a compensation step size of 0.2 nm / ℃ allows the photon energies corresponding to the two components to match the migration trajectory of their intrinsic chemical absorption peak positions in their molecular structures. The residual nonlinear deviation caused by the microscopic hydrogen bond changes is canceled out in the subsequent dual-channel logarithmic difference operation, thereby maintaining the test baseline under temperature change conditions.

[0049] To correct the residual scattering error caused by localized uneven basis weight of the unidirectional reinforcing fiber, scattering probes were installed at three discrete geometrical spatial positions at 30°, 45°, and 60° to the incident optical axis. Each scattering probe synchronously captured the discrete diffuse reflection intensity under the reference wavelength light. The signal analysis unit calculated the difference between the maximum and minimum values ​​of the three discrete diffuse reflection intensities, and this difference was taken as the spatial dispersion range. When the spatial dispersion range is very poor When the stability threshold of 15% is exceeded, the spatial dispersion range will be extremely poor. As a feature input to the coefficient matrix, a weighted correction operator is calculated for the uneven distribution of reinforcing fibers, and this weighted correction operator is applied to the input term of the component differential absorbance to correct the non-common-mode scattering bias caused by the uneven distribution of fibers.

[0050] The signal analysis unit corrects the first characteristic light intensity value. Second characteristic light intensity correction value and reference light intensity value Introducing a signal analysis loop, the first differential absorbance value of the epoxy resin matrix is ​​calculated according to the following formula. The second difference absorbance value between the curing agent and the curing agent : ,in, This is the first difference absorbance value. The second differential absorbance value is calculated by the signal analysis unit using the first differential absorbance value corrected by the light intensity transient degradation rate, the adjusted center emission wavelength light intensity, and the weighted correction operator. The second difference absorbance value Input the preset component correlation matrix, and perform matrix inverse operation on the component correlation matrix under full-rank matrix constraint to decouple and calculate the resin mass percentage content and curing agent molar ratio of the current sampling area of ​​carbon fiber prepreg. Output the two as quality control data. Within the above working conditions, the absolute error of component testing is controlled within 0.4%.

[0051] The maximum adjustment of the center emission wavelength is limited to a physical window of 2 nm. Within this spectral range, the spectral responsivity curve of the indium gallium arsenide photodetector and the substrate reflectivity of the carbon fiber fabric are both in the local linear region, and the absolute change of their differential response fluctuation is less than 0.01%. Therefore, the light intensity change caused by the fine adjustment of the center emission wavelength can be treated as a definite equivalent scalar gain. The signal analysis unit introduces the adjusted center emission wavelength light intensity as a scalar data stream into the signal analysis loop, so that the light intensities corresponding to the reference wavelength, the adjusted first characteristic wavelength, and the adjusted second characteristic wavelength are differentially ratioed in logarithmic space. The small change in the photodetector responsivity and the change in the scattering cross section caused by the wavelength fine adjustment are algebraically canceled in the common-mode comparison of each channel, thereby converting the adjustment result of the center emission wavelength into a feedforward balancing quantity that can be processed by a one-dimensional light intensity data stream, so that the physical quantities and data streams in the signal analysis loop are consistent.

[0052] Example 2: In this example, an offline physical and chemical testing calibration platform with a running speed of 20 m / min was selected for testing. An eccentric wheel mechanism was set at the connection of the drive shaft to introduce periodic mechanical amplitude disturbance to simulate the displacement change caused by uneven winding tension. A spectral acquisition unit was set at the data output end of the test platform. The spectral acquisition unit includes a multi-channel continuous monochromatic light generator and a photodetector with a response bandwidth of 50 MHz. The wavelength sensing range of the photodetector is 1100 nm to 1700 nm, which is used to collect the diffuse reflection light intensity of the carbon fiber prepreg surface under transmission conditions.

[0053] The alternating sampling gap is determined based on the material transport speed, the incident light spot size, and the charge accumulation of the photodetector. When the alternating sampling gap is greater than 50 μs, the anisotropic diffuse reflection of the carbon fiber fabric texture and the material displacement will cause phase deviation between the characteristic absorption channel and the reference channel, forming pseudo-component noise. When the alternating sampling gap is less than 50 μs, the single photon integration time is shortened, and the photoelectric readout noise will cover the measurement signal. When the continuous transport speed of the carbon fiber prepreg is between 15 m / min and 30 m / min, the alternating sampling gap is determined to be 50 μs, with the material surface displacement distance within the alternating sampling gap being less than 5% of the incident light spot diameter and the charge accumulation reaching 70% of the charge well capacity. The true value of the resin content of each group of samples is 34%.

[0054] In sample group 1 of this invention, the carbon fiber prepreg is transported at a speed of 20 m / min, with a jump amplitude of 2 mm and a surface temperature of 25°C. The discrete diffuse reflection intensity collected by the three scattering detection probes forms a spatial dispersion range of 12.1%. The reference light intensity value for the current sampling period The transient degradation rate of light intensity is 452.3 nA, calculated based on the temporal fluctuations of the reference light intensity. The first characteristic light intensity correction value is 0.012 after feedforward balancing. The second characteristic light intensity correction value is 385.4 nA. The value is 312.1 nA. The corrected light intensity is input into the signal analysis circuit to obtain the first differential absorbance value. The second differential absorbance value is 0.16. The value was 0.37; after decoupling by the component correlation matrix, the measured resin content was 34.2%, which was +0.2% relative to the true value of resin content.

[0055] In sample group 2 of this invention, the continuous transmission speed remains at 20 m / min, the fluctuation amplitude increases to 4.5 mm, and the surface temperature is maintained at 25°C. At this time, the spatial dispersion is extremely poor. It is 14.3%, with a reference light intensity value. The calculated transient degradation rate of light intensity is 421.5 nA. The value is 0.068. After feedforward balancing of the characteristic light intensity using this degradation rate, the first characteristic light intensity correction value is... The second characteristic light intensity correction value is 342.6 nA. The first differential absorbance value is 285.3 nA; The second differential absorbance value is 0.207. The value was 0.39. After inputting the two differential absorbance values ​​into the component correlation matrix, the measured resin content was 34.3%, and the measurement deviation was +0.3%.

[0056] In sample group 3 of this invention, the continuous transmission speed is 20 m / min, the fluctuation amplitude is 2 mm, the surface temperature rises to 35°C, and the spatial dispersion is extremely poor. It is 11.8%, with a reference light intensity value. The transient degradation rate of light intensity is 455.1 nA. The value is 0.009. The center emission wavelengths of the first and second characteristic wavelengths are adjusted according to the real-time temperature value. After feedforward balancing of the diffuse reflection light intensity, the first characteristic light intensity correction value is... The second characteristic light intensity correction value is 388.2 nA. The value is 314.5 nA; further calculation yields the first differential absorbance value. The second differential absorbance value is 0.159. The value was 0.368. After decoupling by the component correlation matrix, the measured resin content was 33.8%, and the measurement deviation was -0.2%.

[0057] In another optional operating condition of this embodiment, a continuous transmission speed of 15 m / min is used, the fluctuation amplitude is 2.1 mm, the surface temperature is 24.5℃, and the spatial dispersion of the collected data is extremely poor. It is 8.5%, with a reference light intensity value. The transient degradation rate of light intensity is 458.7 nA. The first characteristic light intensity correction value is -0.005 after light intensity correction. The second characteristic light intensity correction value is 392.3 nA. The value is 318.4 nA; the first differential absorbance value is calculated from this. The second differential absorbance value is 0.156. The value was 0.365, the resin content was 34.1%, and the measurement deviation was +0.1%.

[0058] The sample group 5 of this invention uses a continuous transmission speed of 30 m / min, a fluctuation amplitude of 5 mm, and a surface temperature of 25.5℃. Under these conditions, the spatial dispersion is extremely poor. It is 14.9%, with a reference light intensity value. The transient degradation rate of light intensity is 412.6 nA. The first characteristic light intensity correction value is 0.082, obtained through feedforward balancing. The second characteristic light intensity correction value is 331.4 nA. The first differential absorbance value is 272.8 nA. The second differential absorbance value is 0.219. The value is 0.413. After decoupling by the component correlation matrix, the measured resin content is 34.4%, and the measurement deviation is +0.4%. In the above sample group of the present invention, the temporal fluctuation of the reference light intensity is used to characterize the optical path change during the alternating sampling period, and the transient degradation rate of light intensity is used to balance the path offset between the characteristic absorption channel and the reference channel, so that the first differential absorbance value and the second differential absorbance value correspond to the absorption changes of epoxy groups and amino groups.

[0059] Control group 1 uses the same continuous transmission speed of 20 m / min, 4.5 mm fluctuation amplitude, 25 °C surface temperature, and 14.3% spatial dispersion range as sample group 2 of this invention. and 421.5 nA reference light intensity value However, it blocks feedforward balancing and reduces the transient degradation rate of light intensity. Set to 0, the first characteristic light intensity correction value when the temporal fluctuation of the reference light intensity is not used to correct the characteristic light intensity. The second characteristic light intensity correction value is 320.8 nA. The first differential absorbance value is 267.1 nA. The second differential absorbance value is 0.272. The value is 0.456. After decoupling by the component correlation matrix, the measured resin content is 29.5%, and the measurement deviation is -4.5%. It can be seen that under the condition of the same mechanical runout amplitude, without feedforward leveling, the spatial phase deviation between the characteristic absorption channel and the reference channel will be directly transmitted to the differential absorbance calculation.

[0060] Control group two used a continuous transmission speed of 20 m / min, a fluctuation amplitude of 2 mm, and a surface temperature of 35 °C, exhibiting extremely poor spatial dispersion. It is 11.8%, with a reference light intensity value. The transient degradation rate of light intensity is 455.1 nA. The value is 0.009, indicating that dynamic tracking of the emission wavelength is disabled for this group, and the first characteristic light intensity correction value is [value missing]. The second characteristic light intensity correction value is 388.2 nA. The first differential absorbance value is 314.5 nA, when the fixed center emission wavelength is not adjusted with changes in the absorption peak position. The second differential absorbance value decreased to 0.105. The value dropped to 0.298, and the resin content measured after decoupling was 37.8%, with a measurement deviation of +3.8%. This result corresponds to the migration of the characteristic absorption band caused by the change in surface temperature. The center wavelength of the fixed filter channel deviates from the position of the migrated absorption peak, causing the two differential absorbance values ​​to deviate from the calibration input of the component correlation matrix.

[0061] Control group three used a continuous transmission speed of 20 m / min, a fluctuation amplitude of 2 mm, and a surface temperature of 25 °C, which enhanced the uneven local basis weight of the fibers, resulting in extremely poor spatial dispersion. Increased to 18.5%, reference light intensity value The transient degradation rate of light intensity is 448.2 nA. The first characteristic light intensity correction value is 0.015 after feedforward balancing. The second characteristic light intensity correction value is 381.5 nA. The absorbance is 308.6 nA; the first differential absorbance value is [value missing] when the weighted correction operator is not used to correct the component differential absorbance. The second differential absorbance value is 0.161. The value was 0.372, the resin content was 35.9%, and the measurement deviation was +1.9%. This shows that when the spatial dispersion exceeds the stability threshold, the non-common mode scattering residue caused by the change in the unidirectional reinforcing fiber arrangement density will enter the decoupling results.

[0062] The out-of-range control group used a continuous transmission speed of 20 m / min, a fluctuation amplitude of 7.5 mm, and a surface temperature of 25 °C, exhibiting extremely poor spatial dispersion. The value is 13.5%, because the transient displacement of the material surface exceeds the geometric cross-sectional boundary of the photosensitive surface of the photodetector, and the reference light intensity value... The transient degradation rate of light intensity decreased to 210.4 nA. Increasing it to 0.352, at this point, the first characteristic light intensity correction value... The second characteristic light intensity correction value is 140.2 nA. The first differential absorbance value is 112.5 nA. The second differential absorbance value is 0.406. The value was 0.626, the resin content was 28.2%, and the measurement deviation was -5.8%. Under this condition, the photoelectric signal exhibited nonlinear shear saturation, the transient degradation rate of light intensity lost its stable benchmark for characterizing changes in the optical path, and the differential absorbance could no longer be used as an effective input for the component correlation matrix.

[0063] The out-of-range control group two used a continuous transmission speed of 20 m / min, a fluctuation amplitude of 2 mm, and a surface temperature of 55℃, exhibiting extremely poor spatial dispersion. It is 12.2%, with a reference light intensity value. The transient degradation rate of light intensity is 453.6 nA. The value is 0.011, which is the first characteristic light intensity correction value after light intensity correction. The second characteristic light intensity correction value is 386.1 nA. The first differential absorbance value is 313.2 nA. The second differential absorbance value is 0.052. The value was 0.184, the resin content was measured to be 40.5%, and the measurement deviation was +6.5%. Under this temperature condition, the change in the condensed structure of the matrix caused the spectral characteristic bands to overlap, exceeding the adjustment window for linear tracking with a compensation step size of 0.2 nm / ℃. Consequently, the eigenvalues ​​of the component correlation matrix degraded.

[0064] Within the working range of 2mm to 5mm mechanical fluctuation amplitude, the feedforward balancing, dynamic tracking of the center emission wavelength, and spatial dispersion range correction formed by the temporal fluctuation of the reference light intensity are input to the component differential absorbance calculation loop, so that the absolute value of the resin content measurement deviation is kept within 0.4%. When the aforementioned parameters exceed the working range, the nonlinear distortion of the photoelectric signal or the overlap of spectral characteristic bands will destroy the decoupling conditions of the component correlation matrix. During the test, the surface of the carbon fiber prepreg does not contact the rigid limiting mechanism, thereby avoiding slippage and scratches during continuous transmission, and completing the non-destructive quantitative determination of the components of the polymer multiphase composite material.

[0065] Example 3: In the current continuous production line and large roll re-inspection process, carbon fiber prepreg is transported at a speed of 15m / min to 30m / min. The local stacking of carbon fiber unidirectional reinforcement on the material surface will cause thickness deviation. After the spectral acquisition unit has been running continuously for more than 500 hours, the radiation efficiency of the light source emitter gradually decreases. A condensation film formed by polymer volatiles will also be deposited on the surface of the optical window, causing the measurement baseline to shift. Within the 50μs alternating sampling interval, the reference light intensity value collected by the photodetector will thus produce a non-stationary drift. If a fixed fluctuation deviation threshold is still used, the slow drift caused by light source attenuation or window contamination will be identified as material fluctuation, which will cause the first differential absorbance value and the second differential absorbance value to be superimposed with a deviation unrelated to chemical absorption, resulting in a tilt in the composition measurement results during the continuous production process.

[0066] During system startup or idle transmission intervals, the signal analysis unit continuously reads reference light intensity values ​​for 100 sampling periods to construct a baseline fluctuation sequence. The sample standard deviation of this sequence is used as the jump deviation threshold. During continuous measurement, 10,000 sampling periods are used as the update sliding window period, and a time decay factor of 0.95 is used to adjust the weight of historical reference light intensity values ​​within the sliding window. The time decay factor is a dimensionless quantity. After each update sliding window is completed, the signal analysis unit recalculates the jump deviation threshold, so that the threshold is updated synchronously with the changes in light source radiation efficiency and light path transmission state, thereby distinguishing between slow light intensity decay caused by continuous operation and transient light intensity changes caused by material mechanical jump.

[0067] When the spatial dispersion is extremely poor When the stability threshold exceeds 15%, the signal analysis unit will calculate the spatial dispersion range before calculating the component differential absorbance. As a feature input to the coefficient matrix, and according to the spatial dispersion range The one-dimensional linear relationship between the operator and the constant scalar 0.05 is used to calculate the weighted correction operator, which is a dimensionless quantity with a spatial dispersion range. The percentage values ​​are used in the calculation. Subsequently, the weighted correction operator is compared with the real-time sampled value of the first characteristic light intensity. and the real-time sampled value of the second characteristic light intensity Multiplying these two values ​​yields the corrected first characteristic light intensity and the corrected second characteristic light intensity, which compensate for the baseline shift caused by the anisotropic scattering variation of the carbon fiber unidirectional reinforcement.

[0068] The coefficient matrix used to calculate the weighted correction operator adopts a diagonal structure, representing the spatial dispersion range of data collected from multiple detection channels. The weighted correction operator, obtained from a one-dimensional linear relationship, is mapped to two characteristic light intensity channels respectively. It corresponds to the main diagonal element of the diagonal coefficient matrix in the characteristic light intensity channel. The subsequent signal analysis loop performs a natural logarithmic difference operation on the corrected characteristic light intensity. Therefore, the multiplicative correction performed at the light intensity input is equivalent to superimposing the corresponding correction amount on the first difference absorbance value and the second difference absorbance value in logarithmic space. Using this processing method, it is not necessary to repeatedly perform multidimensional matrix operations during real-time signal analysis to correct the non-common mode scattering residual error caused by uneven local arrangement of reinforcing fibers.

[0069] The signal analysis unit inputs the light intensity data, corrected by the weighted correction operator, into the absorbance calculation loop to obtain the corrected first differential absorbance value. Second differential absorbance value Then, the two are input into a preset component correlation matrix. Under the constraint of full-rank matrixization, the component correlation matrix is ​​subjected to matrix inverse operation. The resin mass percentage content and curing agent molar ratio of the current sampling area of ​​the carbon fiber prepreg are calculated simultaneously. The calculation results are output as quality control data. The radiation efficiency of the continuous monochromatic light generator is attenuated by 15%, and the spatial dispersion is extremely poor. Under interference conditions of 18.5%, the absolute error of component testing remained within 0.4%, and the surface of the carbon fiber prepreg did not come into contact with the rigid limiting mechanism during the test.

[0070] During continuous production line quality monitoring, the signal analysis unit generates resin component anomaly index and material evaluation index based on the output quality control data. After obtaining the resin mass percentage content, it compares it with the preset standard range of 30% to 38% material composition. When the resin mass percentage content exceeds this range, the absolute difference between the measured value and the center value of the standard range is taken as the dimensionless resin component anomaly index. Subsequently, the resin component anomaly index is multiplied by a dynamic feedback adjustment coefficient of 0.1, and the resulting adjustment amount is superimposed on the input zero-point voltage bias of the signal analysis loop. The analysis baseline is then dynamically calibrated at zero point during the subsequent material feeding idle gap.

[0071] Simultaneously, the signal analysis unit continuously records the resin mass percentage content of each acquisition area with a spatial sampling step size of 1m, constructs a trend sequence of resin mass percentage content changing with winding length, calculates the first derivative of this trend sequence through the first-order finite difference between adjacent sampling points, and obtains the composition spatial evolution rate. When the composition spatial evolution rate continuously exceeds the preset fluctuation threshold of 0.5% / m, and the corresponding continuous cumulative length reaches more than 5m, it outputs a material evaluation index with a value of 1 to the control terminal, and marks the location of non-uniform defects along the winding inside the carbon fiber prepreg roll and provides audible and visual alarms.

[0072] Example 4: This example combines Figures 1 to 2 A method for testing the resin composition of a carbon fiber prepreg is described, such as... Figure 1 As shown, multiple characteristic wavelengths and reference wavelengths of monochromatic light are alternately projected onto the surface of the carbon fiber prepreg in a continuous transmission state. Diffuse reflection light is captured and the corresponding characteristic light intensity and reference light intensity are output. Real-time temperature is collected simultaneously, and discrete diffuse reflection intensity under reference wavelength light is collected through multiple detection channels as step S1. The transient degradation rate of light intensity is calculated based on the temporal fluctuation of reference light intensity. When the real-time temperature value deviates from the reference temperature and the absolute value of the deviation reaches the preset temperature deviation threshold, a wavelength fine-tuning command is generated. The center emission wavelength of multiple characteristic wavelengths is adjusted according to the preset compensation step size as step S2. The spatial dispersion range is calculated based on the discrete diffuse reflection intensity. When the spatial dispersion range exceeds the stability threshold, a weighted correction operator for uneven fiber distribution is determined as step S3. The transient degradation rate of light intensity, the adjusted light intensity of each characteristic wavelength, and the weighted correction operator are introduced into the signal analysis loop to correct the differential absorbance of each component. The resin content and curing agent ratio of the current collection area of ​​the carbon fiber prepreg are decoupled and calculated as step S4.

[0073] like Figure 2 As shown, the projection light source module, optical wavelength adjuster, and infrared temperature measurement channel are all located in the transmitter. The photodetector contains a scattering detection probe, and the signal analysis unit contains a signal analysis loop, an absorbance calculation loop, and a resident memory chip. The real-time temperature value output by the infrared temperature measurement channel, the diffuse reflection light intensity output by the photodetector, and the continuous transmission speed output by the speed tracking unit are all input to the signal analysis loop. The absorbance calculation loop outputs material evaluation indicators to the control terminal.

[0074] Example 5: Before determining the resin composition, five groups of standard carbon fiber prepreg samples prepared under the same curing conditions were selected. Each group of standard samples was placed in turn at the center of the diverging optical axis of the projection light source module. In a constant temperature chamber at 25°C, the diffuse reflection intensity corresponding to the reference wavelength, the first characteristic wavelength, and the second characteristic wavelength was collected by the photodetector. After completing the spectral measurement, each group of standard samples was graded and extracted to separate the epoxy resin matrix and the curing agent components and weigh them to determine the resin mass percentage content and curing agent molar ratio of each group of samples.

[0075] The signal analysis unit converts the diffuse reflectance intensity of each set of standard samples into first differential absorbance values ​​according to the logarithmic difference relationship between the reference light intensity and the two characteristic light intensities. Second differential absorbance value Then, the differential absorbance data, resin mass percentage content and curing agent molar ratio of the five standard samples were combined to form a multivariate linear first-order overdetermined equation system. The least squares multivariate linear regression algorithm was used to solve the equation system to obtain the component correlation matrix corresponding to the first characteristic wavelength of 1430nm and the second characteristic wavelength of 1510nm.

[0076] The component correlation matrix is ​​a 2x2 full-rank real matrix. The elements in the first row and first column of the matrix represent the intrinsic absorption contribution coefficient of the epoxy resin matrix to the first characteristic wavelength of 1430nm, and the elements in the first row and second column represent the cross-absorption perturbation contribution coefficient of the curing agent to this wavelength. The elements in the second row and first column represent the cross-interference contribution coefficient of the epoxy resin matrix to the second characteristic wavelength of 1510nm, and the elements in the second row and second column represent the intrinsic absorption contribution coefficient of the curing agent to this wavelength. After solving the multivariate linear first-order overdetermined equation system, the component correlation matrix is ​​subjected to matrix inverse operation under full-rank matrix constraints to obtain the inverse matrix coefficients composed of 4 independent decoupling coefficients.

[0077] During real-time detection, the signal analysis unit will use the first differential absorbance value. Second differential absorbance value The components are combined into a one-dimensional row vector and multiplied with the inverse of the component correlation matrix to eliminate the overlapping and cross-interference of absorption bands between the epoxy resin matrix and the curing agent. The resin mass percentage content and the molar ratio of the curing agent are calculated simultaneously. The coefficients of the inverse of the component correlation matrix are stored in the resident memory chip of the signal analysis unit for use during continuous detection.

[0078] Before starting the drive of the continuous coating production line, the carbon fiber prepreg is replaced with a high-reflectivity ceramic target plate. The photodetector collects the air light intensity values ​​of the high-reflectivity ceramic target plate at the first characteristic wavelength, the second characteristic wavelength and the reference wavelength at an alternating sampling interval of 50μs. The geometric average value of the air light intensity value of each wavelength is calculated over 500 consecutive sampling cycles and used as the static reference baseline value of the corresponding channel under the current working condition.

[0079] After the continuous production line starts up and the carbon fiber prepreg enters the detection area, the signal analysis unit retrieves the static reference baseline values ​​of each channel and the real-time acquired reference light intensity values. First characteristic light intensity correction value Second characteristic light intensity correction value The ratios of the two characteristic channels are calculated separately with the corresponding static reference baseline values. Then, using the reference channel as a reference, a differential logarithmic transformation is performed on the ratios of the two characteristic channels to obtain the corresponding first-order partial differential increments, which are then used to correct the first differential absorbance value. Second differential absorbance value The two corrected differential absorbance values ​​are input into the component correlation matrix, and the resin mass percentage content and curing agent molar ratio are calculated through matrix inverse operation. Under multiple batch material switching conditions, the absolute error of component testing is kept within 0.4%.

[0080] Example 6: When materials are changed, the areal density of the unidirectional continuous carbon fiber fabric changes abruptly from 100 g / m² to 200 g / m². The optical transmission conditions and basic scattering cross section of the carbon fiber prepreg surface change anisotropically, causing an amplitude shift in the reference light intensity sequence collected by the photodetector. This alters the fixed component correlation matrix input baseline, leading to decoupling zero-point drift. Without correction, the first differential absorbance value... Second differential absorbance value The deviation from the original working baseline resulted in a system measurement deviation exceeding the allowable error limit of 0.4% in the output resin mass percentage content and curing agent molar ratio.

[0081] During the idle interval of batch feeding, the projection light source module projects monochromatic light of 1300nm reference wavelength onto the surface of pure carbon fiber dry yarn. The photodetector collects the diffuse reflection light intensity on the surface to obtain the reference light intensity value of the current batch. The signal analysis unit reads the reference light intensity value and calculates its ratio with the static reference baseline value measured by the high-reflectivity ceramic target plate to obtain the batch drift coefficient.

[0082] The signal analysis unit compares the batch drift coefficient with the standard value 1. When the absolute deviation between the two is greater than or equal to 0.03, it determines that the fabric surface density change has altered the basic optical attenuation. The reciprocal of the batch drift coefficient is then used as a correction multiplier to sample the first characteristic light intensity in real time. Second characteristic light intensity real-time sampling value Scaling is performed to correct the characteristic light intensity entering the signal analysis loop.

[0083] After scaling, the signal analysis unit calculates the reconstructed first difference absorbance value based on the logarithmic difference relationship between the reference light intensity and the two characteristic light intensities. Second differential absorbance value Then, the two are input into the component correlation matrix, and the matrix inverse operation under the full-rank matrix constraint is used to simultaneously calculate the resin mass percentage content and curing agent molar ratio of the current sampling area of ​​the carbon fiber prepreg. Under the condition of large-scale changes in fabric surface density, the absolute error of component determination is controlled within 0.4%. The detection process adopts non-contact spectral acquisition, and the surface of the carbon fiber prepreg is not subjected to friction from rigid mechanisms.

[0084] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for testing the resin composition of carbon fiber prepreg, characterized in that, Includes the following steps: Step S1: Multiple monochromatic lights of characteristic wavelengths and reference wavelengths are alternately projected onto the surface of the carbon fiber prepreg in continuous transmission state, diffuse reflection light is captured and the corresponding characteristic light intensity and reference light intensity are output, real-time temperature is collected simultaneously, and discrete diffuse reflection intensity under reference wavelength light is collected through multiple detection channels. Step S2: Calculate the transient degradation rate of light intensity based on the temporal fluctuations of the reference light intensity; When the real-time temperature value deviates from the reference temperature and the absolute value of the deviation reaches the preset temperature deviation threshold, a wavelength fine-tuning command is generated to adjust the center emission wavelength of multiple characteristic wavelengths according to the preset compensation step size. Step S3: Calculate the spatial dispersion range based on the discrete diffuse reflection intensity. When the spatial dispersion range exceeds the stability threshold, determine the weighted correction operator for uneven fiber distribution. Step S4: The transient degradation rate of light intensity, the light intensity corresponding to each characteristic wavelength after adjustment, and the weighted correction operator are introduced into the signal analysis loop to correct the differential absorbance of each component and decouple the resin content and curing agent ratio of the current collection area of ​​the carbon fiber prepreg.

2. The method for testing the resin composition of carbon fiber prepreg according to claim 1, characterized in that, In step S1, the surface temperature of the carbon fiber prepreg is captured in real time by adding an infrared temperature measurement channel on the same side of the transmitter, which is used as the real-time temperature value. Multiple characteristic wavelengths include a first characteristic wavelength for responding to the resin and a second characteristic wavelength for responding to the curing agent. Each characteristic light intensity includes a first characteristic light intensity corresponding to the first characteristic wavelength and a second characteristic light intensity corresponding to the second characteristic wavelength. In step S2, the reference temperature is 25°C. The preset temperature deviation threshold is 0.

5. The preset compensation step size is 0.2nm / The optical wavelength modulator inside the transmitter receives the wavelength fine-tuning command and correspondingly corrects the emission driving current parameters of the first characteristic wavelength and the second characteristic wavelength to lock the adjusted center emission wavelength at the intrinsic chemical absorption peak of the molecular structure of the carbon fiber prepreg. In step S4, the differential absorbance of each component includes the first differential absorbance value corresponding to the resin and the second differential absorbance value corresponding to the curing agent. The resin content and curing agent ratio are the resin mass percentage content and the curing agent molar ratio, respectively.

3. The method for testing the resin composition of carbon fiber prepreg according to claim 1, characterized in that, In step S1, the discrete diffuse reflection intensity under the reference wavelength light is simultaneously captured by scattering probes deployed at three discrete geometric spatial positions at 30°, 45° and 60° to the incident optical axis, and used as the discrete diffuse reflection intensity; in step S3, the calculation of the spatial dispersion range and the determination of the weighted correction operator based on the discrete diffuse reflection intensity includes the following sub-steps: Step S31, calculate the difference between the maximum and minimum values ​​of the discrete diffuse reflection intensity at the three captured discrete geometric spatial positions to obtain the spatial dispersion range; Step S32: When the spatial dispersion range exceeds the stability threshold, the spatial dispersion range is used as a feature input of the coefficient matrix to calculate a weighted correction operator to correct the differential absorbance of each component.

4. The method for testing the resin composition of carbon fiber prepreg according to claim 2, characterized in that, The method also includes a step performed before step S1: step S401, controlling the projection light source module to alternately emit a first characteristic wavelength, a second characteristic wavelength and a reference wavelength, and performing pulse code modulation on the emitted light intensity to form monochromatic light with coding characteristics, which is input to the surface of the carbon fiber prepreg in step S1.

5. The method for testing the resin composition of carbon fiber prepreg according to claim 1, characterized in that, In step S4, before introducing the light intensity into the signal analysis loop, the following sub-steps are included: Step S41, acquiring the transient power fluctuation data of the transmitter, quantitatively correcting the transient degradation rate of the light intensity by combining the temporal fluctuation of the reference light intensity, and generating a feedforward balancing gain coefficient, which is applied to the captured diffuse reflection light intensity to correct the transient attenuation of the light intensity caused by the continuous operation of the transmitter.

6. The method for testing the resin composition of carbon fiber prepreg according to claim 2, characterized in that, In step S4, the decoupling calculation of the resin mass percentage content and curing agent molar ratio of the current sampling area of ​​the carbon fiber prepreg includes the following sub-steps: Step S42, the corrected first difference absorbance value and second difference absorbance value are input into the preset component correlation matrix, and the resin mass percentage content and curing agent molar ratio are calculated simultaneously by performing matrix inverse operation under full-rank square matrix constraint on the component correlation matrix.

7. The method for testing the resin composition of carbon fiber prepreg according to claim 2, characterized in that, The method also includes a step after step S4: Step S701, after calculating the resin mass percentage content and the curing agent molar ratio, compare the resin mass percentage content with the preset material composition standard range. When the resin mass percentage content exceeds the preset material composition standard range, output the resin composition anomaly index to perform zero-point dynamic calibration of the analysis baseline in the signal analysis loop.

8. The method for testing the resin composition of carbon fiber prepreg according to claim 2, characterized in that, In step S4, the decoupling calculation of the differential absorbance of each component includes the following sub-steps: Step S43, read the first characteristic light intensity, the second characteristic light intensity, and the reference light intensity, perform component differential decoupling calculation, and obtain the first differential absorbance value of the epoxy resin matrix. The second difference absorbance value with the curing agent : ,in, For the test component number, in hour, The first difference absorbance value and The first characteristic light intensity after correction, in hour, The second differential absorbance value and The corrected second characteristic light intensity. This is a real-time sampled value of the reference light intensity.

9. The method for testing the resin composition of carbon fiber prepreg according to claim 2, characterized in that, The method also includes a step of evaluating the uniformity of materials produced continuously: Step S901, continuously record the resin mass percentage content of each collection area, construct a trend sequence of the resin mass percentage content as the winding length evolves, calculate the first derivative of the trend sequence to obtain the component space evolution rate, and when the component space evolution rate continues to exceed the preset fluctuation threshold, output a material evaluation index characterizing the non-uniformity defects along the winding process inside the carbon fiber prepreg roll.

10. The method for testing the resin composition of carbon fiber prepreg according to claim 1, characterized in that, The method leverages the accuracy of velocity tracking unit matching analysis and includes the following steps: Step S1001: Collect the continuous transmission speed of the carbon fiber prepreg. Based on the change in the continuous transmission speed, adjust the transmission frequency of the transmitter proportionally. When the continuous transmission speed is in the range of 15m / min to 30m / min, make the transmission frequency dynamically stable in the frequency domain range corresponding to the current transmission speed, so as to keep the spatial sampling step size of the current collection area constant.

Images