Optimized preparation method of hydrogen-rich water and multi-cycle generation system

Abstract

The present invention discloses a method for optimizing the preparation of hydrogen-rich water and a multi-cycle generation system; the present invention relates to the technical field of hydrogen-rich water preparation; input the dependent variable y into the regression function f(y), including the hydrogen content h, the ORP value O, and the oxygen content o, and use the RBF kernel function to utilize its dot product operation and its support vector x i , process the non-linear relationship between the independent variable x and the dependent variable y, and perform error constraint on the regression function f(y) by the ε-insensitive loss function; for the preparation of large batches of products on an industrial scale, the solution of the present invention enables manufacturers to no longer be limited to fixed process indicators; when manufacturers need to complete the production tasks of hydrogen-rich water products with different hydrogen contents, ORP values, and oxygen contents, the technical solution of the present invention can help them flexibly control the temperature and the soaking time of the hydrogen rod to meet the production tasks. That is, whether it is a product with a high hydrogen content or a specific ORP value, the corresponding production parameters can be obtained through this solution and automated control can be carried out.

Description

Technical Field

[0001] The present invention relates to the technical field of hydrogen-rich water preparation, specifically to the technical field of preparing HRW by electrolyzing water, and particularly to an optimized preparation method of hydrogen-rich water and a multi-cycle generation system. Background Art

[0002] Hydrogen-rich water (HRW), that is, "water rich in dissolved hydrogen molecules", is commonly known as "hydrogen molecule bubble water" or "hydrogen water" in the market. Its hydrogen content, ORP value (oxidation-reduction potential), and oxygen content are its main process indicators:

[0003] 1) Hydrogen content: The main efficacy of hydrogen-rich water comes from the dissolved hydrogen molecules. Therefore, it is necessary to ensure that the hydrogen-rich water contains a sufficient and stable hydrogen molecule concentration.

[0004] 2) ORP value: A lower ORP value means that hydrogen-rich water has stronger reducibility, which is directly related to its antioxidant ability.

[0005] 3) Oxygen content (dissolved oxygen): The dissolved oxygen in hydrogen-rich water plays a certain role in maintaining the ecological balance and biological activity of the water body.

[0006] HRW not only plays important roles in the biomedical field, such as reducing tissue oxidative damage after ischemia-reperfusion, inhibiting inflammatory reactions, preventing and treating metabolic diseases, preventing and treating arteriosclerosis, and treating ophthalmic diseases, but also plays important roles in the field of crops, such as promoting biosynthesis, preventing oxidative damage, and increasing the level of growth hormone.

[0007] Currently, there are mainly three methods for preparing HRW. One is to introduce H2 into water to prepare saturated water, the second is to use the reaction between metallic magnesium and water to produce hydrogen to prepare HRW, and the third is to produce hydrogen by electrolyzing water to prepare HRW.

[0008] For the electrolytic preparation of HRW by the third method and the principle of its hydrogen production unit, reference can be made to the content of the literature [3-4] ; however, the latest research shows [5] , in the hydrogen production unit, the influencing factors for preparing HRW by using a hydrogen rod are as Figure 2 shown: When the temperature remains unchanged, the hydrogen content in HRW increases with the prolonging of the hydrogen production time of the hydrogen rod immersed in water, the ORP value decreases with the prolonging of the immersion time, and the oxygen content decreases with the increase of temperature. There is a non-linear relationship between these independent variables and dependent variables.

[0009] Although an experimentally relatively optimized preparation process has been given [5]: When preparing HRW using a container with a diameter of 25 mm, the hydrogen content in the HRW of the group without removing the hydrogen rod remained constant after dropping to 0.5 mg / L, and the hydrogen content in the group with the hydrogen rod removed decreased to 0 mg / L at 72 h. However, this undoubtedly restricts the large-scale industrial production of hydrogen-rich water with ideal and accurate hydrogen content, ORP value, and oxygen content. Therefore, some existing technologies attempt to solve this technical problem. For example:

[0010] An existing technology discloses a hydrogen-rich water preparation device with optimized electrolysis [1] : The hydrogen discharge and pressure limiting mechanism can control the above independent variables, thereby realizing the regulation of the dependent variables. However, in essence, it does not solve the above non-linear relationship, and subjective debugging by the user is still required in practical applications, so the practicality is not high;

[0011] Another existing technology discloses electro-thermal cold hydrogen storage technology [2] : The optimal operating conditions of the system under multiple sets of planning schemes are optimized at the operation level. The planning results in the upper-layer configuration optimization stage serve as the input parameters for the lower-layer operation optimization stage, directly affecting the lower-layer objective function and constraint conditions. The operation results of the lower-layer operation optimization stage will be timely fed back to the upper layer to verify the accuracy of the upper-layer planning results, and prompt the planning stage model to re-find the optimal planning scheme and continuously cycle to achieve the overall optimum. However, in essence, it belongs to a linear programming technology based on prior experience, and it is not very applicable to the non-linear relationship between the above independent variables and dependent variables;

[0012] Therefore, the present invention proposes a method for optimizing the preparation of hydrogen-rich water and a multi-cycle generation system.

[0013] The cited documents in the present invention are as follows:

[0014] [1] Yang Xiaoxi, Huang Nan, Cheng Longfei, etc. An electrolytic cell with the ability to prepare hydrogen-rich water [P]. CN202210237297.8: 2022-05-06.

[0015] [2] Yang Jiaqi, Zhang Shunyu, Zhou Zhenling, etc. An integrated electro-thermal cold hydrogen storage comprehensive energy system and method for production and storage [P]. CN202311088386.1: 2023-11-24.

[0016] [3] Park Yincheol, Kim Ilbong. Portable HRW manufacturing device:

[0017] CN201180022475.5 [P]. 2013-04-10.

[0018] [4] Zhang Minghui. Design and research of a hydrogen-rich water cup for the elderly based on the concept of service design [D]. Qingdao University, 2019.

[0019] [5] Huang Qingjian, Zhang Shuangshuang, Sha Jibin, et al. Experimental study on preparation of hydrogen-rich water by hydrogen rod [J]. Military Medical Sciences, 2016, 40(8): 646-650. Summary of the Invention

[0020] In view of this, embodiments of the present invention hope to provide an optimized preparation method for hydrogen-rich water and a multi-cycle generation system, so as to solve or alleviate the technical problems existing in the prior art, that is: for production tasks with different production requirements and specification indicators, namely production tasks of hydrogen-rich water products with different hydrogen contents, ORP values and oxygen contents, how to flexibly control the temperature and the soaking time of the hydrogen rod to achieve mass production of hydrogen-rich water with ideal and accurate hydrogen content, ORP value and oxygen content. And at least provide a beneficial choice for this; the technical solution of the present invention is realized as follows:

[0021] In the first aspect, an optimized preparation method for hydrogen-rich water:

[0022] (I) Overview:

[0023] This solution aims to solve the above technical problems; by introducing an SVR regression model to perform regression tasks, dealing with the non-linear relationship between independent variables (temperature and soaking time of the hydrogen rod) and dependent variables (hydrogen content, ORP value and oxygen content), and introducing a dynamic compensation mechanism to correct the SVR regression model in each production task; after generating the corresponding error value e, calculating the control error signal e1(t) and the control error signal e2(t), and controlling the multi-cycle generation system of hydrogen-rich water based on the PID controller algorithm to achieve mass production of hydrogen-rich water with ideal and accurate hydrogen content, ORP value and oxygen content.

[0024] (II) PID controller algorithm:

[0025] For this solution, first of all, it is necessary to clarify the method of controlling the independent variable to reach the dependent variable. The PID controller algorithm selected in this solution constitutes the control quantity through the linear combination of three links: proportional (P), integral (I) and derivative (D), and controls the controlled object.

[0026] 2.1 Application of the PID controller:

[0027] For the gas-liquid mixing unit and the hydrogen production unit, the control error signal e1(t) and the control error signal e2(t) need to be applied respectively, then:

[0028] 1) For the gas-liquid mixing unit:

[0029]

[0030] 2) For the hydrogen production unit:

[0031]

[0032] Where: u1(t) and u2(t) are the control output electrical signals of the gas-liquid hybrid unit and the hydrogen production unit respectively; K P1 , K I1 , K D1 are the proportional, integral, and derivative coefficients of the PID controller of the gas-liquid hybrid unit; K P2 , K I2 , K D2 are the proportional, integral, and derivative coefficients of the PID controller of the hydrogen production unit; e1(t) and e2(t) are the control error signals of the gas-liquid hybrid unit and the hydrogen production unit respectively, that is, the difference between the expected output and the actual output; τ is the time variable.

[0033] 2.2 PID control mechanism:

[0034] Proportional link (P): Proportional to the error signal, used to quickly respond to error changes. The larger the proportional coefficient K P , the faster the system response.

[0035] Integral link (I): Integrates past errors to eliminate the steady-state error of the system. The integral coefficient K I determines the strength of the integral action, and an appropriate integral action can improve the control accuracy of the system.

[0036] Derivative link (D): Predicts the trend of error changes, generates an anticipatory control action, used to reduce overshoot, overcome oscillations, and improve the stability of the system. The larger the derivative coefficient K D , the stronger the derivative action.

[0037] 2.3 Technical idea:

[0038] Based on 2.1 - 2.2, it can be known that currently, it is necessary to further clarify the difference between the expected output and the actual output, that is, to obtain the control error signals e1(t) and e2(t); however, for this solution, there is a non-linear relationship between the independent variable and the dependent variable; therefore, how to handle the non-linear relationship between the two, and then predict the dependent variable, and further control the controllable independent variable, can realize the control of the hydrogen-rich water multi-cycle generation system, and achieve mass production of hydrogen-rich water with ideal and accurate hydrogen content, ORP value, and oxygen content.

[0039] (Three) Technical solution:

[0040] To achieve the above goals, this solution introduces the following steps S1 - S4 and executes them in a loop for each preparation batch:

[0041] 3.1 Step S1, read the SVR regression model:

[0042] Input the dependent variable y into the regression function f(y), including the hydrogen content h, the ORP value O, and the oxygen content o, and use the RBF kernel function to perform its dot product operation and its support vector x i , process the non-linear relationship between the independent variable x and the dependent variable y, and use the ε-insensitive loss function to perform error constraint on the regression function f(y); output the corresponding independent variable x by the regression function f(y), including the temperature T and the immersion time S of the hydrogen rod. Step S1 specifically includes the following steps S100 to S101;

[0043] 3.1.1 Step S100, process the regression task:

[0044] First, hand it over to the RBF kernel function K(y, x i ) to perform the dot product operation:

[0045]

[0046] Among them: Υ is the hyperparameter of the RBF kernel function; σ is the width parameter of the RBF kernel function, which controls the radial action range of the function. ||y - x i || 2 is the square of the Euclidean distance between the dependent variable y and the support vector x i ; the RBF kernel function can map the input data to a high-dimensional space to process complex non-linear relationships.

[0047] The goal of the SVR regression model is to find a regression function f(y) such that for a given dependent variable y (input value), performing the regression task can predict the corresponding independent variable x (output value):

[0048]

[0049] Among them: α i and α i * are the Lagrange multipliers obtained through training; b is the bias term; x i is the i-th support vector x i ;; n is the number of support vectors.

[0050] 3.1.2 Step S101, introduce the constraint of the ε-insensitive loss function L ε (x’, f(y)):

[0051]

[0052] Among them, x' is the value of the independent variable x in the previous preparation batch. The radius ε refers to the radius of the insensitive region, which is used to control the tolerance of the model to prediction errors. When the absolute value of the difference between the predicted value and the true value is less than or equal to the radius ε, the SVR regression model considers its prediction result accurate this time, and the loss is 0 at this time. Once the absolute value of the difference between the predicted value and the true value exceeds this range, the loss will accumulate linearly.

[0053] 3.2 Method for training the SVR regression model:

[0054] Although the above step S1 uses the SVR regression model, this model still needs to be trained before it can be used. The methods include:

[0055] 3.2.1 First step, data engineering:

[0056] Collect historical data, record the dependent variable y under different hydrogen contents h, ORP values O, and oxygen contents o, as well as the corresponding temperature T and hydrogen rod immersion time S, that is, the independent variable x; and then construct a data set D:

[0057] D = {[(T1, S1)(h1, O 1, o1)], [(T2, S2)(h2, O 2, o2)],..., [(T m , S m )(h m , O m, o m )]};

[0058] Among them, [(T i , S i )(h i , O i, o i )] is the i-th sample, and m is the number of samples.

[0059] Then remove outliers and noise from the data set D and perform normalization processing.

[0060] 3.2.2 Second step, set the convex optimization objective:

[0061] The goal is to find a hyperplane to divide the data points in the samples of the data set D and form several support vectors x i ; Based on the RBF kernel function K(y, x i ) to perform the kernel trick, implicitly implementing the dot product operation in the high-dimensional space

[0062] 3.2.3 Third step, constrain the convex optimization objective:

[0063] Introduce the ε-insensitive loss function as in step S101 to constrain the convex optimization objective. When the absolute value of the difference between the predicted value and the true value is less than or equal to the radius ε, the SVR regression model considers its current prediction result to be accurate, and the loss is 0 at this time. Once the absolute value of the difference between the predicted value and the true value exceeds this range, the loss will accumulate linearly.

[0064] 3.2.4 The third step, training:

[0065] Training is to continuously solve the above convex optimization problem, and introduce the k-fold cross-validation method during this process; the overall training steps are as follows:

[0066] (1) Randomly divide the dataset D into k subsets of equal size, namely D1, D2,..., D k

[0067] (2) Loop to execute:

[0068] Select D i as the test set D test , and merge the remaining subsets as the training set D train ; in the training set D train , use the QP algorithm (quadratic programming algorithm) to solve the convex optimization objective; it is responsible for finding a set of Lagrange multipliers α i and α i * and requires it to correspond to the data points, and then the normal vector W can be obtained.

[0069] Based on first constructing the Lagrangian function to obtain a dual problem that only contains the Lagrange multipliers α i and α i * and then solve it, and then obtain the optimal Lagrange multipliers α i and α i * values. Furthermore, the normal vector W can be represented by α i and α i * as well as the corresponding support vector x i . At the same time, the value of the bias term b can also be directly obtained through the support vector x i .

[0070] (3) Use the test set D test to test the trained SVR model. Calculate the average value of the k test results as the final evaluation of the model performance.

[0071] (4) Select the model hyperparameters obtained from the round of training that minimizes the mean squared error MSE result as the optimal model hyperparameters; that is, the Lagrange multipliers α i and α i *, the normal vector W, the bias term b, the penalty coefficient C, and the hyperparameter Υ;

[0072] (5) After the model converges or reaches the predetermined number of loop iterations, stop the training.

[0073] 3.3 Step S2, dynamic compensation mechanism:

[0074] For the dependent variable y input in Step S1, minimize the sum of squared residuals S of the dependent variable y through the least squares method to obtain the error feedback coefficient β i , trace back the historical prediction errors for compensation. Form the compensated independent variable x'';

[0075] Error feedback coefficient β i can help the SVR regression model "remember - correct" past prediction errors, thereby improving the accuracy of current predictions.

[0076] 3.3.1 Step S200, calculate the sum of squared residuals S:

[0077] Trace back the previous M historical moments, each moment including the actual value y of the dependent variable at that time i and the predicted value Calculate the sum of squared residuals S to quantify the accuracy of the regression task (prediction):

[0078]

[0079] Calculating the sum of squared residuals S to quantify the gap between the predicted value and the actual value is the objective function that needs to be minimized during the optimization process.

[0080] 3.3.2 Step S201, obtain the error feedback coefficient β i :

[0081] In order to use historical prediction errors to improve current predictions, solve for the error feedback coefficient β through the least squares method i , so that the predicted value after compensation is closer to the actual value. Let ∈ i = y i - y i ’ be the historical prediction error, and construct a linear relationship to represent the predicted value after compensation:

[0082]

[0083] Among them, is the dependent variable after compensation, a i-1 is the prediction error of the previous data point. To solve for β i , substitute the above linear relationship into the sum of squared residuals S and take the derivative of β i and set the derivative equal to 0 to find the β i value that minimizes the sum of squared residuals S.

[0084] The minimum β is obtained. i After obtaining the value, it can be used to update the independent variable x":

[0085] x" = x' + β i Δx;

[0086] where Δx is an adjustment amount related to the historical prediction error. In this way, the SVR regression model can "remember - correct" past prediction errors, thereby improving the accuracy of current predictions.

[0087] 3.4 Step S3, calculate the error value e:

[0088] Based on the independent variable x" obtained in step S2, calculate the difference between it and the independent variable x' in the previous preparation batch, which can be the absolute value difference or the relative error, to obtain the error value e;

[0089] 3.5 Step S4, apply the PID controller:

[0090] Based on the technical solutions described in 2.1 - 2.2; when we obtain the error value e, we can obtain the control error signal e1(t) and the control error signal e2(t):

[0091] e1(t) = |e - u1(t - 1)|;

[0092] e2(t) = |e - u2(t - 1)|;

[0093] where u1(t - 1) and u2(t - 1) are the control output electrical signals of the gas - liquid mixing unit and the hydrogen production unit in the previous production batch respectively.

[0094] Then, input the control output electrical signals u1(t - 1) and u2(t - 1) into the PID controller algorithm described in 2.1 - 2.2. After outputting the control output electrical signals u1(t) and u2(t), perform dynamic control on the gas - liquid mixing unit and the hydrogen production unit to meet the production tasks.

[0095] (IV) Mechanism for solving technical problems:

[0096] The SVR regression model can capture the complex non - linear relationship between the independent variables (temperature and hydrogen rod immersion time) and the dependent variables (hydrogen content, ORP value, and oxygen content). In each production task, by comparing the actual measured values with the predicted values of the SVR model; after generating the error value e through step S3, it can reflect the difference between the model prediction and the actual production;

[0097] Two control error signals, e1(t) and e2(t), are calculated based on the error value e, which reflect the gap between the current production state and the target state. The PID controller adjusts the temperature and the soaking time of the hydrogen rod according to these error signals to reduce the error and drive the system to approach the target state.

[0098] Furthermore, through the real-time feedback and adjustment of the PID controller, the system can flexibly and precisely control the temperature and the soaking time of the hydrogen rod during the production process. Furthermore, the system can be flexibly adjusted according to different production requirements (i.e., different hydrogen content, ORP value, and oxygen content indicators).

[0099] In practical applications, through the cyclic operation and continuous optimization of the above mechanism, the system can maintain a high degree of consistency and accuracy in a large-scale production environment. It ensures that each batch of hydrogen-rich aquatic products can reach the ideal and accurate hydrogen content, ORP value, and oxygen content.

[0100] Second aspect, a hydrogen-rich water multi-cycle generation system:

[0101] It includes a pure water treatment device and a hydrogen production unit;

[0102] It further includes a processor and a memory connected to the processor. Program instructions are stored in the memory. When the program instructions are executed by the processor, the processor executes the hydrogen-rich water optimized preparation method as described above and transmits the control output electrical signals u1(t) and u2(t) to the PLC controller;

[0103] After the PLC controller reads the control output electrical signal u1(t), it controls the soaking time of the hydrogen rod of the hydrogen production unit to generate hydrogen with a corresponding concentration and transmits it to the pressure cutter;

[0104] The pure water treatment device extracts pure water from the pure water storage tank, filters out salts and other impurities, and also sends it to the pressure cutter;

[0105] The pressure cutter adjusts the hydrogen and pure water inside to the corresponding pressures and then sends the hydrogen and pure water to the gas-liquid mixing unit at the same time;

[0106] After the PLC controller reads the control output electrical signal u2(t), it controls the gas-liquid mixing unit to mix and heat to a corresponding temperature to form hydrogen-rich water; the hydrogen-rich water is sent to the gas-mixed water storage bucket for storage;

[0107] When the production volume of the hydrogen-rich water in this batch meets the production index, it is sent to the storage bucket for storage through the water outlet.

[0108] Compared with the prior art, the beneficial effects of the present invention are:

[0109] I. Improving the flexibility of product preparation: For the preparation of large batches of products on an industrial scale, the solution of the present invention enables manufacturers to no longer be limited to fixed process indicators. Based on the non-linear processing function provided by this solution, when manufacturers need to complete the production tasks of hydrogen-rich water products with different hydrogen contents, ORP values, and oxygen contents, the technical solution of the present invention can help them flexibly control the temperature and the soaking time of the hydrogen rods to meet the production tasks. That is, whether it is a product with a high hydrogen content or a specific ORP value, the corresponding production parameters can be obtained through this solution and automated control can be carried out.

[0110] II. Improving the quality of product preparation: Through the precise control of the SVR regression model and the PID controller, the present invention can ensure that the produced hydrogen-rich water products have consistent and high-quality hydrogen content, ORP value, and oxygen content, thus meeting the health needs of consumers.

[0111] III. Enhancing production efficiency: The PID controller used in the present invention has a real-time feedback mechanism, which can reduce the adjustment time and errors in the production process, thereby improving production efficiency. In addition, the prediction ability of the SVR regression model also helps to optimize the production process and reduce unnecessary downtime or adjustment time.

[0112] IV. Reducing costs and resource waste: Precise control means less waste of raw materials and a lower defective rate, which helps to reduce production costs. At the same time, the optimized production process can also reduce energy consumption and further reduce operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0113] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0114] Figure 1 It is a schematic flow chart of the method of the present invention;

[0115] Figure 2 It is a schematic diagram of the change degree of hydrogen production amount and redox potential, wherein part (a) is a broken line graph showing the relationship between the hydrogen content in the hydrogen-rich water prepared by the hydrogen rod and the increase in temperature, and part (b) is a broken line graph showing the relationship between the ORP value in the hydrogen-rich water prepared by the hydrogen rod and the decrease in temperature under the condition of the same time;

[0116] Figure 3 It is a schematic flow chart of the training method of the SVR regression model of the present invention;

[0117] Figure 4Schematic diagram of the generation system of the present invention;

[0118] Figure 5 Schematic diagram of the usage process of the generation system of the present invention;

[0119] Figure 6 Comparison chart of hydrogen content results in the test example of the present invention;

[0120] Figure 7 Comparison chart of oxygen content results in the test example of the present invention. Detailed implementation manners

[0121] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will describe the detailed implementation manners of the present invention in conjunction with the accompanying drawings. Many specific details are set forth in the following description in order to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described herein, and those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below;

[0122] It should be noted that the various embodiments in this specification are described in a progressive manner, and the key points of each embodiment are the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0123] Embodiment 1: Please refer to Figure 1 , this embodiment discloses an optimized preparation method for hydrogen-rich water; when applied, different production batches need to be planned in advance, and the hydrogen content h, ORP value O, and oxygen content o of the hydrogen-rich water products in each production batch should be consistent, so as to facilitate steps S101 and S2 of reading historical data in the following. In each round of preparation batches, the following steps S1 to S4 are sequentially executed by this method.

[0124] In this embodiment, regarding step S1, read the SVR regression model: input the dependent variable y into the regression function f(y), including the hydrogen content h, ORP value O, and oxygen content o, and use the RBF kernel function to utilize its dot product operation and its support vector x i , process the non-linear relationship between the independent variable x and the dependent variable y, and perform error constraint on the regression function f(y) by the ε-insensitive loss function; output the corresponding independent variable x by the regression function f(y), including the temperature T and the hydrogen rod immersion time S. Step S1 specifically includes the following steps S100 to S101;

[0125] Specifically, for step S100, process the regression task: first hand it over to the RBF kernel function K(y, x i)Perform a dot product operation:

[0126]

[0127] where: Υ is the hyperparameter of the RBF kernel function; σ is the width parameter of the RBF kernel function, which controls the radial range of the function. ||y - x i || 2 is the square of the Euclidean distance between the dependent variable y and the support vector x i ; the RBF kernel function can map the input data to a high-dimensional space to handle complex non-linear relationships.

[0128] The goal of the SVR (Support Vector Regression) model is to find a regression function f(y) such that for a given dependent variable (input value) to perform a regression task, it can predict the corresponding independent variable x (output value):

[0129]

[0130] where: α i and α i * are the trained Lagrange multipliers; b is the bias term; x i is the i-th support vector x i ; n is the number of support vectors.

[0131] It can be understood that

[0132] By using the RBF kernel function and the SVR model, complex non-linear relationships between the dependent variable and the independent variable can be captured, thus achieving high-precision prediction. Accurate prediction can ensure the production of products that meet the specifications. The SVR model constructs a regression function by finding support vectors, and these support vectors represent the key features of the data. Therefore, the model has strong generalization ability for the prediction of new data and can maintain good prediction performance even in cases outside the training data.

[0133] Specifically, in step S101, to ensure the prediction accuracy of the SVR regression model and tolerate small prediction errors to a certain extent, step S101 introduces the ε-insensitive loss function L ε (x’, f(y)) for constraint. The characteristic of this loss function is that it tolerates prediction errors within a small range, thus avoiding the model from overfitting the noise in the data:

[0134]

[0135] Among them, x’ is the value of the independent variable x in the previous preparation batch. The radius ε refers to the radius of the insensitive region, which is used to control the tolerance of the model to prediction errors. When the absolute value of the difference between the predicted value and the true value is less than or equal to the radius ε, the SVR regression model considers its current prediction result to be accurate, and the loss is 0 at this time. Once the absolute value of the difference between the predicted value and the true value exceeds this range, the loss will accumulate linearly.

[0136] It can be understood that this step actually needs to be executed at the next time step by obtaining the true value x’ after steps S1 to S4 are completed; therefore, this step is actually a correction measure for the regression function f(y). By introducing the ε-insensitive loss function, the model can tolerate small prediction errors to a certain extent, while paying attention to and optimizing those predictions with larger errors. This helps to improve the overall prediction accuracy of the model. At the same time, it avoids overfitting to small fluctuations in the data, enabling the model to better adapt to new data and new situations, thereby enhancing its generalization ability. And by continuously adjusting and optimizing the SVR regression model according to actual data, the conditions required for producing hydrogen-rich water (such as temperature and hydrogen rod immersion time) can be predicted more accurately, thereby optimizing the preparation process and improving product quality and efficiency.

[0137] Furthermore, the Python execution program for steps S100 to S101 is as follows:

[0138]

[0139]

[0140] In the above program, the principle is as follows:

[0141] (1) Step S100 (predict method): The SVR model is a regression version based on the support vector machine. In this model, the RBF kernel function is used to map the input data into a high-dimensional space, so as to be able to handle complex non-linear relationships. Given a dependent variable y, the SVR model will predict the corresponding independent variable x. The predict method uses the trained SVR model to make this prediction.

[0142] (2) Step S101 (calculate_loss method): In the calculate_loss method, we first use the predict method to obtain the predicted value x_predicted. Then we calculate the difference between each predicted value and the true value x_prime, and calculate the loss according to the definition of the ε-insensitive loss function. If the difference between the predicted value and the true value is less than or equal to ε (here we use the epsilon parameter in the SVR model as the radius of the insensitive region), the loss is 0. If the difference is greater than ε, the loss is the difference minus ε. Finally, we calculate the average loss of all samples as the output of this method.

[0143] In this embodiment, regarding step S2, the dynamic compensation mechanism: During the optimized preparation of hydrogen-rich water, in order to more precisely control the production parameters to obtain products that meet the specifications, we introduce a dynamic compensation mechanism. The core of this mechanism is to use historical prediction errors to optimize the current prediction, thereby improving the accuracy and efficiency of the production process. For the dependent variable y input in step S1, we minimize the sum of squared residuals S of the dependent variable y by the least squares method to obtain the error feedback coefficient β i , and backtrack the historical prediction errors for compensation. The independent variable x” after compensation is formed; the error feedback coefficient β i can help the SVR regression model "remember - correct" past prediction errors, thereby improving the accuracy of the current prediction. It includes the following steps S200 - S201.

[0144] Specifically, in step S200, calculate the sum of squared residuals S: First, backtrack the data of the previous M historical moments. For each historical moment, we have the actual value y of the dependent variable at that time i and the predicted value obtained through the SVR regression model To quantify the accuracy of the prediction, we calculate the sum of squared residuals S between these actual values and predicted values:

[0145]

[0146] Calculating the sum of squared residuals S to quantify the gap between the predicted value and the actual value is the objective function that needs to be minimized in the optimization process.

[0147] Specifically, in step S201, obtain the error feedback coefficient β i : In order to use historical prediction errors to improve the current prediction, we solve the error feedback coefficient β by the least squares method i , so that the predicted value after compensation is closer to the actual value. Let be the historical prediction error, and construct a linear relationship to represent the predicted value after compensation:

[0148]

[0149] Among them, is the dependent variable after compensation, and a i-1 is the prediction error of the previous data point. To solve for β i , the operation method is as follows:

[0150] 1) Substitute into the residual sum of squares formula:

[0151]

[0152] 2) Minimize the residual sum of squares:

[0153] Take the derivative of S with respect to β i , and set the derivative equal to 0 to find the extreme point, and find the β i that makes S reach the minimum value:

[0154]

[0155] Further expand and organize to obtain the solution of β i .

[0156] It can be understood that if there is a clear historical prediction error sequence ∈ i in practice, and a i-1 is the prediction error of the previous moment ∈ i-1 , then a i-1 in the above formula should be replaced by ∈ i-1 ;

[0157] After obtaining the minimum β i value, it can be used to update the independent variable x":

[0158] x" = x' + β i Δx;

[0159] Among them, Δx is an adjustment amount related to the historical prediction error and is a preset fixed value. In this way, the SVR regression model can "remember - correct" past prediction errors, thereby improving the accuracy of current predictions.

[0160] Furthermore, the python execution program for step S2 is as follows:

[0161]

[0162]

[0163]

[0164] In the above program, the calculate_residual_square_sum function obtains the residual sum of squares S by calculating the sum of the squares of the differences between the actual dependent variable values y_actual and the initial predicted dependent variable values y_predicted. This value is used to quantify the accuracy of the prediction. The smaller S is, the more accurate the prediction is. The calculate_beta function uses the least squares method to solve for the error feedback coefficient βi. It constructs a system of linear equations, where the A matrix is composed of the historical prediction errors a_prev and the b vector is composed of the differences between the actual values and the predicted values. The np.linalg.lstsq function is used to solve this system of linear equations to obtain βi. The update_x_predicted function uses the obtained βi and the preset adjustment amount delta_x to correct the initially predicted independent variable x_predicted to obtain the corrected independent variable x”. This correction process takes into account the historical prediction errors and aims to improve the accuracy of the current prediction.

[0165] In this embodiment, regarding step S3, calculate the error value e: Based on the independent variable x” obtained in step S2, calculate the difference between it and the independent variable x’ in the previous preparation batch, which can be the absolute difference or the relative error, to obtain the error value e that quantifies the error:

[0166] 1) Absolute difference: If the absolute difference is used to calculate the error e, then:

[0167] e = |x” - x’|;

[0168] where |·| represents taking the absolute value. The absolute difference can intuitively reflect the actual gap between the independent variables in two rounds of preparation batches, which is convenient for understanding and operation.

[0169] 2) Relative error: If the relative error is selected for calculation, it is:

[0170] e = |(x” - x’) / x’|;

[0171] The relative error takes into account the magnitude of the independent variable itself, so it can better reflect the degree of change of the independent variable in two rounds of preparation batches. This is especially useful when the value of the independent variable is large or small, because it provides a standardized comparison index.

[0172] In this embodiment, regarding step S4, apply the PID controller: After obtaining the error value e based on step S3, the control error signal e1(t) and the control error signal e2(t) can be obtained:

[0173] e1(t) = |e - u1(t - 1)|;

[0174] e2(t) = |e - u2(t - 1)|;

[0175] Among them, u1(t - 1) and u2(t - 1) are respectively the control output electrical signals of the gas-liquid hybrid unit and the hydrogen production unit in the previous production batch.

[0176] Then, the control output electrical signals u1(t - 1) and u2(t - 1) are input into the PID controller algorithm to generate the control output electrical signal u1(t) and the control output electrical signal u2(t), and then dynamically control the gas-liquid hybrid unit and the hydrogen production unit for preparing hydrogen-rich water, so as to meet the production tasks.

[0177] That is, the gas-liquid hybrid unit performs corresponding temperature control, while the hydrogen production unit adjusts its corresponding hydrogen rod soaking time, so as to drive the corresponding independent variables based on the preset dependent variables (i.e., the product parameters we want):

[0178] 1) For the gas-liquid hybrid unit:

[0179]

[0180] 2) For the hydrogen production unit:

[0181]

[0182] Among them: u1(t) and u2(t) are respectively the control output electrical signals of the gas-liquid hybrid unit and the hydrogen production unit; K P1 , K I1 , K D1 are the proportional, integral, and differential coefficients of the PID controller of the gas-liquid hybrid unit; K P2 , K I2 , K D2 are the proportional, integral, and differential coefficients of the PID controller of the hydrogen production unit; e1(t) and e2(t) are respectively the control error signals of the gas-liquid hybrid unit and the hydrogen production unit, that is, the difference between the expected output and the actual output; τ is the time variable;

[0183] It is understandable that the time variable τ enables the PID controller to have a real-time feedback mechanism, which can reduce the adjustment time and error in the production process, thereby improving production efficiency: by integrating the control error signals e1(τ) or e2(τ) from 0 to t, the controller can accumulate the error information at all past moments. This integration effect enables the controller to not only focus on the error at the current moment but also consider the historical cumulative effect of the error, thus more comprehensively evaluating the control effect of the system. Moreover, due to the existence of the integral term, even if the error at the current moment is very small or zero, but if there is a large cumulative error in the past, the controller will still make adjustments according to the value of the integral term. This continuous adjustment ability helps the system approach the set target faster and reduces the adjustment time. At the same time, after the control system reaches a steady state, if there is a steady-state error (i.e., a continuous deviation between the system output and the set target), the integral term will continue to accumulate this error and eliminate it by adjusting the control output.

[0184] Furthermore, the Python execution program for step S4 is as follows:

[0185]

[0186]

[0187]

[0188] In the above program, the serial communication parameters are configured through serial.Serial to communicate with the PLC controller. The pid_controller function calculates the new control output based on the current error, PID parameters, the previous error, and the previous output. It returns the new output value, the updated integral value, and the current error for use in the next loop. The send_to_plc function sends the calculated control output electrical signals u1 and u2 to the PLC controller through the serial port. In the main loop, first, the error value e is obtained. Then, the control error signals e1 and e2 are calculated, and the new control outputs u1 and u2 are calculated through the PID controller function. Finally, we update the control output of the previous round and send the new control signal to the PLC controller through the serial port. The time.sleep in the loop is used to simulate the time interval in a real-time control system.

[0189] Furthermore, through the real-time feedback and adjustment of the PID controller, the system can flexibly and precisely control the temperature and the soaking time of the hydrogen rod in the production process. Furthermore, the system can be flexibly adjusted quickly according to different production requirements (i.e., different hydrogen content, ORP value, and oxygen content indicators).

[0190] In practical applications, through the cyclic operation and continuous optimization of the above mechanism, the system can maintain a high degree of consistency and accuracy in a large-scale production environment, ensuring that each batch of hydrogen-rich aquatic products can achieve the ideal and accurate hydrogen content, ORP value, and oxygen content.

[0191] Embodiment 2: Based on the SVR regression model in step S1 of Embodiment 1, this embodiment further provides a method for training it. The technical solution of this embodiment needs to be implemented between the SVR regression model deployment steps S1. Please refer to Figure 3 :

[0192] (1) Data engineering:

[0193] First, historical data needs to be collected. These data cover the dependent variable y and the corresponding temperature T and hydrogen rod immersion time S under different hydrogen contents h, ORP values O, and oxygen contents o, that is, the independent variable x. These data are obtained through experiments or records during the actual production process. The dataset D is constructed as follows:

[0194] D = {[(T1, S1)(h1, O 1, o1)], [(T2, S2)(h2, O 2, o2)],..., [(T m , S m )(h m , O m, o m )]};

[0195] Among them, [(T i , S i )(h i , O i, o i )] is the i-th sample, and m is the number of samples.

[0196] Then, conventional data processing operations can also be performed on the dataset D, such as removing outliers and noise, performing normalization processing, etc.

[0197] (2) Setting the convex optimization objective:

[0198] Our goal is to find a hyperplane that can best divide the sample points in the dataset D. This hyperplane is determined by the normal vector W and the bias term b. To achieve this goal, first take the regression function f(y) as the training function and set the following convex optimization objective:

[0199]

[0200] Among them: W is the normal vector of the hyperplane. C is the penalty coefficient, which is used to control the trade-off between the model complexity and the fitting of the data. ζii and ζi i * is a slack variable; Φx i is a function that maps the support vector x i to a high-dimensional space (implicitly defined by a kernel function). Furthermore, the task of this convex optimization objective is as follows:

[0201] 1) Minimize The magnitude (length) of the normal vector W of the hyperplane should be as small as possible. The advantage of this is to prevent the model from overfitting, that is, the model will not be too complex, so that it can maintain good generalization ability on new data.

[0202] 2) Minimize All data points should be as close to the hyperplane as possible; by minimizing this term, we hope that as many data points as possible are close to the hyperplane, thereby improving the prediction accuracy of the model.

[0203] 3) The gap between (w·φ(x i )) + b should be within the radius ε plus the slack variables ζi i and ζi i * inside: On the basis of satisfying the first two tasks, some data points are allowed to deviate from the hyperplane to a certain extent. These constraint conditions ensure the robustness of the model, that is, the model can tolerate noise and outliers in the data to a certain extent.

[0204] Since there may be a non-linear relationship between the independent variable and the dependent variable, the RBF kernel function is introduced to perform the kernel trick to handle this complexity. The kernel trick allows us to perform linear partitioning in the high-dimensional space without knowing the i specific form of the high-dimensional mapping function Φx

[0205] K(x i , x j ) = <φ(x i ), φ(x j )>;

[0206] where <·,·> is the dot product operation, and the dot product operation represents determining and partitioning the similarity between any two different support vectors x i and x j in the high-dimensional space. Through the kernel trick, the non-linear problem in the original data space can be transformed into a linear problem in the high-dimensional feature space, and then the SVR model can be used for solution.

[0207] (III) Constrained convex optimization objective:

[0208] Introduce the ε-insensitive loss function as in step S101 to constrain the convex optimization objective. When the absolute value of the difference between the predicted value and the true value is less than or equal to a set radius ε, the SVR regression model considers its prediction result to be accurate, and the loss is 0 at this time. That is, as long as the predicted value falls within the ε range near the true value, we will not penalize the model. However, once the absolute value of the difference between the predicted value and the true value exceeds this ε range, the loss will accumulate linearly. This accumulation method ensures that the model is penalized when the prediction deviation is too large, thus prompting the model to continuously adjust parameters during training to improve the prediction accuracy.

[0209] (IV) Training:

[0210] Training is to continuously solve the above convex optimization problem, and introduce the k-fold cross-validation method during this process; the overall training steps are as follows:

[0211] (1) Randomly divide the dataset D into k subsets of equal size, namely D1, D2,..., D k

[0212] (2) Loop to execute the following steps:

[0213] Select D i as the test set D test , and merge the remaining subsets as the training set D train ; in the training set D train , use the QP algorithm (quadratic programming algorithm) to solve the convex optimization objective; it is responsible for finding a set of Lagrange multipliers α i and α i * and requires them to correspond to the data points, and then the normal vector W can be obtained.

[0214] First, construct the Lagrangian function L and introduce a set of Lagrange multipliers (α i , α' i η i , η i ^), where

[0215]

[0216] Take the partial derivative of L with respect to and set it to 0, we get:

[0217]

[0218] Substitute these relationships into the Lagrangian function to obtain a dual problem that only contains the Lagrange multipliers α i and α i *, and then solve it to obtain the optimal Lagrange multipliers α i and αi * value. Then the normal vector W can be obtained by α i and α i * and the corresponding support vector x i It can also be expressed by the support vector x i To directly obtain the value of the bias term b.

[0219] (3) Use the test set D test Test the trained SVR model. Calculate the average of k test results as the final evaluation of model performance.

[0220] (4) Select the model hyperparameters obtained from the round of training that minimizes the mean square error (MSE) as the optimal model hyperparameters, i.e., the Lagrange multiplier α i and α i *, normal vector W, bias term b, penalty coefficient C and hyperparameter Υ;

[0221] (5) When the model converges or reaches the predetermined number of iterations, stop training. Otherwise, repeat the above training steps.

[0222] It is understandable that by using the k-fold cross-validation method, the model is trained and tested on different subsets, which helps to discover and correct overfitting, thereby improving the model's generalization ability on new data. The optimal parameters of the model are determined by solving the convex optimization problem, which ensures the accuracy and stability of the model in predicting the conditions for preparing hydrogen-rich water. At the same time, selecting the hyperparameters that minimize the MSE further improves the performance of the model.

[0223] Furthermore, the training framework program (python) of the SVR regression model of this embodiment is as follows:

[0224]

[0225]

[0226]

[0227]

[0228]

[0229] The principle of the above procedure is as follows:

[0230] (1) Data Engineering (I): Use KFold from sklearn.model_selection to perform k-fold cross-validation data partitioning.

[0231] (2) Set the convex optimization objective (II) and the constrained convex optimization objective (III): In the SVR model, the convex optimization objective and the constraints have been internally implemented by the SVR class in the scikit-learn library. There is no need to explicitly set these objectives and constraints. One only needs to adjust the parameters of SVR (such as C, epsilon, gamma, etc.) to affect the optimization process.

[0232] (3) Training (IV): Use GridSearchCV to perform a parameter grid search to find the optimal combination of hyperparameters. The search ranges of parameters such as C, epsilon, and gamma are defined.

[0233] GridSearchCV will automatically perform k-fold cross-validation and use the specified scoring function (here it is the negative mean squared error) to evaluate the performance of each set of parameters. After the training is completed, GridSearchCV will save the optimal parameter combination and the corresponding best model.

[0234] (4) Package the SVR regression model: Package the trained SVR model and the optimal parameters in the SVRModel class for convenient program calls in step S1 of Example 1.

[0235] (5) This class provides a predict method for predicting new data and a get_best_params method for obtaining the optimal parameters.

[0236] Example 3: On the basis of Example 1, this example further provides a more intelligent technical solution for step S2.

[0237] In step S2 of Example 1, Δx is an adjustment amount related to the historical prediction error and is a preset fixed value. However, this form based on a preset fixed adjustment amount restricts the flexibility of the model in practical applications. Therefore, this example further provides an exponential smoothing method. As the prediction error (in each preparation batch) changes, Δx will also be adjusted accordingly, so as to more accurately reflect the actual demand of the current preparation batch. It includes the following steps:

[0238] Step S2020, Initialization:

[0239] Select an initial prediction value. It can be the first observation value of the sequence or the average of the first few observation values. It can also be a preset value.

[0240] Step S2021, Recursive calculation:

[0241] For each subsequent time point, use the exponential smoothing formula to calculate the prediction value:

[0242] St = αY t+(1 - α)S t-1 ;

[0243] St is the smoothed value (i.e., the predicted value) at time point t. Y t is the actual observed value at time point t. S t-1 is the smoothed value at time point t - 1. α is the smoothing constant, which is between 0 and 1 and is used to determine the weights of the actual observed value and the previous smoothed value.

[0244] Step S2022, dynamic adjustment of Δx:

[0245] In order to enable Δx to dynamically reflect the change of the prediction error, it is combined with the prediction error of the exponential smoothing method. The specific approach is as follows:

[0246] Step S20200, calculate the prediction error:

[0247] At each time point t, calculate the prediction error et = Y t - S t-1 (the actual observed value minus the predicted value of the previous time point).

[0248] Step S20201, dynamically adjust Δx:

[0249] According to the prediction error et, dynamically adjust Δx. Use the idea of exponential smoothing to update Δx:

[0250] Δb xnew = β·et+(1 - β)Δb xold +(1 - β);

[0251] where β is another smoothing constant. Δx new is the adjusted Δx, and Δx old is the Δx before adjustment. Then the adjusted Δx is used to correct the subsequent predicted values or as the adjustment basis for other relevant parameters.

[0252] It can be understood that the logic of the solution in this embodiment is that the prediction error et reflects the gap between the model prediction and the actual observation. When this gap is large, it indicates that the model prediction deviates from the real situation. By adjusting Δx in a timely manner, this deviation can be quickly fed back into the model to correct the subsequent predictions. Moreover, the dynamic adjustment of Δx means that the model can adapt to the data changes in real time. If the prediction error becomes larger, the adjustment of Δx will also increase accordingly, so that the model can return to the correct prediction track faster. This adjustment mechanism helps to improve the prediction accuracy because it can reduce the situation where the model lags behind the changes in the actual data.

[0253] During the preparation batch process, the actual demand may change due to various factors (such as raw material quality, environmental conditions, etc.). By dynamically adjusting Δx, the model can more closely track these changes and more accurately reflect the actual demand of the current preparation batch. This helps to optimize the production process, reduce waste, and improve product quality.

[0254] Example 4: On the basis of Example 1, this example further introduces an adaptive correction mechanism during the SVM regression model used in step S1:

[0255] Considering that in a production batch, step S1 requires the overall technical solution to be executed multiple rounds for the production batch. The independent variable x regressed in each round can indirectly utilize its historical credible information, and the D-S evidence theory is used to perform a certain amount of correction on the final predicted value at the current time step; the predicted independent variable x' and the corresponding actual independent variable x'' (which needs to be detected at a previous moment) are divided into evidence A and evidence B, and merged through Dempster's combination rule to output the combined belief function Bel(A∪B); the combined belief function Bel(A∪B) is mapped to an interval value between [0,1] as the correction factor δ, and the prediction function f(x) of step S1 is corrected. It includes the following steps S1020~S1023:

[0256] In this example, regarding step S1020, according to the time series, collect evidence A in the form of a set and evidence B in the form of a set:

[0257] Evidence A is the independent variable x' predicted in the previous n time steps (t - 1~t - 5) at the current time step t:

[0258] A = [x'(t - 1), x'(t - 2),..., x'(t - n)];

[0259] Evidence B is the gelatinization degree actually measured in the previous n time steps (t - 1~t - 5) at the current time step t:

[0260] B = [x''(t - 1), x''(t - 2),..., x''(t - n)];

[0261] In this example, regarding step S1021, execute Dempster's combination rule:

[0262] (1) Calculate the combined belief assignment Bel(C):

[0263]

[0264] (2) Calculate the combined uncertainty assignment Pl(C):

[0265]

[0266] Among them, A i and B i are respectively a subset of Evidence A and Evidence B; mass is the weight (it can be assigned a value of 0.5. Of course, if some evidence is special, subjective assignment can be made); C is the "possible set"; represents the empty set.

[0267] (3) Obtain the combined belief function Bel(A∪B):

[0268]

[0269] Bel(A) and Pl(A) are the belief and uncertainty assignments of the set of Evidence A;

[0270] Bel(B) and Pl(B) are the belief and uncertainty assignments of the set of Evidence B;

[0271] (Pl(A)∩Pl(B)) is the uncertainty assignment of the intersection of Evidence A and Evidence B respectively:

[0272]

[0273] Specifically, among them, represents any subset A i in Evidence A and a certain subset B i in Evidence B whose intersection is non-empty. That is, it explores a subset A i whose mass in the set of Evidence A is not zero, and at the same time there exists a subset B i whose mass in B is also not zero, that is, these two subsets have an intersection.

[0274] Furthermore, it can be understood that is responsible for finding some elements in the hypothesis set C that are supported in both Evidence A and Evidence B. If there are then when calculating the combined belief assignment Bel(C), these common elements need to be considered because they contribute to the support of the hypothesis set C.

[0275] Further, after concretizing the above concept, it is equivalent to this step of determining whether any element in evidence A and evidence B at the same time step t has the same "index"; if there is no same "index", it is regarded that the current evidence A and evidence B do not support each other, which means that there may be a certain deviation in the regression of the prediction function f(x), and then the correction factor δ of the subsequent mapping will be further biased and weighted. On the contrary, it proves that there is no deviation in the regression of the prediction function f(x), and the correction factor δ will be kept as 1 as much as possible, that is, no correction is made.

[0276] In this embodiment, regarding step S1022, the combined belief function Bel(A∪B) is mapped to an interval value between [0,1] as the correction factor δ through the sigmoid function:

[0277]

[0278] where e is the base of the natural logarithm;

[0279] Specifically, the sigmoid function is selected to ensure that the correction factor δ can smoothly change between 0 and 1, so as to realize the continuous correction of the predicted value. This mapping method not only simplifies the calculation process, but also makes the correction factor γ have an intuitive interpretability.

[0280] It can be understood that the combined belief function Bel(A∪B) reflects the belief degree distribution between evidence A and evidence B, and it comprehensively considers the matching degree between the predicted value and the actual value. And in the weighted correction process, the correction factor δ is used as a weight to scale the output of the prediction function f(x). Specifically, the correction factor δ can dynamically adjust the weight output by the SVM regression model according to the matching degree between the predicted value and the actual value.

[0281] In this embodiment, regarding step S1023, correct: scale the prediction function f(x):

[0282]

[0283] Specifically, scale the prediction function according to the correction factor δ. This correction method not only considers the contribution degree of different prediction models, but also realizes the dynamic adjustment of the predicted value through the correction factor δ. This adaptive correction helps to improve the accuracy and flexibility of production control.

[0284] Exemplarily, when the predicted value is relatively close to the actual value, the value of the combined belief function Bel(A∪B) will be relatively high, which results in a relatively high value of the correction factor δ. That is, the weight output by the model will be significantly enhanced, making the corrected predicted value closer to the actual value. On the contrary, when the difference between the predicted value and the actual value is large, the value of the combined belief function Bel(A∪B) will be low, and the value of the correction factor δ will also decrease accordingly, thereby reducing the weight output by the model and preventing the corrected predicted value from deviating too far from the variation law of the actual value. The effectiveness of this adaptive correction mechanism lies in its ability to adjust the weight output by the prediction model according to the matching degree between the real-time predicted value and the actual value. This makes the correction process more flexible and accurate, and can adapt to the changes in different production batches and product characteristics.

[0285] Example Five: As Figures 4 - 5 shown, this example discloses a multi-cycle hydrogen-rich water generation system. Its core components include a pure water treatment device, a hydrogen production unit, a processor, a memory, a PLC controller (i.e., the PLC controller described in Example One), a pressure cutter, a gas-liquid mixing unit, and a booster pump.

[0286] (I) Pure Water Treatment Device and Hydrogen Production Unit:

[0287] The pure water storage tank is externally connected to a water source (such as purified water, tap water, mineral water, etc.), stores water, and is equipped with a three-way valve a. The pure water treatment device extracts water from the pure water storage tank through one of the paths in this three-way valve a, and removes salts and other impurities through a filtration system to ensure the purity of the water quality and form pure water.

[0288] The hydrogen production unit extracts water through another path of this three-way valve a and is responsible for electrolyzing to produce oxygen (which is directly discharged) and hydrogen. The other path of the three-way valve a is connected to the pressure cutter for transporting the water there.

[0289] The concentration of the hydrogen it produces will be precisely controlled by the PLC controller in Example One, which is achieved by adjusting the soaking time of the hydrogen rods. The gas outlet of the hydrogen production unit is equipped with a three-way valve b.

[0290] (II) Processor and Memory:

[0291] The processor is the brain of the system and is responsible for executing the program instructions stored in the memory. These instructions are programs for implementing the optimized preparation method of hydrogen-rich water in Examples One to Three above. They can predict and adjust independent variables (such as the soaking time of the hydrogen rods of the hydrogen production unit, the mixing temperature and time of the gas-liquid mixing unit, etc.) according to the input dependent variables (such as the desired hydrogen-rich water parameters). The memory is responsible for storing these program instructions and historical data information.

[0292] (III) PLC Controller:

[0293] The PLC controller is the control center of the system. It receives the control output electrical signals u1(t) and u2(t) from the processor and controls the working states of the hydrogen production unit and the gas-liquid mixing unit according to these signals.

[0294] After the PLC controller reads the control output electrical signal u1(t), it precisely adjusts the soaking time of the hydrogen rods in the hydrogen production unit to generate hydrogen with a corresponding concentration. This step is crucial for ensuring the hydrogen content in the hydrogen-rich water.

[0295] Similarly, after reading the control output electrical signal u2(t), the PLC controller controls the gas-liquid mixing unit to mix and heat it at an appropriate temperature to form hydrogen-rich water that meets the specifications.

[0296] (IV) Pressure cutter and gas-liquid mixing unit:

[0297] One passage of the three-way valve b is connected to the pressure cutter. The pressure cutter is responsible for adjusting the hydrogen and pure water to appropriate pressures to ensure their effective mixing in the gas-liquid mixing unit. The other passage of the three-way valve b is connected to the gas-liquid mixing unit. The gas-liquid mixing unit is the key equipment for realizing the mixing of hydrogen and pure water. Under the precise control of the PLC controller, it performs efficient mixing and heating operations to produce high-quality hydrogen-rich water. After pressure relief treatment, it enters the booster pump.

[0298] Specifically, one passage of the three-way valve b is connected to the pressure cutter. However, in essence, the hydrogen flowing through this passage first mixes in the gas-liquid mixing unit, then enters the main flow path (main pipe) of the pressure cutter, then enters the pressure chamber for cutting, and then enters the gas-mixed water storage bucket.

[0299] (V) Booster pump and storage facilities:

[0300] After the above mixing and heating operations, the hydrogen-rich water has completed "primary cutting". There is already a certain amount of gas-mixed water in the gas-mixed water storage bucket (the remaining water can be pumped back to the pure water storage tank for multiple uses without generating waste water). However, the hydrogen in the water is prone to escape after primary cutting. At this time, a booster pump can repeatedly send the hydrogen-rich water back to the pressure cutter for cutting again. This cycle is repeated multiple times to prevent the hydrogen from escaping.

[0301] Finally, another booster pump sends the produced hydrogen-rich water to the gas-mixed water storage bucket for storage to ensure that the hydrogen-rich water can be supplied to users with sufficient pressure and flow rate when needed. When the production volume of this batch of hydrogen-rich water meets the production index, it will be sent to the storage bucket for long-term storage for subsequent use or sales.

[0302] It should be noted that Figure 4The booster pumps shown are arranged side by side, so only one booster pump is shown; the other booster pump is covered by this booster pump.

[0303] In this embodiment, the working process of the above system is as Figure 5 shown, including:

[0304] P1, Initial preparation and pure water treatment:

[0305] First, the pure water treatment device extracts pure water from the pure water storage tank. This step ensures the basic purity of the water quality and provides a high-quality base liquid for subsequent steps. The extracted pure water will pass through a filtration system, which can remove salts and other impurities to further purify the water quality.

[0306] P2, Hydrogen production and control:

[0307] 1) Hydrogen generation: The hydrogen production unit starts to work and generates hydrogen by electrolysis or other means. During this process, the soaking time of the hydrogen rod is a key parameter, which directly affects the concentration of the generated hydrogen.

[0308] 2) Hydrogen concentration control: The PLC controller precisely controls the soaking time of the hydrogen rod according to the control output electrical signal u1(t) received from the processor. This control method can ensure that the concentration of the generated hydrogen meets the production requirements each time and provides a stable hydrogen source for the subsequent gas-liquid mixing step.

[0309] P3, Gas-liquid mixing and heating:

[0310] 1) Gas-liquid mixing: In the pressure cutter, the pure water and hydrogen with adjusted pressure are sent into the gas-liquid mixing unit. Here, the PLC controller plays a role again and precisely controls the mixing process according to the control output electrical signal u2(t).

[0311] 2) Heating treatment: During the mixing process, the gas-liquid mixing unit also heats at an appropriate temperature. This step helps to improve the mixing effect of hydrogen and pure water and ensures the stable quality of the generated hydrogen-rich water.

[0312] P4, Storage and output:

[0313] 1) Boosting and storage: The mixed hydrogen-rich water is sent to the gas-mixed water storage bucket by the booster pump for storage. This step ensures that the hydrogen-rich water can be supplied with sufficient pressure and flow when needed.

[0314] 2) When the production volume of the hydrogen-rich water in this batch reaches the preset production index, it will be transferred to the storage bucket for long-term storage for subsequent use or sale.

[0315] P5, Recycling generation:

[0316] Once a batch is completed, the system will automatically reset and be ready to start the production of the next batch. This design ensures the efficiency and continuity of production. Through precise control systems and efficient hardware devices, the system realizes the multi-cycle generation of hydrogen-rich water. From the extraction and filtration of pure water to the preparation and control of hydrogen, then to the gas-liquid mixing and heating, and finally to the storage and output, to ensure the quality and stability of the finally generated hydrogen-rich water.

[0317] In this embodiment, further, the embedded C language execution program built in the PLC controller is as follows:

[0318]

[0319]

[0320]

[0321] In the above program, u1_t and u2_t are the control output electrical signals received from the processor, representing the soaking time of the hydrogen rod and the heating temperature of the gas-liquid mixing unit respectively. These variables need to be updated in the communication interface between the PLC controller and the processor. The while loop in the main function ensures that the PLC controller continuously adjusts the parameters of the hydrogen production unit and the gas-liquid mixing unit according to the received control signals.

[0322] set_hydrogen_stick_dip_time and set_mixing_temperature are functions related to the hardware interface, used to convert the control signals into actual hardware operations. The control period is achieved through the delay function, ensuring that the PLC controller updates the control signals at a certain frequency (every 100 milliseconds).

[0323] Test example: This example aims to verify the beneficial degree of the optimized preparation method of hydrogen-rich water (steps S1 to S4) disclosed in Embodiments 1 to 3 compared with the traditional preparation method.

[0324] (I) Experimental purpose:

[0325] This experiment aims to verify the superiority of the optimized preparation method of hydrogen-rich water (steps S1 to S4) compared with the traditional preparation process in maintaining the key parameters of hydrogen-rich water (hydrogen content, ORP value, and oxygen content).

[0326] (II) Experimental equipment and materials:

[0327] Two sets of hydrogen-rich water multi-cycle generation systems as disclosed in Embodiment 5 (one set for the experimental group and one set for the control group).

[0328] Agilent 7890C gas chromatograph and supporting software.

[0329] ORP-2096 detector

[0330] Raw water and other necessary preparation materials

[0331] (III) Experimental procedures:

[0332] 3.1 System preparation:

[0333] Configure a hydrogen-rich water multi-cycle generation system for the experimental group and the control group respectively, and ensure that the initial states of the two systems are the same

[0334] 3.2 Operations for the experimental group:

[0335] Execute the optimized preparation method (S1-S4) on the hydrogen-rich water multi-cycle generation system of the experimental group

[0336] 3.3 Operations for the control group:

[0337] Do not execute the optimized preparation method (S1-S4) on the hydrogen-rich water multi-cycle generation system of the control group to simulate the traditional preparation process

[0338] 3.4 Preparation of hydrogen-rich water:

[0339] The two groups use the same raw materials to prepare hydrogen-rich water with the same hydrogen content, ORP value, and oxygen content parameters, with a total amount of 5 liters each

[0340] 3.5 Sampling and detection:

[0341] Sample 10 ml from the hydrogen-rich water prepared by the two groups every 15 seconds

[0342] Send the sampled water to an Agilent 7890C gas chromatograph to detect its hydrogen content and oxygen content

[0343] Use the supporting software to record the data and draw a graph of the changes in hydrogen content and oxygen content parameters within the 9.45-minute time interval

[0344] Simultaneously use the ORP-2096 detector to detect the ORP values of the hydrogen-rich water prepared by the experimental group and the control group

[0345] (IV) Experimental results and analysis:

[0346] 4.1 Analysis of ORP value:

[0347] The ORP values of the experimental group and the control group are both between -50 mV and +350 mV, both meeting the industry specifications

[0348] 4.2 Analysis of hydrogen content:

[0349] As Figure 6As shown (line 1 in the figure is the experimental group, line 2 is the threshold line, and line 3 is the control group), the hydrogen content of the experimental group can always remain above the preset threshold line, demonstrating the superiority of the optimized preparation method. The hydrogen content of the control group may fluctuate near the threshold line, reflecting the poor stability of the traditional preparation process.

[0350] 4.3 Oxygen content analysis:

[0351] As Figure 7 shown (line 1 in the figure is the experimental group and line 2 is the control group), the change in the oxygen content of the experimental group is more gentle compared to the control group, indicating that the optimized preparation method also has an advantage in maintaining the stability of the oxygen content. The oxygen content of the control group shows significant fluctuations.

[0352] (V) Conclusion:

[0353] Through this experiment, the effectiveness of the optimized preparation method of hydrogen-rich water in maintaining the key parameters of hydrogen-rich water is verified, providing an experimental basis for further optimizing the preparation process and improving the quality of hydrogen-rich water.

[0354] The above specific implementation manners only express the implementation manners of the relevant practical applications of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the appended claims.

[0355] For those skilled in the art, it can be further realized that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Skilled professionals can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0356] Meanwhile, those skilled in the art can understand that all or part of the processes in the methods of all the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided in this application and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Claims

1. An optimized preparation method for hydrogen-rich water, characterized in that: In each batch preparation, the following steps are performed: S1, read the SVR regression model: input the dependent variable y into the regression function f(y), including the hydrogen content h, ORP value O and oxygen content o; use the RBF kernel function to utilize its dot product operation and its support vector x i , dealing with the nonlinear relationship between the independent variable x and the dependent variable y; the regression function f(y) outputs the corresponding independent variable x, including the temperature T and the hydrogen rod immersion time S; S2, execute dynamic compensation mechanism: minimize the residual sum of squares S of the dependent variable y, and obtain the error feedback coefficient β i , look back on the historical forecast error to compensate and form the compensated independent variable x"; S3, for the independent variable x', calculate the difference between it and the independent variable x' in the previous round of preparation batch, and obtain the error value e; S4, after obtaining the control error signal e1(t) and the control error signal e2(t) based on the error value e, input them into the PID controller algorithm, obtain the control output electrical signal u1(t) and the control output electrical signal u2(t), dynamically control the gas-liquid hybrid unit and the hydrogen production unit, and then prepare hydrogen-rich water; In S2, the residual sum of squares S of the dependent variable y is minimized and the error feedback coefficient β is obtained. i The method is: Look back to the previous M historical moments, each moment includes the actual value y of the dependent variable at that time i and predicted values set up For the historical forecast error, the following linear relationship is constructed to represent the forecast value after compensation: in, is the dependent variable after compensation, a i-1 is the prediction error of the previous data point; Substitute the linear relationship into the residual sum of squares S and calculate β i Take the derivative and set the derivative equal to 0 to find the β that minimizes the residual sum of squares S i value; In S2, the method for obtaining the independent variable x" is: x" = x' + β i Δx; Where Δx is the adjustment associated with the historical forecast error; x' is the value of the independent variable x in the previous round of preparation batches.

2. The optimized preparation method according to claim 1, characterized in that: In S1, the regression function f(y) and the RBF kernel function perform the following regression task to process the nonlinear relationship: RBF kernel function K(y,x i ) performs the dot product operation: Where: γ is the hyperparameter of the RBF kernel function; σ is the width parameter of the RBF kernel function; ||yx i || 2 is the dependent variable y and the support vector x i The square of the Euclidean distance between them; The regression function f(y): Where: α i and α i * is the trained Lagrange multiplier; b is the bias term; x i is the i-th support vector x i ; n is the number of support vectors.

3. The optimized preparation method according to claim 2, characterized in that: In S1, the regression function f(y) is replaced by the ε-insensitive loss function L ε (x',f(y)) constrained by: Where x' is the value of the independent variable x in the previous preparation batch; ε is the radius of the insensitive region.

4. The optimized preparation method according to claim 1 or 2, characterized in that: In S3, the difference is an absolute value difference or a relative error.

5. The optimized preparation method according to claim 1 or 2, characterized in that: In S4, the control error signal e1(t) and the control error signal e2(t) are obtained by: e1(t)=|e-u1(t-1)|; e2(t)=|e-u2(t-1)|; Among them, u1(t-1) and u2(t-1) are the control output electrical signals of the gas-liquid hybrid unit and the hydrogen production unit in the previous production batch, respectively.

6. The optimized preparation method according to claim 5, characterized in that: In S4, the methods of dynamically controlling the gas-liquid hybrid unit and the hydrogen production unit using the PID controller algorithm are respectively: For the gas-liquid hybrid unit: For the hydrogen production unit: Wherein: u1(t) and u2(t) are the control output electrical signals of the gas-liquid hybrid unit and the hydrogen production unit respectively; K P1 , K I1 and K D1 are the proportional, integral and differential coefficients of the PID controller of the gas-liquid hybrid unit; K P2 , K I2 and K D2 are the proportional, integral and differential coefficients of the PID controller of the hydrogen production unit; τ is the time variable.

7. A multi-cycle hydrogen-rich water generation system, including a pure water treatment device, characterized in that: The method comprises the hydrogen production unit and the gas-liquid hybrid unit in the optimized preparation method according to any one of claims 1 to 6; It also includes a processor and a memory connected to the processor, wherein program instructions are stored in the memory. When the program instructions are executed by the processor, the processor executes the optimization preparation method as described in any one of claims 1 to 6, and transmits the control output electrical signal u1(t) and the control output electrical signal u2(t) to the PLC controller, thereby controlling the gas-liquid hybrid unit and the hydrogen production unit.

8. The multi-cycle generation system according to claim 7, characterized in that: After reading the control output electrical signal u2(t), the PLC controller controls the hydrogen rod immersion time of the hydrogen production unit, generates hydrogen of corresponding concentration, and transmits it to the pressure cutter; The pure water treatment device extracts pure water from the pure water storage tank and filters the salt, and then sends it to the pressure cutter; After the pressure cutter adjusts the hydrogen and pure water therein to the corresponding pressures, the hydrogen and pure water are simultaneously delivered to the gas-liquid mixing unit; after the PLC controller reads the control output electrical signal u1(t), it controls the gas-liquid mixing unit to mix and heat to the corresponding temperature to form hydrogen-rich water.

Images