Method and system for generating multi-component dynamic impedance fingerprint spectrum of fermented grains

Abstract

The invention relates to the field of fermentation monitoring, and provides a fermented grain multi-component dynamic impedance fingerprint spectrum generation method and system in order to improve the accuracy of impedance detection. A temperature gradient program is set, dual-channel impedance spectrum synchronous acquisition is carried out, and component relaxation parameters are extracted based on a relaxation model; and a temperature-relaxation correlation model and a parameter database are constructed in combination with an Arrhenius equation. On the basis, correction and decoupling of the main body impedance spectrum are executed, and finally the three-dimensional dynamic impedance fingerprint spectrum is generated according to real-time weight and standard fingerprint reconstruction. Preliminary physical separation of components is realized through temperature gradient and a specific adsorbent, and multi-component aliasing is reduced from the source; in cooperation with a decoupling algorithm based on concentration constraint, the problem that relaxation peaks of multiple components in a traditional impedance spectrum are overlapped and difficult to distinguish is effectively solved, and therefore the accuracy of impedance detection is improved.

Description

Technical Field

[0001] This invention relates to the field of fermentation monitoring, specifically a method and system for generating multi-component dynamic impedance fingerprints of fermented mash. Background Technology

[0002] Solid-state fermentation (such as in the brewing of baijiu and soy sauce) produces various volatile and semi-volatile components, including water, ethanol, organic acids, and esters. The concentration, proportion, and variation patterns of these components directly affect the fermentation progress and the quality of the final product. Existing impedance spectroscopy techniques suffer from severe aliasing of relaxation peaks among multiple components: the dielectric relaxation response frequency ranges of components such as water, ethanol, acids, and esters in the solid-state fermentation system overlap, making it difficult to accurately decompose the peak values ​​of each component in the impedance spectrum and hindering component separation and identification. Summary of the Invention

[0003] To improve the accuracy of impedance detection, this invention provides a method and system for generating dynamic impedance fingerprint spectra of multi-component fermented mash.

[0004] The technical solution adopted by the present invention to solve the above problems is:

[0005] A method for generating multi-component dynamic impedance fingerprints of fermented mash includes:

[0006] Step 1: Set the temperature gradient program according to the volatilization temperature of the mash components;

[0007] Step 2: Perform simultaneous acquisition of dual-channel impedance spectra based on a temperature gradient program.

[0008] Channel A: During the temperature change process, the impedance of the main mash is continuously measured. ;

[0009] Channel B: Based on the volatilization temperature of each component, each component is introduced into a different adsorption point. When the adsorption point reaches adsorption equilibrium, the impedance of component i is measured at the current temperature. ;

[0010] Step 3, based on impedance The relaxation parameters of each component are extracted using the relaxation model, including the relaxation time constant, relaxation intensity, and relaxation time distribution coefficient.

[0011] Step 4: Establish a temperature-relaxation correlation model based on the relaxation time constant and the Arrhenius equation;

[0012] Step 5: Conduct temperature change experiments based on the temperature-relaxation correlation model to obtain the activation energy of each component;

[0013] Step 6: Correction and Decoupling of the Main Impedance Spectrum: Obtain the activation energy of each component based on the current real-time temperature, calculate the relaxation time constant at that temperature using the Arrhenius equation, and substitute it into the relaxation model to generate the standard impedance fingerprint of the component corresponding to the current temperature. ;

[0014] Obtain temperature sensitivity correction factor Based on temperature sensitivity correction factor The drift correction of the main impedance is obtained. ;

[0015] Construct the following objective function to solve for the real-time weights of each component in the principal component. :

[0016] ,

[0017] in: This is the scaling factor. For concentration trends, F represents the Frobenius norm;

[0018] Step 7: Based on the real-time weights of each component and its corresponding standard impedance fingerprint The reconstructed three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated.

[0019] Furthermore, the components include water, ethanol, and ethyl acetate, and the temperature gradient program is set as follows: 30–50℃, rate 0.3℃ / min; 50–70℃, rate 0.5℃ / min; 70–85℃, rate 0.5℃ / min.

[0020] Furthermore, the adsorption sites corresponding to water are modified. The adsorption sites for ethanol were determined using ZIF-8@graphene; the adsorption sites for ethyl acetate were determined using MOF adsorbents.

[0021] Furthermore, the adsorption equilibrium determination method is as follows: the rate of change of the impedance modulus at the adsorption point within a continuous time is less than a preset value.

[0022] Furthermore, step 3 uses the least squares method to convert the measured impedance... The relaxation parameters are obtained by fitting the model with the relaxation parameters.

[0023] Furthermore, the relaxation model is either the Cole-Cole model or the Havriliak–Negami model.

[0024] Furthermore, the temperature sensitivity correction factor The calculation formula is: ,in, Let be the relaxation strength of the i-th component at the adsorption site. The total relaxation intensity of the main fermented mash, This is the partial derivative of the impedance with respect to temperature.

[0025] Furthermore, based on the temperature sensitivity correction factor The formula for correcting the drift of the main impedance is: , where n is the total number of components.

[0026] Furthermore, the concentration trend The calculation formula is: , where k is the adsorption rate constant.

[0027] A system for generating dynamic impedance fingerprint spectra of multiple components of fermented mash, used to implement a method for generating dynamic impedance fingerprint spectra of multiple components of fermented mash, including:

[0028] The temperature setting unit is used to set the temperature gradient program according to the volatilization temperature of the mash components;

[0029] The impedance acquisition unit is used for synchronous acquisition of dual-channel impedance spectra based on a temperature gradient program, including the impedance of the main mash. and the impedance of component i ;

[0030] Relaxation parameter extraction unit, based on impedance The relaxation parameters of each component are extracted using the relaxation model, including the relaxation time constant, relaxation intensity, and relaxation time distribution coefficient.

[0031] The model building unit establishes a temperature-relaxation correlation model based on the relaxation time constant and the Arrhenius equation.

[0032] The activation energy acquisition unit uses a temperature-relaxation correlation model to conduct temperature change experiments to obtain the activation energy of each component.

[0033] The correction and decoupling unit for the main impedance spectrum obtains the activation energy of each component based on the current real-time temperature, calculates the relaxation time constant at that temperature using the Arrhenius equation, and substitutes it into the relaxation model to generate the standard impedance fingerprint of the component at the current temperature. ;

[0034] Obtain temperature sensitivity correction factor Based on temperature sensitivity correction factor The drift correction of the main impedance is obtained. The objective function is constructed as follows to solve for the real-time weights of each component in the principal component. :

[0035] ,

[0036] in, This is the scaling factor. For concentration trends, F represents the Frobenius norm;

[0037] The three-dimensional dynamic impedance fingerprint generation unit generates the spectrum based on the real-time weights of each component. and its corresponding standard impedance fingerprint The reconstructed three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated.

[0038] The advantages of this invention compared to the prior art are:

[0039] It achieves non-destructive, in-situ, and real-time monitoring of the internal components of the mash: by combining temperature-controlled induction with impedance measurement, it eliminates the need for destructive sampling and can continuously output impedance information during fermentation, overcoming the shortcomings of traditional chromatographic methods such as detection lag and cumbersome operation.

[0040] The dual mechanism of "physical separation + algorithm decoupling" significantly improves resolution: preliminary physical separation of components is achieved through temperature gradients and specific adsorbents, reducing multi-component mixing at the source; combined with concentration-constrained... The decoupling algorithm effectively solves the problem of overlapping and difficult-to-distinguish relaxation peaks of multiple components in traditional impedance spectra.

[0041] This invention effectively eliminates temperature drift interference by correcting for adsorption points: The adsorption point is introduced as a "reference channel," and a calculated correction factor is used... The main signal is subjected to baseline correction. This not only eliminates baseline drift caused by changes in ambient temperature, but also eliminates non-specific background noise through weighted correction, significantly improving the signal-to-noise ratio of the measurement data and the repeatability of the results.

[0042] The three-dimensional dynamic fingerprint spectrum has strong characterization capabilities: the generated frequency-time-temperature three-dimensional spectrum (tensor) contains rich fermentation kinetic information, which can more comprehensively reflect the migration law, reaction rate and structural changes of each component during fermentation, and provide high-value visualization data for the quality control of solid-state fermentation systems such as baijiu. Attached Figure Description

[0043] Figure 1 Flowchart of a method for generating dynamic impedance fingerprints of multiple components in fermented mash;

[0044] Figure 2 This is a schematic diagram of the impedance acquisition unit structure;

[0045] Figure 3 This is a schematic diagram of a three-dimensional dynamic impedance fingerprint spectrum;

[0046] Figure labels: 1 is nitrogen source, 2 is mass flow controller, 3 is temperature control reaction unit, 4 is multi-way switching valve, 5 is adsorption column, and 6 is impedance analyzer and data processing unit. Detailed Implementation

[0047] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0048] The method provided by this invention first sets a temperature gradient program based on the volatilization temperature of the mash components. Next, dual-channel impedance spectroscopy synchronous acquisition was performed: channel A continuously measured the mixing impedance of the main mash during the temperature change process. Channel B utilizes a temperature gradient to introduce components such as water, ethanol, and esters into the modification process according to their volatilization temperature ranges. Specific adsorption sites such as ZIF-8@graphene and MOFs were used to measure the quasi-pure state impedance of each component after adsorption equilibrium. Subsequently, component relaxation parameters were extracted based on the Cole-Cole relaxation model, and a temperature-relaxation correlation model and parameter database were constructed using the Arrhenius equation. Based on this, correction and decoupling of the bulk impedance spectrum were performed: the temperature sensitivity correction factor was calculated. The drift correction of the main impedance is obtained. Concentration trends retrieved using adsorption channels As a physical constraint, the real-time weights of each component in the host are solved using an optimization algorithm. Finally, a three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated based on real-time weights and standard fingerprint reconstruction. Preliminary physical separation of components is achieved through temperature gradients and specific adsorbents, reducing multi-component mixing at the source; this is combined with concentration-constrained... The decoupling algorithm effectively solves the problem of overlapping and difficult-to-distinguish relaxation peaks of multiple components in traditional impedance spectra, thereby improving the accuracy of impedance detection.

[0049] like Figure 1 As shown, the method for generating multi-component dynamic impedance fingerprints of fermented mash includes:

[0050] Step 1: Set the temperature gradient program according to the volatilization temperature of the mash components;

[0051] Step 2: Perform simultaneous acquisition of dual-channel impedance spectra based on a temperature gradient program.

[0052] Channel A: During the temperature change process, the impedance of the main mash is continuously measured. ;

[0053] Channel B: Based on the volatilization temperature of each component, each component is introduced into a different adsorption point. When the adsorption point reaches adsorption equilibrium, the impedance of component i is measured at the current temperature. ;

[0054] Step 3, based on impedance The relaxation parameters of each component are extracted using the relaxation model, including the relaxation time constant, relaxation intensity, and relaxation time distribution coefficient.

[0055] Step 4: Establish a temperature-relaxation correlation model based on the relaxation time constant and the Arrhenius equation;

[0056] Step 5: Conduct temperature change experiments based on the temperature-relaxation correlation model to obtain the activation energy of each component;

[0057] Step 6: Correction and Decoupling of the Main Impedance Spectrum: Obtain the activation energy of each component based on the current real-time temperature, calculate the relaxation time constant at that temperature using the Arrhenius equation, and substitute it into the relaxation model to generate the standard impedance fingerprint of the component corresponding to the current temperature. ;

[0058] Obtain temperature sensitivity correction factor Based on temperature sensitivity correction factor The drift correction of the main impedance is obtained. The objective function is constructed as follows to solve for the real-time weights of each component in the principal component. :

[0059] ,

[0060] in, This is the scaling factor. For concentration trends, F represents the Frobenius norm;

[0061] Step 7: Based on the real-time weights of each component and its corresponding standard impedance fingerprint The reconstructed three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated.

[0062] This paper uses in-situ non-destructive testing of key components such as water, ethanol, and ethyl acetate in the fermentation mash of baijiu (Chinese liquor) as an example to illustrate the process, specifically including:

[0063] 1. Temperature gradient-induced component volatilization and separation

[0064] Experimental sample obtained: solid mash taken from the 12th day of fermentation of distiller's grains, with a moisture content of about 58%, an ethanol mass fraction of about 7.5%, and a total acid content of about 0.45g / 100g.

[0065] By setting a controllable temperature gradient program T(t), physical separation is achieved by utilizing the differences in volatilization kinetics among the components. Based on selective separation due to volatilization kinetic differences, a small amount of cross-release may still occur between different components in adjacent temperature zones. However, through adsorbent selectivity and subsequent relaxation spectrum correction algorithms, the interference of non-target components on the detection results within the set temperature range can be effectively suppressed. By using selective adsorbents, the target component is made to dominate the impedance response at the adsorption point, with its corresponding relaxation peak amplitude accounting for no less than 80%, thereby obtaining the impedance characteristics of the quasi-pure component that can be used for model correction.

[0066] The target components were enriched into different channels: water ( ): Volatilization temperature range 30-50℃, from modified Adsorption; Ethanol: volatile temperature range 50-70℃, adsorbed by ZIF-8@graphene; Ethyl acetate: volatile temperature range 75-85℃, adsorbed by MOF adsorbents, adsorption point properties are shown in the table below:

[0067]

[0068] 2. Simultaneous acquisition of dual-channel impedance spectrum

[0069] The impedance acquisition unit used in this embodiment is as follows: Figure 2 As shown, the system includes a nitrogen source 1, a mass flow controller 2, a temperature-controlled reaction unit 3, a multi-way switching valve 4, an adsorption column 5, and an impedance analyzer and data processing unit 6. The system contains two parallel detection channels, which are used to detect the impedance of the main mash. and the impedance of each component :

[0070] Channel A (Main Fermented Grains): During the temperature change process t, the broadband impedance of the main fermented grains is continuously measured. Under engineering conditions of small-signal excitation, weak coupling, and distinct characteristic relaxation frequency bands of each component, the complex impedance of the main fermented grains can be approximately expressed as a weighted superposition of the relaxation responses of multiple components. Since the main state changes simultaneously with time and temperature, its impedance is expressed as a ternary function:

[0071] ,

[0072] in: : Complex impedance modulus of the main fermented mash (unit: ); Angular frequency ,in Test frequency (range 100Hz-10MHz); : Time variable during fermentation or testing process; T: Real-time temperature (unit: K); The real part of the impedance (resistive component, unit: ); The imaginary part of impedance (reactance component, unit: ), where j is the imaginary unit.

[0073] Channel B (Adsorption point-induced component volatilization): After the adsorption point reaches saturation equilibrium, the impedance of component i is measured at the current temperature T. At this point, the component properties are stable, and the impedance is only a bivariate function of frequency and temperature. ,in: These represent water, ethanol, and ethyl acetate, respectively.

[0074] The adsorption equilibrium determination method is as follows: the rate of change of the impedance modulus of the adsorption point within a continuous time is less than a preset value. In this embodiment, it is set to a rate of change of less than 1% within 5 consecutive minutes. Other values ​​can also be set according to the actual situation, and there is no restriction here.

[0075] 3. Relaxation parameter extraction based on relaxation model

[0076] In order to obtain the measured component impedance spectrum To extract physical features, this invention uses the Cole-Cole dielectric relaxation model as the theoretical fitting formula, or the Havriliak–Negami model. The Cole-Cole fitting model formula is as follows:

[0077] ,

[0078] in, High-frequency limiting resistance (unit: ), that is, when The real part of the impedance at that time;

[0079] Relaxation strength (unit: ), defined as ,in This is the low-frequency limiting resistor;

[0080] Characteristic relaxation time constant at temperature T (unit: s);

[0081] Relaxation time distribution coefficient (or broadening coefficient), range of values Characterizes the non-uniformity of the system.

[0082] Parameter extraction steps: Use the least squares method to extract the measured data. Compared with the above model By performing a fitting process, the key characteristic parameters of each component i at temperature T can be calculated: , , .

[0083] 4. Construct a temperature-relaxation correlation model

[0084] Based on extraction The functional relationship between temperature T and the Arrhenius equation can be established: ,in: : Pre-exponential factor of the i-th component (theoretical limit relaxation time, unit: s); : Dielectric relaxation activation energy of the i-th component (unit: eV or J); Boltzmann constant ( ).

[0085] 5. Temperature variation experiments were conducted based on a temperature-relaxation correlation model to obtain the activation energies of each component.

[0086] The activation energies of each component can be obtained by fitting a temperature-relaxation correlation model through variable-temperature experiments. , of which ethanol: ;Moisture: Because water molecules are being modified The surface forms a confined adsorption state, and the energy barrier for hydrogen bond network rearrangement is lowered, thus exhibiting a lower characteristic activation energy than free water; Ethyl acetate: Lipids have relatively weak polarity, and their dielectric relaxation behavior is mainly dominated by dipole rotation and changes in intermolecular forces.

[0087] To facilitate subsequent data retrieval, this embodiment constructs a component parameter database based on currently known data and preset parameters. This database specifically stores the following three types of key feature information:

[0088] 1. Thermodynamic kinetic parameters: including the dielectric relaxation activation energy of each component ( ) and pre-exponential factors ( This is used to predict the characteristic relaxation time of the component at any real-time temperature T using the Arrhenius equation. ;

[0089] 2. Electrochemical morphology parameters: These include the Cole-Cole model eigenvalues ​​of each component under standard conditions, i.e., the relaxation time distribution coefficients. Relaxation strength temperature response coefficient and high-frequency limiting resistance ;

[0090] 3. Systems engineering factors: including pre-calibrated adsorption rate constant k and temperature sensitivity correction factor. This is used for subsequent concentration inversion and signal decoupling calculations.

[0091] 6. Correction and decoupling of the main impedance spectrum

[0092] Since the temperature T changes constantly, a fixed fingerprint cannot be used for matching. Instead, the activation energy parameter from step 5 can be used to calculate the theoretical standard impedance fingerprint at the current temperature in real time. Then, decoupling is performed. Therefore, before performing correction and decoupling, the system first retrieves the activation energies of each component from the "component parameter database" constructed in step 5, based on the current real-time temperature T. and pre-exponential factors Using the Arrhenius equation The characteristic relaxation time at this temperature is calculated and substituted into the Cole-Cole model to generate the component standard impedance fingerprint corresponding to the current temperature. This fingerprint will serve as a baseline template in subsequent optimization algorithms.

[0093] Subsequently, using the quasi-pure component characteristic parameters obtained from channel B (adsorption point), the mixing impedance measured in channel A (main tank) was analyzed. A step-by-step processing method is employed to eliminate temperature drift background and achieve quantitative separation of multi-component signals. The processing flow consists of two stages:

[0094] Step 1: Temperature drift correction (denoising) based on sensitivity differences

[0095] Since the main mash and the adsorption point are in the same variable temperature environment but have different physical structures, it is necessary to first eliminate the non-specific system baseline drift caused by temperature changes.

[0096] Call the pre-calibrated temperature sensitivity correction factor from the component parameter database As a correction weighting factor for each component, this factor is used to quantify and eliminate inconsistencies in temperature response caused by physical differences in the detection system (main tank and adsorption points):

[0097] ,

[0098] Among them, temperature sensitivity correction factor This is an engineering correction parameter, a dimensionless weighting factor, whose value is obtained through preliminary experiments with standard samples. Specifically, it involves measuring the sensitivity of the bulk impedance to temperature changes under known target component concentrations and fitting it to the impedance changes of the corresponding components at adsorption points for calibration. This factor characterizes the sensitivity of component i to the contribution of overall impedance to temperature changes. Characteristic relaxation strength of the i-th component at the adsorption point (unit: ), defined as That is, the difference between the low-frequency limiting resistance and the high-frequency limiting resistance measured at the adsorption point at this temperature, which characterizes the polarization response amplitude capability of the component in the "quasi-pure state". Total relaxation intensity of the main fermented mash (unit: ), defined as , which is the difference between the low-frequency limiting resistance and the high-frequency limiting resistance of the main mash at the corresponding temperature, and it characterizes the background of the polarization response amplitude of the entire mixing system. This ratio, as a dimensionless normalization coefficient, is used to "map" the weak pure signal measured at the adsorption point to the order of magnitude of the bulk mixed signal, thereby correctly assessing the relative weight of the component in the bulk impedance change. The partial derivative of impedance with respect to temperature reflects the thermal drift rate at that temperature.

[0099] Using temperature sensitivity correction factor Perform reference correction on the main impedance:

[0100] ,

[0101] in: : During characteristic relaxation time The impedance value of the i-th component obtained by actual measurement at the adsorption point at the corresponding frequency point; The main impedance at the characteristic relaxation time The impedance value at the corresponding frequency point, where n is the total number of components.

[0102] At this time, it was obtained It is still a total impedance signal of a multi-component mixture, and the independent separation of each component has not yet been achieved, but the baseline shift caused by temperature has been calibrated.

[0103] Step 2: Quantitative decoupling of multi-components based on physical constraints (component separation):

[0104] Obtain the corrected pure total impedance Subsequently, a multi-component linear superposition model was established, mathematically decomposing the mixed signal into the dynamic contributions of water, ethanol, and esters.

[0105] 1. Constructing physical concentration constraints: utilizing concentration trends provided by adsorption channels. Construct a constrained optimization problem.

[0106] Before constructing the constrained optimization problem, it is necessary to clarify The physical meaning and calculation path of [the gas concentration]. In this invention, the system does not rely on expensive gas concentration sensors for direct measurement. Instead, it is based on an inversion estimation using the principles of adsorption kinetics. This is because the component impedance modulus at adsorption point i... It is positively correlated with the mass of adsorbate accumulated on its surface, that is... Therefore, the relative concentration trend of the i-th component at time t It can be obtained by calculating the first time derivative (i.e., adsorption rate) of the measured impedance data at the adsorption point: Wherein, k is the adsorption rate constant related to the carrier gas flow rate and adsorbent sensitivity, retrieved from the component parameter database described in step 5. Its value is calibrated through adsorption experiments of standard components of known mass. The calibration process involves: introducing a series of standard component samples of known mass into the adsorption channel, recording the total change in impedance modulus corresponding to adsorption saturation of each sample (i.e., the integral of the impedance change rate curve over time), constructing a linear fitting curve between the adsorbed mass and the impedance response amplitude, and determining the slope of this fitting curve as the adsorption rate constant k. The curve reflects the peak shape (such as peak time and peak width) of each component during the heating process, and is a weighted average. The optimized solution provides physical constraint boundaries, thereby ensuring that the decomposition results conform to the actual volatilization law.

[0107] 2. With the goal of minimizing error, solve for the real-time weights of each component in the subject to ensure the uniqueness and accuracy of the decomposition results:

[0108] ,

[0109] in: : The component weights to be solved, representing the real-time concentration contribution of the i-th component in the host; st(subjectto): Represents the constraint conditions; F represents the Frobenius norm; : Indicates weight The range of variation is limited by the concentration trend measured by the adsorption channel. That is, when the adsorption channel does not detect the volatilization of component i. The algorithm is forced to be constrained to This eliminates multiple solutions and spurious signal interference in mathematical fitting; : Scaling factor, a system tolerance adjustment parameter, whose value range is limited to 0.9–1.1, used to compensate for errors introduced by adsorption kinetic hysteresis and environmental disturbances.

[0110] 7. Dynamic fingerprint generation and reconstruction

[0111] Based on the real-time weights of each component obtained in the previous step Combined with component standard impedance fingerprint The reconstructed three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated. The calculation formula is as follows: Where: frequency axis ω: 20Hz–10MHz (101 logarithmic intervals), time axis t: the entire temperature control process (0–120min), temperature axis T: 30–85℃ (strictly corresponding to t).

[0112] Define the mathematical representation of the spectrum and construct an m×n dimensional impedance magnitude matrix (frequency point m, time point n), where the matrix elements The reconstructed impedance value:

[0113] ,in: This spectrum visually reflects the actual evolution trajectory of the superposition of each component after mathematical decoupling. After removing background interference, a schematic diagram of the three-dimensional dynamic impedance fingerprint spectrum is shown below. Figure 3 As shown.

[0114] Correspondingly, the present invention also provides a system for generating dynamic impedance fingerprint spectra of multi-component fermented mash, used to realize a method for generating dynamic impedance fingerprint spectra of multi-component fermented mash, including:

[0115] The temperature setting unit is used to set the temperature gradient program according to the volatilization temperature of the mash components;

[0116] The impedance acquisition unit is used for synchronous acquisition of dual-channel impedance spectra based on a temperature gradient program, including the impedance of the main mash. and the impedance of component i ;

[0117] Relaxation parameter extraction unit, based on impedance The relaxation parameters of each component are extracted using the relaxation model, including the relaxation time constant, relaxation intensity, and relaxation time distribution coefficient.

[0118] The model building unit establishes a temperature-relaxation correlation model based on the relaxation time constant and the Arrhenius equation.

[0119] The activation energy acquisition unit uses a temperature-relaxation correlation model to conduct temperature change experiments to obtain the activation energy of each component.

[0120] The correction and decoupling unit for the main impedance spectrum obtains the activation energy of each component based on the current real-time temperature, calculates the relaxation time constant at that temperature using the Arrhenius equation, and substitutes it into the relaxation model to generate the standard impedance fingerprint of the component at the current temperature. ;

[0121] Obtain temperature sensitivity correction factor Based on temperature sensitivity correction factor The drift correction of the main impedance is obtained. The objective function is constructed as follows to solve for the real-time weights of each component in the principal component. :

[0122] ,

[0123] in, This is the scaling factor. For concentration trends, F represents the Frobenius norm;

[0124] The three-dimensional dynamic impedance fingerprint generation unit generates the spectrum based on the real-time weights of each component. and its corresponding standard impedance fingerprint The reconstructed three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated.

Claims

1. A method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash, characterized in that, include: Step 1: Set the temperature gradient program according to the volatilization temperature of the mash components; Step 2: Perform simultaneous acquisition of dual-channel impedance spectra based on a temperature gradient program. Channel A: During the temperature change process, the impedance of the main mash is continuously measured. ; Channel B: Based on the volatilization temperature of each component, each component is introduced into a different adsorption point. When the adsorption point reaches adsorption equilibrium, the impedance of component i is measured at the current temperature. ; Step 3, based on impedance The relaxation parameters of each component are extracted using the relaxation model, including the relaxation time constant, relaxation intensity, and relaxation time distribution coefficient. Step 4: Establish a temperature-relaxation correlation model based on the relaxation time constant and the Arrhenius equation; Step 5: Conduct temperature change experiments based on the temperature-relaxation correlation model to obtain the activation energy of each component; Step 6: Correction and Decoupling of the Main Impedance Spectrum: Obtain the activation energy of each component based on the current real-time temperature, calculate the relaxation time constant at that temperature using the Arrhenius equation, and substitute it into the relaxation model to generate the standard impedance fingerprint of the component corresponding to the current temperature. ; Obtain temperature sensitivity correction factor Based on temperature sensitivity correction factor The drift correction of the main impedance is obtained. ; Construct the following objective function to solve for the real-time weights of each component in the principal component. : ， in: This is the scaling factor. For concentration trends, F represents the Frobenius norm; Step 7: Based on the real-time weights of each component and its corresponding standard impedance fingerprint The reconstructed three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated.

2. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to claim 1, characterized in that, The components include water, ethanol, and ethyl acetate. The temperature gradient program is set as follows: 30–50℃, rate 0.3℃ / min; 50–70℃, rate 0.5℃ / min; 70–85℃, rate 0.5℃ / min.

3. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to claim 2, characterized in that, The adsorption points of water are modified. The adsorption sites for ethanol were determined using ZIF-8@graphene; the adsorption sites for ethyl acetate were determined using MOF adsorbents.

4. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to claim 1, characterized in that, The adsorption equilibrium is determined by the following method: the rate of change of the impedance modulus at the adsorption point over a continuous period of time is less than a preset value.

5. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to claim 1, characterized in that, Step 3: Use the least squares method to convert the measured impedance The relaxation parameters are obtained by fitting the model with the relaxation parameters.

6. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to claim 1, characterized in that, The relaxation model is either the Cole-Cole model or the Havriliak–Negami model.

7. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to claim 1, characterized in that, The temperature sensitivity correction factor The calculation formula is: ,in, Let be the relaxation strength of the i-th component at the adsorption site. The total relaxation intensity of the main fermented mash, This is the partial derivative of the impedance with respect to temperature.

8. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to claim 1, characterized in that, Based on temperature sensitivity correction factor The formula for correcting the drift of the main impedance is: , where n is the total number of components.

9. The method for generating a multi-component dynamic impedance fingerprint spectrum of fermented mash according to any one of claims 1-8, characterized in that, The concentration trend The calculation formula is: , where k is the adsorption rate constant.

10. A system for generating dynamic impedance fingerprint spectra of multi-component fermented mash, used to implement the method for generating dynamic impedance fingerprint spectra of multi-component fermented mash as described in any one of claims 1-9, characterized in that, include: The temperature setting unit is used to set the temperature gradient program according to the volatilization temperature of the mash components; The impedance acquisition unit is used for synchronous acquisition of dual-channel impedance spectra based on a temperature gradient program, including the impedance of the main mash. and the impedance of component i ; Relaxation parameter extraction unit, based on impedance The relaxation parameters of each component are extracted using the relaxation model, including the relaxation time constant, relaxation intensity, and relaxation time distribution coefficient. The model building unit establishes a temperature-relaxation correlation model based on the relaxation time constant and the Arrhenius equation. The activation energy acquisition unit uses a temperature-relaxation correlation model to conduct temperature change experiments to obtain the activation energy of each component. The correction and decoupling unit for the main impedance spectrum obtains the activation energy of each component based on the current real-time temperature, calculates the relaxation time constant at that temperature using the Arrhenius equation, and substitutes it into the relaxation model to generate the standard impedance fingerprint of the component corresponding to the current temperature. ; Obtain temperature sensitivity correction factor Based on temperature sensitivity correction factor The drift correction of the main impedance is obtained. The objective function is constructed as follows to solve for the real-time weights of each component in the principal component. : ， in, This is the scaling factor. For concentration trends, F represents the Frobenius norm; The three-dimensional dynamic impedance fingerprint generation unit generates the spectrum based on the real-time weights of each component. and its corresponding standard impedance fingerprint The reconstructed three-dimensional dynamic impedance fingerprint spectrum containing frequency-time-temperature dimensions is generated.

Images