Physical and chemical parameter prediction and production process recommendation method and device for fermented grains in wine cellar

Abstract

The invention relates to the technical field of wine brewing, discloses a physical and chemical parameter prediction and production process recommendation method and device for fermented grains in a wine cellar, and aims to solve the problems of low efficiency and poor accuracy of an existing scheme. According to the scheme, the method mainly comprises the following steps: acquiring key index characteristics of the out-cellar fermented grains of the multi-retort fermented grains in different groups and corresponding physical and chemical parameter characteristics of the in-cellar fermented grains; respectively constructing a group pit-entering fermented grain physical and chemical prediction model and a comprehensive pit-entering fermented grain physical and chemical prediction model corresponding to each group; inputting the cellar-out fermented grain key index characteristics of the to-be-predicted fermented grains into the corresponding group cellar-in fermented grain physical and chemical prediction model and the comprehensive cellar-in fermented grain physical and chemical prediction model respectively to obtain first cellar-in fermented grain physical and chemical parameters and second cellar-in fermented grain physical and chemical parameters of the to-be-predicted fermented grains; and determining the final physical and chemical parameters of the cellar-entering fermented grains according to the first physical and chemical parameters of the cellar-entering fermented grains and the second physical and chemical parameters of the cellar-entering fermented grains. The determination efficiency and accuracy of the physical and chemical parameters of the fermented grains in the cellar are improved, and the method is suitable for producing Luzhou-flavor liquor.

Description

Technical Field

[0001] The present invention relates to the technical field of brewing, and in particular to a method and device for predicting the physical and chemical parameters of fermented grains in a wine cellar and recommending production processes. Background Art

[0002] In the production process of Luzhou-flavor liquor, taking out the fermented grains from the cellar and putting them into the cellar are two important links. Taking out the fermented grains from the cellar means taking out the fermented grains (mother grains) from the cellar during the liquor brewing process. This process usually includes: dripping the cellar to reduce acidity, removing the surface grains, and removing the mother grains. Putting the fermented grains into the cellar means putting the fermented grains taken out of the cellar back into the cellar for the next round of fermentation. The specific steps include: adding raw materials and auxiliary materials, cooking and gelatinizing, measuring water, spreading and cooling for inoculation, and sealing for fermentation. Among them, the physical and chemical parameters of the fermented grains put into the cellar are crucial for the formation of the liquor flavor. Reasonable physical and chemical parameters of the fermented grains put into the cellar can ensure the best metabolic activities of microorganisms during the fermentation process, thereby increasing the liquor yield and the quality of the finished product.

[0003] There are usually two traditional methods for determining the physical and chemical parameters of the fermented grains put into the cellar. The first one is to conduct corresponding tests on the fermented grain samples when the fermented grains are put into the cellar, so as to obtain the physical and chemical parameters of the fermented grains put into the cellar. This method can only be carried out after the addition of raw materials and auxiliary materials, cooking and gelatinizing, measuring water, spreading and cooling for inoculation of the fermented grains, and it is cumbersome to operate and has low detection efficiency. The second one is that the winemaker estimates the physical and chemical parameters of the fermented grains put into the cellar according to indicators such as the physical and chemical parameters, weight, and production process of the fermented grains based on winemaking experience. This method relies on the winemaker's experience, is easily affected by subjectivity, and due to the complexity of the fermentation process and the particularity of the process, the consistency and accuracy are poor. Summary of the Invention

[0004] The present invention aims to solve the problems of low efficiency and poor accuracy in the existing solutions for determining the physical and chemical parameters of the fermented grains put into the cellar, and proposes a method and device for predicting the physical and chemical parameters of the fermented grains in a wine cellar and recommending production processes.

[0005] The technical solutions adopted by the present invention to solve the above technical problems are as follows:

[0006] In a first aspect, the present invention provides a method for predicting the physical and chemical parameters of fermented grains in a wine cellar, the method comprising:

[0007] Obtaining the key index characteristics of the fermented grains taken out of the cellar and the corresponding physical and chemical parameter characteristics of the fermented grains put into the cellar for multiple batches of fermented grains under different groups, wherein the key index characteristics of the fermented grains taken out of the cellar include the physical and chemical characteristics of the fermented grains taken out of the cellar, the weight characteristics of the fermented grains taken out of the cellar, and the production process characteristics of the fermented grains taken out of the cellar;

[0008] Respectively constructing a physical and chemical prediction model for the fermented grains put into the cellar corresponding to each group according to the key index characteristics of the fermented grains taken out of the cellar and the corresponding physical and chemical parameter characteristics of the fermented grains put into the cellar under each group;

[0009] Construct a comprehensive physical and chemical prediction model for the fermented grains put into the cellar based on the key index characteristics of the fermented grains taken out of the cellar under all groups and their corresponding physical and chemical parameter characteristics of the fermented grains put into the cellar;

[0010] Determine the group of the fermented grains to be predicted and the key index characteristics of the fermented grains taken out of the cellar, and input the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted into the corresponding physical and chemical prediction model of the fermented grains put into the cellar of the group to obtain the first physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted;

[0011] Input the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted into the comprehensive physical and chemical prediction model of the fermented grains put into the cellar to obtain the second physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted;

[0012] Determine the final physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted according to the first physical and chemical parameters of the fermented grains put into the cellar and the second physical and chemical parameters of the fermented grains put into the cellar.

[0013] Furthermore, the physical and chemical characteristics of the fermented grains taken out of the cellar include: the moisture content, acidity, starch content and residual sugar content of the fermented grains taken out of the cellar;

[0014] The weight characteristics of the fermented grains taken out of the cellar include: the weight of the fermented grains taken out of the cellar;

[0015] The production process characteristics of the fermented grains taken out of the cellar include: the amount of water added for moistening grains, the amount of grains added, the amount of bran added, the amount of water added for sizing, the amount of koji added and the steaming time of grains;

[0016] The physical and chemical parameter characteristics of the fermented grains put into the cellar include: the moisture content, acidity and starch content of the fermented grains put into the cellar.

[0017] Furthermore, the group is divided according to the shift, workshop, season or month.

[0018] Furthermore, the construction method of the physical and chemical prediction model of the fermented grains put into the cellar of the group includes:

[0019] For the key index characteristics of the fermented grains taken out of the cellar under the same group and their corresponding physical and chemical parameter characteristics of the fermented grains put into the cellar, use the key index characteristics of the fermented grains taken out of the cellar as the input characteristics of the neural network model, and use the physical and chemical parameter characteristics of the fermented grains put into the cellar as the true output labels of the neural network model to construct the first sample dataset;

[0020] Divide the first sample dataset into a first training set and a first test set, train the neural network model with the first training set, during the training process, compare the predicted value of the neural network model with the corresponding true output label, calculate the corresponding loss function, and optimize the model through the backpropagation of the neural network; verify the neural network model with the first test set, when the corresponding loss function is less than the loss function threshold, use the corresponding neural network model as the physical and chemical prediction model of the fermented grains put into the cellar of the corresponding group;

[0021] The construction method of the comprehensive in-pit fermented grains physical and chemical prediction model includes:

[0022] For the key index characteristics of the out-pit fermented grains and the corresponding physical and chemical parameter characteristics of the in-pit fermented grains under all groups, taking the key index characteristics of the out-pit fermented grains as the input characteristics of the neural network model, and taking the physical and chemical parameter characteristics of the in-pit fermented grains as the true output labels of the neural network model, a second sample data set is constructed;

[0023] Dividing the second sample data set into a second training set and a second test set, training the neural network model with the second training set. During the training process, comparing the predicted values of the neural network model with the corresponding true output labels, calculating the corresponding loss function, and optimizing the model through the backpropagation of the neural network; validating the neural network model with the second test set. When the corresponding loss function is less than the loss function threshold, taking the corresponding neural network model as the corresponding comprehensive in-pit fermented grains physical and chemical prediction model.

[0024] Furthermore, the loss function is SmoothL1Loss, and the calculation formula is as follows:

[0025]

[0026] Among them, Loss represents the loss function, Loss(x i ,y i ) represents the loss function of the i-th sample data, x i represents the predicted value of the i-th sample data, y i represents the true output label of the i-th sample data, β represents the smoothing parameter, and N represents the number of sample data.

[0027] Furthermore, the method further includes:

[0028] Normalizing the key index characteristics of the out-pit fermented grains and the corresponding physical and chemical parameter characteristics of the in-pit fermented grains of multiple batches of fermented grains under different groups, performing corresponding normalization processing on the key index characteristics of the out-pit fermented grains to be predicted, and performing corresponding inverse normalization processing on the first in-pit fermented grains physical and chemical parameters and the second in-pit fermented grains physical and chemical parameters obtained by prediction.

[0029] In a second aspect, the present invention provides a device for predicting the physical and chemical parameters of fermented grains in a wine cellar. The device includes:

[0030] An acquisition unit for acquiring the key index characteristics of the out-pit fermented grains and the corresponding physical and chemical parameter characteristics of the in-pit fermented grains of multiple batches of fermented grains under different groups. The key index characteristics of the out-pit fermented grains include the physical and chemical characteristics of the out-pit fermented grains, the weight characteristics of the out-pit fermented grains, and the production process characteristics of the out-pit fermented grains;

[0031] A construction unit for respectively constructing a physical and chemical prediction model of the pit-entry fermented grains for each group according to the key index characteristics of the fermented grains taken out of the pit under each group and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains; constructing a comprehensive physical and chemical prediction model of the pit-entry fermented grains according to the key index characteristics of the fermented grains taken out of the pit under all groups and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains;

[0032] A prediction unit for determining the group of the fermented grains to be predicted and the key index characteristics of the fermented grains taken out of the pit, inputting the key index characteristics of the fermented grains to be predicted into the corresponding physical and chemical prediction model of the pit-entry fermented grains for the group to obtain the first physical and chemical parameters of the pit-entry fermented grains of the fermented grains to be predicted; inputting the key index characteristics of the fermented grains to be predicted into the comprehensive physical and chemical prediction model of the pit-entry fermented grains to obtain the second physical and chemical parameters of the pit-entry fermented grains of the fermented grains to be predicted; determining the final physical and chemical parameters of the pit-entry fermented grains of the fermented grains to be predicted according to the prediction results of the first physical and chemical parameters of the pit-entry fermented grains and the prediction results of the second physical and chemical parameters of the pit-entry fermented grains.

[0033] In a third aspect, the present invention provides a method for recommending a production process of fermented grains in a wine cellar, the method comprising:

[0034] Obtaining the key index characteristics of the fermented grains taken out of the pit and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains of multiple batches of fermented grains under different groups, wherein the key index characteristics of the fermented grains taken out of the pit include the physical and chemical characteristics of the fermented grains taken out of the pit, the weight characteristics of the fermented grains taken out of the pit, and the production process characteristics of the fermented grains taken out of the pit;

[0035] Respectively constructing a corresponding physical and chemical prediction model of the pit-entry fermented grains for each group according to the key index characteristics of the fermented grains taken out of the pit under each group and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains;

[0036] Constructing a comprehensive physical and chemical prediction model of the pit-entry fermented grains according to the key index characteristics of the fermented grains taken out of the pit under all groups and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains;

[0037] Constructing a production process recommendation model for the fermented grains taken out of the pit for the group based on the physical and chemical prediction model of the pit-entry fermented grains for the group, and constructing a comprehensive production process recommendation model for the fermented grains taken out of the pit based on the comprehensive physical and chemical prediction model of the pit-entry fermented grains;

[0038] Determining the group of the fermented grains to be recommended, the physical and chemical characteristics of the fermented grains taken out of the pit, the weight characteristics of the fermented grains taken out of the pit, and the target physical and chemical parameters of the pit-entry fermented grains, inputting the physical and chemical characteristics of the fermented grains taken out of the pit, the weight characteristics of the fermented grains taken out of the pit, and the target physical and chemical parameters of the pit-entry fermented grains of the fermented grains to be recommended into the corresponding production process recommendation model for the fermented grains taken out of the pit for the group for optimization to obtain the first recommended production process of the fermented grains taken out of the pit for the fermented grains to be recommended;

[0039] Inputting the physical and chemical characteristics of the fermented grains taken out of the pit, the weight characteristics of the fermented grains taken out of the pit, and the target physical and chemical parameters of the pit-entry fermented grains of the fermented grains to be recommended into the comprehensive production process recommendation model for the fermented grains taken out of the pit for optimization to obtain the second recommended production process of the fermented grains taken out of the pit for the fermented grains to be recommended;

[0040] Determine the recommended production process of the finally discharged fermented grains to be recommended according to the recommended production process of the first discharged fermented grains and the recommended production process of the second discharged fermented grains.

[0041] Furthermore, the recommended production process model of the fermented grains discharged from the group and the comprehensive recommended production process model of the fermented grains discharged from the cellar are constructed based on an evolutionary algorithm, and the evolutionary algorithm is a genetic algorithm, a simulated annealing algorithm or an evolutionary strategy algorithm;

[0042] The fitness function of the evolutionary algorithm is as follows:

[0043]

[0044] where Fitness represents fitness, and obj j represents the physicochemical parameters of the j-th target fermented grains put into the cellar, and pre j represents the physicochemical parameters of the j-th fermented grains put into the cellar predicted by the physicochemical prediction model of the fermented grains put into the cellar of the group or the comprehensive physicochemical prediction model of the fermented grains put into the cellar of the cellar, and M represents the number of physicochemical parameters of the fermented grains put into the cellar.

[0045] Fourthly, the present invention provides a device for recommending the production process of fermented grains in a wine cellar, and the device includes:

[0046] An acquisition unit for acquiring the key index characteristics of the discharged fermented grains and the corresponding physicochemical parameter characteristics of the fermented grains put into the cellar of multiple retorts under different groups, and the key index characteristics of the discharged fermented grains include the physicochemical characteristics of the discharged fermented grains, the weight characteristics of the discharged fermented grains and the production process characteristics of the discharged fermented grains;

[0047] A construction unit for respectively constructing corresponding physicochemical prediction models of the fermented grains put into the cellar of the group according to the key index characteristics of the discharged fermented grains and the corresponding physicochemical parameter characteristics of the fermented grains put into the cellar of each group; constructing a comprehensive physicochemical prediction model of the fermented grains put into the cellar according to the key index characteristics of the discharged fermented grains and the corresponding physicochemical parameter characteristics of the fermented grains put into the cellar of all groups; constructing a recommended production process model of the fermented grains discharged from the group based on the physicochemical prediction model of the fermented grains put into the cellar of the group, and constructing a comprehensive recommended production process model of the fermented grains discharged from the cellar based on the comprehensive physicochemical prediction model of the fermented grains put into the cellar;

[0048] A recommendation unit, which is used to determine the group of the fermented grains to be recommended, the physical and chemical characteristics of the fermented grains taken out of the cellar, the weight characteristics of the fermented grains taken out of the cellar, and the physical and chemical parameters of the target fermented grains put into the cellar. The physical and chemical characteristics of the fermented grains taken out of the cellar, the weight characteristics of the fermented grains taken out of the cellar, and the physical and chemical parameters of the target fermented grains put into the cellar of the fermented grains to be recommended are input into the corresponding production process recommendation model of the fermented grains taken out of the cellar for each group to obtain the first recommended production process of the fermented grains taken out of the cellar for the fermented grains to be recommended; the physical and chemical characteristics of the fermented grains taken out of the cellar, the weight characteristics of the fermented grains taken out of the cellar, and the physical and chemical parameters of the target fermented grains put into the cellar of the fermented grains to be recommended are input into the comprehensive production process recommendation model of the fermented grains taken out of the cellar to obtain the second recommended production process of the fermented grains taken out of the cellar for the fermented grains to be recommended; the final recommended production process of the fermented grains taken out of the cellar for the fermented grains to be recommended is determined according to the first recommended production process of the fermented grains taken out of the cellar and the second recommended production process of the fermented grains taken out of the cellar.

[0049] The beneficial effects of the present invention are as follows: The method and device for predicting the physical and chemical parameters of the fermented grains in the wine cellar provided by the present invention can predict the physical and chemical parameters of the fermented grains put into the cellar according to the key index characteristics of the fermented grains taken out of the cellar before the treatment of the fermented grains taken out of the cellar is completed, improving the prediction efficiency and accuracy of the physical and chemical parameters of the fermented grains put into the cellar. At the same time, the present invention respectively conducts predictions based on the group-based physical and chemical prediction model of the fermented grains put into the cellar and the comprehensive physical and chemical prediction model of the fermented grains put into the cellar, and determines the final physical and chemical parameters of the fermented grains put into the cellar according to the prediction results of the two prediction models. By considering the group characteristics of the fermented grains, the influence of different groups on the prediction of the physical and chemical parameters of the fermented grains put into the cellar is avoided, further improving the accuracy of the prediction of the physical and chemical parameters of the fermented grains put into the cellar. In addition, the production process recommendation method and device for the fermented grains in the wine cellar provided by the present invention construct a production process recommendation model for the fermented grains taken out of the cellar based on the physical and chemical prediction model of the fermented grains put into the cellar. Through the production process recommendation model of the fermented grains taken out of the cellar, after setting the target physical and chemical parameters of the fermented grains put into the cellar, a recommended production process for the fermented grains taken out of the cellar that meets the target physical and chemical parameters of the fermented grains put into the cellar can be sought, realizing the optimization of the production process of the fermented grains taken out of the cellar, and also improving the efficiency and accuracy of the optimization of the production process of the fermented grains taken out of the cellar. At the same time, the present invention respectively conducts production process optimization based on the group-based production process recommendation model of the fermented grains taken out of the cellar and the comprehensive production process recommendation model of the fermented grains taken out of the cellar, and determines the final recommended production process of the fermented grains taken out of the cellar according to the optimization results of the two recommendation models. By considering the group characteristics of the fermented grains, the influence of different groups on the recommended production process is avoided, further improving the accuracy of the optimization of the production process of the fermented grains taken out of the cellar. Through the present invention, the stability of the fermentation of the fermented grains and the consistency of the product quality can be ensured, thereby enhancing the flavor and market competitiveness of the Luzhou-flavor liquor. Description of the Drawings

[0050] Figure 1 It is a schematic flowchart of a method for predicting the physical and chemical parameters of the fermented grains in a wine cellar provided by an embodiment;

[0051] Figure 2 It is a schematic structural diagram of a device for predicting the physical and chemical parameters of the fermented grains in a wine cellar provided by an embodiment;

[0052] Figure 3 Flow schematic diagram of a method for recommending the production process of fermented grains in a wine cellar provided for an embodiment;

[0053] Figure 4 Structural schematic diagram of a device for recommending the production process of fermented grains in a wine cellar provided for an embodiment. Detailed implementation manners

[0054] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the present embodiment will be clearly and completely described below in conjunction with the accompanying drawings in the present embodiment.

[0055] In some processes described in the specification of the present invention and the above-mentioned accompanying drawings, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear in this article or may be executed in parallel. The serial numbers of the operations are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel.

[0056] Since the current methods for determining the physical and chemical parameters of fermented grains entering the cellar mainly detect the physical and chemical parameters of fermented grains entering the cellar by detecting the fermented grain samples after the cellar is emptied, or the winemakers estimate the physical and chemical parameters of fermented grains entering the cellar based on experience. The inventors have found through research that these methods have at least one or more of the following problems: unable to detect before the cellar is emptied, low efficiency, strong subjectivity, and poor accuracy.

[0057] Based on this, the technical solution of the present invention is proposed. In the present invention, the key index characteristics of the fermented grains leaving the cellar and the corresponding physical and chemical parameter characteristics of the fermented grains entering the cellar under different groups are obtained. The key index characteristics of the fermented grains leaving the cellar include the physical and chemical characteristics of the fermented grains leaving the cellar, the weight characteristics of the fermented grains leaving the cellar, and the production process characteristics of the fermented grains leaving the cellar; the physical and chemical prediction models of the fermented grains entering the cellar corresponding to each group are respectively constructed according to the key index characteristics of the fermented grains leaving the cellar and the corresponding physical and chemical parameter characteristics of the fermented grains entering the cellar under each group; a comprehensive physical and chemical prediction model of the fermented grains entering the cellar is constructed according to the key index characteristics of the fermented grains leaving the cellar and the corresponding physical and chemical parameter characteristics of the fermented grains entering the cellar under all groups; the group and the key index characteristics of the fermented grains leaving the cellar of the fermented grains to be predicted are determined, and the key index characteristics of the fermented grains leaving the cellar of the fermented grains to be predicted are input into the corresponding physical and chemical prediction model of the fermented grains entering the cellar to obtain the first physical and chemical parameters of the fermented grains entering the cellar of the fermented grains to be predicted; the key index characteristics of the fermented grains leaving the cellar of the fermented grains to be predicted are input into the comprehensive physical and chemical prediction model of the fermented grains entering the cellar to obtain the second physical and chemical parameters of the fermented grains entering the cellar of the fermented grains to be predicted; the final physical and chemical parameters of the fermented grains entering the cellar of the fermented grains to be predicted are determined according to the first physical and chemical parameters of the fermented grains entering the cellar and the second physical and chemical parameters of the fermented grains entering the cellar.

[0058] Specifically, the present invention constructs a group-based physical and chemical prediction model for cellar-entry fermented grains that represents the relationship between the key index characteristics of cellar-exit fermented grains and the physical and chemical parameters of cellar-entry fermented grains, and a comprehensive physical and chemical prediction model for cellar-entry fermented grains. Then, predictions are made based on the group-based physical and chemical prediction model for cellar-entry fermented grains and the comprehensive physical and chemical prediction model for cellar-entry fermented grains respectively, and the final physical and chemical parameters of cellar-entry fermented grains are determined according to the prediction results of the two prediction models. Before the treatment of cellar-exit fermented grains is completed, the physical and chemical parameters of cellar-entry fermented grains can be predicted based on the key index characteristics of cellar-exit fermented grains, which improves the determination efficiency and accuracy of the physical and chemical parameters of cellar-entry fermented grains. At the same time, by considering the group characteristics of fermented grains, the influence of different groups on the prediction of the physical and chemical parameters of cellar-entry fermented grains is avoided, further improving the accuracy of the prediction of the physical and chemical parameters of cellar-entry fermented grains.

[0059] Next, the technical solutions in this embodiment will be clearly and completely described in conjunction with the accompanying drawings in this embodiment. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments.

[0060] Figure 1 The flowchart of a method for predicting the physical and chemical parameters of fermented grains in a wine cellar is shown. Please refer to Figure 1 , and this method includes the following steps:

[0061] Step 101: Obtain the key index characteristics of cellar-exit fermented grains of multiple batches of fermented grains under different groups and their corresponding physical and chemical parameter characteristics of cellar-entry fermented grains.

[0062] Among them, the key index characteristics of cellar-exit fermented grains include the physical and chemical characteristics of cellar-exit fermented grains, the weight characteristics of cellar-exit fermented grains, and the production process characteristics of cellar-exit fermented grains.

[0063] In this embodiment, the physical and chemical characteristics of cellar-exit fermented grains include: the moisture content, acidity, starch content, and residual sugar content of cellar-exit fermented grains; the weight characteristics of cellar-exit fermented grains include: the weight of cellar-exit fermented grains; the production process characteristics of cellar-exit fermented grains include: the amount of added moistening grains, the amount of added grains, the amount of added bran, the amount of added water, the amount of added koji, and the steaming duration of grains; the physical and chemical parameter characteristics of cellar-entry fermented grains include: the moisture content, acidity, and starch content of cellar-entry fermented grains.

[0064] It should be noted that the physical and chemical characteristics of cellar-exit fermented grains, the weight characteristics of cellar-exit fermented grains, and the production process characteristics of cellar-exit fermented grains may also include other characteristics concerned by experts and engineers in the field.

[0065] In this embodiment, the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains can be determined according to the data of the in-cellar and out-cellar fermented grains. For example, a set of in-cellar and out-cellar fermented grain data is as follows: the physical and chemical characteristics of the out-cellar fermented grains are [62.1, 4.46, 14.1, 0.60], the weight characteristics of the out-cellar fermented grains are [872 kg], and the production process characteristics of the out-cellar fermented grains are [6 barrels, 200 kg, 55.9 kg, 210 L, 42 kg, 85 min]; the physical and chemical parameter characteristics of the in-cellar fermented grains are [56.0, 2.02, 22.6]. Then the key index characteristics of the out-cellar fermented grains are [62.1, 4.46, 14.1, 0.60, 872, 6, 200, 55.9, 210, 42, 85], and the physical and chemical parameter characteristics of the in-cellar fermented grains are [56.0, 2.02, 22.6].

[0066] In the embodiment, it further includes normalizing the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains under different groups, including using normalization methods such as min-max normalization and Z-score normalization for normalization. For example, the sklearn.preprocessing.MinMaxScaler function in the scikit-learn library can be used for automatic normalization. This normalization method is the min-max normalization method, and the maximum and minimum parameter values of the data during the normalization process can be recorded, enabling convenient restoration of the normalized data.

[0067] Among them, the groups can be divided according to the shift, workshop, season or month, that is, the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains under different shifts, workshops, seasons or months are obtained. Among them, the workshop is divided into the first aisle, the second aisle, and the third aisle. The fermentation conditions of the fermented grains under different groups are different, so the corresponding relationship between the key index characteristics of the out-cellar fermented grains and the physical and chemical parameter characteristics of the in-cellar fermented grains may vary.

[0068] In practical applications, an in-cellar and out-cellar fermented grain database can be constructed according to the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains under different groups, and the data in the in-cellar and out-cellar fermented grain database can be continuously supplemented and improved.

[0069] Step 102: Respectively construct a group-specific in-cellar fermented grain physical and chemical prediction model according to the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains under each group;

[0070] In this embodiment, the construction method of the group-specific in-cellar fermented grain physical and chemical prediction model includes:

[0071] For the key index features of the pit-out fermented grains under the same group and their corresponding physical and chemical parameter features of the pit-in fermented grains, the key index features of the pit-out fermented grains are used as the input features of the neural network model, and the physical and chemical parameter features of the pit-in fermented grains are used as the true output labels of the neural network model to construct the first sample dataset;

[0072] Divide the first sample dataset into a first training set and a first test set. Train the neural network model with the first training set. During the training process, compare the predicted values of the neural network model with the corresponding true output labels, calculate the corresponding loss function, and optimize the model through the backpropagation of the neural network; Verify the neural network model with the first test set. When the corresponding loss function is less than the loss function threshold, use the corresponding neural network model as the physical and chemical prediction model of the pit-in fermented grains for the corresponding group.

[0073] The physical and chemical prediction model of the pit-in fermented grains for each group is a deep learning model constructed based on neural network models suitable for processing one-dimensional data, including MLP, LSTM, and RNN. In this embodiment, the physical and chemical prediction model of the pit-in fermented grains is an MLP network, which is constructed using the PyTorch library. The model construction code is as follows:

[0074]

[0075] Among them, the input feature dimension input_size of the MLP network is 11, the hidden layer feature dimension hidden_size is 8, and the output layer feature dimension output_size is 3.

[0076] In practical applications, the physical and chemical prediction model of the pit-in fermented grains for each group is trained separately based on the key index features of the pit-out fermented grains under each group and their corresponding physical and chemical parameter features of the pit-in fermented grains. Specifically: Use the key index features of the pit-out fermented grains as the input features of the physical and chemical prediction model of the pit-in fermented grains for each group, and use the physical and chemical parameter features of the pit-in fermented grains as the true output labels of the neural network model to construct the first sample dataset. Calculate the loss based on the predicted values and true output labels of the physical and chemical prediction model of the pit-in fermented grains, and optimize the model through the backpropagation of the neural network model. During the training process of the model, the first sample dataset is randomly divided into a first training set and a second test set. The first training set is used for the training of the neural network model, and the second test set is used to evaluate the training effect of the neural network model. When the loss function value corresponding to the sample data in the second test set is lower than the pre-set loss function threshold, the training of the neural network model is completed, and the physical and chemical prediction model of the pit-in fermented grains for each group is obtained.

[0077] In this embodiment, the proportion of the first test set is 20%. The first sample dataset in this embodiment contains a total of 1000 sample data, and the data in the first test set accounts for 200. The pre-set loss function threshold in this embodiment is 0.01.

[0078] During the training process of the neural network model, the loss function is SmoothL1Loss, and its calculation formula is as follows:

[0079]

[0080] Among them, Loss represents the loss function, Loss(x i ,y i ) represents the loss function of the i-th sample data, x i represents the predicted value of the i-th sample data, y i represents the true output label of the i-th sample data, β represents the smoothing parameter, and N represents the number of sample data.

[0081] Step 103: Construct a comprehensive physical and chemical prediction model for the pit-entry fermented grains based on the key index characteristics of the pit-exit fermented grains and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains under all groups.

[0082] In the embodiment of the present application, the construction method of the comprehensive physical and chemical prediction model for the pit-entry fermented grains includes:

[0083] For the key index characteristics of the pit-exit fermented grains and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains under all groups, use the key index characteristics of the pit-exit fermented grains as the input characteristics of the neural network model, and use the physical and chemical parameter characteristics of the pit-entry fermented grains as the true output labels of the neural network model to construct a second sample data set;

[0084] Divide the second sample data set into a second training set and a second test set. Use the second training set to train the neural network model. During the training process, compare the predicted value of the neural network model with the corresponding true output label, calculate the corresponding loss function, and optimize the model through the backpropagation of the neural network; use the second test set to verify the neural network model. When the corresponding loss function is less than the loss function threshold, use the corresponding neural network model as the corresponding comprehensive physical and chemical prediction model for the pit-entry fermented grains.

[0085] It can be understood that the training processes of the comprehensive physical and chemical prediction model for the pit-entry fermented grains and the physical and chemical prediction model for the pit-entry fermented grains of each group are the same, and the training process will not be elaborated here. For the relevant parts, refer to the training process of the physical and chemical prediction model for the pit-entry fermented grains of each group. The difference is only in the training data, that is, the comprehensive physical and chemical prediction model for the pit-entry fermented grains constructs a second sample data set using the key index characteristics of the pit-exit fermented grains and the corresponding physical and chemical parameter characteristics of the pit-entry fermented grains under all groups without considering the groups, and trains the neural network model based on the second sample data set to obtain the comprehensive physical and chemical prediction model for the pit-entry fermented grains.

[0086] Step 104: Determine the group of the fermented grains to be predicted and the key index characteristics of the fermented grains taken out of the cellar. Input the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted into the corresponding physical and chemical prediction models of the fermented grains put into the cellar for each group, and obtain the first physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted.

[0087] In practical applications, after obtaining the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted, it is necessary to perform normalization processing on the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted using the same normalization method as in Step 101. Then, input the key index characteristics of the fermented grains taken out of the cellar after normalization processing into the corresponding physical and chemical prediction models of the fermented grains put into the cellar for each group, and the first physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted can be obtained. Finally, perform inverse normalization processing on the first physical and chemical parameters of the fermented grains put into the cellar corresponding to the normalization method in Step 101 to restore the actual numerical range of the obtained first physical and chemical parameters of the fermented grains put into the cellar.

[0088] Step 105: Input the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted into the comprehensive physical and chemical prediction model of the fermented grains put into the cellar, and obtain the second physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted.

[0089] Similarly, after obtaining the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted, it is necessary to perform normalization processing on the key index characteristics of the fermented grains taken out of the cellar of the fermented grains to be predicted using the same normalization method as in Step 101. Then, input the key index characteristics of the fermented grains taken out of the cellar after normalization processing into the comprehensive physical and chemical prediction model of the fermented grains put into the cellar, and the second physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted can be obtained. Finally, perform inverse normalization processing on the second physical and chemical parameters of the fermented grains put into the cellar corresponding to the normalization method in Step 101 to restore the actual numerical range of the obtained second physical and chemical parameters of the fermented grains put into the cellar.

[0090] Step 106: Determine the final physical and chemical parameters of the fermented grains put into the cellar of the fermented grains to be predicted according to the first physical and chemical parameters of the fermented grains put into the cellar and the second physical and chemical parameters of the fermented grains put into the cellar.

[0091] After obtaining the first physical and chemical parameters of the fermented grains put into the cellar and the second physical and chemical parameters of the fermented grains put into the cellar respectively, the final physical and chemical parameters of the fermented grains put into the cellar can be determined according to the two prediction results. In practical applications, the final physical and chemical parameters of the fermented grains put into the cellar can be determined according to the actual production requirements. For example, the first physical and chemical parameters of the fermented grains put into the cellar and the second physical and chemical parameters of the fermented grains put into the cellar can be subjected to corresponding weighted averaging to obtain the final physical and chemical parameters of the fermented grains put into the cellar, or the first physical and chemical parameters of the fermented grains put into the cellar or the second physical and chemical parameters of the fermented grains put into the cellar can be used alone as the final physical and chemical parameters of the fermented grains put into the cellar according to the actual production requirements.

[0092] In this embodiment, by separately constructing a group-based physical and chemical prediction model for the relationship between the key index characteristics of the out-cellar fermented grains and the physical and chemical parameters of the in-cellar fermented grains, and a comprehensive physical and chemical prediction model for the in-cellar fermented grains, then making predictions based on the group-based physical and chemical prediction model for the in-cellar fermented grains and the comprehensive physical and chemical prediction model for the in-cellar fermented grains respectively, and determining the final physical and chemical parameters of the in-cellar fermented grains according to the prediction results of the two prediction models. In this embodiment, before the completion of the out-cellar process, the physical and chemical parameters of the in-cellar fermented grains can be predicted based on the key index characteristics of the out-cellar fermented grains, improving the determination efficiency and accuracy of the physical and chemical parameters of the in-cellar fermented grains. At the same time, by considering the group characteristics of the fermented grains, the influence of different groups on the prediction of the physical and chemical parameters of the in-cellar fermented grains is avoided, further improving the accuracy of the prediction of the physical and chemical parameters of the in-cellar fermented grains.

[0093] Figure 2 Fig. shows a device for predicting the physical and chemical parameters of fermented grains in a wine cellar. Please refer to Figure 2 , and the device includes:

[0094] An acquisition unit, configured to acquire the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains of multiple steamers of fermented grains under different groups, where the key index characteristics of the out-cellar fermented grains include the physical and chemical characteristics of the out-cellar fermented grains, the weight characteristics of the out-cellar fermented grains, and the production process characteristics of the out-cellar fermented grains;

[0095] A construction unit, configured to separately construct a group-based physical and chemical prediction model for the in-cellar fermented grains corresponding to each group according to the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains under each group; construct a comprehensive physical and chemical prediction model for the in-cellar fermented grains according to the key index characteristics of the out-cellar fermented grains and the corresponding physical and chemical parameter characteristics of the in-cellar fermented grains under all groups;

[0096] A prediction unit, configured to determine the group of the fermented grains to be predicted and the key index characteristics of the out-cellar fermented grains, input the key index characteristics of the fermented grains to be predicted into the corresponding group-based physical and chemical prediction model for the in-cellar fermented grains to obtain the first physical and chemical parameters of the in-cellar fermented grains of the fermented grains to be predicted; input the key index characteristics of the fermented grains to be predicted into the comprehensive physical and chemical prediction model for the in-cellar fermented grains to obtain the second physical and chemical parameters of the in-cellar fermented grains of the fermented grains to be predicted; determine the final physical and chemical parameters of the in-cellar fermented grains of the fermented grains to be predicted according to the prediction results of the first physical and chemical parameters of the in-cellar fermented grains and the prediction results of the second physical and chemical parameters of the in-cellar fermented grains.

[0097] It can be understood that since the device for predicting the physical and chemical parameters of fermented grains in a wine cellar described in this embodiment is a device for implementing the method for predicting the physical and chemical parameters of fermented grains in a wine cellar described in the embodiment, for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, please refer to the partial description of the method, and details are not described here again.

[0098] Figure 3 Fig. shows a schematic flowchart of a method for recommending the production process of fermented grains in a wine cellar. Please refer toFigure 3 , the method includes the following steps:

[0099] Step 301, obtain the key index features of the out-cell fermented grains of multiple distillation pots under different groups and the corresponding physical and chemical parameter features of the in-cell fermented grains.

[0100] Among them, the key index features of the out-cell fermented grains include the physical and chemical features of the out-cell fermented grains, the weight features of the out-cell fermented grains, and the production process features of the out-cell fermented grains.

[0101] Step 302, respectively construct the corresponding group in-cell fermented grains physical and chemical prediction models according to the key index features of the out-cell fermented grains under each group and the corresponding physical and chemical parameter features of the in-cell fermented grains.

[0102] Step 303, construct a comprehensive in-cell fermented grains physical and chemical prediction model according to the key index features of the out-cell fermented grains under all groups and the corresponding physical and chemical parameter features of the in-cell fermented grains.

[0103] It can be understood that the implementation principles and processes of Step 301 to Step 303 in this embodiment are the same as those of Step 101 to Step 103 in this embodiment, and this embodiment will not elaborate on this.

[0104] Step 304, construct a group out-cell fermented grains production process recommendation model based on the group in-cell fermented grains physical and chemical prediction model, and construct a comprehensive out-cell fermented grains production process recommendation model based on the comprehensive in-cell fermented grains physical and chemical prediction model.

[0105] In this embodiment, the group out-cell fermented grains production process recommendation model and the comprehensive out-cell fermented grains production process recommendation model are constructed based on an evolutionary algorithm, and the evolutionary algorithm is a genetic algorithm, a simulated annealing algorithm, or an evolutionary strategy algorithm.

[0106] Step 305, determine the group of the fermented grains to be recommended, the physical and chemical features of the out-cell fermented grains, the weight features of the out-cell fermented grains, and the target physical and chemical parameters of the in-cell fermented grains, and input the physical and chemical features of the out-cell fermented grains, the weight features of the out-cell fermented grains, and the target physical and chemical parameters of the in-cell fermented grains of the fermented grains to be recommended into the corresponding group out-cell fermented grains production process recommendation model for optimization to obtain the first out-cell fermented grains recommended production process of the fermented grains to be recommended.

[0107] Among them, the physical and chemical features of the out-cell fermented grains and the weight features of the out-cell fermented grains of the fermented grains to be recommended can be obtained by measurement, and the target physical and chemical parameters of the in-cell fermented grains of the fermented grains to be recommended are the set expected physical and chemical parameters of the in-cell fermented grains. In practical applications, after inputting the physical and chemical features of the out-cell fermented grains, the weight features of the out-cell fermented grains, and the target physical and chemical parameters of the in-cell fermented grains of the fermented grains to be recommended into the corresponding group out-cell fermented grains production process recommendation model, the group out-cell fermented grains production process recommendation model performs optimization based on the group in-cell fermented grains physical and chemical prediction model, so as to obtain the first out-cell fermented grains recommended production process of the fermented grains to be recommended.

[0108] The optimization process of the physical and chemical prediction model for the fermented grains in the cellar by group is specifically as follows: Obtain the physical and chemical characteristics and weight characteristics of the fermented grains out of the cellar of the fermented grains to be recommended, and then input them into the physical and chemical prediction model for the fermented grains in the cellar by group respectively together with the production process characteristics of all possible fermented grains out of the cellar. Predict the physical and chemical parameters of the fermented grains put into the cellar, and then compare the predicted physical and chemical parameters of the fermented grains put into the cellar with the target physical and chemical parameters of the fermented grains put into the cellar. Calculate the fitness according to the fitness function, and determine the optimal physical and chemical parameters of the predicted fermented grains put into the cellar according to the fitness. The production process characteristics of the fermented grains out of the cellar corresponding to the optimal physical and chemical parameters of the fermented grains put into the cellar are the optimal production process characteristics of the fermented grains out of the cellar. Finally, use them as the recommended production process of the first fermented grains out of the cellar for the fermented grains to be recommended.

[0109] In this embodiment, the fitness function of the evolutionary algorithm is as follows:

[0110]

[0111] Among them, Fitness represents the fitness, obj j represents the j-th target physical and chemical parameter of the fermented grains put into the cellar, and pre j represents the j-th physical and chemical parameter of the fermented grains put into the cellar predicted by the physical and chemical prediction model for the fermented grains in the cellar by group or the comprehensive physical and chemical prediction model for the fermented grains in the cellar. M represents the number of physical and chemical parameters of the fermented grains put into the cellar.

[0112] Step 306: Input the physical and chemical characteristics, weight characteristics of the fermented grains out of the cellar, and target physical and chemical parameters of the fermented grains to be recommended into the comprehensive production process recommendation model for the fermented grains out of the cellar for optimization, and obtain the recommended production process of the second fermented grains out of the cellar for the fermented grains to be recommended.

[0113] Similarly, in practical applications, after inputting the physical and chemical characteristics, weight characteristics of the fermented grains out of the cellar, and target physical and chemical parameters of the fermented grains to be recommended into the comprehensive production process recommendation model for the fermented grains out of the cellar, the comprehensive production process recommendation model for the fermented grains out of the cellar performs optimization based on the comprehensive physical and chemical prediction model for the fermented grains in the cellar, so as to obtain the recommended production process of the second fermented grains out of the cellar for the fermented grains to be recommended. The optimization process of the comprehensive production process recommendation model for the fermented grains out of the cellar is the same as that of the production process recommendation model for the fermented grains out of the cellar by group. The only difference is that the production process recommendation model for the fermented grains out of the cellar by group performs optimization based on the physical and chemical prediction model for the fermented grains in the cellar by group, while the comprehensive production process recommendation model for the fermented grains out of the cellar performs optimization based on the comprehensive physical and chemical prediction model for the fermented grains in the cellar. The optimization process and fitness function of the comprehensive production process recommendation model for the fermented grains out of the cellar will not be elaborated in this embodiment.

[0114] Step 307: Determine the final recommended production process of the fermented grains out of the cellar for the fermented grains to be recommended according to the first recommended production process of the fermented grains out of the cellar and the second recommended production process of the fermented grains out of the cellar.

[0115] After obtaining the recommended production processes for the first and second out-cell fermented grains respectively, the final recommended production process for the out-cell fermented grains can be determined based on the two recommended results. In practical applications, the final recommended production process for the out-cell fermented grains can be determined according to the actual production requirements. For example, the recommended production processes for the first and second out-cell fermented grains can be weighted and averaged correspondingly to obtain the final recommended production process for the out-cell fermented grains, or the recommended production process for the first out-cell fermented grains or the second out-cell fermented grains can be used alone as the physical and chemical parameters of the in-cell fermented grains according to the actual production requirements.

[0116] In this embodiment, a recommended production process model for out-cell fermented grains is constructed based on the physical and chemical prediction model of in-cell fermented grains. Through the recommended production process model for out-cell fermented grains, after setting the target physical and chemical parameters of in-cell fermented grains, a recommended production process for out-cell fermented grains that meets the target physical and chemical parameters of in-cell fermented grains can be sought, realizing the optimization of the production process of out-cell fermented grains, and also improving the efficiency and accuracy of the optimization of the production process of out-cell fermented grains. At the same time, in this embodiment, the production process optimization is carried out respectively based on the group-based recommended production process model for out-cell fermented grains and the comprehensive recommended production process model for out-cell fermented grains, and the final recommended production process for out-cell fermented grains is determined according to the optimization results of the two recommended models. By considering the group characteristics of the fermented grains, the influence of different groups on the recommended production process is avoided, and the accuracy of the optimization of the production process of out-cell fermented grains is further improved.

[0117] Figure 4 A recommended production process device for fermented grains in a wine cellar is shown. Please refer to Figure 4 , the device includes:

[0118] An acquisition unit, configured to acquire the key index characteristics of out-cell fermented grains and the corresponding physical and chemical parameter characteristics of in-cell fermented grains of multiple batches of fermented grains under different groups, where the key index characteristics of out-cell fermented grains include the physical and chemical characteristics of out-cell fermented grains, the weight characteristics of out-cell fermented grains, and the production process characteristics of out-cell fermented grains;

[0119] A construction unit, configured to respectively construct corresponding group-based physical and chemical prediction models for in-cell fermented grains according to the key index characteristics of out-cell fermented grains and the corresponding physical and chemical parameter characteristics of in-cell fermented grains under each group; construct a comprehensive physical and chemical prediction model for in-cell fermented grains according to the key index characteristics of out-cell fermented grains and the corresponding physical and chemical parameter characteristics of in-cell fermented grains under all groups; construct a group-based recommended production process model for out-cell fermented grains based on the group-based physical and chemical prediction model for in-cell fermented grains, and construct a comprehensive recommended production process model for out-cell fermented grains based on the comprehensive physical and chemical prediction model for in-cell fermented grains;

[0120] A recommendation unit is configured to determine the group of the fermented grains to be recommended, the physical and chemical characteristics of the fermented grains taken out of the cellar, the weight characteristics of the fermented grains taken out of the cellar, and the physical and chemical parameters of the target fermented grains to be put into the cellar. The physical and chemical characteristics of the fermented grains taken out of the cellar, the weight characteristics of the fermented grains taken out of the cellar, and the physical and chemical parameters of the target fermented grains to be put into the cellar of the fermented grains to be recommended are input into the corresponding group of production process recommendation models for the fermented grains taken out of the cellar to obtain the first recommended production process for the fermented grains taken out of the cellar of the fermented grains to be recommended. The physical and chemical characteristics of the fermented grains taken out of the cellar, the weight characteristics of the fermented grains taken out of the cellar, and the physical and chemical parameters of the target fermented grains to be put into the cellar are input into the comprehensive production process recommendation model for the fermented grains taken out of the cellar to obtain the second recommended production process for the fermented grains taken out of the cellar of the fermented grains to be recommended. The final recommended production process for the fermented grains taken out of the cellar of the fermented grains to be recommended is determined according to the first recommended production process for the fermented grains taken out of the cellar and the second recommended production process for the fermented grains taken out of the cellar.

[0121] It can be understood that since the production process recommendation device for the fermented grains in the wine cellar described in this embodiment is a device for implementing the production process recommendation method for the fermented grains in the wine cellar described in the embodiment, for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple. For the relevant parts, reference can be made to the partial description of the method, and details are not elaborated here.

Claims

1. A method for predicting the physical and chemical parameters of fermented grains in a wine cellar, characterized in that: The method comprises: Obtaining key indicator characteristics of the uncellared glutinous rice grains of multiple steamers under different groups and their corresponding physical and chemical parameter characteristics of the uncellared glutinous rice grains, wherein the key indicator characteristics of the uncellared glutinous rice grains include physical and chemical characteristics of the uncellared glutinous rice grains, weight characteristics of the uncellared glutinous rice grains, and production process characteristics of the uncellared glutinous rice grains; According to the key index characteristics of the fermented grains out of the cellar under each group and the corresponding physical and chemical parameter characteristics of the fermented grains entering the cellar, the physical and chemical prediction models of the fermented grains entering the cellar corresponding to each group are respectively constructed; According to the key index characteristics of the uncellared mash under all groups and the corresponding physicochemical parameter characteristics of the mash entering the cellar, a comprehensive physicochemical prediction model of the mash entering the cellar was constructed; Determine the group of the to-be-predicted glutinous rice and the key indicator characteristics of the glutinous rice out of the cellar, input the key indicator characteristics of the glutinous rice out of the cellar of the to-be-predicted glutinous rice into the corresponding group of physicochemical prediction model of the glutinous rice entering the cellar, and obtain the first physicochemical parameters of the glutinous rice entering the cellar of the glutinous rice to be predicted; Inputting the key indicator characteristics of the out-cellar fermented grains of the fermented grains to be predicted into the comprehensive physicochemical prediction model of the fermented grains entering the cellar, and obtaining the physicochemical parameters of the second fermented grains entering the cellar of the fermented grains to be predicted; The final physicochemical parameters of the fermented grains entering the cellar are determined according to the first physicochemical parameters of the fermented grains entering the cellar and the second physicochemical parameters of the fermented grains entering the cellar.

2. The method for predicting the physical and chemical parameters of fermented grains in a wine cellar according to claim 1, characterized in that: The physical and chemical characteristics of the uncellared fermented grains include: moisture, acidity, starch content and residual sugar content of the uncellared fermented grains; The weight characteristics of the uncellared fermented grains include: the weight of the uncellared fermented grains; The production process characteristics of the fermented mash out of the cellar include: the amount of water added to moisten the fermented mash, the amount of grain added, the amount of bran added, the amount of water added, the amount of koji added and the duration of steaming the fermented mash; The physical and chemical parameter characteristics of the fermented grains entering the cellar include: moisture, acidity and starch content of the fermented grains entering the cellar.

3. The method for predicting the physical and chemical parameters of fermented grains in a wine cellar according to claim 1, characterized in that: The groups are divided according to work groups, workshops, seasons or months.

4. The method for predicting the physical and chemical parameters of fermented grains in a wine cellar according to claim 1, characterized in that: The method for constructing the physical and chemical prediction model of the fermented grains entering the cellar of the group comprises: For the key indicator characteristics of the uncellared fermented grains and their corresponding physicochemical parameter characteristics of the fermented grains entering the cellar under the same group, the key indicator characteristics of the uncellared fermented grains are used as the input characteristics of the neural network model, and the physicochemical parameter characteristics of the fermented grains entering the cellar are used as the true output labels of the neural network model to construct a first sample data set; The first sample data set is divided into a first training set and a first test set, and the neural network model is trained with the first training set. During the training process, the predicted value of the neural network model is compared with the corresponding true output label, the corresponding loss function is calculated, and the model is optimized through the back propagation of the neural network; the neural network model is verified with the first test set, and when the corresponding loss function is less than the loss function threshold, the corresponding neural network model is used as the physical and chemical prediction model of the corresponding group of fermented grains entering the cellar; The method for constructing the comprehensive physical and chemical prediction model of the fermented grains entering the cellar comprises: According to the key indicator characteristics of the uncellared mash and its corresponding physicochemical parameter characteristics of the mash entering the cellar under all groups, the key indicator characteristics of the uncellared mash are used as the input characteristics of the neural network model, and the physicochemical parameter characteristics of the mash entering the cellar are used as the true output labels of the neural network model to construct the second sample data set; The second sample data set is divided into a second training set and a second test set. The neural network model is trained with the second training set. During the training process, the predicted value of the neural network model is compared with the corresponding true output label, the corresponding loss function is calculated, and the model is optimized through the back propagation of the neural network. The neural network model is verified with the second test set. When the corresponding loss function is less than the loss function threshold, the corresponding neural network model is used as the corresponding comprehensive physical and chemical prediction model for the mash entering the cellar.

5. The method for predicting the physical and chemical parameters of the fermented grains in a wine cellar according to claim 4, characterized in that: The loss function is SmoothL1Loss, and the calculation formula is as follows: Among them, Loss represents the loss function, Loss(x i ,y i ) represents the loss function of the i-th sample data, x i Represents the predicted value of the i-th sample data, y i represents the true output label of the i-th sample data, β represents the smoothing parameter, and N represents the number of sample data.

6. The method for predicting the physical and chemical parameters of fermented grains in a wine cellar according to claim 1, characterized in that: The method further comprises: The key indicator characteristics of the outgoing mash from multiple steamers in different groups and their corresponding physicochemical parameter characteristics of the incoming mash are normalized, the key indicator characteristics of the outgoing mash to be predicted are normalized accordingly, and the predicted physicochemical parameters of the first incoming mash and the second incoming mash are inversely normalized accordingly.

7. A device for predicting the physical and chemical parameters of fermented grains in a wine cellar, characterized in that: The device comprises: An acquisition unit is used to acquire key index characteristics of uncellared glutinous rice fermented grains of multiple steamers under different groups and their corresponding physicochemical parameter characteristics of uncellared glutinous rice fermented grains, wherein the key index characteristics of uncellared glutinous rice fermented grains include physicochemical characteristics of uncellared glutinous rice fermented grains, weight characteristics of uncellared glutinous rice fermented grains and production process characteristics of uncellared glutinous rice fermented grains; A construction unit is used to construct a group-by-group physicochemical prediction model of the glutinous grains entering the cellar according to the key indicator characteristics of the glutinous grains out of the cellar under each group and the corresponding physicochemical parameter characteristics of the glutinous grains entering the cellar; and to construct a comprehensive physicochemical prediction model of the glutinous grains entering the cellar according to the key indicator characteristics of the glutinous grains out of the cellar under all groups and the corresponding physicochemical parameter characteristics of the glutinous grains entering the cellar; A prediction unit is used to determine the group of the mash to be predicted and the key indicator characteristics of the mash out of the cellar, input the key indicator characteristics of the mash to be predicted into the corresponding group mash physicochemical prediction model for entering the cellar, and obtain the first physicochemical parameters of the mash to be predicted; input the key indicator characteristics of the mash to be predicted into the comprehensive physicochemical prediction model for entering the cellar, and obtain the second physicochemical parameters of the mash to be predicted; determine the final physicochemical parameters of the mash to be predicted according to the prediction results of the first physicochemical parameters of the mash and the second physicochemical parameters of the mash.

8. The recommended method for producing mash in a wine cellar is characterized in that: The method comprises: Obtaining key indicator characteristics of the uncellared glutinous rice grains of multiple steamers under different groups and their corresponding physical and chemical parameter characteristics of the uncellared glutinous rice grains, wherein the key indicator characteristics of the uncellared glutinous rice grains include physical and chemical characteristics of the uncellared glutinous rice grains, weight characteristics of the uncellared glutinous rice grains, and production process characteristics of the uncellared glutinous rice grains; According to the key index characteristics of the fermented grains out of the cellar in each group and the corresponding physicochemical parameter characteristics of the fermented grains entering the cellar, the corresponding group's physicochemical prediction model of the fermented grains entering the cellar is constructed respectively; A comprehensive physical and chemical prediction model for the mash entering the cellar is constructed based on the key indicator characteristics of the mash leaving the cellar under all groups and the corresponding physical and chemical parameter characteristics of the mash entering the cellar; Based on the physical and chemical prediction model of the group of fermented grains entering the cellar, a group fermented grains production process recommendation model is constructed; based on the comprehensive physical and chemical prediction model of the fermented grains entering the cellar, a comprehensive fermented grains production process recommendation model is constructed; Determine the group of the recommended fermented grains, the physicochemical characteristics of the fermented grains out of the cellar, the weight characteristics of the fermented grains out of the cellar, and the target physicochemical parameters of the fermented grains entering the cellar, input the physicochemical characteristics of the fermented grains out of the cellar, the weight characteristics of the fermented grains out of the cellar, and the target physicochemical parameters of the fermented grains entering the cellar into the corresponding group fermented grains out of the cellar production process recommendation model for optimization, and obtain the first recommended production process of the fermented grains out of the cellar for the fermented grains to be recommended; Inputting the physical and chemical characteristics of the uncellared fermented grains, the weight characteristics of the uncellared fermented grains and the target physical and chemical parameters of the uncellared fermented grains into the comprehensive uncellared fermented grains production process recommendation model for optimization, and obtaining the second uncellared fermented grains production process recommended for the uncellared fermented grains; The final recommended production process for the uncellared fermented grains of the uncellared fermented grains to be predicted is determined according to the first recommended production process for the uncellared fermented grains and the second recommended production process for the uncellared fermented grains.

9. The recommended method for producing fermented grains in a wine cellar according to claim 8, characterized in that: The group-based recommended model for producing fermented grains out of the cellar and the comprehensive recommended model for producing fermented grains out of the cellar are constructed based on an evolutionary algorithm, and the evolutionary algorithm is a genetic algorithm, a simulated annealing algorithm or an evolutionary strategy algorithm; The fitness function of the evolutionary algorithm is as follows: Among them, Fitness represents fitness, obj j represents the physical and chemical parameters of the jth target mash entering the cellar, pre j It represents the jth physicochemical parameter of the fermented grains entering the cellar obtained by the group physicochemical prediction model of the fermented grains entering the cellar or the comprehensive physicochemical prediction model of the fermented grains entering the cellar, and M represents the number of physicochemical parameters of the fermented grains entering the cellar.

10. The recommended device for the production process of mash in a wine cellar is characterized in that: The device comprises: An acquisition unit is used to acquire key index characteristics of uncellared glutinous rice fermented grains of multiple steamers under different groups and their corresponding physicochemical parameter characteristics of uncellared glutinous rice fermented grains, wherein the key index characteristics of uncellared glutinous rice fermented grains include physicochemical characteristics of uncellared glutinous rice fermented grains, weight characteristics of uncellared glutinous rice fermented grains and production process characteristics of uncellared glutinous rice fermented grains; A construction unit is used to construct corresponding group-entering glutinous grains physical and chemical prediction models according to key indicator characteristics of uncellared glutinous grains under each group and corresponding physicochemical parameter characteristics of uncellared glutinous grains; to construct a comprehensive uncellaring glutinous grains physical and chemical prediction model according to key indicator characteristics of uncellared glutinous grains under all groups and corresponding physicochemical parameter characteristics of uncellared glutinous grains; to construct a group-entering glutinous grains production process recommendation model based on the group-entering glutinous grains physical and chemical prediction model, and to construct a comprehensive uncellaring glutinous grains production process recommendation model based on the comprehensive uncellaring glutinous grains physical and chemical prediction model; The recommendation unit is used to determine the group of the mash to be recommended, the physicochemical characteristics of the mash out of the cellar, the weight characteristics of the mash out of the cellar and the target physicochemical parameters of the mash entering the cellar, input the physicochemical characteristics of the mash out of the cellar, the weight characteristics of the mash out of the cellar and the target physicochemical parameters of the mash entering the cellar of the mash to be predicted into the corresponding group mash production process recommendation model to obtain the first recommended production process for the mash out of the cellar; input the physicochemical characteristics of the mash out of the cellar, the weight characteristics of the mash out of the cellar and the target physicochemical parameters of the mash entering the cellar of the mash to be predicted into the comprehensive mash production process recommendation model to obtain the second recommended production process for the mash out of the cellar; determine the final recommended production process for the mash out of the cellar of the mash to be predicted based on the first recommended production process for the mash out of the cellar and the second recommended production process for the mash.

Images