White spirit distillation system with filtering structure and distillation method of white spirit distillation system

Abstract

The invention relates to the technical field of spirit distillation, in particular to a white spirit distillation system with a filtering structure and a distillation method.According to the white spirit distillation system with the filtering structure and the distillation method.Before heating is started, the liquid volume value and the alcohol concentration value of mash in a raw material tank are collected in real time, and traceable raw material parameters are formed through time labels and batch numbers; basic data with batch consistency is established for subsequent heat distribution, so that thermal power output and adjustment of steam pressure can be directly associated with a matching relationship between a tower plate liquid level and a temperature difference ratio, a matching result is converted into a staged power sequence by means of a genetic algorithm, and the flow of a main pipe and branch pipes is synchronously recorded, so that the flow of the main pipe and the branch pipes is calculated. According to the method, the difference between the top temperature and the bottom temperature and the branch flow ratio are subjected to deviation value calculation, valve opening adjustment is executed on an abnormal branch with the flow and temperature corresponding relation, and flow distribution considering stability and energy consumption performance is screened out through Bayesian optimization, so that steam resources obtain higher balance in multiple channels.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of liquor distillation, and in particular to a liquor distillation system with a filtering structure and a distillation method thereof. BACKGROUND

[0002] The technical field of liquor distillation is a process technology that vaporizes ethanol and aromatic components by heating fermented mash and obtains high-concentration liquor through condensation. The liquor purity, flavor stability, and product safety are improved by introducing impurity removal links between and after distillation and condensation.

[0003] A liquor distillation system with a filtering structure separates the vapor and the condensed liquor by solid-liquid and liquid-liquid separation through the filtering structure provided in the equipment during the liquor distillation process, reduces suspended particles, fusel oil, fine solid substances, and other components that affect the taste in the distillation link, and the purpose is to achieve preliminary purification of the liquor in the distillation link, reduce the pressure of the post-processing process, and at the same time improve the flavor purity and taste balance of the product and enhance the storage stability of the liquor body.

[0004] In the traditional liquor distillation technology, the raw material heating is usually operated with fixed heating power and constant steam pressure, and lacks pre-parameter recording and batch management required by subsequent processes, which makes it difficult to obtain data support that fits the characteristics of different batches of raw materials in the heat treatment stage, the heat transfer process is not stable enough for the components of the mash, and the steam distribution depends on the static setting of the equipment structure and the flow regulating valve. When the temperature difference and flow ratio of different positions of the distillation column change, it is difficult to correct the distribution strategy in time, which may cause flow rate fluctuations in some steam passages, affect the mass transfer in the column, and affect the liquor quality. The temperature control process usually uses single-point and a small number of measurement points as the basis for adjustment, and lacks linkage judgment for the comprehensive state between the temperature difference of multiple sections, the alcohol concentration, and the fusel oil components, which may cause uneven heating or ineffective inhibition of impurities. In the product collection stage, only liquid level or flow control is used for storage management, and there is no system record of the fluctuation trend of alcohol and impurity concentration, which may amplify the quality difference between batches in subsequent blending and storage, and weaken the guarantee ability of liquor quality stability. SUMMARY

[0005] The present application aims to solve the shortcomings in the prior art and proposes a liquor distillation system with a filtering structure and a distillation method thereof.

[0006] To achieve the above-mentioned purpose, the present application adopts the following technical solutions: a liquor distillation system with a filtering structure comprises: Raw material heating module: obtain the raw material tank body wine mash liquid volume value and alcohol concentration value through the sensor, the heating temperature controller outputs power in stages and adjusts the steam distributor pressure, the tower plate liquid level and the temperature difference proportion match the reflux pipeline opening degree, the genetic algorithm is used to adjust the power distribution according to the matching result, and the staged heating power sequence is obtained; Steam distribution module: based on the staged heating power sequence, the main pipe and branch flow are recorded synchronously, the top and bottom temperature difference is combined with the branch flow ratio, the branch valve is adjusted when the offset exceeds the limit, the stable distribution ratio is selected by using Bayesian optimization, and the preferred steam flow distribution group is generated; Temperature dynamic control module: based on the preferred steam flow distribution group, the temperature difference of each heating section is compared with the previous cycle value, the outlet wine alcohol concentration and fusel oil concentration are matched with the power adjustment amplitude, and the power time of adjacent high temperature section is delayed to obtain the temperature adjustment time sequence table; Filter switching module: based on the temperature adjustment time sequence table, the filter membrane assembly inlet and outlet flow rate are compared and the differential pressure change rate is determined, the combined value is switched when it exceeds the limit, the pore size flow rate differential pressure is optimally matched, and the filter operation parameter set is formed; Finished product collection module: based on the filter operation parameter set, the cumulative flow of the storage tank inlet is compared with the liquid level, the sampling alcohol and impurity concentration ratio is sorted, the ratio fluctuation and liquid surface amplitude are associated and recorded, and the finished product quality quantization record set is formed.

[0007] As a further scheme of the present application, the staged heating power sequence includes the heating power, heating duration and reflux pipeline opening degree of each stage, the preferred steam flow distribution group includes the steam flow, corresponding steam pressure and tower plate temperature difference range of each branch, the temperature adjustment time sequence table specifically includes the adjustment time, power adjustment amplitude and execution sequence of each section, the filter operation parameter set includes the filter membrane pore size, filter flow rate range and filter differential pressure range, and the finished product quality quantization record set specifically includes the alcohol concentration, impurity concentration and concentration ratio fluctuation range.

[0008] As a further scheme of the present application, the raw material heating module includes: Raw material parameter measurement submodule: obtain the raw material tank body wine mash liquid volume value and alcohol concentration value through the sensor, record and number archive the volume value and concentration value according to the time label, and generate a structured table corresponding to the record result and batch number to generate a raw material parameter set; Heating pressure matching submodule: based on the raw material parameter set, the heating temperature controller outputs power according to time period and adjusts the steam distributor outlet pressure to the target value, collects the tower plate liquid level height and stores it in pairs with the temperature difference value at the same position, writes the paired results into the matching record table, and generates a heating pressure matching table; The power distribution adjustment submodule: based on the heating pressure matching table, the time series of the temperature difference proportion of each heating section is compared with the return line opening degree in turn by using a genetic algorithm, and the matching numerical combination is selected, the selected data is converted into the power output value of the heating section, and the power distribution record table is updated to obtain the staged heating power sequence.

[0009] As a further scheme of the present application, the genetic algorithm, the temperature difference proportion value of each heating section is compared with the time series of the return line opening degree value in turn, the difference of each pair is calculated and a difference set is formed, the pairs exceeding the set threshold are filtered out in the difference set, the remaining pairs are sorted in order of difference from small to large, and a plurality of pairs are randomly selected from the sorted sequence for cross operation, then the cross-paired record is compared with the original pair set, the pair with smaller difference is retained, and the extraction, cross and retention operations are repeated until the difference converges to the set range, the temperature difference proportion and the return line opening degree value corresponding to the difference value are converted into the heating section power output value, and the converted power is updated to the staged heating power sequence.

[0010] As a further scheme of the present application, the steam distribution module comprises: The flow record submodule: based on the staged heating power sequence, the total flow of the steam main pipe and the flow of each branch are collected and stored according to the branch number, and the flow value and the time stamp are combined to generate a flow data set; The temperature-flow comparison submodule: based on the flow data set, the top and bottom temperatures of the distillation column are collected and cross-matched with the flow values of each branch, the temperature difference offset is calculated, and the abnormal branch is marked in the record table to generate a temperature difference-flow comparison table; The valve adjustment and screening submodule: based on the temperature difference-flow comparison table, the opening degree of the corresponding valve is adjusted in the branch with an out-of-limit temperature difference, the adjusted flow value and energy consumption data are recorded, the steam distribution stability and energy consumption performance of different opening degree combinations are predicted by using Bayesian optimization, and the flow combination with high stability and low energy consumption is selected from the prediction results to generate an optimal steam flow distribution group.

[0011] As a further aspect of the present invention, the Bayesian optimization involves adjusting the valve opening of branches with excessive temperature difference deviations, and recording multiple sets of flow rate and energy consumption data for each branch corresponding to different valve openings after the adjustment is completed. The recorded flow rate and energy consumption data are then paired to form a combination set. A comprehensive score is calculated for the combination set based on the branch flow stability index and energy consumption index. The score results are sorted according to the score value. Several combinations are selected as candidate sets based on the score. New opening parameters are added sequentially to the candidate sets, and the comprehensive score is updated in real time. The updated score is compared with the historical score, and combinations with higher scores are retained. This evaluation, comparison, and retention process is repeated until the combination score no longer significantly improves. Based on the score, the preferred steam flow distribution group is determined.

[0012] As a further aspect of the present invention, the temperature dynamic control module includes: Temperature difference comparison submodule: Based on the optimized steam flow distribution group, the temperature above the liquid surface of the current tray measured in each heating section is compared with the temperature value at the same position in the previous cycle one by one. The alcohol content and fusel oil content measured in the outlet wine are respectively associated with the corresponding temperature difference data. The temperature difference value and the component content value are combined into a pair record to generate a temperature difference component mapping set. Power timing adjustment submodule: Based on the temperature difference component mapping set, calculate the corresponding power increase or decrease value for each pair of data, arrange the power change values ​​of each heating section in sequence according to the temperature section sequence, so that the power adjustment time of the high temperature section lags behind the power adjustment time of the low temperature section, and obtain the temperature adjustment timing table by forming a recording unit through the delay time and power value.

[0013] As a further aspect of the present invention, the filtering switching module includes: Flow velocity and pressure difference monitoring submodule: Based on the temperature adjustment time table, the instantaneous liquid flow rate at the inlet and outlet of the filter membrane module is measured and the ratio is calculated. At the same time, the pressure difference value at the corresponding position is recorded. The two sets of data are compared in amplitude according to the same time sequence. The ratio change value and the pressure difference change value are combined into a unified data unit to generate a flow velocity and pressure difference combination set. Filtration path selection submodule: Based on the set of flow rate and pressure difference combinations, compare the combined values ​​with preset limits to determine whether the limits are exceeded, and switch the conduction path of the filter switching valve when the limits are exceeded. Set the filter membrane pore size, the flow rate of the liquid and the pressure difference generated in the switching state as a set of matching parameters to form a set of filtration operation parameters.

[0014] As a further aspect of the present invention, the finished product collection module includes: Liquid volume and level comparison submodule: Based on the filter operation parameter set, record the cumulative inlet flow rate and corresponding liquid level height of the finished product storage tank over a fixed time period, obtain the difference curve by subtracting the two sets of values, and arrange them into a difference table according to the measurement time order to generate a liquid volume and height association set; Concentration fluctuation recording submodule: Based on the liquid volume high correlation set, the alcohol concentration and impurity concentration of the wine sample in the storage tank are collected during each measurement period. The ratio of the two concentration values ​​is arranged in chronological order, and the fluctuation range of the ratio and the change of liquid level are used to form corresponding data records, forming a quantitative record set of finished product quality.

[0015] A method for distilling baijiu (Chinese liquor) with a filtration structure, wherein the method is performed based on the aforementioned baijiu distillation system with a filtration structure, and includes the following steps: S1: The volume and alcohol concentration of the mash in the raw material tank are obtained by the sensor. The power output of the heating temperature controller and the pressure of the steam distributor are adjusted in segments. The change sequence of the liquid level on the tray and the temperature difference between the upper and lower trays are measured. The sequence is compared with the power segment sequence. The power distribution is adjusted by the genetic algorithm to generate a segmented heating power sequence. S2: Based on the segmented heating power sequence, record the main flow rate and the flow rate of multiple branches, compare the temperature difference between the top and bottom with the branch flow rate, calculate the offset, adjust the valve opening of the over-limit branch, use Bayesian optimization to screen the stable ratio, and generate the preferred steam flow distribution group. S3: Based on the preferred steam flow distribution group, compare the current temperature difference of each heating section with the previous cycle value, match the difference with the alcohol concentration value and fusel oil concentration value of the outlet wine, adjust the power amplitude of the section and the delay time of the adjacent high temperature section to obtain the temperature adjustment timing table. S4: Based on the temperature adjustment timing table, compare the inlet and outlet flow rate values ​​of the filter membrane module and measure the differential pressure change rate, combine them into a joint value and determine its relationship with the limit value, perform switching for cases exceeding the limit and sort the pore size conditions to generate a set of filter operation parameters; S5: Based on the filter operation parameter set, record the cumulative flow rate at the tank inlet and compare it with the liquid level, sample and measure the ratio of alcohol concentration to impurity concentration, sort by time and record the correlation between fluctuation and liquid level fluctuation, and generate a quantitative record set of finished product quality.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, the liquid volume and alcohol concentration of the mash in the raw material tank are collected in real time before heating begins. Traceable raw material parameters are formed by time tags and batch numbers, which establishes basic data with batch consistency for subsequent heat distribution. This allows the adjustment of heat power output and steam pressure to be directly related to the matching relationship between tray liquid level and temperature difference ratio. The matching result is transformed into a phased power sequence by using a genetic algorithm to achieve dynamic adaptation of heat and steam distribution during the heating stage. In this invention, by synchronously recording the flow rates of the main pipe and the branches, the deviation between the temperature difference between the top and bottom and the flow rate ratio of the branches is calculated. For branches with abnormal flow rate and temperature correspondence, the valve opening is adjusted. Bayesian optimization is used to select a flow distribution that takes into account both stability and energy consumption performance, so that the steam resources can achieve a higher balance in multiple channels. In this invention, the real-time temperature difference changes of each heating zone are compared with the previous cycle data, and the power amplitude and the delay sequence of the high temperature zone are adjusted in combination with the alcohol concentration and fusel oil concentration of the outlet wine, so that the wine undergoes a more reasonable heat exchange process in different zones. In this invention, the flow rate difference and pressure difference change rate between the inlet and outlet of the filter membrane assembly are combined for judgment. When the difference exceeds the set range, the flow rate and pore size are switched to the optimal matching state, so that the removal process of suspended solids and impurities can be maintained under the best conditions. Attached Figure Description

[0017] Figure 1 This is a system flowchart of the present invention; Figure 2 This is a schematic diagram of the method steps of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] Example 1 Please see Figure 1 The present invention provides a technical solution: a liquor distillation system with a filtration structure includes: Raw material heating module: The volume and alcohol concentration of the mash in the raw material tank are obtained by sensors. The heating temperature controller outputs power in stages and adjusts the pressure of the steam distributor. The opening of the reflux pipeline is matched with the ratio of the tray liquid level and temperature difference. The power distribution is adjusted according to the matching result by a genetic algorithm to obtain a staged heating power sequence. Steam distribution module: Based on the phased heating power sequence, the main pipe and branch flow are recorded synchronously. The temperature difference between the top and bottom is compared with the branch flow ratio. When the deviation exceeds the limit, the branch valve is adjusted. Bayesian optimization is used to screen the stable distribution ratio and generate the preferred steam flow distribution group. Temperature dynamic control module: Based on the optimized steam flow distribution group, the temperature difference of each heating section is compared with the previous cycle value, the power adjustment range is matched with the alcohol concentration and fusel oil concentration of the outlet wine, and the power time of the adjacent high temperature section is delayed to obtain the temperature adjustment timing table. Filter switching module: Based on the temperature adjustment timing table, the inlet and outlet flow rates of the filter membrane module are compared and the differential pressure change rate is measured. When the combined value exceeds the limit, the filter switching valve path is switched, and the optimal matching of orifice diameter, flow rate and differential pressure is selected to form a set of filter operation parameters. Finished product collection module: Based on the filter operation parameter set, the cumulative flow rate and liquid level at the tank inlet are compared, the ratio of sampled alcohol and impurity concentrations is sorted, and the ratio fluctuation is correlated with the liquid level amplitude to form a quantitative record set of finished product quality.

[0020] The phased heating power sequence includes the heating power, heating duration, and reflux line opening for each phase. The preferred steam flow distribution group includes the steam flow rate of each branch, the corresponding steam pressure, and the temperature difference range of the trays. The temperature adjustment sequence table specifically includes the adjustment time, power adjustment range, and execution order for each section. The filtration operation parameter set includes the filter membrane pore size, filtration flow rate range, and filtration pressure difference range. The finished product quality quantification record set specifically includes the alcohol concentration, impurity concentration, and concentration ratio fluctuation range.

[0021] The raw material heating module includes: Raw material parameter measurement submodule: The volume and alcohol concentration of the mash in the raw material tank are obtained through sensors. The volume and concentration values ​​are recorded and archived with time tags and numbered. The recorded results are matched with the batch number to generate a structured table and generate a raw material parameter set. Heating pressure matching submodule: Based on the raw material parameter set, the heating temperature controller outputs power according to time periods and adjusts the steam distributor outlet pressure to the target value, collects the tray liquid level height and pairs it with the temperature difference at the same position and stores it, writes the pairing result into the matching record table, and generates a heating pressure matching table. Power distribution adjustment submodule: Based on the heating pressure matching table, a genetic algorithm is used to compare the temperature difference ratio of each heating section with the time series of the return pipe opening, select the numerical combinations that can match each other, convert the selected data into the power output value of the heating section, update the power distribution record table, and obtain the staged heating power sequence. Raw material parameter measurement submodule: Based on the liquid level sensor and alcohol concentration sensor connected to the raw material tank, the 4-20mA current signal output by the liquid level sensor is converted by AD using the Modbus RTU protocol. The conversion parameters include sampling frequency of 1Hz, quantization bits of 16 bits, and a range of 0 to 2000 liters. The RS-485 data frame output by the alcohol concentration sensor is read using the Modbus RTU protocol. The command parameters include function code 03, starting register address 0000, and number of registers of 2. The data is parsed into an alcohol concentration value. An additional acquisition time stamp is added in the format of year, month, day, hour, minute, and second, and the numbering rule is the first six digits of the batch number plus the last three digits of the serial number. The liquid volume value and alcohol concentration value are recorded sequentially to a data file. The file fields include number, time stamp, liquid volume value, alcohol concentration value, and batch number, generating a raw material parameter set. Heating pressure matching submodule: Based on the raw material parameter set, a PID control algorithm is used to adjust the power output of the heating temperature controller. The input parameters include the target temperature value, the current temperature value, the proportional coefficient 2.5, the integral coefficient 0.1, the derivative coefficient 0.05, and the control cycle 1 second. The output is converted into the temperature controller's PWM duty cycle. The steam distributor outlet pressure is adjusted using a proportional valve pressure adjustment method. The adjustment command includes a step size of 5% and an execution cycle of 0.5 seconds. The liquid level sensor data of the tray is read to obtain the liquid level height value, and the K-type thermocouple data is read to obtain the corresponding temperature value. The two types of data are paired according to the location number to form a data item, which is stored in a JSON key-value structure. The key is the location number, and the value is an array containing the liquid level height value and the temperature difference value. The data is recorded in the matching record table to generate a heating pressure matching table. Power Allocation Adjustment Submodule: Based on the heating pressure matching table, a genetic algorithm is used to match the heating section temperature difference ratio sequence with the reflux pipe opening time sequence. The population size is set to 100, the chromosome length is equal to the number of heating sections, the crossover probability is 0.8, the mutation probability is 0.1, and the maximum number of iterations is 500. When initializing the population, each gene locus is defined as the target power ratio value of the heating section. A roulette wheel selection operation is performed to obtain parent chromosome pairs. A single-point crossover operation is performed at randomly selected gene locus positions. After the crossover, a mutation operation is performed on the chromosome to increase or decrease the power ratio value of a certain gene locus by 5%. After the iteration, the chromosome with the highest matching degree is output. The power ratio value of the gene locus is multiplied by the rated power of the section to obtain the power output value, which is written into the power allocation record table to obtain the staged heating power sequence.

[0022] The genetic algorithm compares the temperature difference ratio of each heating section with the time series of return pipe opening values ​​in the sequence one by one, calculates the difference of each pair and forms a difference set, filters out the pair exceeding a set threshold in the difference set, sorts the remaining pair according to the difference from smallest to largest, randomly selects multiple pairs in the sorted sequence for crossover operation, compares the difference of the crossover pair record with the original pair set, retains the pair with smaller difference, repeats the sampling, crossover and retention operation until the difference converges to the set range, converts the temperature difference ratio corresponding to the difference value and the return pipe opening value into the power output value of the heating section, and updates the staged heating power sequence with the converted power. Genetic algorithm, according to the formula:

[0023] in: This represents the overall fitness value. This indicates the number of heating sections in the distillation system that participate in the control of the heating and filtration processes. This represents the temperature difference ratio of the i-th heating section. This indicates the opening degree of the return pipe, which is the percentage of the opening degree of the pipe corresponding to the i-th heating section at a specific time point. This represents the relative humidity of the i-th heating section. Indicates a humidity reference value. This represents the specific heat capacity of the fluid in the i-th heating section. Indicates the target specific heat capacity. This represents the pipeline pressure in the i-th heating section. Indicates the safe operating pressure threshold. Indicates the temperature difference matching weight. Indicates humidity matching weight. Indicates the specific heat capacity matching weight. Indicates the pressure matching weight; Execution process: First, collect the actual temperature of each heating section according to the heating pressure matching table of the distillation process and calculate the temperature difference ratio. Then, the percentage of the opening of the return pipeline at each time point is collected to obtain... The squared difference between the temperature difference ratio and the opening degree of the return pipe is calculated and multiplied by the temperature difference matching weight. To determine the impact of temperature difference deviation on overall fitness, the relative humidity of each heating zone was subsequently measured using a humidity sensor. And compared with the preset humidity reference value Compare the squared differences and multiply the results by the humidity matching weights. To reflect the impact of humidity control on the accuracy of distillation and separation, the specific heat capacity of the liquor mash in each heating zone is calculated using a fluid composition analysis module. And compared with the target specific heat capacity obtained by experimental calibration Compare the squared differences and multiply them by the specific heat capacity matching weight. To reflect the impact of changes in fluid properties on heating energy consumption, pressure sensors are then used to detect the pipeline pressure in each heating zone. and with safe operating pressure threshold Compare the squared differences and multiply the results by the pressure matching weights. To ensure system operational safety, the entire weighting coefficient , , , The fitness level is determined by performing multiple Monte Carlo random sampling simulations on historical system operating data and combining this with the output fitness value, to ensure that the deviation of each parameter affects the overall fitness. The influence conforms to the optimal control objective. When the weighted deviations of all heating sections are summed and the negative sign is taken as the fitness value, the genetic algorithm maximizes the fitness value in the process of multiple generations of iteration, thereby obtaining the optimal combination of temperature difference, humidity, specific heat capacity and pressure of the heating section, and converting it into the corresponding power output value to update the power allocation record table to form the final staged heating power sequence.

[0024] The steam distribution module includes: Flow recording submodule: Based on the staged heating power sequence, it collects the total flow of the steam main and the flow of each branch and stores them according to the branch number. It merges the flow value with the timestamp to generate a flow dataset. Temperature-flow comparison submodule: Based on the flow dataset, it collects the temperature at the top and bottom of the distillation column and cross-matches it with the flow values ​​of each branch, calculates the temperature difference offset, marks abnormal branches in the record table, and generates a temperature-flow comparison table. Valve adjustment and screening submodule: Based on the temperature difference flow rate comparison table, adjust the opening of the corresponding valve in the branch where the temperature difference deviation exceeds the limit, record the adjusted flow rate value and energy consumption data, use Bayesian optimization to predict the steam distribution stability and energy consumption performance of different opening combinations, select the flow rate combination with high stability and low energy consumption from the prediction results, and generate the preferred steam flow rate distribution group. Flow recording submodule: Based on the phased heating power sequence, the Modbus TCP data acquisition method is used to read the instantaneous total flow data of the vortex flowmeter installed on the steam main pipe. The read command is function code 04, starting register address 0000, register quantity 2, and sampling frequency 1Hz. The cumulative flow register value output by the flowmeter is converted according to the proportional coefficient. The Modbus TCP data acquisition method is used to read the instantaneous flow value of each branch electromagnetic flowmeter according to the branch number. The read command is function code 04, and the register address is the branch number multiplied by the address offset. The acquired instantaneous flow value is appended with a system timestamp in the format of year, month, day, hour, minute, and second. The branch number, instantaneous flow value, total flow value, and timestamp are recorded into a CSV data file. The file fields include branch number, instantaneous flow value, total flow value, and timestamp, generating a flow dataset. Temperature-flow comparison submodule: Based on the flow dataset, a multi-channel temperature acquisition and control method is used to collect temperature data from the K-type thermocouples installed at the top and bottom of the distillation column. The sampling frequency is 1Hz and the conversion accuracy is 0.1℃. The top and bottom temperature data are archived by time label. The flow value is obtained at the same time point and cross-matched using the branch number index function. The temperature difference threshold comparison method is used to calculate the difference between the top and bottom temperatures. The temperature difference offset comparison threshold is set to 5 degrees Celsius. For branches with excessive offset, the abnormal status is marked in the abnormal flag field of the record table. The record table fields are timestamp, branch number, top temperature, bottom temperature, flow value, temperature difference value, and abnormal status, generating a temperature-flow comparison table. Valve regulation and screening submodule: Based on the temperature difference flow rate reference table, a step valve opening control method is used to adjust the valve opening in branches where the temperature difference deviation exceeds the limit. The adjustment command is an opening change step of 5%, an adjustment interval of 30 seconds, and a maximum allowable adjustment range of 50%. The instantaneous flow rate and energy consumption data after each opening adjustment are recorded. The energy consumption data is collected by connecting to a power meter at a sampling frequency of 1Hz. The Bayesian optimization method is used to predict the steam distribution stability and energy consumption performance under different valve opening combinations. The parameter settings include an initial sample size of 20, an iteration count of 100, a Gaussian kernel kernel, and an expected improvement acquisition function. The predicted valve opening combinations are sorted according to stability score and energy consumption score. The combination with the highest stability score and the lowest energy consumption score is selected to generate the optimal steam flow distribution group.

[0025] Bayesian optimization involves adjusting the valve opening of branches with excessive temperature difference deviations. After the adjustment, multiple sets of flow rate and energy consumption data for each branch with different valve openings are recorded. The recorded flow rate and energy consumption data are paired to form a combination set. A comprehensive score is calculated for the combination set based on the branch flow stability index and energy consumption index. The score results are sorted according to the score value. Several combinations are selected as candidate sets based on the score. New opening parameters are added to the candidate set in turn, and the comprehensive score is updated in real time. The updated score is compared with the historical score, and the combination with the higher score is retained. The evaluation, comparison, and retention steps are repeated until the combination score no longer improves significantly. The optimal steam flow distribution group is determined based on the score. Bayesian optimization, according to the formula:

[0026] in: This indicates the valve opening combination in the baijiu distillation system. The expected improvement value is as follows. This represents a vector representing the combination of valve openings in each branch of the baijiu distillation system. Indicates the valve opening combination Predicted steam distribution stability score under the given conditions This indicates the highest steam distribution stability score obtained from the currently evaluated valve opening combinations. Indicates the exploration parameters, Indicates the valve opening combination Down, The cumulative distribution function represents the standard normal distribution. The probability density function representing the standard normal distribution; Execution process: First, calculate the temperature difference offset of each branch based on the temperature difference flow rate comparison table and determine whether it exceeds the allowable range. For branches with excessive temperature difference offset, obtain the valve position and prepare for opening adjustment. Combine the initial opening percentage of valves in each branch to form a valve opening combination vector. Then, the stability score is obtained by predicting the steam distribution stability of the opening combination under distillation operating conditions. Simultaneously calculate the standard deviation of the uncertainty of the predicted values. Extract the steam distribution stability score of all evaluated valve opening combinations from historical operating data and determine their maximum value. Set exploration parameters To control the balance between exploration and exploitation strategies, and then... , , Substituting into the formula, the expected improvement value is calculated. As a criterion, the combination with the largest expected improvement value is selected from multiple candidate valve opening combinations as the preferred target. The valve opening of each branch is adjusted according to the steam flow distribution scheme corresponding to the preferred combination. The adjusted flow and energy consumption data are measured in real time and updated to the flow and energy consumption record table for use in the next round of optimization, thereby obtaining a preferred steam flow distribution group with high stability and low energy consumption.

[0027] The temperature dynamic control module includes: Temperature difference comparison submodule: Based on the optimized steam flow distribution group, the temperature above the liquid surface of the current tray measured in each heating section is compared with the temperature value at the same position in the previous cycle one by one. The alcohol content and fusel oil content measured in the outlet wine are respectively associated with the corresponding temperature difference data. The temperature difference value and the component content value are combined into a pair record to generate a temperature difference component mapping set. Power timing adjustment submodule: Based on the temperature difference component mapping set, calculate the corresponding power increase or decrease value for each pair of data, arrange the power change value of each heating section in sequence according to the temperature section sequence, so that the power adjustment time of the high temperature section lags behind the power adjustment time of the low temperature section, and obtain the temperature adjustment timing table by forming a recording unit through the delay time and power value. Temperature difference comparison submodule: Based on the optimized steam flow distribution group, the time-synchronized temperature comparison method is used to synchronously compare the temperature above the liquid surface of the current tray measured in each heating section with the temperature value at the same position in the previous acquisition cycle. During the comparison process, the heating section number is used as the index order, and the current temperature value and the temperature value of the previous cycle are arranged one by one according to the time tag. The output data frame of the alcohol content sensor of the outlet wine is read using the serial port acquisition and analysis method. The parameters are set as baud rate 9600, data bits 8 bits, stop bits 1 bit, and parity check none. After parsing, the alcohol content value is obtained. The value output by the fusel oil content sensor is read in the same way. The alcohol content value and fusel oil content value are associated with different fields of the current temperature difference data entry, and stored as a record unit according to the acquisition time tag order. The record unit includes heating section number, current temperature value, previous cycle temperature value, temperature difference value, alcohol content value, and fusel oil content value. All record units are combined into a data file to generate a temperature difference component mapping set. Power timing adjustment submodule: Based on the temperature difference component mapping set, the partitioned power adjustment sequence generation method is used to calculate the corresponding power increase or decrease value for each paired record. The calculation rule for the power increase or decrease value is to read the corresponding power change value according to the fixed adjustment range table set by the temperature difference range, and arrange the power change value in the order of the temperature segment number to which the heating segment belongs. A delay parameter is set so that the power adjustment execution time of the high temperature segment is later than the power adjustment execution time of the low temperature segment. The delay parameter is in seconds and can be set from 0 to 120 seconds. The power adjustment value and the corresponding delay time are combined into a recording unit. The recording unit includes the heating segment number, the power adjustment value, and the delay time. All recording units are written into the data file in the order of the segment number to obtain the temperature adjustment timing table.

[0028] The filter switching module includes: Flow velocity and differential pressure monitoring submodule: Based on the temperature adjustment time table, the instantaneous liquid flow rate at the inlet and outlet of the filter membrane module is measured and the ratio is calculated. At the same time, the differential pressure value at the corresponding location is recorded. The two sets of data are compared in amplitude according to the same time sequence. The ratio change value and the differential pressure change value are combined into a unified data unit to generate a flow velocity and differential pressure combination set. Filtration path selection submodule: Based on the combination set of flow rate and pressure difference, it compares the combined value with the preset limit to determine whether the limit is exceeded. If the limit is exceeded, it switches the conduction path of the filter switching valve. The filter membrane pore size, the flow rate of the liquid and the pressure difference generated in the switching state are set as a set of matching parameters to form a set of filter operation parameters. The flow velocity and differential pressure monitoring submodule, based on a temperature adjustment timetable, uses a dual-channel instantaneous flow acquisition method to read data from the liquid flow sensors at the inlet and outlet of the filter membrane module. The acquisition command is set to a sampling frequency of 1Hz, a continuous data output mode, and a decimal numerical stream data format. After reading the inlet flow value, the outlet flow value is immediately read through a buffer queue, and the two sets of data are stored in memory in the order of acquisition time at the corresponding index positions. The inlet flow value and the outlet flow value are calculated using a data ratio calculation method. The differential pressure sensor acquisition method records the differential pressure value at the corresponding position. The differential pressure sensor reading command parameters are a sampling frequency of 1Hz, a range of 0 to 1MPa, and an output accuracy of 0.001MPa. The ratio value and the differential pressure value are compared in amplitude according to the same time series. A synchronous sequence comparison method is used to generate a recording unit containing the acquisition time, the change in the inlet flow ratio, and the change in differential pressure. All recording units are combined into a structured table file to generate a flow velocity and differential pressure combination set. The filtration path selection submodule, based on a combination of flow rate and differential pressure, uses a limit comparison method to compare the combined value of each data unit with the preset limit. The limit parameter file is stored in CSV format, with two columns: ratio limit and differential pressure limit. When the combined value exceeds either limit, a switching path command is sent to the filter switching valve controller via an electronic valve switching method. The switching command is a digital output high level for 200 milliseconds. After the switching is completed, the filter membrane pore size parameter in the switching state is immediately read. The parameter is read from the electronic tag attached to the filter membrane frame using an RFID tag reader. The read command baud rate is 19200 and the data block address is 0002. The flow rate data of the liquid passing through in the switching state is acquired using a flow sensor, and the differential pressure data generated is acquired using a differential pressure sensor. The filter membrane pore size, flow rate value, and differential pressure value are combined into a matching parameter unit. All matching parameter units are stored in a tabular data file indexed by the branch number, forming a filter operation parameter set.

[0029] The finished product collection module includes: Liquid volume and level comparison submodule: Based on the filter operation parameter set, it records the cumulative inlet flow rate and corresponding liquid level height of the finished product storage tank over a fixed time period. It obtains the difference curve by subtracting the two sets of values ​​and arranges them into a difference table according to the measurement time order, generating a liquid volume and level association set. Concentration fluctuation recording submodule: Based on the liquid volume high correlation set, the alcohol concentration and impurity concentration of the wine sample in the storage tank are collected at each measurement time period. The ratio of the two concentration values ​​is arranged in chronological order, and the fluctuation range of the ratio and the change of liquid level are used to form corresponding data records, forming a quantitative record set of finished product quality. Liquid volume and level comparison submodule: Based on the filter operation parameter set, a timed data acquisition method is used to record the inlet cumulative flow value of the finished product storage tank at a set acquisition time interval of 60 seconds. The inlet cumulative flow is read by an electromagnetic flow meter. The command parameters include sampling frequency of 1Hz, range of 0 to 20000 liters, and quantization accuracy of 0.1 liters. At the same time, the liquid level height value at the corresponding time is collected using an ultrasonic liquid level measurement method. The measurement command parameters include sampling frequency of 1Hz, range of 0 to 5000 mm, and quantization accuracy of 1 mm. The cumulative flow value and liquid level height value are matched in the memory with the time tag as the index. The difference sequence generation method is used to subtract the cumulative flow value and liquid level height value at the same time point in turn and record the difference. The difference data are arranged into a difference table according to the measurement time order to generate a liquid volume and height association set. Concentration Fluctuation Recording Submodule: Based on the high correlation set of liquid volume, a two-component concentration measurement method is used to collect the alcohol concentration and impurity concentration of the wine sample in the storage tank at each measurement time period. The alcohol concentration is collected by an infrared spectrometer, and the acquisition command includes an integration time of 200 milliseconds, 10 scans, and output as a numerical percentage. The impurity concentration is collected by a high-performance liquid chromatograph, and the acquisition command includes an injection volume of 5 μL, a mobile phase flow rate of 1 mL / min, and a detection wavelength of 210 nm. The two concentration values ​​are divided sequentially using a numerical ratio processing method to obtain the concentration ratio. The concentration ratio values ​​are arranged into a recording sequence in chronological order. The fluctuation range value corresponding to the concentration ratio value and the liquid level change amplitude value are bound and stored in the same recording unit. The recording unit fields include time tag, concentration ratio value, ratio fluctuation range, and liquid level change amplitude, forming a quantitative record set of finished product quality.

[0030] A method for distilling baijiu (Chinese liquor) with a filtration structure, the method being performed based on the aforementioned baijiu distillation system with a filtration structure, includes the following steps: S1: The volume and alcohol concentration of the mash in the raw material tank are obtained by the sensor. The power output of the heating temperature controller and the pressure of the steam distributor are adjusted in segments. The change sequence of the liquid level on the tray and the temperature difference between the upper and lower trays are measured. The sequence is compared with the power segment sequence. The power distribution is adjusted by the genetic algorithm to generate a segmented heating power sequence. S2: Based on the segmented heating power sequence, record the main flow rate and the flow rate of multiple branches, compare the temperature difference between the top and bottom with the branch flow rate, calculate the offset, adjust the valve opening of the over-limit branch, use Bayesian optimization to screen the stable ratio, and generate the preferred steam flow distribution group. S3: Based on the preferred steam flow distribution group, the current temperature difference of each heating section is compared with the previous cycle value. The difference is matched with the alcohol concentration value and fusel oil concentration value of the outlet wine. The power amplitude of the section and the delay time of the adjacent high temperature section are adjusted to obtain the temperature adjustment timing table. S4: Based on the temperature adjustment timing table, compare the inlet and outlet flow rates of the filter membrane module and measure the differential pressure change rate, combine them into a joint value and determine its relationship with the limit value. For cases exceeding the limit, perform switching and sort the pore size conditions to generate a set of filter operation parameters. S5: Based on the filter operation parameter set, record the cumulative flow rate at the tank inlet and compare it with the liquid level, sample and measure the ratio of alcohol concentration to impurity concentration, sort by time and record the correlation between fluctuation and liquid level fluctuation, and generate a quantitative record set of finished product quality.

[0031] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A liquor distillation system with a filtration structure, characterized in that, The system includes: Raw material heating module: The volume and alcohol concentration of the mash in the raw material tank are obtained by sensors. The heating temperature controller outputs power in stages and adjusts the pressure of the steam distributor. The opening of the reflux pipeline is matched with the ratio of the tray liquid level and temperature difference. The power distribution is adjusted according to the matching result by a genetic algorithm to obtain a staged heating power sequence. Steam distribution module: Based on the phased heating power sequence, the main pipe and branch flow are recorded synchronously. The temperature difference between the top and bottom is compared with the branch flow ratio. When the deviation exceeds the limit, the branch valve is adjusted. Bayesian optimization is used to screen stable distribution ratios and generate an optimal steam flow distribution group. Temperature dynamic control module: Based on the preferred steam flow distribution group, the temperature difference of each heating section is compared with the previous period value, the power adjustment range of the outlet wine alcohol concentration and fusel oil concentration is matched, the power time of adjacent high temperature sections is delayed, and a temperature adjustment timing table is obtained. Filter switching module: Based on the temperature adjustment timing table, the inlet and outlet flow rates of the filter membrane module are compared and the pressure difference change rate is measured. When the combined value exceeds the limit, the filter switching valve path is switched, and the optimal matching of orifice flow rate and pressure difference is selected to form a set of filter operation parameters. Finished product collection module: Based on the set of filtration operation parameters, the cumulative flow rate at the tank inlet is compared with the liquid level, the ratio of sampled alcohol to impurity concentration is sorted, and the ratio fluctuation is correlated with the liquid level amplitude to form a set of quantitative records of finished product quality.

2. The liquor distillation system with a filtration structure according to claim 1, characterized in that, The phased heating power sequence includes the heating power, heating duration, and reflux line opening for each phase. The preferred steam flow distribution group includes the steam flow rate, corresponding steam pressure, and tray temperature difference range for each branch. The temperature adjustment timing table specifically includes the adjustment time, power adjustment range, and execution order for each section. The filtration operation parameter set includes the filter membrane pore size, filtration flow rate range, and filtration pressure difference range. The finished product quality quantification record set specifically includes the alcohol concentration, impurity concentration, and concentration ratio fluctuation range.

3. The liquor distillation system with a filtration structure according to claim 1, characterized in that, The raw material heating module includes: Raw material parameter measurement submodule: The volume and alcohol concentration of the mash in the raw material tank are obtained through sensors. The volume and concentration values ​​are recorded and archived with time tags and numbered. The recorded results are matched with the batch number to generate a structured table and generate a raw material parameter set. Heating pressure matching submodule: Based on the raw material parameter set, the heating temperature controller outputs power according to time periods and adjusts the steam distributor outlet pressure to the target value, collects the tray liquid level height and pairs it with the temperature difference at the same position for storage, writes the pairing result into the matching record table, and generates a heating pressure matching table. Power distribution adjustment submodule: Based on the heating pressure matching table, a genetic algorithm is used to compare the temperature difference ratio of each heating section with the time series of the return pipe opening, select the numerical combinations that can match each other, convert the selected data into the power output value of the heating section, update the power distribution record table, and obtain the staged heating power sequence.

4. The liquor distillation system with a filtration structure according to claim 3, characterized in that, The genetic algorithm compares the temperature difference ratio of each heating section with the time series of the return pipe opening value in pairs. It calculates the difference of each pair and forms a difference set. Pairs exceeding a set threshold are filtered out from the difference set. The remaining pairs are sorted by difference from smallest to largest. Multiple pairs are randomly selected from the sorted sequence for crossover. The crossover pair records are then compared with the original pair set for difference. Pairs with smaller differences are retained. The sampling, crossover and retention operations are repeated until the difference converges to a set range. The temperature difference ratio corresponding to the difference value and the return pipe opening value are combined to convert into the power output value of the heating section. The converted power is then updated to the staged heating power sequence.

5. The liquor distillation system with a filtration structure according to claim 1, characterized in that, The steam distribution module includes: Flow recording submodule: Based on the phased heating power sequence, collect the total flow rate of the steam main pipe and the flow rate of each branch and store them according to the branch number. Merge the flow rate value with the timestamp to generate a flow rate dataset. Temperature-flow comparison submodule: Based on the flow dataset, the temperature at the top and bottom of the distillation column is collected and cross-matched with the flow values ​​of each branch, the temperature difference offset is calculated, and abnormal branches are marked in the record table to generate a temperature-flow comparison table. Valve adjustment and screening submodule: Based on the temperature difference flow rate comparison table, adjust the opening degree of the corresponding valve in the branch where the temperature difference deviation exceeds the limit, record the adjusted flow rate value and energy consumption data, use Bayesian optimization to predict the steam distribution stability and energy consumption performance of different opening degree combinations, select the flow rate combination with high stability and low energy consumption from the prediction results, and generate the preferred steam flow rate distribution group.

6. The liquor distillation system with a filtration structure according to claim 5, characterized in that, The Bayesian optimization process involves adjusting the valve opening of branches with excessive temperature difference deviations. After the adjustment, multiple sets of flow rate and energy consumption data for each branch with different valve openings are recorded. The recorded flow rate and energy consumption data sets are paired to form a combination set. A comprehensive score is calculated for the combination set based on the branch flow stability index and energy consumption index. The score results are sorted according to the score value. Several combinations are selected as candidate sets based on the score. New opening parameters are added to the candidate sets in sequence, and the comprehensive score is updated in real time. The updated score is compared with the historical score, and the combination with the higher score is retained. The evaluation, comparison, and retention steps are repeated until the combination score no longer improves significantly. The optimal steam flow distribution group is determined based on the score.

7. The liquor distillation system with a filtration structure according to claim 1, characterized in that, The temperature dynamic control module includes: Temperature difference comparison submodule: Based on the optimized steam flow distribution group, the temperature above the liquid surface of the current tray measured in each heating section is compared with the temperature value at the same position in the previous cycle one by one. The alcohol content and fusel oil content measured in the outlet wine are respectively associated with the corresponding temperature difference data. The temperature difference value and the component content value are combined into a pair record to generate a temperature difference component mapping set. Power timing adjustment submodule: Based on the temperature difference component mapping set, calculate the corresponding power increase or decrease value for each pair of data, arrange the power change values ​​of each heating section in sequence according to the temperature section sequence, so that the power adjustment time of the high temperature section lags behind the power adjustment time of the low temperature section, and obtain the temperature adjustment timing table by forming a recording unit through the delay time and power value.

8. The liquor distillation system with a filtration structure according to claim 1, characterized in that, The filtering switching module includes: Flow velocity and pressure difference monitoring submodule: Based on the temperature adjustment time table, the instantaneous liquid flow rate at the inlet and outlet of the filter membrane module is measured and the ratio is calculated. At the same time, the pressure difference value at the corresponding position is recorded. The two sets of data are compared in amplitude according to the same time sequence. The ratio change value and the pressure difference change value are combined into a unified data unit to generate a flow velocity and pressure difference combination set. Filtration path selection submodule: Based on the set of flow rate and pressure difference combinations, compare the combined values ​​with preset limits to determine whether the limits are exceeded, and switch the conduction path of the filter switching valve when the limits are exceeded. Set the filter membrane pore size, the flow rate of the liquid and the pressure difference generated in the switching state as a set of matching parameters to form a set of filtration operation parameters.

9. The liquor distillation system with a filtration structure according to claim 1, characterized in that, The finished product collection module includes: Liquid volume and level comparison submodule: Based on the filter operation parameter set, record the cumulative inlet flow rate and corresponding liquid level height of the finished product storage tank over a fixed time period, obtain the difference curve by subtracting the two sets of values, and arrange them into a difference table according to the measurement time order to generate a liquid volume and height association set; Concentration fluctuation recording submodule: Based on the liquid volume high correlation set, the alcohol concentration and impurity concentration of the wine sample in the storage tank are collected during each measurement period. The ratio of the two concentration values ​​is arranged in chronological order, and the fluctuation range of the ratio and the change of liquid level are used to form corresponding data records, forming a quantitative record set of finished product quality.

10. A method for distilling baijiu (Chinese liquor) with a filtration structure, characterized in that, The liquor distillation system with a filtration structure according to any one of claims 1-9 shall be implemented. Includes the following steps: S1: The volume and alcohol concentration of the mash in the raw material tank are obtained by the sensor. The power output of the heating temperature controller and the pressure of the steam distributor are adjusted in segments. The change sequence of the liquid level on the tray and the temperature difference between the upper and lower trays are measured. The sequence is compared with the power segment sequence. The power distribution is adjusted by the genetic algorithm to generate a segmented heating power sequence. S2: Based on the segmented heating power sequence, record the main flow rate and the flow rate of multiple branches, compare the temperature difference between the top and bottom with the branch flow rate, calculate the offset, adjust the valve opening of the over-limit branch, use Bayesian optimization to screen the stable ratio, and generate the preferred steam flow distribution group. S3: Based on the preferred steam flow distribution group, compare the current temperature difference of each heating section with the previous cycle value, match the difference with the alcohol concentration value and fusel oil concentration value of the outlet wine, adjust the power amplitude of the section and the delay time of the adjacent high temperature section to obtain the temperature adjustment timing table. S4: Based on the temperature adjustment timing table, compare the inlet and outlet flow rate values ​​of the filter membrane module and measure the differential pressure change rate, combine them into a joint value and determine its relationship with the limit value, perform switching for cases exceeding the limit and sort the pore size conditions to generate a set of filter operation parameters; S5: Based on the filter operation parameter set, record the cumulative flow rate at the tank inlet and compare it with the liquid level, sample and measure the ratio of alcohol concentration to impurity concentration, sort by time and record the correlation between fluctuation and liquid level fluctuation, and generate a quantitative record set of finished product quality.

Images