A wine body water feeling evaluation method and system based on low-field nuclear magnetic resonance

Abstract

The present application relates to the technical field of liquor flavor evaluation, and discloses a liquor body water sensitivity evaluation method and system based on low-field nuclear magnetic resonance, aiming to solve the problems of strong subjectivity, low efficiency and lack of quantitative indicators in the existing method, and the scheme mainly comprises the following steps: placing a to-be-tested liquor sample in a low-field nuclear magnetic resonance instrument for constant temperature balancing; using a CPMG sequence to collect the transverse relaxation echo train signal of the to-be-tested liquor sample at a preset constant temperature balancing temperature, wherein the echo time TE is 1 ms, the number of echoes NECH is 18000, and the repetition waiting time TW is 15000 ms; performing T2 inversion on the echo train to obtain a T2 relaxation time distribution spectrum; determining whether there is a free water related long T2 component with a peak top position greater than 3000 ms, and if so, calculating the peak area ratio as the free water proportion; inputting the free water proportion into a pre-established water sensitivity evaluation model to obtain a water sensitivity index and compare it with a preset threshold value, and outputting the liquor body water sensitivity grade. The present application realizes objective, quantitative and rapid evaluation of liquor water sensitivity.

Description

Technical Field

[0001] This invention relates to the field of baijiu flavor evaluation technology, specifically to a method and system for evaluating the water content of baijiu based on low-field nuclear magnetic resonance. Background Technology

[0002] The "watery" or "watery" taste in baijiu is a crucial dimension influencing consumer experience, especially in low-alcohol products. An overly strong watery taste is often described as "obvious wateriness and a bland body," while the watery taste in high-alcohol baijiu is relatively weaker. Currently, the baijiu industry primarily relies on sensory evaluations by professional tasters. This method has inherent drawbacks: First, it is highly subjective and lacks repeatability; sensory evaluation results are influenced by various factors such as the taster's physiological state, psychological factors, environmental conditions, and personal preferences, making it difficult to reach consistent conclusions for the same sample at different times or among different tasters. Second, it lacks quantitative indicators; descriptions of watery taste typically use vague qualitative terms such as "obvious," "weak," and "moderate," failing to generate traceable and comparable numerical data, which is detrimental to batch-to-batch stability control. Third, it is difficult to guide process optimization; in processes such as alcohol reduction, blending, and storage, there is a lack of real-time, rapid, and objective detection methods, leading to process adjustments often relying on trial and error, resulting in low efficiency and high costs.

[0003] Low-field nuclear magnetic resonance (LF-NMR) technology can reflect the molecular motion state of water molecules and their interactions with surrounding substances by detecting the transverse relaxation time (T2) distribution of hydrogen protons in a sample. This technology has been successfully used in the food and agricultural product fields to distinguish between bound water and free water. However, when applied to the evaluation of the water-feel of baijiu (Chinese liquor), the following technical difficulties arise, making it difficult for those skilled in the art to establish a correlation between T2 relaxation components and water-feel: First, baijiu contains ethanol, water, and hundreds of flavor substances (esters, acids, alcohols, aldehydes, etc.). Water molecules and ethanol molecules form a complex hydrogen bond network, and the addition of flavor substances further interferes with the relaxation behavior of water molecules; the T2 relaxation spectrum shows overlapping peaks of multiple components, making it difficult to clearly attribute specific relaxation components to "free water". Second, existing studies mostly use ethanol-water standard solutions. This type of system only shows a single T2 peak, and the peak position changes with ethanol concentration but is always less than 3000 ms. Based on this simulation system, those skilled in the art are prone to forming a technical bias that "there is no independent free water peak in baijiu", thus failing to further search for relaxation components related to free water. Third, "water sensitivity" is a subjective sensory indicator, and there is a lack of readily available technical pathways for establishing a quantifiable regression model based on it and physical parameters (such as relaxation time and peak area ratio). Fourth, T2 relaxation time is extremely sensitive to temperature; even small temperature fluctuations can cause relaxation peak shifts, masking differences between real samples. At the same time, the relaxation time of free water-related long T2 components can be as long as several seconds, requiring a sufficiently long acquisition window and a suitable inversion algorithm for stable detection, which conventional NMR detection parameters cannot meet.

[0004] In summary, traditional sensory evaluation is highly subjective, lacks quantitative indicators, and is difficult to guide process optimization. While low-field nuclear magnetic resonance (NMR) technology can detect moisture state, it is limited by the complexity of the baijiu system, the technical bias resulting from the reliance of existing research on ethanol-water simulation systems, the lack of a modeling path between sensory indicators and physical parameters, and stringent detection conditions. Therefore, current technologies have not yet been able to establish an effective correlation between T2 relaxor components and the water-feel of baijiu. These reasons prevent the objective, quantitative, and rapid evaluation of the water-feel of baijiu. Summary of the Invention

[0005] This invention aims to address the problems of strong subjectivity, low efficiency, and lack of quantitative indicators in existing methods for evaluating the water quality of baijiu (Chinese liquor). It proposes a method and system for evaluating the water quality of baijiu based on low-field nuclear magnetic resonance.

[0006] The technical solution adopted by the present invention to solve the above-mentioned technical problems is as follows:

[0007] In a first aspect, the present invention provides a method for evaluating the water content of wine based on low-field nuclear magnetic resonance, the method comprising:

[0008] Step 1: Place the liquor sample to be tested in a low-field nuclear magnetic resonance spectrometer for isothermal equilibrium until the preset isothermal equilibrium temperature is reached. The liquor sample to be tested contains ethanol, water and flavor substances.

[0009] Step 2: Using the Carr-Purcell-Meiboom-Gill sequence, the transverse relaxation echo train signal of the liquor sample to be tested is acquired at the preset isothermal equilibrium temperature; the acquisition parameters of the low-field nuclear magnetic resonance spectrometer include: echo time TE=1ms, number of echoes NECH=18000, and repetition waiting time TW=15000ms.

[0010] Step 3: Perform T2 inversion on the transverse relaxation echo train signal to obtain the T2 relaxation time distribution spectrum;

[0011] Step 4: In the T2 relaxation time distribution spectrum, determine whether there is a free water-related long T2 component with a peak position greater than 3000 ms:

[0012] If present, obtain the peak area of ​​the free water-related long T2 component. and total area of ​​the full spectrum and in accordance with Calculate the proportion of free water ;

[0013] If it does not exist, then determine the proportion of free water. ;

[0014] Step 5: Adjust the free water ratio Input a pre-established water sensitivity evaluation model to obtain a water sensitivity index. Compare the water sensitivity index with a preset threshold to determine the water sensitivity level of the liquor sample to be tested.

[0015] Furthermore, in step 1, the preset constant temperature equilibrium temperature is 32.00±0.01℃;

[0016] In step 2, the acquisition parameters of the low-field nuclear magnetic resonance spectrometer also include: sampling bandwidth SW=200kHz, resonance frequency SF=19MHz, radio frequency center frequency offset O1=452387.13Hz, 90° radio frequency pulse width P1=7.8μs, 180° radio frequency pulse width P2=12.64μs, and accumulation count NS=2.

[0017] Furthermore, the water-sensory evaluation model was established by performing regression analysis on the free water ratio and sensory water-sensory scores of multiple baijiu samples with known sensory water-sensory scores.

[0018] Furthermore, in step 3, the T2 inversion uses the SIRT algorithm, and the inversion model is a multi-component inversion model or a continuous distribution inversion model.

[0019] Furthermore, in step 4, the free water-related long T2 component appears as an independent peak in the T2 relaxation time distribution spectrum, distinct from the single-peak spectrum of the pure ethanol-water control sample. The pure ethanol-water control sample exhibits only a single peak under the same detection conditions, and the peak position is less than 3000 ms.

[0020] Furthermore, prior to step 4, the method includes: establishing a single-peak T2 spectrum reference database based on at least one set of pure ethanol-water control samples to assist in determining whether additional free water-related long T2 components appear in the liquor sample to be tested.

[0021] Furthermore, the T2 relaxation time distribution spectrum of the tested liquor sample exhibits a bimodal structure, with the peak position of the short T2 component ranging from 722ms to 1262ms, and the peak position of the long T2 component ranging from 3511ms to 7391ms.

[0022] Furthermore, the wateriness level of the liquor includes weak wateriness, medium wateriness, or strong wateriness. When the wateriness index is lower than the first threshold, it corresponds to weak wateriness; when it is between the first threshold and the second threshold, it corresponds to medium wateriness; and when it is higher than the second threshold, it corresponds to strong wateriness.

[0023] Furthermore, the liquor sample to be tested is a strong-aroma type of liquor.

[0024] Secondly, the present invention provides a wine body water-sensitivity evaluation system based on low-field nuclear magnetic resonance, for implementing the wine body water-sensitivity evaluation method based on low-field nuclear magnetic resonance as described in the first aspect, the system comprising:

[0025] A low-field nuclear magnetic resonance spectrometer is used to equilibrate the baijiu sample to be tested at a constant temperature, and the transverse relaxation echo train signal of the baijiu sample to be tested is acquired using the Carr-Purcell-Meiboom-Gill sequence. The baijiu sample to be tested contains ethanol, water and flavor substances.

[0026] The data processing module is used to perform T2 inversion on the transverse relaxation echo train signal to obtain the T2 relaxation time distribution spectrum, and to determine whether there is a free water-related long T2 component with a peak position greater than 3000ms in the T2 relaxation time distribution spectrum.

[0027] If present, obtain the peak area of ​​the free water-related long T2 component. and total area of ​​the full spectrum and in accordance with Calculate the proportion of free water ;

[0028] If it does not exist, then determine the proportion of free water. ;

[0029] The evaluation module is used to determine the proportion of free water. Input a pre-established water sensitivity evaluation model to obtain a water sensitivity index. Compare the water sensitivity index with a preset threshold to determine the water sensitivity level of the liquor sample to be tested.

[0030] The beneficial effects of this invention are as follows: The low-field nuclear magnetic resonance-based method and system for evaluating the water content of liquor provided by this invention, through precise isostatic control and optimized CPMG (Carr-Purcell-Meiboom-Gill) acquisition parameters, stably identifies free water-related long T2 components with peaks greater than 3000 ms in real baijiu, overcoming the technical bias in the field that baijiu does not contain independent free water peaks; secondly, based on the proportion of free water... As an objective physical indicator, it transforms the subjective description of water sensation into a calculable and reproducible numerical value, avoiding interference from human factors; furthermore, by... By inputting a pre-established water sensitivity evaluation model, the water sensitivity index is automatically obtained and compared with a threshold to output the water sensitivity level, realizing full automation of the detection, calculation, and evaluation process. This method is highly efficient and the results are traceable. Finally, the method has a fast detection speed and is easy to operate. It can be directly used for online monitoring of process links such as blending, dilution, and storage, providing an objective and quantitative technical means for the quality control of liquor and significantly improving the scientificity and practicality of water sensitivity evaluation of baijiu. Attached Figure Description

[0031] Figure 1 A schematic flowchart of a wine body water-sensitivity evaluation method based on low-field nuclear magnetic resonance provided for an embodiment;

[0032] Figure 2 A schematic diagram of the T2 relaxation time distribution spectrum provided for an embodiment;

[0033] Figure 3 A schematic diagram of the structure of a wine body water-sensitivity evaluation system based on low-field nuclear magnetic resonance provided for an embodiment. Detailed Implementation

[0034] To address the problem of the inability to objectively, quantitatively, and rapidly evaluate the water-like characteristics of baijiu (Chinese liquor), this invention proposes a technical solution. In this invention, the baijiu sample to be tested is first equilibrated to a preset temperature to eliminate the influence of temperature on relaxation time drift. Then, a CPMG sequence is used to acquire transverse relaxation echo train signals. By setting optimized parameters such as echo time TE=1ms, number of echoes NECH=18000, and repetition waiting time TW=15000ms, a sufficiently long acquisition window and complete spin recovery are ensured, thereby fully capturing the attenuation information of long relaxation components. Next, T2 inversion is performed on the echo train to obtain the T2 relaxation time distribution spectrum. Based on whether the peak position is greater than 3000ms, long T2 components related to free water are identified, and the proportion of free water is then calculated. Finally, The system inputs a pre-established water-sensitivity evaluation model, obtains a water-sensitivity index, compares it with a preset threshold, and outputs the water-sensitivity grade of the liquor. This invention, through precise temperature control and optimized acquisition parameters, is the first to stably detect free water-related long T2 components with peak values ​​exceeding 3000 ms in real baijiu, overcoming the technical bias of conventional parameters that miss long relaxation signals due to insufficient acquisition windows; it transforms the subjective water-sensitivity into an objectively quantifiable proportion of free water. It automatically outputs the water sensitivity level by combining a regression model, realizing a rapid, non-destructive, and reproducible evaluation, which significantly improves the scientific nature and process applicability of water sensitivity testing for baijiu.

[0035] The technical solutions in this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0036] Figure 1 A flowchart illustrating a method for evaluating the water content of wine based on low-field nuclear magnetic resonance is shown. Please refer to [link / reference]. Figure 1 The method includes the following steps:

[0037] Step 1: Place the liquor sample to be tested in a low-field nuclear magnetic resonance spectrometer for isothermal equilibrium until the preset isothermal equilibrium temperature is reached. The liquor sample to be tested contains ethanol, water and flavor substances.

[0038] Specifically, this step involves placing the baijiu sample to be tested into a low-field nuclear magnetic resonance spectrometer and equilibrating it to a preset temperature, such as 32.00±0.01℃. Setting this temperature range approximates the typical ambient temperature for baijiu sensory evaluation and eliminates the influence of temperature fluctuations on the transverse relaxation time T2, ensuring the stability of the relaxation peak position and comparability between different samples. The baijiu sample contains ethanol, water, and flavor compounds (such as esters, acids, alcohols, and aldehydes). This composition is key to distinguishing real baijiu from a pure ethanol-water simulation system: the presence of flavor compounds disrupts the homogeneity of the ethanol-water hydrogen bond network, allowing some water molecules to gain higher degrees of freedom, resulting in a long T2 component with a peak apex greater than 3000 ms in the T2 relaxation spectrum. Isothermal equilibration is a necessary prerequisite for the subsequent stable detection of this long T2 component; otherwise, even small temperature fluctuations can cause the relaxation peak position to drift by hundreds of milliseconds, leading to misjudgment or missed detection of the free water peak.

[0039] In this embodiment, the liquor sample to be tested is a strong-aroma liquor (such as Wuliangye, Wuliangchun, Wuliangtequ, etc.). Those skilled in the art will understand that this strong-aroma liquor is only an example, and the method is also applicable to other types of liquor such as sauce-aroma and light-aroma liquor. As long as the sample contains ethanol, water and flavor substances, the water sense evaluation can be achieved through the same detection conditions and steps.

[0040] Step 2: Using the Carr-Purcell-Meiboom-Gill sequence, the transverse relaxation echo train signal of the liquor sample to be tested is acquired at the preset isothermal equilibrium temperature; the acquisition parameters of the low-field nuclear magnetic resonance spectrometer include: echo time TE=1ms, number of echoes NECH=18000, and repetition waiting time TW=15000ms.

[0041] Specifically, this step employs the Carr-Purcell-Meiboom-Gill (CPMG) sequence to acquire transverse relaxation echo train signals of the baijiu sample under test at an isothermal equilibrium temperature. To stably capture the long-relaxation free water components present in real baijiu, the acquisition parameters of the low-field NMR spectrometer were optimized as follows: echo time TE = 1ms, which is short enough to avoid signal attenuation caused by diffusion while ensuring complete recording of fast-relaxing components; number of echoes NECH = 18000, which, combined with TE = 1ms, provides an acquisition window of up to 18 seconds, sufficient to completely record the entire process of free water-related long T2 components with peaks greater than 3000ms from the initial peak to complete decay; repetition wait time TW = 15000ms, ensuring that all spin systems in the sample (especially long T1 components) fully recover to thermal equilibrium, avoiding underestimation of the long T2 peak area due to saturation effects.

[0042] In this embodiment, the following acquisition parameters are further set to ensure signal quality and inversion accuracy: sampling bandwidth SW = 200kHz, resonant frequency SF = 19MHz, RF center frequency offset O1 = 452387.13Hz, 90° RF pulse width P1 = 7.8μs, 180° RF pulse width P2 = 12.64μs (meeting the CPMG sequence requirement of P2≈2×P1), and accumulation count NS = 2. These acquisition parameters can shorten the total detection time (approximately 30-40 seconds) while ensuring the signal-to-noise ratio, enabling rapid evaluation. Through the above combination of acquisition parameters, the acquired echo train data contains complete relaxation information, providing a data foundation for subsequent T2 inversion and stable identification of free water-related long T2 components.

[0043] Step 3: Perform T2 inversion on the transverse relaxation echo train signal to obtain the T2 relaxation time distribution spectrum.

[0044] Specifically, this step performs T2 inversion on the acquired transverse relaxation echo train signal to obtain the T2 relaxation time distribution spectrum. T2 inversion refers to using mathematical algorithms to convert the original echo train data into a distribution relationship between relaxation time and signal amplitude. The SIRT algorithm (Simultaneous Iterative Reconstruction Technique) is usually used, and the inversion model can be a multi-component inversion model or a continuous distribution inversion model.

[0045] In this embodiment, the inversion parameters were set as follows: minimum relaxation time Tmin = 10 ms, maximum relaxation time Tmax = 100,000 ms, and 100,000 iterations. Through inversion, water molecule populations with different relaxation times exhibited different peaks on the T2 relaxation time distribution spectrum: short T2 components corresponded to water molecules that were more tightly bound to ethanol and flavor compounds, while long T2 components corresponded to free water molecules with higher degrees of freedom of motion. This distribution spectrum is used for subsequent identification of free water-related long T2 components (peak position greater than 3000 ms) and calculation of the free water proportion. The basic data. Table 1 shows a schematic diagram of detection conditions for low-field NMR.

[0046] Table 1. Schematic diagram of detection conditions for low-field NMR.

[0047]

[0048] In this embodiment, before step 4 is executed, a single-peak T2 spectrum reference database can be established based on at least one set of pure ethanol-water control samples to help determine the peak structure of the liquor sample to be tested, that is, whether there are additional free water-related long T2 components in the liquor sample to be tested, so as to avoid misjudging noise or false peaks as free water-related long T2 components.

[0049] Specifically, at least one set of pure ethanol-water control samples (e.g., ethanol-water standard solutions with alcohol content of 5%–93%) are first selected and tested under the same isothermal equilibrium temperature and CPMG acquisition parameters to obtain their T2 relaxation time distribution spectra. A single-peak T2 spectrum reference database is then established based on this data. The purpose of this database is to compare the T2 relaxation time spectrum of the tested baijiu sample with the single-peak spectrum in the database when determining the free water-related long T2 component. If the tested baijiu sample exhibits an additional independent peak (i.e., a double-peak structure) at a position greater than 3000 ms at the peak apex, this peak can be confirmed as a free water-related long T2 component, rather than a false peak caused by temperature fluctuations or instrument noise. This reference database effectively improves the accuracy and reliability of identification, avoiding the misidentification of relaxation signals not present in the pure ethanol-water system as free water.

[0050] Figure 2 A schematic diagram of the T2 relaxation time distribution spectrum of a pure ethanol-water control sample and a baijiu sample is shown, where the horizontal axis represents the relaxation time T2 and the vertical axis represents the signal amplitude.

[0051] exist Figure 2 In the table, samples 1-8 are the T2 relaxation time distribution spectra of the ethanol-water control samples under the detection conditions in Table 1. Their alcohol content (vol%) is as follows: Sample 1: 5%; Sample 2: 11%; Sample 3: 16%; Sample 4: 30%; Sample 5: 38%; Sample 6: 52%; Sample 7: 68%; Sample 8: 93%.

[0052] from Figure 2 It can be seen that samples 1-8 all exhibit a single main peak structure, with the peak T2 located in the range of 1384-2420 ms, all less than 3000 ms. Table 2 shows the single-peak T2 results of samples 1-8. It is evident that there is no independent long T2 free water peak in the pure ethanol-water simulation system. It is precisely for this reason that those skilled in the art are prone to forming the technical bias that "there is no independent free water peak in baijiu."

[0053] Table 2. Schematic diagram of single-peak T2 results for samples 1-8

[0054]

[0055] exist Figure 2 In the table, samples 9-14 are the T2 relaxation time distribution spectra of real baijiu samples under the detection conditions in Table 1, namely: Sample 9: 39° Wuliangye; Sample 10: 45° Wuliangchun; Sample 11: 42° Wuliangtequ; Sample 12: 52° Wuliangye; Sample 13: 42° Wuliangtequ-2; Sample 14: 52° Wuliangtequ-2.

[0056] from Figure 2 It can be seen that samples 9-14 exhibit a distinct bimodal structure: the peak position of the short T2 component is in the range of 722~1262ms, and the peak position of the long T2 component is in the range of 3511~7391ms. Therefore, under the detection conditions of this embodiment, compared with the pure ethanol-water control sample, the presence of an additional independent peak with a peak position greater than 3000ms in the T2 relaxation time distribution spectrum of the tested liquor sample confirms that this independent peak is a free water-related long T2 component, and not a spurious peak caused by temperature fluctuations or instrument noise.

[0057] Step 4: In the T2 relaxation time distribution spectrum, determine whether there is a free water-related long T2 component with a peak position greater than 3000 ms:

[0058] If present, obtain the peak area of ​​the free water-related long T2 component. and total area of ​​the full spectrum and in accordance with Calculate the proportion of free water ;

[0059] If it does not exist, then determine the proportion of free water. .

[0060] according to Figure 2It is known that since the peak position of the pure ethanol-water standard solution sample is less than 3000ms at all alcohol contents, while the peak position of the long T2 component in the real liquor sample is distributed in the range of 3511ms to 7391ms, based on this, this embodiment uses 3000ms as the effective boundary to distinguish the free water-related long T2 component from the conventional ethanol-water relaxation peak.

[0061] In practical applications, if an independent peak with a apex greater than 3000 ms exists in the T2 relaxation time distribution spectrum (distinguishing it from the single-peak spectrum of the pure ethanol-water control sample), this peak is identified as a free water-related long T2 component, and the peak area of ​​this component is obtained. and total area of ​​the full spectrum (That is, the sum of the peak areas of all relaxor components). Then follow the formula Calculate the proportion of free water This ratio reflects the proportion of highly mobile free water molecules in the tested liquor sample relative to the total hydrogen proton signal.

[0062] If there is no component with a peak apex greater than 3000 ms in the T2 relaxation time distribution spectrum (i.e., the entire spectrum shows only a single peak, and the peak position is less than 3000 ms), then it is determined that the sample does not contain any free water-related long T2 components. In this case, the following setting is used: This situation may occur in some high-proof spirits or specially treated spirits, indicating that the water is basically tightly bound to the ethanol and flavor substances, with no obvious free water present.

[0063] Table 3 shows... Figure 2 Results of two-component T2 (short T2 component and long T2 component) of samples 9-14 and The diagram shows the peak position of the short T2 component, the peak position of the long T2 component, the peak area of ​​the short T2 component, and the peak area of ​​the long T2 component.

[0064] Table 3. Two-component T2 results for samples 9-14 and Schematic table

[0065]

[0066] As can be seen from Table 3, the proportion of free water in samples 9-14 The concentration ranged from 0.004 to 0.026. As the alcohol content of the samples increased (samples 12 and 14 were 52° samples),... The values ​​decreased significantly (0.004 and 0.006 respectively), indicating a negative correlation between the free water ratio and alcohol content. This is consistent with the sensory experience that high-proof liquor has a weaker watery taste and low-proof liquor has a stronger watery taste. This demonstrates that the method in this embodiment can distinguish and quantify the degree of free water correlation in different liquor samples.

[0067] Step 5: Adjust the free water ratio Input a pre-established water sensitivity evaluation model to obtain a water sensitivity index. Compare the water sensitivity index with a preset threshold to determine the water sensitivity level of the liquor sample to be tested.

[0068] Specifically, the calculated free water ratio Input a pre-established water sensitivity evaluation model, which outputs a continuous numerical value as the water sensitivity index. This water sensitivity evaluation model is established by performing regression analysis on the free water ratio and sensory water sensitivity scores of multiple baijiu samples with known sensory water sensitivity scores. The regression analysis can employ linear or nonlinear regression, using the free water ratio as the starting point. Using the sensory water sensitivity scores of multiple baijiu samples with known sensory water sensitivity scores as the dependent variable, a mapping function is obtained through regression analysis. The continuous value output by this mapping function is the water sensitivity index. For example, for a batch of baijiu samples covering different alcohol contents and aroma types, professional tasters first score their water sensitivity (e.g., 0-10 points, with higher scores indicating stronger water sensitivity). Simultaneously, the water sensitivity index of each sample is measured under the same testing conditions. A regression analysis was used to establish the relationship between the water perception index and the sensory water perception score. After the water perception evaluation model was established, for unknown samples, only the required input... You can then obtain the corresponding water feel index.

[0069] After obtaining the water sensitivity index, it is compared with a preset threshold to determine the water sensitivity level of the liquor sample to be tested. In this embodiment, the preset threshold includes a first threshold and a second threshold, and the water sensitivity level includes weak water sensitivity, medium water sensitivity, and strong water sensitivity. Specifically, when the water sensitivity index is lower than the first threshold, it is determined to be weak water sensitivity; when it is between the first threshold and the second threshold, it is determined to be medium water sensitivity; and when it is higher than the second threshold, it is determined to be strong water sensitivity. The specific values ​​of the first threshold and the second threshold can be set according to the distribution of training samples when the model is built and the actual application requirements. For example, the dividing point can be ±0.5 times the standard deviation of the mean water sensitivity score in the training samples.

[0070] If the proportion of free water This indicates that the T2 relaxation time distribution spectrum of the tested baijiu sample does not contain a free water-related long T2 component with a peak greater than 3000 ms, meaning that all water molecules are tightly bound to ethanol and flavor substances, and there is no highly mobile free water present. After inputting the pre-established water-sensory evaluation model, the water-sensory index output by the model should be an extremely low value (approaching the minimum predicted value corresponding to the intercept term of the regression equation). When comparing this water-sensory index with a preset threshold, since the water-sensory index is lower than the first threshold (the upper limit of weak water-sensory), the corresponding water-sensory level of the wine is judged as weak water-sensory. This judgment is consistent with sensory experience. Commonly found in high-proof spirits or specially treated liquors, it has very little wateriness and an inconspicuous watery taste, resulting in a mellow and full-bodied flavor.

[0071] Through the above steps, the objectively measured proportion of free water will be obtained. Ultimately, this is transformed into an intuitive and comparable level of wateriness in the wine, achieving the quantification, automation, and standardization of wateriness evaluation.

[0072] In summary, the low-field nuclear magnetic resonance-based method for evaluating the water quality of liquor provided in this embodiment, through precise isostatic control and optimized CPMG acquisition parameters, for the first time stably identifies free water-related long T2 components with peaks greater than 3000 ms in real baijiu, and calculates their area ratio as the free water proportion. This transforms the traditional sensory evaluation of water quality into an objective and quantifiable physical indicator, avoiding interference from human factors and improving the repeatability and traceability of the evaluation. Simultaneously, the water quality evaluation model established by regression analysis automatically outputs the water quality index and water quality level, achieving rapid automation of the entire process of detection, calculation, and evaluation. It can be directly used for online monitoring of process stages such as blending, dilution, storage, and batch quality control, significantly improving the scientific nature and process applicability of baijiu water quality evaluation, and overcoming the technical bias in this field that "there is no independent free water peak in baijiu" based on the ethanol-water simulation system.

[0073] Based on the above technical solution, this embodiment also provides a wine body water-sensitivity evaluation system based on low-field nuclear magnetic resonance, used to implement the wine body water-sensitivity evaluation method based on low-field nuclear magnetic resonance described in the embodiment. Please refer to [link to relevant documentation]. Figure 3 The system includes:

[0074] A low-field nuclear magnetic resonance spectrometer is used to equilibrate the baijiu sample to be tested at a constant temperature, and the transverse relaxation echo train signal of the baijiu sample to be tested is acquired using the Carr-Purcell-Meiboom-Gill sequence. The baijiu sample to be tested contains ethanol, water and flavor substances.

[0075] The data processing module is used to perform T2 inversion on the transverse relaxation echo train signal to obtain the T2 relaxation time distribution spectrum, and to determine whether there is a free water-related long T2 component with a peak position greater than 3000ms in the T2 relaxation time distribution spectrum.

[0076] If present, obtain the peak area of ​​the free water-related long T2 component. and total area of ​​the full spectrum and in accordance with Calculate the proportion of free water ;

[0077] If it does not exist, then determine the proportion of free water. ;

[0078] The evaluation module is used to determine the proportion of free water. Input a pre-established water sensitivity evaluation model to obtain a water sensitivity index. Compare the water sensitivity index with a preset threshold to determine the water sensitivity level of the liquor sample to be tested.

[0079] It is understood that the wine body water quality evaluation system based on low-field nuclear magnetic resonance described in this embodiment is a system used to implement the wine body water quality evaluation method based on low-field nuclear magnetic resonance described in the embodiment. As the system disclosed in the embodiment corresponds to the method disclosed in the embodiment, the description is relatively simple. For relevant parts, please refer to the description of the method. It will not be repeated here.

Claims

1. A method for evaluating the water content of wine based on low-field nuclear magnetic resonance, characterized in that, The method includes: Step 1: Place the liquor sample to be tested in a low-field nuclear magnetic resonance spectrometer for isothermal equilibrium until the preset isothermal equilibrium temperature is reached. The liquor sample to be tested contains ethanol, water and flavor substances. Step 2: Using the Carr-Purcell-Meiboom-Gill sequence, the transverse relaxation echo train signal of the liquor sample to be tested is acquired at the preset isothermal equilibrium temperature; the acquisition parameters of the low-field nuclear magnetic resonance spectrometer include: echo time TE=1ms, number of echoes NECH=18000, and repetition waiting time TW=15000ms. Step 3: Perform T2 inversion on the transverse relaxation echo train signal to obtain the T2 relaxation time distribution spectrum; Step 4: In the T2 relaxation time distribution spectrum, determine whether there is a free water-related long T2 component with a peak position greater than 3000 ms: If present, obtain the peak area of ​​the free water-related long T2 component. and total area of ​​the full spectrum and in accordance with Calculate the proportion of free water ; If it does not exist, then determine the proportion of free water. ; Step 5: Adjust the free water ratio Input a pre-established water sensitivity evaluation model to obtain a water sensitivity index. Compare the water sensitivity index with a preset threshold to determine the water sensitivity level of the liquor sample to be tested.

2. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, In step 1, the preset constant temperature equilibrium temperature is 32.00±0.01℃; In step 2, the acquisition parameters of the low-field nuclear magnetic resonance spectrometer also include: sampling bandwidth SW=200kHz, resonance frequency SF=19MHz, radio frequency center frequency offset O1=452387.13Hz, 90° radio frequency pulse width P1=7.8μs, 180° radio frequency pulse width P2=12.64μs, and accumulation count NS=2.

3. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, The water sensitivity evaluation model was established by performing regression analysis on the free water ratio and sensory water sensitivity scores of multiple baijiu samples with known sensory water sensitivity scores.

4. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, In step 3, the T2 inversion uses the SIRT algorithm, and the inversion model is a multi-component inversion model or a continuous distribution inversion model.

5. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, In step 4, the free water-related long T2 component appears as an independent peak in the T2 relaxation time distribution spectrum, distinct from the single peak spectrum of the pure ethanol-water control sample. The pure ethanol-water control sample only exhibits a single peak under the same detection conditions, and the peak position is less than 3000ms.

6. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, Before step 4, the method further includes: establishing a single-peak T2 spectrum reference database based on at least one set of pure ethanol-water control samples to help determine whether additional free water-related long T2 components appear in the liquor sample to be tested.

7. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, The T2 relaxation time distribution spectrum of the tested liquor sample exhibits a bimodal structure, with the peak position of the short T2 component ranging from 722ms to 1262ms, and the peak position of the long T2 component ranging from 3511ms to 7391ms.

8. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, The wateriness level of the liquor includes weak wateriness, medium wateriness, or strong wateriness. When the wateriness index is below the first threshold, it corresponds to weak wateriness; when it is between the first threshold and the second threshold, it corresponds to medium wateriness; and when it is above the second threshold, it corresponds to strong wateriness.

9. The method for evaluating the water content of wine based on low-field nuclear magnetic resonance according to claim 1, characterized in that, The liquor sample to be tested was a strong-aroma type of liquor.

10. A wine body water-sensitivity evaluation system based on low-field nuclear magnetic resonance, characterized in that, For implementing the low-field nuclear magnetic resonance-based wine body water-sensitivity evaluation method as described in any one of claims 1 to 9, the system comprises: A low-field nuclear magnetic resonance spectrometer is used to equilibrate the baijiu sample to be tested at a constant temperature, and the transverse relaxation echo train signal of the baijiu sample to be tested is acquired using the Carr-Purcell-Meiboom-Gill sequence. The baijiu sample to be tested contains ethanol, water and flavor substances. The data processing module is used to perform T2 inversion on the transverse relaxation echo train signal to obtain the T2 relaxation time distribution spectrum, and to determine whether there is a free water-related long T2 component with a peak position greater than 3000ms in the T2 relaxation time distribution spectrum. If present, obtain the peak area of ​​the free water-related long T2 component. and total area of ​​the full spectrum and in accordance with Calculate the proportion of free water ; If it does not exist, then determine the proportion of free water. ; The evaluation module is used to determine the proportion of free water. Input a pre-established water sensitivity evaluation model to obtain a water sensitivity index. Compare the water sensitivity index with a preset threshold to determine the water sensitivity level of the liquor sample to be tested.

Images