Remote Intelligent Beer Fermentation Monitoring Platform Based on the Internet of Things

Abstract

The present invention discloses a remote intelligent beer fermentation monitoring platform based on the Internet of Things, which relates to the field of fermentation control technology. By setting a fermentation sample collection module to collect fermentation characteristic data and progress label data of fermented beer samples in a fermentation test environment, setting a progress model training module to train a progress prediction model, setting a progress judgment module to obtain the expected progress after the p-th moment output by the progress prediction model, and collecting the actual fermentation progress after the p-th moment, a fermentation regulation module, constructing a state space, an action space and a reward function of a fermentation parameter decision model, and based on the expected progress and the actual fermentation progress, judging whether parameter regulation is required. If parameter regulation is required, through the fermentation parameter decision model, an action decision for the subsequent fermentation process is generated for the fermenter to be monitored; realizing the automation and intelligence of the fermentation process, improving the stability of beer quality and reducing the labor cost.

Description

Technical Field

[0001] The present invention relates to the technical field of fermentation control, specifically a remote intelligent beer fermentation monitoring platform based on the Internet of Things. Background Art

[0002] With the rapid development of Internet of Things (IoT) technology, the traditional beer fermentation process is expected to achieve more intelligent and automated monitoring. Beer fermentation is a complex biochemical process involving dynamic changes in multiple key parameters such as temperature, pressure, pH value, sugar content, etc. Minor fluctuations in these parameters will directly affect the quality of the final product.

[0003] In the traditional fermentation process, operators usually need to manually collect fermentation data regularly and conduct manual analysis and judgment. This is not only time-consuming and laborious, but also difficult to ensure the real-time and accuracy of the data. Manual operation is also difficult to detect and regulate abnormal situations in a timely manner, which may lead to out-of-control fermentation processes and uneven product quality. In addition, the beer brewing industry has strict requirements for the fermentation environment and needs to operate in a clean and enclosed fermentation environment. Manual operation inevitably increases the risk of contamination.

[0004] The traditional manual monitoring and regulation methods can no longer meet the requirements of modern beer brewing for quality stability and production efficiency. To solve this problem, IoT technology has emerged. By deploying various sensors in the fermentation tank, key parameters such as temperature, pressure, pH value, sugar content, and alcohol concentration during the fermentation process can be monitored in real time, and these parameter data can be transmitted to the control center in real time through a wireless network. The control center can conduct intelligent analysis and prediction on the fermentation process through big data analysis and machine learning algorithms, detect abnormal situations in a timely manner, and make regulation decisions.

[0005] However, existing IoT monitoring systems mainly focus on the real-time collection of various parameters during the fermentation process and lack the ability to accurately predict the fermentation progress and intelligent regulation. The fermentation progress is an important indicator comprehensively reflecting the fermentation state. Only by accurately predicting and dynamically regulating the fermentation progress can the automatic intelligent control of the fermentation process be truly achieved.

[0006] A Chinese patent with the publication number CN105467940A discloses a monitoring method for a beer fermentation tank, which includes the following steps: Step 1, real-time detection of the temperature, CO2 gas concentration, CO2 gas flow rate, refrigerant flow rate, and yeast concentration of the beer fermentation tank; Step 2, using a data acquisition and processing device to collect and process the temperature signal, gas concentration signal, gas flow rate signal, liquid flow rate signal, and yeast concentration signal of the beer fermentation tank; Step 3, data display and fault alarm; Step 4, data recording and printing; Step 5, execution of control commands, adjustment of gas flow rate and liquid flow rate, capable of automatically monitoring the beer fermentation tank with high control accuracy. However, this method does not consider the impact of various parameters on the fermentation progress during the fermentation process, resulting in the problem that the progress is not as expected.

[0007] Therefore, the present invention proposes a remote intelligent beer fermentation monitoring platform based on the Internet of Things. Summary of the Invention

[0008] The present invention aims to solve at least one of the technical problems existing in the prior art. For this reason, the present invention proposes a remote intelligent beer fermentation monitoring platform based on the Internet of Things, which realizes the automation and intelligence of the fermentation process, improves the stability of beer quality, and reduces the labor cost.

[0009] To achieve the above object, a remote intelligent beer fermentation monitoring platform based on the Internet of Things is proposed, including a fermentation sample collection module, a progress model training module, a progress judgment module, and a fermentation regulation module; among them, each module is connected electrically.

[0010] The fermentation sample collection module pre-collects the fermentation characteristic data and progress label data of the fermented beer sample in the fermentation test environment, and sends the fermentation characteristic data and progress label data to the progress model training module.

[0011] The progress model training module takes the fermentation characteristic data as input and the progress label data as output, trains the progress prediction model, and sends the trained progress prediction model to the progress judgment module.

[0012] The progress judgment module collects the parameter values of various fermentation parameters in the fermentation tank to be monitored in real time, forms the actual input features, inputs the actual input features into the progress prediction model, obtains the expected progress after the p-th moment output by the progress prediction model, then collects the actual fermentation progress after the p-th moment, and sends the expected progress and the actual fermentation progress to the fermentation regulation module.

[0013] The fermentation regulation module constructs the state space, action space, and reward function of the fermentation parameter decision model, and based on the expected progress and the actual fermentation progress, judges whether parameter regulation is required. If parameter regulation is required, through the fermentation parameter decision model, it generates the action decision for the subsequent fermentation process of the fermentation tank to be monitored.

[0014] The fermentation test environment is set up as follows:

[0015] Prepare N fermentation tanks as test fermentation tanks for parallel multi-batch fermentation experiments; N is the number of selected fermentation tanks;

[0016] Install parameter sensors in each fermentation tank and connect them to the Internet of Things data acquisition device; the parameter sensors are used to collect the fermentation parameters in the test fermentation tank in real time;

[0017] Set up an automatic control system for each test fermentation tank to remotely adjust the control parameters in the fermentation tank;

[0018] The control parameters are adjustable parameters during the fermentation of beer in the fermentation tank.

[0019] The fermentation beer sample is constructed as follows:

[0020] Manually set the raw material ratio for each test fermentation tank;

[0021] Prepare a set of raw materials for fermenting beer for each test fermentation tank according to the raw material ratio;

[0022] Set different fermentation parameter combinations for each test fermentation tank, and use the fermentation parameter combination to start the fermentation process of the test fermentation tank to conduct fermentation experiments to prepare the corresponding beer wort;

[0023] During the fermentation experiment, collect the parameter values and fermentation progress of various fermentation parameters in each test fermentation tank in real time.

[0024] The method for collecting the fermentation characteristic data and progress label data of the fermentation beer sample is as follows:

[0025] Preset the input sequence duration L, the sliding window step size w, and the prediction duration p;

[0026] For each test fermentation tank, set the start time of the fermentation experiment as t0;

[0027] Mark each fermentation parameter as c;

[0028] For any fermentation parameter c, collect the parameter value sequence Qc sorted in chronological order during the fermentation experiment;

[0029] Then for any parameter value sequence Qc, starting from time t0, construct a set of input parameter sequences every sliding window step size w, and the length of the input parameter sequence is L;

[0030] For all fermentation parameters, the input parameter sequences constructed at the same time together form a set of sample input features;

[0031] The sample input features constructed by all test fermenters form the fermentation feature data;

[0032] For each set of sample input features, the fermentation progress after a duration of L + p from its corresponding time is used as the progress label data corresponding to the sample input features.

[0033] The method for training the progress prediction model is as follows:

[0034] In the fermentation feature data, each sample input feature of each test fermenter is used as the input of the progress prediction model. The prediction model takes the predicted value of the fermentation progress at the time p + L after this moment as the output, takes the progress label corresponding to this moment as the prediction target, takes the difference between the predicted value of the fermentation progress and the progress label as the prediction error, and takes minimizing the sum of the prediction errors as the training target; the progress prediction model is trained until the sum of the prediction errors reaches convergence and then the training stops; the progress prediction model is a time series prediction model.

[0035] The method for collecting the parameter values of various fermentation parameters in the fermenter to be monitored in real time and forming the actual input features is as follows:

[0036] Install parameter sensors corresponding to various fermentation parameters in the fermenter to be monitored, and collect the parameter values of various fermentation parameters in real time through the parameter sensors;

[0037] Mark the start time of fermentation of the fermenter to be monitored as t1;

[0038] Represent any moment of t1 + k×w as the prediction moment, where k is an integer between k0 and k1. Here, k0 is a positive integer and simultaneously satisfies t1 + k0×w < L and t1+(k0 - 1)×w ≥ L, and k1 is a positive integer and satisfies that at the moment of t1 + k1×w, the fermentation has ended, while at the moment of t1+(k1 - 1)×w, the fermentation has not ended;

[0039] Mark the prediction moment as t2. For any prediction moment t2, collect the sequence of parameter values of various fermentation parameters changing with time from time t2 - L to t2 to form the actual input features.

[0040] The method for constructing the state space, action space, and reward function of the fermentation parameter decision model is as follows:

[0041] The fermentation parameter decision model is set as an Actor - Critic model;

[0042] Represent each time period within each sliding window step time as the acquisition period;

[0043] Mark the state space as S. The state space S reflects the real-time state of the entire fermentation process, including the parameter value time series of various fermentation parameters in each collection period. Mark the state of each collection period in the state space S as s.

[0044] Mark the action space as A. The action space A includes the adjustment amounts of adjustable parameters output by the Actor model, including the temperature adjustment amount and the stirring frequency adjustment amount. Mark the parameter adjustment action to be taken in the next collection period of fermentation as a.

[0045] The reward function is marked as Q. The reward function Q = r(s, a), which is used to measure the effect of executing action a in the current state s.

[0046] The intelligent agent for designing the fermentation parameter decision model is an Actor-Critic structure based on RNN.

[0047] The input of the intelligent agent is the state s.

[0048] The state value V(s) is output by the Critic network.

[0049] The action probability π(a|s) is output by the Actor network.

[0050] The method for judging whether parameter regulation is needed is as follows: If the difference between the expected progress and the actual fermentation progress is greater than the preset progress preset threshold, it is judged that parameter regulation is needed.

[0051] A fermentation monitoring method for a remote intelligent beer fermentation monitoring platform based on the Internet of Things is proposed, including the following steps:

[0052] Step 1: Collect the fermentation characteristic data and progress label data of the fermented beer sample in the fermentation test environment in advance.

[0053] Step 2: Use the fermentation characteristic data as the input and the progress label data as the output to train the progress prediction model.

[0054] Step 3: Collect the parameter values of various fermentation parameters in the fermenter to be monitored in real time to form the actual input features. Input the actual input features into the progress prediction model to obtain the expected progress after the p-th moment output by the progress prediction model, and then collect the actual fermentation progress after the p-th moment.

[0055] Step 4: Construct the state space, action space, and reward function of the fermentation parameter decision model. Based on the expected progress and the actual fermentation progress, judge whether parameter regulation is needed. If parameter regulation is needed, use the fermentation parameter decision model to generate the action decision for the subsequent fermentation process of the fermenter to be monitored.

[0056] An electronic device is provided, comprising: a processor and a memory, wherein a computer program that can be called by the processor is stored in the memory;

[0057] By calling the computer program stored in the memory, the processor executes the fermentation monitoring method of the above-mentioned remote intelligent beer fermentation monitoring platform based on the Internet of Things.

[0058] A computer-readable storage medium is provided, on which a rewritable computer program is stored;

[0059] When the computer program runs on a computer device, the computer device is caused to execute the fermentation monitoring method of the above-mentioned remote intelligent beer fermentation monitoring platform based on the Internet of Things.

[0060] Compared with the prior art, the beneficial effects of the present invention are:

[0061] The present invention first collects a large amount of fermentation sample data to train a fermentation progress prediction model; then it predicts the fermentation progress in real time and compares it with the actual progress. If an abnormal deviation occurs, regulation is triggered; during regulation, based on the fermentation parameter decision model of reinforcement learning, adjustment strategies for fermentation parameters such as temperature and stirring frequency are automatically generated to bring the fermentation progress back to the normal state. Through modeling and intelligent regulation, the problems of fermentation out of control and product quality fluctuations in the traditional beer brewing process are effectively solved, realizing the automation and intelligence of the fermentation process, improving the stability of beer quality, reducing labor costs, and enhancing production efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 is a module connection relationship diagram of the remote intelligent beer fermentation monitoring platform based on the Internet of Things in Embodiment 1 of the present invention;

[0063] Figure 2 is a flowchart of the fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things in Embodiment 2 of the present invention;

[0064] Figure 3 is a schematic structural diagram of the electronic device in Embodiment 3 of the present invention;

[0065] Figure 4 is a schematic structural diagram of the computer-readable storage medium in Embodiment 4 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0066] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of them. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0067] Example 1

[0068] As Figure 1 shown, the remote intelligent beer fermentation monitoring platform based on the Internet of Things includes a fermentation sample collection module, a progress model training module, a progress judgment module, and a fermentation regulation module; among them, each module is connected electrically;

[0069] The fermentation sample collection module pre-collects the fermentation characteristic data and progress label data of the fermented beer samples in the fermentation test environment, and sends the fermentation characteristic data and progress label data to the progress model training module;

[0070] Specifically, the setting method of the fermentation test environment is as follows:

[0071] Prepare N fermentation tanks as test fermentation tanks for parallel multi-batch fermentation experiments; N is the number of selected fermentation tanks;

[0072] Install parameter sensors in each fermentation tank and connect to the Internet of Things data acquisition device; the parameter sensors are used to collect the fermentation parameters in the test fermentation tank in real time;

[0073] The fermentation parameters include several physical parameters that affect the beer fermentation efficiency, including but not limited to temperature, raw material ratio, pressure, pH value, sugar content, alcohol concentration, carbon dioxide, etc. Each physical parameter is collected in real time using a corresponding parameter sensor. For example, temperature is collected in real time using a temperature sensor;

[0074] Set an automatic control system for each test fermentation tank to remotely adjust the control parameters in the fermentation tank;

[0075] The control parameters are adjustable parameters when fermenting beer in the fermentation tank, including but not limited to stirring frequency, temperature, etc.

[0076] Furthermore, the construction method of the fermented beer sample is as follows:

[0077] Manually set the raw material ratio for each test fermentation tank;

[0078] Prepare a set of raw materials for fermenting beer for each test fermentation tank according to the raw material ratio; the raw materials include malt, yeast strains, etc.;

[0079] Set different fermentation parameter combinations for each test fermentation tank, and use the fermentation parameter combination to start the fermentation process of the test fermentation tank and conduct fermentation experiments to prepare the corresponding beer wort;

[0080] During the fermentation experiment, collect the parameter values and fermentation progress of each fermentation parameter in each test fermentation tank in real time.

[0081] In a preferred embodiment, the fermentation progress can be inferred by measuring the alcohol content, that is, first determining the theoretical final alcohol concentration at the end of the fermentation process according to the proportion of each component in the raw material, and then measuring the ratio of the current alcohol concentration to the final concentration in real time as a measure of the fermentation progress;

[0082] In another preferred embodiment, the fermentation progress can also be evaluated based on the residual amount of reactants, that is, by measuring the residual concentration of glucose and comparing it with the initial glucose concentration, and the consumption ratio obtained reflects the degree of fermentation progress;

[0083] In another preferred embodiment, the fermentation progress can also be inferred by establishing a quantitative relationship model between the metabolite and the fermentation progress, and measuring the concentration or production amount of the metabolite in real time.

[0084] Furthermore, the method for collecting the fermentation characteristic data and progress label data of the fermented beer sample is as follows:

[0085] Preset the input sequence duration L, the sliding window step size w, and the prediction duration p;

[0086] For each test fermentation tank, set the start time of the fermentation experiment as t0;

[0087] Mark each fermentation parameter as c;

[0088] For any fermentation parameter c, collect the parameter value sequence Qc sorted in chronological order during the fermentation experiment;

[0089] Then for any parameter value sequence Qc, starting from time t0, construct a set of input parameter sequences at intervals of the sliding window step size w, and the length of the input parameter sequence is L;

[0090] All the input parameter sequences constructed for all fermentation parameters at the same time together form a set of sample input features.

[0091] As an example, the first input parameter sequence of the fermentation parameter of temperature is the temperature value sequence in the time interval from t0 to t0 + L, and the second input parameter sequence is the temperature value sequence in the time interval from t0 + w to t0 + w + L, and so on;

[0092] Then the first sample input feature contains the parameter value sequences of all fermentation parameters in the time interval from t0 to t0 + L, and the second sample input feature contains the parameter value sequences of all fermentation parameters in the time interval from t0 + w to t0 + w + L;

[0093] The sample input features constructed for all test fermentation tanks form the fermentation characteristic data;

[0094] For each set of sample input features, the fermentation progress after a duration of L + p from its corresponding time is used as the progress label data corresponding to the sample input features.

[0095] For example, for the first set of sample input features, the corresponding progress label is the fermentation progress at time t0 + L + p.

[0096] It can be understood that using the real-time sequence of various fermentation parameter values as the input and the fermentation progress after the prediction duration p as the prediction target to train the progress prediction model, the expected fermentation progress in the fermenter after the future prediction duration p can be obtained. Thus, after the future prediction duration p, the actual fermentation progress can be collected and compared with the expected fermentation progress to determine whether there are problems such as slow fermentation or too fast fermentation.

[0097] The progress model training module takes the fermentation feature data as the input, the progress label data as the output, trains the progress prediction model, and sends the trained progress prediction model to the progress judgment module.

[0098] Specifically, the method for training the progress prediction model is as follows:

[0099] In the fermentation feature data, each sample input feature of each test fermenter is used as the input of the progress prediction model. The prediction model takes the predicted value of the fermentation progress at the p + L moment after this moment as the output, the progress label corresponding to this moment as the prediction target, the difference between the predicted value of the fermentation progress and the progress label as the prediction error, and minimizing the sum of the prediction errors as the training target. The progress prediction model is trained until the sum of the prediction errors reaches convergence and then the training stops. The progress prediction model is a time series prediction model. The time series prediction model is any one of the RNN model or the LSTM model.

[0100] The progress judgment module collects in real time the parameter values of various fermentation parameters in the fermenter to be monitored to form the actual input features, inputs the actual input features into the progress prediction model to obtain the expected progress after the p moment output by the progress prediction model, then collects the actual fermentation progress after the p moment, and sends the expected progress and the actual fermentation progress to the fermentation control module.

[0101] Specifically, the method for collecting in real time the parameter values of various fermentation parameters in the fermenter to be monitored to form the actual input features is as follows:

[0102] Install parameter sensors corresponding to various fermentation parameters in the fermenter to be monitored, and collect the parameter values of various fermentation parameters in real time through the parameter sensors.

[0103] Mark the start time of fermentation of the fermenter to be monitored as t1.

[0104] Any moment of \(t1 + k\times w\) is represented as a prediction moment, where \(k\) is an integer between \(k0\) and \(k1\), \(k0\) is a positive integer, and at the same time \(t1 + k0\times w\lt L\), \(t1+(k0 - 1)\times w\geq L\), and \(k1\) is a positive integer, and at the moment of \(t1 + k1\times w\), the fermentation has ended, while at the moment of \(t1+(k1 - 1)\times w\), the fermentation has not ended;

[0105] Mark the prediction moment as \(t2\). For any prediction moment \(t2\), collect the parameter value sequence of various fermentation parameters changing with time between time \(t2 - L\) and \(t2\) to form the actual input features.

[0106] It can be understood that the collection of the actual input features reflects the fermentation state in the current fermenter, so as to predict the fermentation progress after the future prediction duration \(p\) according to this fermentation state. The predicted fermentation progress can be regarded as the expected progress. After the time reaches the future prediction duration \(p\), the actual fermentation progress can be obtained through a preset fermentation progress calculation method.

[0107] Fermentation control module, construct the state space, action space and reward function of the fermentation parameter decision model. Based on the expected progress and the actual fermentation progress, judge whether parameter adjustment is needed. If parameter adjustment is needed, through the fermentation parameter decision model, generate action decisions for the fermenter to be monitored in the subsequent fermentation process;

[0108] Specifically, the ways to construct the state space, action space and reward function of the fermentation parameter decision model are as follows:

[0109] The fermentation parameter decision model is set as the Actor-Critic model;

[0110] Represent the time period within each sliding window step time as the acquisition period;

[0111] Mark the state space as \(S\). The state space \(S\) reflects the real-time state of the entire fermentation process, including the parameter value time series of various fermentation parameters within each acquisition period; mark the state of each acquisition period in the state space \(S\) as \(s\);

[0112] Mark the action space as \(A\). The action space \(A\) includes the adjustment amounts of adjustable parameters output by the Actor model, such as temperature adjustment amount, stirring frequency adjustment amount, etc. Mark the parameter adjustment action taken by the fermentation in the next acquisition period as \(a\);

[0113] Mark the reward function as \(Q\). The reward function \(Q = r(s,a)\) is used to measure the effect of executing action \(a\) in the current state \(s\);

[0114] Specifically, the reward function \(Q\) can be specifically set as:

[0115] Q = a1×(1 - |P_t - P|) - a2×∑ i hi;

[0116] Wherein, P_t represents the actual fermentation progress at the current moment, P represents the expected progress, which is used to measure whether the deviation between the fermentation progress and the theoretical progress is reduced after performing the action a. The smaller the deviation, the larger the reward Q.

[0117] i is the number of adjustable parameters, hi is the specific adjustment amount of the i-th adjustable parameter. The smaller the adjustment amount, the smaller the perturbation to the fermentation process, the better the execution effect, and the larger the reward Q;

[0118] Wherein, both a1 and a2 are preset proportionality coefficients.

[0119] It can be understood that the setting of the above reward function Q is only a simple example provided by this embodiment. In practical applications, the reward function Q can be adaptively modified, deleted, or added.

[0120] The intelligent agent for designing the fermentation parameter decision model is an Actor-Critic structure based on RNN;

[0121] The input of the intelligent agent is the state s;

[0122] The state value V(s) is output by the Critic network;

[0123] The action probability π(a|s) is output by the Actor network;

[0124] Furthermore, the method for determining whether parameter regulation is required is: if the difference between the expected progress and the actual fermentation progress is greater than the preset progress threshold, it is determined that parameter regulation is required;

[0125] Furthermore, the method for generating the action decision for the subsequent fermentation process of the fermentation tank to be monitored through the fermentation parameter decision model is:

[0126] Taking the real-time actual input features of the fermentation tank to be monitored as the current state s; inputting the current state s into the Actor model of the fermentation parameter decision model, and the Actor model outputs an action in the action space A to maximize the state-action value output by the Critic model, and outputs the action decision of the adjustment amount of each adjustable parameter in the fermentation tank to be monitored;

[0127] Execute the action decision to obtain the next state s' and the actual reward value;

[0128] The Critic network outputs the state-action value after executing the action decision in the current state, calculates the TD error δ of the Critic network, and updates the Critic network parameters according to δ to minimize the Critic loss function; the state-action value output by the Critic model is calculated by the state-action value function generated by the temporal difference optimization method;

[0129] Specifically, the state-action value function generated by the temporal difference optimization method can be:

[0130] Q(s, a) = r(s, a) + γ × V(s'), where γ is the discount factor and V(s') is the estimated value in state s'; generally, the estimated value can be replaced by Q(s', a'), that is, V(s') represents the estimated value of the state-action value function corresponding to the next state s' and action a', so as to ensure that Q(s, a) can consider more in the long term and not only consider the current state;

[0131] Then the TD error δ is δ = (r(s, a) + γ × V(s′) - Q(s, a)) 2 ;

[0132] The Actor model aims to maximize the state-action value output by the Critic model and updates the parameters of the Actor network model using the policy gradient update method.

[0133] Embodiment 2

[0134] As Figure 2 shown, the fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things includes the following steps:

[0135] Step 1: Collect the fermentation characteristic data and progress label data of the fermented beer sample in the fermentation test environment in advance;

[0136] Step 2: Train a progress prediction model with the fermentation characteristic data as the input and the progress label data as the output;

[0137] Step 3: Collect the parameter values of various fermentation parameters in the fermenter to be monitored in real time to form actual input features, input the actual input features into the progress prediction model to obtain the expected progress after p moments output by the progress prediction model, and then collect the actual fermentation progress after p moments; p is the preset prediction duration;

[0138] Step 4: Construct the state space, action space, and reward function of the fermentation parameter decision model. Based on the expected progress and the actual fermentation progress, determine whether parameter regulation is required. If parameter regulation is required, generate an action decision for the subsequent fermentation process of the fermenter to be monitored through the fermentation parameter decision model.

[0139] Embodiment 3

[0140] Figure 3 It is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 3 shown, according to another aspect of the present application, an electronic device 100 is further provided. The electronic device 100 may include one or more processors and one or more memories. Among them, computer-readable code is stored in the memory, and when the computer-readable code is run by one or more processors, it can execute the fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things as described above.

[0141] The method or device according to the embodiment of the present application can also be implemented by means of Figure 3 the architecture of the electronic device shown. As Figure 3 shown, the electronic device 100 may include a bus 101, one or more CPUs 102, a ROM 103, a RAM 104, a communication port 105 connected to the network, an input / output component 106, a hard disk 107, etc. The storage device in the electronic device 100, such as the ROM 103 or the hard disk 107, can store the fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things provided by the present application.

[0142] Furthermore, the electronic device 100 may further include a user interface 108. Of course, Figure 3 the architecture shown is only exemplary, and when implementing different devices, one or more components in the electronic device shown may be omitted according to actual needs. Figure 3

[0143] Embodiment 4

[0144] Figure 4 It is a schematic structural diagram of a computer-readable storage medium provided by an embodiment of the present application. As Figure 4 shown, it is a computer-readable storage medium 200 according to an embodiment of the present application. Computer-readable instructions are stored on the computer-readable storage medium 200. When the computer-readable instructions are run by a processor, the fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things according to the embodiment of the present application described with reference to the above drawings can be executed. The computer-readable storage medium 200 includes but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory, etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0145] ​In addition, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium storing machine-readable instructions that can be run by a processor to execute instructions corresponding to the method steps provided by the present application. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0146] The method, device, and equipment of the present application can be implemented in many ways. For example, the method, device, and equipment of the present application can be implemented by software, hardware, firmware, or any combination of software, hardware, and firmware. The above order of steps for the method is only for illustration, and the steps of the method of the present application are not limited to the specific order described above, unless otherwise specifically stated. In addition, in some embodiments, the present application can also be implemented as a program recorded in a recording medium, and these programs include machine-readable instructions for implementing the method according to the present application. Therefore, the present application also covers a recording medium storing a program for executing the method according to the present application.

[0147] In addition, parts of the above technical solutions provided in the embodiments of the present application that are consistent with the implementation principles of the corresponding technical solutions in the prior art are not described in detail to avoid excessive elaboration.

[0148] As described above in the specific embodiments, the purpose, technical solution, and beneficial effects of the present invention are further described in detail. It should be understood that the above is only the specific embodiment of the present invention and is not used to limit the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0149] The above preset parameters or preset thresholds are all set by those skilled in the art according to the actual situation or obtained through a large amount of data simulation.

[0150] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A remote intelligent beer fermentation monitoring platform based on the Internet of Things, characterized in that, Including: A fermentation sample collection module, a progress model training module, a progress judgment module, and a fermentation regulation module; among them, each module is connected electrically; The fermentation sample collection module pre-collects the fermentation characteristic data and progress label data of the fermented beer sample in the fermentation test environment, and sends the fermentation characteristic data and progress label data to the progress model training module; The progress model training module takes the fermentation characteristic data as input and the progress label data as output, trains the progress prediction model, and sends the trained progress prediction model to the progress judgment module; The progress judgment module collects the parameter values of various fermentation parameters in the fermentation tank to be monitored in real time to form the actual input features, inputs the actual input features into the progress prediction model to obtain the expected progress after the pth moment output by the progress prediction model, then collects the actual fermentation progress after the pth moment, and sends the expected progress and the actual fermentation progress to the fermentation regulation module; p is the preset prediction duration; The fermentation regulation module constructs the state space, action space, and reward function of the fermentation parameter decision model. Based on the expected progress and the actual fermentation progress, it judges whether parameter regulation is needed. If parameter regulation is needed, through the fermentation parameter decision model, it generates the action decision for the subsequent fermentation process of the fermentation tank to be monitored.

2. The remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 1, characterized in that The setting method of the fermentation test environment is as follows: Prepare N fermentation tanks as test fermentation tanks for parallel multi-batch fermentation experiments; N is the number of selected fermentation tanks; Install parameter sensors in each fermentation tank and connect them to the Internet of Things data acquisition device; the parameter sensors are used to collect the fermentation parameters in the test fermentation tank in real time; Set an automatic control system for each test fermentation tank to remotely adjust the control parameters in the fermentation tank; The control parameters are adjustable parameters when fermenting beer in the fermentation tank.

3. The remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 2, characterized in that, The construction method of the fermented beer sample is as follows: Manually set the raw material ratio for each test fermentation tank; Prepare a set of raw materials for fermenting beer for each test fermentation tank according to the raw material ratio; Set different fermentation parameter combinations for each test fermentation tank, and use the fermentation parameter combination to start the fermentation process of the test fermentation tank to conduct fermentation experiments to prepare the corresponding beer paste sugar; During the fermentation experiment process, collect the parameter values and fermentation progress of various fermentation parameters in each test fermentation tank in real time.

4. The remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 3, characterized in that The method of collecting the fermentation characteristic data and progress label data of the fermented beer sample is as follows: Preset the input sequence duration L and the sliding window step size w; For each test fermentation tank, set the start time of the fermentation experiment as t0; Mark each fermentation parameter as c; For any fermentation parameter c, collect the parameter value sequence Qc composed of its values in chronological order during the fermentation experiment; Then for any parameter value sequence Qc, starting from the t0 time, construct a set of input parameter sequences at intervals of the sliding window step size w, and the length of the input parameter sequence is L; For all fermentation parameters, the input parameter sequences constructed at the same time together form a set of sample input features; The sample input features constructed by all test fermentation tanks form the fermentation characteristic data; For each set of sample input features, the fermentation progress after a duration of L + p from its corresponding time is used as the progress label data corresponding to the sample input features.

5. The remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 4, characterized in that, The method for training the progress prediction model is as follows: In the fermentation feature data, each sample input feature of each test fermenter is used as the input of the progress prediction model. The prediction model takes the predicted value of the fermentation progress at the p + L moment after this moment as the output, takes the progress label corresponding to this moment as the prediction target, takes the difference between the predicted value of the fermentation progress and the progress label as the prediction error, and takes minimizing the sum of the prediction errors as the training target; the progress prediction model is trained until the sum of the prediction errors reaches convergence and then the training stops; the progress prediction model is a time series prediction model.

6. The remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 5, characterized in that The method for collecting the parameter values of various fermentation parameters in the fermenter to be monitored in real time and forming the actual input features is as follows: Install parameter sensors corresponding to various fermentation parameters in the fermenter to be monitored, and collect the parameter values of various fermentation parameters in real time through the parameter sensors; Mark the start time of fermentation of the fermenter to be monitored as t1; Represent any moment of t1 + k×w as the prediction moment, where k is an integer between k0 and k1. Here, k0 is a positive integer and simultaneously satisfies t1 + k0×w < L and t1 + (k0 - 1)×w ≥ L, and k1 is a positive integer and satisfies that at the moment of t1 + k1×w, the fermentation has ended, while at the moment of t1 + (k1 - 1)×w, the fermentation has not ended; Mark the prediction moment as t2. For any prediction moment t2, collect the sequence of parameter values of various fermentation parameters changing with time between the time t2 - L and t2, and form the actual input features.

7. The remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 6, characterized in that The method for constructing the state space, action space, and reward function of the fermentation parameter decision model is as follows: The fermentation parameter decision model is set as an Actor-Critic model; Represent the time period within each sliding window step time as the acquisition period; Mark the state space as S. The state space S reflects the real-time state of the entire fermentation process, including the time series of the parameter values of various fermentation parameters within each acquisition period; mark the state of each acquisition period within the state space S as s; Mark the action space as A. The action space A includes the adjustment amounts of the adjustable parameters output by the Actor model, including the temperature adjustment amount and the stirring frequency adjustment amount. Mark the parameter adjustment action taken by the fermentation in the next acquisition period as a; Mark the reward function as Q. The reward function Q = r(s, a), which is used to measure the effect of performing the action a in the current state s; Design the intelligent agent of the fermentation parameter decision model as an Actor-Critic structure based on RNN; The input of the intelligent agent is the state s; Output the state value V(s) by the Critic network; Output the action probability π(a|s) by the Actor network.

8. The fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things is implemented based on the remote intelligent beer fermentation monitoring platform based on the Internet of Things described in any one of claims 1 - 7, and includes the following steps: Step 1: Collect fermentation characteristic data and progress label data of the fermented beer sample in the fermentation test environment in advance; Step 2: Use the fermentation characteristic data as the input and the progress label data as the output to train the progress prediction model; Step 3: Collect the parameter values of various fermentation parameters in the fermenter to be monitored in real time to form the actual input features, input the actual input features into the progress prediction model to obtain the expected progress after the p-th moment output by the progress prediction model, and then collect the actual fermentation progress after the p-th moment; Step 4: Construct the state space, action space and reward function of the fermentation parameter decision model. Based on the expected progress and the actual fermentation progress, judge whether parameter regulation is needed. If parameter regulation is needed, use the fermentation parameter decision model to generate action decisions for the subsequent fermentation process of the fermenter to be monitored.

9. An electronic device, characterized in that, Including: A processor and a memory, wherein, The memory stores a computer program that can be called by the processor; The processor executes the fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 8 in the background by calling the computer program stored in the memory.

10. A computer-readable storage medium, characterized in that, On which a rewritable computer program is stored; When the computer program runs on the computer device, the computer device executes the fermentation monitoring method of the remote intelligent beer fermentation monitoring platform based on the Internet of Things according to claim 8 in the background.

Images