A liquor component on-line monitoring device and method based on multi-array surface acoustic wave interference

Abstract

The present application relates to a kind of liquor component online monitoring device and method based on multi-array acoustic surface wave interference, liquor component online monitoring device based on multi-array acoustic surface wave interference includes flow base, multi-array acoustic surface wave interference detection module, flow guide component, signal acquisition and solution module, summary analysis module, feedback output module, signal acquisition and solution module are used to gather array response and execute compensation and concentration solution, summary analysis module is used to carry out aggregation and comparison to concentration solution result of multiple samples, feedback output module is used to output early warning, suggestion or control instruction according to the comparison result of summary analysis module.The beneficial effects of the present application are that: both can be completed multi-array acoustic surface wave detection under liquid phase condition, and can be formed by reference compensation, parameter switching and sample summary analysis, the quantitative evaluation result to liquor fermentation state, pit deviation and the risk of later stage process, and further output feedback information to production end.

Description

Technical Field

[0001] This invention relates to the field of baijiu brewing technology, specifically to an online monitoring device and method for baijiu components based on multi-array surface acoustic wave interferometry. Background Technology

[0002] The production process of baijiu involves multiple stages, including fermentation, distillation, spirit extraction, storage, blending, and pre-bottling verification. The alcohol content, temperature, viscosity, conductivity, and concentrations of key compounds such as esters, acids, and alcohols continuously change at each stage. These changes affect not only the stability of the baijiu's style but also production decisions such as spirit extraction segmentation, base liquor diversion, and blending selection. Current baijiu quality testing primarily relies on offline testing. While this provides relatively accurate analytical results, it typically suffers from problems such as long sampling cycles, complex pretreatment, difficulty in continuously tracking fermentation changes within the same fermentation pit, difficulty in rapidly comparing multiple pits from the same batch, and difficulty in timely outputting control information to the production side. Especially for scenarios requiring rapid response, such as spirit extraction segmentation, abnormal pit warnings, and pre-bottling verification, relying solely on offline laboratory testing is insufficient to meet the requirements of online and continuous testing.

[0003] Furthermore, existing online detection methods are mostly limited to single-point detection or single-sample discrimination, making it difficult to simultaneously consider stable flow conditions, matrix interference suppression, and multi-component difference identification in the complex liquid-phase system of baijiu. Especially when alcohol content jumps, temperature drifts, and liquid phase loading changes coexist, it is difficult to obtain stable concentration information or classification results that can be directly used for process control if judgment is based solely on the original frequency shift or amplitude signal.

[0004] In the production of baijiu (Chinese liquor), significant differences can occur within the same fermentation pit at different fermentation points, between different pits within the same batch, and between different distillation stages. Existing test results are mostly limited to single-sample levels, lacking the ability to continuously summarize and analyze data from multiple time points, multiple fermentation pits, and multiple distillation stages. It is also difficult to further generate feedback information that can be directly used for fermentation control, distillation stage segmentation, base liquor diversion, and pre-bottling verification. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide an online monitoring device and method for baijiu components based on multi-array surface acoustic wave interferometry. This device can stably complete multi-array surface acoustic wave detection under liquid phase conditions and can also form quantitative evaluation results on the fermentation status of baijiu, cellar deviation and downstream process risks through reference compensation, parameter switching and sample summary analysis, and further output feedback information to the production end.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] An online monitoring device for the components of baijiu (Chinese liquor) based on multi-array surface acoustic wave interferometry includes:

[0008] The flow base has a flow channel inside, which includes a buffer section, a flow stabilizing section, a detection section and a discharge section connected in sequence.

[0009] A flow guiding component is located upstream of the detection section, and a laminar flow gap is formed between the flow guiding component and the top inner surface of the detection section;

[0010] A multi-array surface acoustic wave (SAW) interferometric detection module is hermetically sealed on a flow base and corresponds to the detection section. The acoustic wave conduction surface of the multi-array SAW interferometric detection module faces the laminar flow gap. The multi-array SAW interferometric detection module includes at least two detection subarrays and at least one reference subarray. The surface of the detection subarrays is provided with a sensitive recognition layer for selectively responding to target compounds in baijiu (Chinese liquor). The reference subarray is used to output a compensation signal characterizing background interference factors.

[0011] The signal acquisition and calculation module and the multi-array surface acoustic wave interferometry detection module are connected to the signal acquisition and calculation module through a signal connection component. The signal acquisition and calculation module is used to acquire the array response and perform compensation and concentration calculation.

[0012] The summary analysis module is connected to the signal acquisition and calculation module. The summary analysis module is used to aggregate and compare the concentration calculation results of multiple samples.

[0013] The feedback output module is connected to the summary analysis module. The feedback output module is used to output warnings, suggestions or control commands based on the comparison results of the summary analysis module.

[0014] Furthermore, the multi-array surface acoustic wave interferometry detection module is a shear horizontal wave surface acoustic wave structure or a Love wave structure.

[0015] Furthermore, the surface of the detection subarray is provided with different sensitive recognition layers, including one or more of the following: a molecularly imprinted polymer layer for enriching ethyl hexanoate, a polyisobutylene layer for enhancing the overall response of esters, a polar polymer layer for adjusting the adsorption characteristics of acids, and an inert protective layer for forming a stable background reference.

[0016] Furthermore, the reference subarray is divided into a viscosity compensation region and a temperature drift calibration region. The viscosity compensation region is used to output compensation signals related to fluid damping or matrix viscosity changes, while the temperature drift calibration region is used to output calibration signals related to ambient temperature or detection chip temperature changes.

[0017] Furthermore, the flow guiding component is an asymmetric convex arc-shaped flow guiding plate with an arc surface deflection angle of 15°±2°. The vertical distance between the flow guiding component and the surface of the detection section is 0.8mm-1.5mm. The flow guiding component is made of food-grade 304 stainless steel, and the surface roughness of the flow guiding component is Ra≤0.8μm.

[0018] Furthermore, the present invention also provides an online monitoring method for components of baijiu (Chinese liquor) based on multi-array surface acoustic wave interferometry, comprising the following steps:

[0019] S1. The liquor to be tested is sampled online during the liquor production process and introduced into the flow channel inside the flow base. The liquor to be tested passes through the buffer section and the flow stabilization section in sequence, and then passes through the detection section in a laminar flow state. Finally, it is discharged through the discharge section at the end of the detection section.

[0020] S2. Acquire at least one of the following response parameters of the wine in the detection segment: frequency shift, phase shift, amplitude change, and insertion loss change, through a multi-array surface acoustic wave interferometry detection module.

[0021] S3. The signal acquisition and calculation module performs at least one of alcohol compensation, temperature compensation, viscosity compensation, background compensation and drift correction on the response of the detection subarray based on the compensation signal output by the reference subarray, obtains the compensated feature quantity, and outputs the concentration feature vector of the target compound based on the compensated feature quantity and the preset calculation model.

[0022] S4. The summary analysis module aggregates, compares, and identifies the concentration feature vectors of samples from different time points, different wine sections of the same batch, or corresponding samples from multiple cellars. It calculates the slope and fluctuation amplitude of changes in adjacent time windows and outputs at least one of the following results: cellar abnormal trend identification, cellar deviation level, concentration feature vector deviation score, and downstream process risk warning.

[0023] S5. The feedback output module outputs at least one of the following based on the output results of the summary analysis module: fermentation control suggestions, distillation and segmentation suggestions, base liquor diversion suggestions, storage risk warnings, blending and screening suggestions, or pre-bottling review instructions.

[0024] Furthermore, the target compounds include one or more of ethyl hexanoate, ethyl acetate, ethyl lactate, ethyl butyrate, ethanol, acetic acid, lactic acid, butyric acid, and hexanoic acid.

[0025] Furthermore, the signal acquisition and calculation module constructs a compensation model based on the output of the detection subarray and the output of the reference subarray, and combines the temperature parameter and alcohol content parameter to obtain at least one compensated feature quantity, and then calculates the concentration feature vector of the target compound based on the compensated feature quantity.

[0026] Furthermore, when establishing the compensation model, a set of standard wine samples covering different alcohol content, different temperatures, and different target component concentration ranges are prepared. The original response parameters of the detection subarray and reference subarray corresponding to the standard wine samples are collected. Then, the known alcohol content, temperature, and target component concentration values ​​in the standard wine samples are used as calibration references to establish the compensation model.

[0027] Furthermore, the signal acquisition and processing module has a preset parameter mapping matrix. The parameter mapping matrix associates the fluid damping range, alcohol content range or temperature range with the excitation frequency, filter coefficient, compensation coefficient and processing model. When the output of the reference subarray exceeds the preset threshold, it automatically switches to the parameter group that matches the current sample operating conditions.

[0028] Compared with the prior art, the present invention has the following beneficial effects:

[0029] 1. This invention reduces the interference of turbulence, wall adhesion, and microbubble attachment on the array response by setting a flow guiding component and laminar flow gap inside the flow base, so that the wine to be tested forms a more stable flow state before entering the detection area.

[0030] 2. This invention sets up multiple detection subarrays and reference subarrays, and performs joint compensation for alcohol content, temperature, viscosity and background drift, so that the original high-frequency response can be stably solved into the concentration feature vector of the target compound.

[0031] 3. By introducing a parameter mapping matrix and adaptive switching logic, this invention enables test samples with different alcohol content or different matrix load conditions to call different working frequencies and solution coefficients, thereby improving the problem that a set of parameters cannot cover all liquor samples.

[0032] 4. This invention can not only test a single sample, but also summarize and analyze the results of continuous sampling in the same pit or samples from multiple pits in the same batch. Based on this, it can output abnormal trends, deviation levels, style matching degree and subsequent risk warning results, thus forming a closed-loop solution of "online sampling - steady flow detection - reference compensation - concentration calculation - pit summary - production feedback". Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the overall structure of the online monitoring device for baijiu components in this invention;

[0034] Figure 2 This is a schematic cross-sectional view of the detection channel in this invention;

[0035] Figure 3 This is a schematic diagram of the subarray layout of the multi-array surface acoustic wave interferometry detection module in this invention;

[0036] Figure 4This is a sample summary analysis curve in this invention.

[0037] In the figure: 1. Flow base; 2. Multi-array surface acoustic wave interferometry detection module; 4. Upper mounting assembly; 5. Signal connection assembly; 31. First signal line; 32. Second signal line; 33. Reference signal line; 11. Flow guiding component; 12. Flow channel; 21. First detection subarray; 22. Second detection subarray; 23. Reference subarray; 41. Trend baseline; 42. Abnormal fluctuation curve. Detailed Implementation

[0038] The present invention will be further described below with reference to the accompanying drawings, but the scope of protection of the present invention is not limited to the following description.

[0039] like Figure 1 , Figure 2 As shown, an online monitoring device for the components of baijiu (Chinese liquor) based on multi-array surface acoustic wave interferometry includes a flow base 1, a multi-array surface acoustic wave interferometry detection module 2, a flow guiding component 11, a signal acquisition and calculation module, a summary analysis module, and a feedback output module. The flow base 1 is provided with a flow channel 12, which includes a buffer section, a flow stabilization section, a detection section, and a discharge section connected in sequence. The multi-array surface acoustic wave interferometry detection module 2 is hermetically sealed on the flow base 1 and corresponds to the detection section. The acoustic wave conduction surface of the multi-array surface acoustic wave interferometry detection module 2 faces the laminar flow gap. The flow guiding member 11 is located upstream of the detection section, and a laminar flow gap is formed between the flow guiding member 11 and the top inner surface of the detection section. The multi-array surface acoustic wave interferometry detection module 2 is connected to the signal acquisition and calculation module through the signal connection component 5. The signal acquisition and calculation module is used to acquire the frequency offset and phase offset of the array response and perform compensation calculation. The summary analysis module is connected to the signal acquisition and calculation module. The summary analysis module is used to aggregate and compare the concentration calculation results of multiple samples. The feedback output module is connected to the summary analysis module. The feedback output module is used to output warnings, suggestions, or control commands based on the comparison results of the summary analysis module.

[0040] The flow base 1 is made of metal or high-strength inert material. The inlet end of the flow channel 12 is connected to the sampling pipeline, and the outlet end of the flow channel 12 is connected to the drain or return pipeline. The wine to be tested is introduced into the flow channel 12 through the inlet end of the flow channel 12, and then passes through the buffer section and the flow stabilization section for buffering and flow stabilization. After passing through the detection section, it passes through the discharge section and is discharged from the outlet end of the flow channel 12. In the detection section, the multi-array surface acoustic wave interferometry detection module 2 collects the array response of the wine to be tested.

[0041] The wine to be tested can come from the online bypass of the distillation section, the sampling branch of the fermentation pit, the sampling port of the storage tank, or the verification branch before blending. The flow guiding component 11 is an asymmetrical convex arc-shaped flow guiding plate. The arc angle of the flow guiding component 11 is 15°±2°. The vertical distance between the flow guiding component 11 and the surface of the detection section is 0.8mm-1.5mm. The vertical distance between the lowest point of the flow guiding component 11 and the surface of the detection area is preferably controlled at about 1.2mm.

[0042] The flow guiding component 11 can be made of food-grade corrosion-resistant metal materials or food-grade high-performance polymer materials. The food-grade corrosion-resistant metal materials include food-grade 304 stainless steel, 316L stainless steel, duplex stainless steel (such as 2205) or high-purity titanium alloy (such as TA1, TA2). The food-grade high-performance polymer materials include polytetrafluoroethylene (PTFE), fusible polytetrafluoroethylene (PFA), polyether ether ketone (PEEK), or ultra-high molecular weight polyethylene (UHMW-PE). 304 or 316L stainless steel has excellent machinability, ensuring that the flow guide component 11 has a precise arc angle of 15°±2°. Its high hardness helps to achieve a stable surface roughness of Ra≤0.8μm through surface polishing, thereby reducing local turbulence during liquor rinsing. Since baijiu is rich in organic acid components such as hexanoic acid and acetic acid, and online monitoring may be accompanied by automatic channel cleaning (such as backwashing with dilute alkaline or acidic solutions), the above-mentioned food-grade high-performance polymer materials have extremely high chemical inertness and acid and alkali resistance, as well as low surface energy and strong hydrophobicity. They can effectively prevent trace macromolecular substances (such as higher fatty acid esters) in the liquor from physical adsorption or scaling on the surface of the flow guide component, thereby maintaining a stable laminar flow state of the liquor within a laminar flow gap of 0.8mm-1.5mm for a long time. In one specific embodiment, the flow guiding component 11 of the present invention is made of food-grade 304 stainless steel. The surface roughness of the flow guiding component 11 is Ra≤0.8μm. When the wine to be tested flows through the flow guiding component 11, the liquid flow gradually transitions from a local disturbance state to a relatively stable laminar flow state, thereby reducing the adhesion and impact on the surface of the detection area.

[0043] The multi-array surface acoustic wave interferometry detection module 2 is a shear horizontal wave surface acoustic wave structure or a Love wave structure. The multi-array surface acoustic wave interferometry detection module 2 is positioned, pressed and packaged by the upper mounting component 4. For easy maintenance, the multi-array surface acoustic wave interferometry detection module 2 can adopt a replaceable structure. After replacement, it can be put back into use after one calibration with a standard wine sample.

[0044] The multi-array surface acoustic wave interferometry detection module 2 includes at least two detection subarrays and at least one reference subarray 23. The surface of the detection subarrays is provided with a sensitive recognition layer for selectively responding to target compounds in baijiu (Chinese liquor). The reference subarray 23 is used to output a compensation signal characterizing background interference factors. More specifically, the surface of the detection subarrays is provided with different sensitive recognition layers, including one or more of the following: a molecularly imprinted polymer layer for enriching ethyl hexanoate, a polyisobutylene layer for enhancing the overall response of esters, a polar polymer layer for adjusting the adsorption characteristics of acids, and an inert protective layer for forming a stable background reference. The reference subarray 23 is divided into a viscosity compensation region and a temperature drift calibration region. The viscosity compensation region is used to output a compensation signal related to fluid damping or matrix viscosity changes, and the temperature drift calibration region is used to output a calibration signal related to ambient temperature or detection chip temperature changes.

[0045] For the detection of free organic acids such as hexanoic acid, acetone, lactic acid, and butyric acid in baijiu (Chinese liquor), a sensitive recognition layer can be made using a chemotherapy polymer material—poly-4-vinylpyridine (P4VP)—based on hydrogen-bonded solutions. Specifically, 20 mg of 160,000 g / mol P4VP is weighed and completely dissolved in 10 mL of ultrapure anhydrous ethanol. The solution is then shaken for 30 minutes under anesthesia to prepare a homogeneous solution with a mass concentration of 0.2 wt%. Precise localization deposition is then performed using a micro / nano dispensing system or piezoelectric diode printing equipment, with the deposition location on the third acoustic delay line between the interdigitated surface acoustic wave (IDT) generators. The material is then placed in a vacuum drying oven and heat-annealed at 100°C for 1 hour to facilitate polymer chain rearrangement and eliminate internal damage and necrosis of the film, ultimately forming a dense, sensitive film with a uniform thickness controlled between 60 nm and 80 nm on the surface of the acoustic wave propagation region.

[0046] like Figure 1 , Figure 3As shown, in a specific embodiment, the multi-array surface acoustic wave interferometry detection module 2 of the present invention includes a first detection subarray 21, a second detection subarray 22, and a reference subarray 23. The first detection subarray 21 is used to enhance the response to a first type of target component in baijiu (Chinese liquor), the second detection subarray 22 is used to enhance the response to a second type of target component in baijiu, and the reference subarray 23 is used to characterize compensation signals for background interference factors. The surface of the first detection subarray 21 is provided with a sensitive recognition layer exhibiting differentiated adsorption characteristics for ethyl hexanoate and ethyl lactate; the surface of the second detection subarray 22 is provided with a sensitive recognition layer that is more sensitive to overall changes in ethyl acetate, acetic acid, and ethanol. The signal connection component 5 includes a first signal line 31, a second signal line 32, and a reference signal line 33. The first signal line 31 and the second signal line 32 are used to output electrical signals from the first and second detection subarrays, respectively, and the reference signal line 33 is used to output electrical signals from the reference subarray.

[0047] The interdigital transducers of the multi-array surface acoustic wave interferometry detection module 2 preferably have an interdigital width of 2μm-5μm and an interdigital spacing of 2μm-5μm. For different detection subarrays, different operating frequencies or different sensitive layer thicknesses can be set to form complementary array response matrices. The operating frequency can be selected in the range of 100MHz-500MHz. High-alcohol samples can preferentially use higher frequency bands, while low-alcohol samples or samples with strong water-based backgrounds can switch to more stable low-frequency bands.

[0048] When monitoring the components of baijiu using the above-mentioned online monitoring device based on multi-array surface acoustic wave interferometry, the following steps S1-S5 are included:

[0049] S1. The liquor to be tested is sampled online during the liquor production process and introduced into the flow channel 12 inside the flow base 1. The liquor to be tested passes through the buffer section and the stabilizing section in sequence and then passes through the detection section in a laminar flow state. At the end of the detection section, it is discharged to the outlet of the flow channel through the discharge section.

[0050] S2. The multi-array surface acoustic wave interferometry detection module 2 collects at least one response parameter among the frequency shift, phase shift, amplitude change and insertion loss change of the wine to be tested in the detection segment.

[0051] S3. The signal acquisition and calculation module performs at least one of alcohol compensation, temperature compensation, viscosity compensation, background compensation and drift correction on the response of the detection subarray based on the compensation signal output by the reference subarray 23, obtains the compensated feature quantity, and outputs the concentration feature vector of the target compound based on the compensated feature quantity and the preset calculation model.

[0052] The target compounds include one or more of the following: ethyl hexanoate, ethyl acetate, ethyl lactate, ethyl butyrate, ethanol, acetic acid, lactic acid, butyric acid, and hexanoic acid. The concentration feature vector includes two or more of the following: absolute concentration values, relative concentration ratios, slope of change, stability scores, and multi-array response coupling indices for multiple target compounds.

[0053] The signal acquisition and calculation module constructs a compensation model based on the output of the detection subarray and the output of the reference subarray 23, and combines the temperature parameter and alcohol content parameter to obtain at least one compensated feature quantity. Then, the concentration feature vector of the target compound is calculated based on the compensated feature quantity.

[0054] When establishing the compensation model, a set of standard wine samples covering different alcohol content, temperatures, and target component concentration ranges is prepared. The original response parameters of the corresponding detection subarray and reference subarray 23 are collected. Then, the known alcohol content, temperature, and target component concentration values ​​in the standard wine samples are used as calibration references to establish the compensation model. Taking the first detection subarray 21, the second detection subarray 22, and the reference subarray 23 as examples, the original response parameters corresponding to the first detection subarray 21, the second detection subarray 22, and the reference subarray 23 are collected, such as frequency offsets Δf1, Δf2, and Δfr, phase offsets Δp1, Δp2, and Δpr, and insertion loss changes Δl1, Δl2, and Δlr. Then, the known alcohol content, temperature, and target component concentration values ​​in the standard wine samples are used as calibration references to establish the compensation model. In a specific exemplary expression, the following compensated feature quantities can be formed first:

[0055]

[0056]

[0057] Where T represents the temperature parameter, A represents the alcohol content parameter, and a1, b1, c1, d1, a2, b2, c2, and d2 are calibration coefficients. These calibration coefficients can be obtained through multiple linear regression, partial least squares regression, or other fitting methods. The compensation features F1 and F2, along with other array features, are input into the concentration calculation model to obtain the concentration feature vector of the target compound.

[0058] The temperature drift calibration region of the reference subarray 23 uses an inert quartz delay line structure without a sensitive film covering. Since the temperature coefficient of quartz (TCF) is known, its high-frequency output phase shift Δpr is negatively correlated with the liquid phase temperature T with high linearity. The temperature calculation unit calculates the current liquid temperature T in real time using the formula T = k_temp·Δpr + T, where k_temp is the temperature calibration coefficient, determined by the temperature frequency coefficient (TCF) of the quartz substrate and the operating frequency, representing the temperature change corresponding to a unit phase shift. Similarly, the insertion loss change Δlr output by the viscosity compensation region of the reference subarray 23 is mainly determined by the fluid viscosity damping. Using a pre-stored family of three-dimensional fitting curves of "acoustic damping-alcohol content-temperature", under the condition of known temperature T, Δlr is substituted as the independent variable into the fitting formula A = f(Δlr, T) to calculate the estimated alcohol content A online. Thus, there is no need to set up an additional metal temperature probe or alcohol density meter in the flow channel, avoiding the introduction of secondary stray wave interference.

[0059] For samples with different alcohol content ranges or different aroma types, a first parameter group, a second parameter group, and a third parameter group can be established respectively. During system operation, the corresponding parameter group is automatically selected based on the fluid damping signal, temperature signal, and initial background signal output by the reference subarray 23.

[0060] To avoid the difficulty of covering all liquor samples with a fixed set of parameters, a preset parameter mapping matrix is ​​used in the signal acquisition and calculation module. The parameter mapping matrix associates the fluid damping range, alcohol content range or temperature range with the excitation frequency, filter coefficient, compensation coefficient and calculation model. When the output of the reference subarray 23 exceeds the preset threshold, it automatically switches to the parameter group that matches the current sample conditions.

[0061] In one specific implementation, when the reference subarray 23 detects that the change in fluid acoustic damping exceeds a preset threshold, or detects that the alcohol content estimate crosses a preset grading boundary, the system automatically performs parameter switching. For example, when the sample switches from the high alcohol content region to the low alcohol content region, the system can switch the excitation frequency from 200MHz to 120MHz and simultaneously call the compensation coefficient and solution model that match the range to reduce the adverse effects of water-based background enhancement on the accuracy of target component identification. To make adaptive switching of the excitation frequency physically feasible, a high-frequency detection channel (corresponding to at least one detection subarray, with a center frequency of 180MHz to 220MHz) and a low-frequency detection channel (corresponding to another detection subarray, with a center frequency of 100MHz to 130MHz) can be arranged side-by-side on the chip of the multi-array surface acoustic wave interferometric detection module 2. An RF switch is included in the signal acquisition and processing module. When the insertion loss change Δlr output from the viscosity compensation region of the reference subarray 23 exceeds a set threshold (indicating excessive viscosity of the solids in the liquor or large-molecule sugars), the RF switch automatically switches the excitation signal from the high-frequency detection channel to the low-frequency detection channel. This utilizes the characteristic of low-frequency surface acoustic waves having relatively low acoustic energy attenuation in dense liquid phases to maintain the signal-to-noise ratio at the receiver. In another embodiment, if the sample is liquor with an alcohol content above 50°, the first processing strategy is adopted; if the sample is liquor with an alcohol content below 50°, a third reference subarray 23 is added or a water-based background compensation parameter group is enabled to improve the recognition stability in low-alcohol liquor scenarios. By automatically switching the parameters as described above, this invention can be better adapted to online detection needs under different working conditions.

[0062] S4. The summary analysis module aggregates, compares, and identifies the deviations of concentration feature vectors from different time points, different wine segments of the same batch, or corresponding samples from multiple cellars. It calculates the slope and fluctuation amplitude of changes in adjacent time windows and outputs at least one of the following results: cellar abnormal trend identification, cellar deviation level, concentration feature vector deviation score, and downstream process risk warning.

[0063] Anomaly trend identification: The samples in the same group are grouped according to the cellar number, sampling time and wine section number. Then, anomaly trend identification is performed based on the change slope, fluctuation amplitude, offset and similarity coefficient of adjacent time windows. The cellar fermentation offset warning is output. The classification results include mild offset, moderate offset and severe offset.

[0064] Cellar Deviation Determination: Concentration feature vectors of multiple cellars in the same batch at the same sampling time point are constructed into a horizontal comparison matrix. Based on the degree of deviation from the batch mean, historical benchmark template or target style template, the cellar deviation level and outlier cellar identifier are output.

[0065] Concentration feature vector deviation score: The numerical score is calculated based on the Euclidean distance, Mahalanobis distance or cosine similarity between the preset standardized feature template and the real-time concentration feature vector. The score is used as the quantitative basis for blending screening, base wine classification and post-processing quality review.

[0066] Downstream process risk warning: The index of each target compound concentration or concentration combination in the current sample concentration feature vector is compared with the preset threshold range corresponding to the target product. When any index exceeds the corresponding threshold range, at least one action is triggered, such as re-inspection, diversion, temporary storage, blending and remixing, or pre-filling interception.

[0067] S5. The feedback output module outputs at least one of the following based on the output results of the summary analysis module: fermentation control suggestions, distillation and segmentation suggestions, base liquor diversion suggestions, storage risk warnings, blending and screening suggestions, or pre-bottling review instructions.

[0068] Specifically, the feedback output module communicates with fermentation control equipment, distillation segment valve group, base liquor diversion equipment, storage tank switching terminal, blending control terminal or quality management server to achieve real-time output or closed-loop control of analysis results.

[0069] This invention, based on the aforementioned online monitoring device and method for baijiu components using multi-array surface acoustic wave interferometry, monitored multiple samples and obtained the following results: Figure 4 The sample summary analysis curve is shown below. For the same pit, samples can be continuously collected at preset sampling intervals, forming a concentration feature vector sequence for that pit at multiple time points. The system calculates the slope, fluctuation amplitude, offset, and similarity of changes in adjacent time windows to obtain the degree of deviation between the trend baseline and the current abnormal fluctuation curve.

[0070] When an abnormal fluctuation curve deviates continuously from the trend baseline to a threshold within a preset time window, the system outputs a fermentation offset warning. This warning can be a simple anomaly marker or a graded result, such as mild offset, moderate offset, and severe offset.

[0071] For multiple fermentation pits in the same batch, the concentration feature vectors of the multiple pits at the same sampling time point can be constructed into a horizontal comparison matrix. Based on the distance between the matrix and the batch mean, historical template, or target style template, the pit deviation level and outlier pit identifier can be output. In this way, outlier pits can be identified in advance during fermentation without waiting for terminal testing for remedial action.

[0072] In the scenario of segmented distillation, the system can compare the real-time concentration feature vector with the preset distillation segment boundary template. When a certain switching threshold is reached, the system outputs a segmentation suggestion or valve position adjustment command to the distillation segmentation unit.

[0073] In the scenario of base liquor diversion or storage, if the style matching degree of a batch of samples is lower than the target threshold, or if the combination of its key components deviates significantly from the target range, the system will output diversion, temporary storage or re-inspection prompts to avoid unsuitable samples from directly entering the unified storage tank.

[0074] In the pre-blending screening scenario, the system can sort multiple tank samples and output the candidate sample set that best matches the target product style. During pre-filling verification, if the concentration feature vector of a current sample exceeds the allowable range, the system will issue a pre-filling interception or manual verification prompt.

[0075] To ensure long-term operational stability, this invention allows for a periodic calibration process. After a period of online operation, the system can automatically introduce standard wine samples or cleaning solutions for baseline correction. When the multi-array surface acoustic wave interferometry detection module 2 is replaced, it can be restored to working condition by performing a standard sample calibration once. If a continuous baseline shift is detected in the reference subarray 23, the system can prompt for recalibration or maintenance.

[0076] In a preferred embodiment, system parameter updates do not require rebuilding the entire model; instead, only local adjustments are needed to the compensation coefficients corresponding to the current alcohol content and temperature ranges. This facilitates engineering implementation and helps reduce equipment maintenance costs.

[0077] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. An online monitoring device for the components of baijiu (Chinese liquor) based on multi-array surface acoustic wave interferometry, characterized in that, include: The flow base (1) is provided with a flow channel (12), which includes a buffer section, a flow stabilizing section, a detection section and a discharge section connected in sequence. A flow guiding member (11) is located upstream of the detection section, and a laminar flow gap is formed between the flow guiding member (11) and the top inner surface of the detection section. A multi-array surface acoustic wave interferometry detection module (2) is hermetically sealed on a flow base (1) and corresponds to the detection section. The acoustic wave conduction surface of the multi-array surface acoustic wave interferometry detection module (2) faces the laminar flow gap. The multi-array surface acoustic wave interferometry detection module (2) includes at least two detection subarrays and at least one reference subarray. The surface of the detection subarray is provided with a sensitive recognition layer for selectively responding to target compounds in liquor. The reference subarray is used to output a compensation signal characterizing background interference factors. The signal acquisition and calculation module and the multi-array surface acoustic wave interferometry detection module (2) are connected to the signal acquisition and calculation module through the signal connection component (5). The signal acquisition and calculation module is used to acquire the frequency offset and phase offset of the array response and perform compensation calculation. The summary analysis module is connected to the signal acquisition and calculation module. The summary analysis module is used to aggregate and compare the concentration calculation results of multiple samples. The feedback output module is connected to the summary analysis module. The feedback output module is used to output warnings, suggestions or control commands based on the comparison results of the summary analysis module.

2. The online monitoring device for baijiu components based on multi-array surface acoustic wave interferometry according to claim 1, characterized in that: The multi-array surface acoustic wave interferometry detection module (2) is a shear horizontal wave surface acoustic wave structure or a Love wave structure.

3. The online monitoring device for baijiu components based on multi-array surface acoustic wave interferometry according to claim 2, characterized in that: The surface of the detection subarray is provided with different sensitive recognition layers, including one or more of the following: a molecularly imprinted polymer layer for enriching ethyl hexanoate, a polyisobutylene layer for enhancing the overall response of esters, a polar polymer layer for adjusting the adsorption characteristics of acids, and an inert protective layer for forming a stable background reference.

4. The online monitoring device for baijiu components based on multi-array surface acoustic wave interferometry according to claim 2, characterized in that: The reference subarray is divided into a viscosity compensation region and a temperature drift calibration region. The viscosity compensation region is used to output compensation signals related to fluid damping or matrix viscosity changes, while the temperature drift calibration region is used to output calibration signals related to ambient temperature or detection chip temperature changes.

5. The online monitoring device for liquor components based on multi-array surface acoustic wave interferometry according to claim 1, characterized in that: The flow guide component (11) is an asymmetrical convex arc-shaped flow guide plate. The arc surface deflection angle of the flow guide component (11) is 15°±2°. The vertical distance between the flow guide component (11) and the surface of the detection section is 0.8mm-1.5mm. The flow guide component (11) is made of food-grade 304 stainless steel. The surface roughness of the flow guide component (11) is Ra≤0.8μm.

6. A method for online monitoring of components in baijiu (Chinese liquor) based on multi-array surface acoustic wave interferometry, characterized in that, Includes the following steps: S1. The liquor to be tested is sampled online during the liquor production process and introduced into the flow channel (12) inside the flow base (1). The liquor to be tested passes through the buffer section and the steady flow section in sequence and then passes through the detection section in a laminar flow state. It is discharged through the discharge section at the end of the detection section. S2. The multi-array surface acoustic wave interferometry detection module (2) collects at least one of the following response parameters of the wine to be tested in the detection segment: frequency shift, phase shift, amplitude change and insertion loss change. S3. The signal acquisition and calculation module performs at least one of alcohol compensation, temperature compensation, viscosity compensation, background compensation and drift correction on the response of the detection subarray based on the compensation signal output by the reference subarray, obtains the compensated feature quantity, and outputs the concentration feature vector of the target compound based on the compensated feature quantity and the preset calculation model. S4. The summary analysis module aggregates, compares, and identifies the concentration feature vectors of samples from different time points, different wine sections of the same batch, or corresponding samples from multiple cellars. It calculates the slope and fluctuation amplitude of changes in adjacent time windows and outputs at least one of the following results: cellar abnormal trend identification, cellar deviation level, concentration feature vector deviation score, and downstream process risk warning. S5. The feedback output module outputs at least one of the following based on the output results of the summary analysis module: fermentation control suggestions, distillation and segmentation suggestions, base liquor diversion suggestions, storage risk warnings, blending and screening suggestions, or pre-bottling review instructions.

7. The method for online monitoring of liquor components based on multi-array surface acoustic wave interferometry according to claim 6, characterized in that: The target compounds include one or more of ethyl hexanoate, ethyl acetate, ethyl lactate, ethyl butyrate, ethanol, acetic acid, lactic acid, butyric acid, and hexanoic acid.

8. The method for online monitoring of liquor components based on multi-array surface acoustic wave interferometry according to claim 6, characterized in that: The signal acquisition and calculation module constructs a compensation model based on the output of the detection subarray and the output of the reference subarray, and combines the temperature parameter and alcohol content parameter to obtain at least one compensated feature quantity. Then, the concentration feature vector of the target compound is calculated based on the compensated feature quantity.

9. The method for online monitoring of liquor components based on multi-array surface acoustic wave interferometry according to claim 8, characterized in that: When establishing the compensation model, a set of standard wine samples covering different alcohol content, different temperatures, and different target component concentration ranges are prepared. The original response parameters of the detection subarray and reference subarray corresponding to the standard wine samples are collected. Then, the known alcohol content, temperature, and target component concentration values ​​in the standard wine samples are used as calibration references to establish the compensation model.

10. The method for online monitoring of baijiu components based on multi-array surface acoustic wave interferometry according to claim 9, characterized in that: The signal acquisition and processing module has a preset parameter mapping matrix. The parameter mapping matrix associates the fluid damping range, alcohol content range or temperature range with the excitation frequency, filter coefficient, compensation coefficient and processing model. When the output of the reference subarray exceeds the preset threshold, it automatically switches to the parameter group that matches the current sample operating conditions.

Images