Adjusting method for preparation process of gypsum

Abstract

The invention relates to the technical field of gypsum preparation, and provides a preparation process adjusting method for gypsum. The method comprises the following steps: collecting particle size characteristics and acidic characteristics of a gypsum particle raw material; preparing a coating solution, and collecting solution attribute information and solution composition information; establishing a twinborn model according to the particle size characteristics, the acidity characteristics, the solution attribute information and the solution composition information; connecting the twin model with an optimization model, wherein a prediction network is embedded in the optimization model; performing optimization according to the prediction network to obtain an optimization parameter meeting a preset effect; and controlling the coating equipment to carry out coating adjustment on the gypsum particle raw materials according to the optimization parameters. The technical problem that in the process of producing gypsum from citric acid gypsum, due to the characteristics of many impurities, small crystal grains and the like, the slow release effect of a coating layer is difficult to accurately control is solved, and the effects of performing harmless treatment on the citric acid gypsum through intelligent process adjustment, improving the slow release performance of building gypsum powder and ensuring the gypsum preparation efficiency are achieved.

Description

Technical Field

[0001] The present application relates to the technical field of building material preparation, specifically to the technical field of gypsum preparation, and in particular to a method for adjusting the preparation process of gypsum. Background Art

[0002] With the rapid development of industrialization, the output of citric gypsum, as an important industrial by-product, has been increasing year by year. However, due to its unique chemical and physical properties, such as containing a large amount of organic impurities such as mycelium, fine grains, flaky crystals and low pH value, it is extremely difficult to directly process citric gypsum into building gypsum. These characteristics not only affect the preparation efficiency and product quality of building gypsum, but also have a serious negative impact on the environment. As the accumulation problem of citric gypsum becomes increasingly prominent, these untreated citric gypsum not only occupies a large amount of land resources, but also may pollute the soil, water and air through infiltration, volatilization and other means, posing a serious threat to the ecological environment. Summary of the invention

[0003] The present application provides a method for adjusting the preparation process of gypsum, aiming to solve the technical problem that in the process of generating gypsum from citric acid gypsum, due to the characteristics of many impurities and small grains, the sustained-release effect of the coating layer is difficult to accurately control.

[0004] In view of the above problems, the present application provides a method for adjusting the preparation process of gypsum.

[0005] The present application provides a method for adjusting the preparation process of gypsum, the method comprising: preparing a gypsum particle raw material, collecting the particle size characteristics and acidity characteristics of the gypsum particle raw material; preparing a coating solution for coating the gypsum particle raw material, collecting solution property information and solution composition information of the coating solution; establishing a gypsum preparation twin model according to the particle size characteristics, the acidity characteristics, the solution property information and the solution composition information; connecting the gypsum preparation twin model with a coating process optimization model, wherein the coating process optimization model is embedded with a sustained release effect prediction network; performing sustained release effect optimization according to the sustained release effect prediction network embedded in the coating process optimization model, and obtaining optimized coating parameters that meet a preset sustained release effect, wherein the optimized coating parameters include the number of coating layers and the thickness of each coating layer; conveying the gypsum particle raw material to a coating device, and controlling the coating device with the optimized coating parameters to perform multi-layer coating adjustment on the gypsum particle raw material.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: The above-mentioned method for adjusting the preparation process of gypsum, the method first collects and measures the particle size and acidity level of the gypsum particle raw material. Then prepare a coating solution and collect the properties and composition information of the solution. Subsequently, a gypsum preparation twin model is established based on the particle size, acidity level, solution property information and solution composition information. This model simulates the behavior of gypsum particles during the coating process and helps predict and optimize the coating effect. After that, this twin model is connected to the coating process optimization model. This coating process optimization model has a built-in slow-release effect prediction network that can predict the slow-release effect of gypsum under different coating processes. Using this prediction network, the slow-release effect can be optimized. By continuously adjusting the coating parameters, a set of optimal parameters can be obtained so that the gypsum has an ideal slow-release effect after coating. Then, the gypsum particle raw material is transported to the coating equipment, and the operation of the equipment is controlled according to the obtained optimal coating parameters. In this way, the gypsum particles can be accurately coated and adjusted in multiple layers, so as to obtain a final product with an expected slow-release effect. Through intelligent and refined control of the entire process, the problem of poor sustained-release effect of citric acid gypsum during the preparation process is effectively solved, and the quality and performance of gypsum are improved.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0009] Figure 1 A schematic diagram of a process flow of a method for adjusting a preparation process of gypsum in one embodiment; Figure 2 It is a schematic diagram of the process after collecting solution information of a method for adjusting the preparation process of gypsum in one embodiment. DETAILED DESCRIPTION

[0010] The embodiments of the present application provide a method for adjusting the preparation process of gypsum to solve the technical problem that in the process of generating gypsum from citric acid gypsum, the sustained-release effect of the coating layer is difficult to accurately control due to the characteristics of many impurities and small grains.

[0011] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0012] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules that are not explicitly listed or inherent to these processes, methods, products or devices.

[0013] Example like Figure 1 As shown, the present application provides a method for adjusting the preparation process of gypsum, the method comprising: Prepare gypsum particle raw materials, collect the particle size characteristics and acidity characteristics of the gypsum particle raw materials; prepare a coating solution for coating the gypsum particle raw materials, and collect solution property information and solution composition information of the coating solution.

[0014] With the continuous development of the building materials market and the increasing environmental protection requirements, the performance and quality requirements for gypsum materials are becoming higher and higher. However, the traditional gypsum coating process often makes it difficult to accurately control the sustained release effect of the coating layer, resulting in unstable performance of the final product and failure to meet market demand. In addition, the impurities in citric acid gypsum have an adverse effect on the preparation process of gypsum and also reduce the preparation efficiency of gypsum.

[0015] In the embodiment of the present application, during the preparation of gypsum, the system terminal obtains citric acid gypsum as a raw material. Citric acid gypsum is a by-product produced during the production of citric acid. It contains certain impurities and has the characteristics of small grains and acidity. In order to deal with these characteristics and improve the performance of citric acid gypsum, the system terminal collects the particle size characteristics of the particles of the gypsum particle raw material generated by the citric acid gypsum raw material. The particle size directly affects the dispersibility, fluidity and processing performance of citric acid gypsum. At the same time, its acidity characteristics are collected because the acidity will affect the effect and stability of the subsequent coating treatment. Subsequently, a coating solution for coating the gypsum particle raw material is prepared. This coating solution is intended to cover the surface of the gypsum particle raw material to form one or more layers of covering layers. These covering layers can effectively cover the gypsum particle raw material and slowly release the ability of the internal substance to ensure that the effective ingredients in the gypsum particle raw material are released in a stable and continuous manner when needed. When preparing the coating solution, the system terminal collects and records the property information of the solution, such as the pH, concentration, viscosity, etc. of the solution, which are crucial to the stability and effect of the coating process. At the same time, the composition information of the solution is also collected, that is, what chemical components are contained in the solution and their respective contents, so as to adjust the formula according to actual needs. These steps are to ensure that the gypsum particle raw materials can meet specific performance and application requirements after coating.

[0016] Further, if Figure 2 As shown, the present application provides that after collecting the solution property information and solution composition information of the coating solution, the method further includes: The coating solution and the gypsum particle raw material are subjected to a compatibility analysis, and a compatibility detection index is output, wherein the compatibility analysis includes chemical reaction rate, solution adhesion, acid tolerance rate, and impurity precipitation rate; when the compatibility detection index is greater than a preset adaptation threshold, the coating process optimization model is activated; when the compatibility detection index is less than the preset adaptation threshold, a solution replacement reminder is output.

[0017] Preferably, in order to ensure good compatibility and adaptability between the coating solution and the gypsum particle raw material, the system terminal performs a series of adaptability analyses. These analyses cover key indicators such as chemical reaction rate, solution adhesion, acid tolerance rate, and impurity precipitation rate. Among them, the chemical reaction rate can evaluate whether an adverse chemical reaction will occur when the coating solution contacts the gypsum particles. Solution adhesion can evaluate whether the coating solution can be tightly attached to the surface of the gypsum particles to form an effective covering layer. The acid tolerance rate can evaluate the stability of the coating solution in an acidic environment, that is, whether it can maintain the original coating performance. The impurity precipitation rate can evaluate whether the coating solution will introduce unnecessary impurities when in contact with the gypsum particles, thereby affecting the quality of the final product. For the chemical reaction rate, the system terminal analyzes the solution composition information of the coating solution and the chemical composition information of the gypsum particle raw material based on historical experience, chemical properties, and reaction mechanism to see whether an adverse reaction will occur. If a reaction occurs, the system terminal calculates the probability of the reaction, i.e., the chemical reaction rate, based on historical data. This is calculated by using the ratio of the number of times the same coating solution reacts when generating the coating layer in the historical data to the total number of times the coating solution is used. If no reaction occurs, the system terminal counts the chemical reaction rate as 0. For solution adhesion, the system terminal evaluates the integrity of the coating layer generated by the same coating solution in the historical data based on professional knowledge and historical experience, and evaluates the adhesion of the solution. For the acid tolerance rate, the system terminal calculates the acid tolerance rate by analyzing whether the coating solution and the gypsum particle raw materials react violently or generate precipitates in acidic environments with different pH values, and whether changes in chemical composition occur. This calculation process is similar to the above, and is based on historical data. For the impurity precipitation rate, the system terminal analyzes the historical data, after the same coating solution is mixed with gypsum particles, and the solution and particles separated by centrifugation, filtration, etc., to see if there are impurities, and calculates the impurity precipitation rate. Through the above steps, the system terminal can obtain a fitness detection index. If each of these indicators is higher than the preset adaptation threshold, it means that the coating solution and the gypsum particle raw material are well adapted to each other, and the coating process optimization model can be activated to further optimize the coating parameters. Otherwise, the coating solution needs to be replaced. At this time, the system terminal will output a solution replacement reminder to remind the operator to take corresponding measures to ensure the smooth progress of the subsequent coating process.

[0018] A gypsum preparation twin model is established based on the particle size characteristics, the acidity characteristics, the solution property information and the solution composition information.

[0019] In one embodiment, the system constructs a digital twin model of gypsum preparation based on the particle size characteristics and acidity characteristics of the gypsum particle raw material, as well as the solution property information and solution composition information of the coating solution. This twin model is a virtual, digital simulation of the gypsum preparation process, which can be used to simulate various situations and changes that gypsum may encounter in the actual preparation process, so as to predict and optimize the preparation conditions of gypsum. Specifically, the system terminal establishes a physical model based on the physical process of gypsum preparation, such as the mixing of particles and the flow of solutions. This includes the simulation of physical phenomena such as the dispersion of gypsum particles, the stirring and mixing of solutions. Subsequently, the particle size characteristics and solution property information are analyzed to understand the physical parameters of the gypsum particle raw material such as particle size, shape, density, and the operating parameters such as the flow rate and stirring speed of the coating solution. Afterwards, the system terminal analyzes the reaction mechanism between gypsum particles and the chemical substances in the solution according to the chemical principles of gypsum preparation. This includes chemical processes such as dissolution, precipitation, and crystallization. Then, a chemical model is established based on the chemical reaction mechanism. The model needs to be able to describe the interaction and transformation relationship between various chemical substances in the gypsum preparation process. Then, the acid characteristics and solution composition information are analyzed to understand the chemical parameters such as the pH value of the gypsum particle raw materials, as well as the chemical parameters such as the ion concentration, reactant type and concentration of the coating solution. Finally, the system terminal sets the parameters of the constructed physical model and chemical model according to these physical and chemical parameters. After the model parameters are set, the system terminal integrates the physical model and the chemical model to form a complete twin model of gypsum preparation. This model can simultaneously simulate the physical and chemical phenomena in the gypsum preparation process, allowing the system terminal to predict the performance and reaction mechanism of gypsum under different preparation conditions without actual experimental operations, which helps to improve the preparation efficiency and quality of gypsum.

[0020] The gypsum preparation twin model is connected with the coating process optimization model, wherein the coating process optimization model is embedded with a sustained-release effect prediction network.

[0021] In one embodiment, in the process of generating gypsum with citric acid gypsum, a digital gypsum preparation twin model is constructed to simulate the physical and chemical processes of gypsum preparation. In order to further optimize the preparation conditions of gypsum and improve its performance, the system terminal connects the gypsum preparation twin model with the coating process optimization model. The coating process optimization model is obtained by training historical coating parameters and is also established by integrating the slow-release effect prediction network. Specifically, the system terminal extracts a large number of historical coating parameters from historical data, which include the number of historical coating layers and the historical coating thickness of each layer. Subsequently, the number of neurons in the input layer is designed according to the specific parameters of the coating process. The number of hidden layers and the number of neurons in each layer are determined according to the complexity of the problem and the characteristics of the data. After that, the weights and bias items of the neural network are randomly initialized. The historical coating parameters are forward propagated through the neural network to calculate the output of each layer. The output results are then evaluated by the integrated slow-release effect prediction network. If the evaluation result does not meet the expected expectations, continue training until it meets expectations. Then, the system terminal embeds the pre-constructed slow-release effect prediction network into the coating process optimization model. The output layer of the coating process optimization model is connected to the input layer of the slow-release effect prediction network. This slow-release effect prediction network is a neural network model trained based on the monitoring data of the slow-release process without environmental interference and the monitoring data of the slow-release process with environmental interference. It can predict and optimize the slow-release effect in the coating process, thereby ensuring that the gypsum products have better performance and use effects. The slow-release effect prediction network predicts the stability and time of citric acid gypsum when releasing specific components by analyzing the slow-release data generated by the gypsum preparation twin model. By working in conjunction with the gypsum preparation twin model, this network can guide the system terminal to optimize the coating process so that the gypsum products can meet specific performance requirements while maintaining a stable slow-release effect.

[0022] Furthermore, the present application provides a method for training a sustained-release effect prediction network, comprising: Obtaining initialization parameters of the sustained-release effect prediction network; obtaining a first sustained-release data sample according to the gypsum preparation twin model, wherein the first sustained-release data sample is monitoring data of the sustained-release process without environmental interference.

[0023] Preferably, in order to optimize the coating process and predict its slow-release effect, the system terminal determines the network structure of the long short-term memory network (LSTM) according to the complexity of the problem and the characteristics of the data, including a feature extraction layer, an LSTM layer and a fully connected layer. The long short-term memory network is suitable for processing time series data and can capture long-term dependencies in the data. Subsequently, the system terminal sets initialization parameters for each layer of the long short-term memory network, namely, the initial weights and bias values. These initial values ​​are obtained by random initialization. After that, the system terminal integrates the slow-release stability rate deviation calculation formula and the slow-release time deviation calculation formula into the feature extraction layer. These two calculation formulas are pre-constructed. Then, the system terminal simulates the slow-release process according to the constructed gypsum preparation twin model to obtain the first slow-release data sample. These data samples are monitoring data of the slow-release process obtained by the twin model simulation without actual environmental interference, including the thickness of the coating layer, the slow-release speed, etc. They represent the slow-release performance of gypsum products under ideal conditions and are an important basis for subsequent network model training and prediction.

[0024] According to the sustained-release effect prediction network, sample prediction is performed on the first sustained-release data sample to obtain a first sustained-release stability rate sample and a first sustained-release duration sample; training is performed according to the first sustained-release stability rate sample and the first sustained-release duration sample until a preset accuracy rate is reached, and the parameters of the sustained-release effect prediction network are updated.

[0025] Preferably, in the process of constructing and training the slow-release effect prediction network, the system terminal first predicts the first slow-release data sample according to the initialized long short-term memory network. During the prediction process, the long short-term memory network performs weighted calculation on the first slow-release data sample according to the initial weight, and calculates the first slow-release stability rate sample. At the same time, the thickness of each coating layer in the first slow-release data sample is used to calculate the ratio of the slow-release rate to generate the first slow-release duration sample. The long short-term memory network is then trained using the calculated first slow-release stability rate sample and the first slow-release duration sample. Specifically, the system terminal uses forward propagation, combined with the calculation formula integrated in the feature extraction layer, to perform deviation prediction on the first slow-release stability rate sample and the first slow-release duration sample. Subsequently, the deviation prediction result is compared with the preset accuracy rate. This preset accuracy rate is set according to the accuracy requirement and is the maximum allowable deviation. If the deviation prediction result does not meet the preset accuracy requirement, it means that there is a loss in the long short-term memory network. At this time, the system terminal calculates the loss between the deviation prediction result and the preset accuracy rate through the mean square error. After that, the gradient of the loss to the parameters of the long short-term memory network is calculated through back propagation, and the weight of the long short-term memory network is updated according to this gradient. Repeat the steps of forward propagation, loss calculation and back propagation weight update until the preset accuracy requirement is met, and output the long short-term memory network to generate a slow-release effect prediction network. Through this process, it can be ensured that the slow-release effect prediction network can accurately predict the slow-release effect in the gypsum preparation process, providing strong support for the treatment of citric acid gypsum.

[0026] Furthermore, the present application provides a method for obtaining a first sustained-release stability rate sample, the method further comprising: Decompose the first sustained-release data sample, and output a first initial sustained-release data sample and a first final sustained-release data sample; for the first initial sustained-release data sample and the first final sustained-release data sample, output a first initial sustained-release stability rate sample and a first final sustained-release stability rate sample; perform weight calculation on the first initial sustained-release stability rate sample and the first final sustained-release stability rate sample, and output a first sustained-release stability rate sample.

[0027] Optionally, when using the slow-release effect prediction network to predict the slow-release effect of citric acid gypsum during the treatment process, the system terminal first processes and analyzes the first slow-release data sample. This data sample contains monitoring data from the beginning of the slow-release of gypsum to the end of the slow-release. In order to train the slow-release effect prediction network more accurately, the system terminal divides the first slow-release data sample into two stages, namely, the first initial slow-release data sample and the first final slow-release data sample. These two stages are based on actual needs and are divided according to the physical or chemical characteristics of the slow-release process. For example, the initial and final stages are defined according to the change in the slow-release rate. Subsequently, for the data samples of these two stages, the system terminal calculates their slow-release stability rates. The stability rate is an indicator to measure the degree of time fluctuation in the slow-release process. For the initial and final stages, due to different slow-release conditions and environments, their stability rates may also be different. Therefore, the system terminal calculates the corresponding stability rate deviations respectively through the slow-release stability rate deviation calculation formula integrated in the feature extraction layer. The two stability rate deviations are the stability rates of these two stages, namely, the first initial slow-release stability rate sample and the first final slow-release stability rate sample. Afterwards, the system terminal performs weighted calculation on the first initial slow-release stability rate sample and the first final slow-release stability rate sample according to the weight currently configured by the slow-release effect prediction network, and merges the stability rates of the two stages into one to generate the first slow-release stability rate sample. This sample reflects the comprehensive stability rate of the entire slow-release process. This sample will serve as one of the important input data for the slow-release effect prediction network training, and is used to guide the slow-release effect prediction network to learn how to predict the slow-release effect during the citric acid gypsum treatment process.

[0028] Furthermore, the present application provides a training sustained-release effect prediction network, and the method also includes: According to the gypsum preparation twin model, a second sustained release data sample is obtained, wherein the second sustained release data sample is the monitoring data of the sustained release process under environmental interference; according to the sustained release effect prediction network, sample prediction is performed on the second sustained release data sample to obtain a second sustained release stability rate sample and a second sustained release duration sample; training is performed according to the second sustained release stability rate sample and the second sustained release duration sample until the preset accuracy rate is reached, and the parameters of the sustained release effect prediction network are updated.

[0029] Optionally, during the treatment of citric acid gypsum, the system terminal also obtains a second slow-release data sample based on the gypsum preparation twin model simulation. These data are monitoring data of the slow-release process under various interference factors that may exist. These interference factors include temperature, humidity, changes in raw material quality, etc. Using the second slow-release data sample to train the slow-release effect prediction network can help the slow-release effect prediction network learn how to adapt to these interferences, thereby improving the generalization ability of the slow-release effect prediction network in a real environment. Subsequently, the system terminal uses the same method as the first slow-release data sample to train the slow-release effect prediction network to predict the second slow-release data sample and obtain the second slow-release stability rate sample and the second slow-release duration sample. These two samples respectively reflect the stability and duration of the gypsum slow-release process under the simulated interference environment. The slow-release effect prediction network is then trained with these two samples to make the weights and biases in the slow-release effect prediction network closer to actual needs. By continuously training and updating the model parameters, the system terminal can gradually improve the predictive ability of the slow-release effect prediction network for the slow-release process, which helps to better understand the slow-release mechanism of citric acid gypsum treatment during gypsum preparation.

[0030] The sustained-release effect is optimized according to the sustained-release effect prediction network embedded in the coating process optimization model to obtain the optimal coating parameters that meet the preset sustained-release effect, wherein the optimal coating parameters include the number of coating layers and the thickness of each coating layer.

[0031] In one embodiment, in order to release organic matter such as mycelium and trace amounts of oil, pigment and other organic matter generated during the citric acid gypsum coating process at a stable rate, the system terminal uses the slow-release effect prediction network embedded in the coating process optimization model to optimize the slow-release effect. The system terminal continuously tries and adjusts the coating parameters, that is, changes the number of coating layers or the thickness of each layer, to minimize the slow-release effect deviation output by the slow-release effect prediction network, thereby finding a set of parameter combinations that can meet the preset slow-release effect, that is, optimizing the coating parameters.

[0032] Furthermore, the present application provides a method for optimizing the sustained release effect according to the sustained release effect prediction network embedded in the coating process optimization model to obtain the optimal coating parameters that meet the preset sustained release effect, and the method also includes: The first set of input data and the second set of input data are analyzed according to the coating process optimization model to obtain initial optimization coating parameters; the initial optimization coating parameters are predicted using the sustained release effect prediction network to output a sustained release effect deviation, wherein the sustained release effect deviation includes a sustained release stability rate deviation and sustained release duration deviation ; By minimizing the sustained-release effect deviation, the optimal coating parameters that meet the preset sustained-release effect are obtained.

[0033] Preferably, the system terminal inputs two sets of input data into the constructed coating process optimization model. These two sets of data contain the key parameters of the coating process, such as the number of coating layers and the thickness of each coating layer. The model will generate initial optimal coating parameters based on these data. These initial optimal coating parameters are a set of coating parameters that are set by the model based on the learned knowledge and are in line with the current conditions. Subsequently, the system terminal inputs these initial optimal coating parameters into the sustained-release effect prediction network. The sustained-release effect prediction network predicts the sustained-release effect of gypsum products based on the initial optimal coating parameters, and outputs a sustained-release effect deviation. This deviation represents the difference between the predicted result and the preset sustained-release effect target, including the sustained-release stability rate deviation. and sustained release duration deviation . Afterwards, the system terminal compares whether the sustained-release effect deviation meets the requirements of the preset sustained-release effect. If not, the system terminal inputs the initial optimal coating parameters into the coating process optimization model for parameter optimization to minimize the sustained-release effect deviation, that is, the sustained-release effect deviation output by the sustained-release effect prediction network tends to be stable. At this time, the system terminal compares whether the sustained-release effect deviation meets the requirements of the preset sustained-release effect. If it still does not meet the requirements, the system terminal re-initializes the parameters and repeats the above process until a set of optimal coating parameters that meet the preset sustained-release effect is found.

[0034] Further, the present application provides the sustained release stability rate deviation The calculation formula and method also include: The sustained release stability rate deviation The calculation formula is as follows: ; in, is the number of coating layers, shows the sustained release time of the i-th coating layer, is the average value of the sustained-release time of each layer, The standard deviation of the sustained-release stability rate was calculated for all the coating layers to evaluate the dispersion of the sustained-release time.

[0035] Optionally, the calculation formula of the sustained-release stability rate deviation integrated by the feature extraction layer of the sustained-release effect prediction network is intended to evaluate the discrete degree of sustained-release time of each coating layer in the coating process. The calculation formula is as follows: ; Wherein, N represents the number of coating layers, that is, the number of coating layers performed on the citric acid gypsum particles in the entire coating process. represents the sustained release time of the i-th coating layer, The thickness of the i-th layer, is the sustained release rate constant of the i-th layer. The sustained-release stability rate deviation is the result of the standard deviation of the sustained-release time of all coating layers, which is used to evaluate the dispersion degree of the sustained-release time, that is, the quality of the sustained-release stability rate. The higher the dispersion degree, the greater the difference in the sustained-release time of each layer, and the relatively poorer the sustained-release stability rate.

[0036] Further, the present application provides the sustained release time deviation The calculation formula of , the method also includes: The sustained release time deviation The calculation formula is as follows: ; in, Indicates the actual average sustained-release time and the target sustained-release time The absolute deviation between It represents the average sustained release time of all coating layers. represents the sustained release time of the i-th coating layer, is the thickness of the i-th layer, is the sustained release rate constant of the i-th layer, is the expected release time, is the number of coating layers.

[0037] Optionally, the calculation formula of the sustained-release duration deviation integrated by the feature extraction layer of the sustained-release effect prediction network is used to measure the difference between the actual average sustained-release duration and the target sustained-release time. The calculation formula is as follows: ; in, The sustained-release duration deviation represents the absolute deviation between the actual average sustained-release duration and the target sustained-release duration. It represents the average sustained-release time of all coating layers. is the sustained release time of the ith coating layer after adjustment by the sustained release rate constant ki. is the thickness of the i-th coating layer, is the sustained release rate constant of the i-th layer, which is used to describe the sustained release rate of this layer. The expected sustained release time is determined according to specific application requirements. is the number of coating layers, which indicates the number of coating layers constructed in the entire citric gypsum coating process. By calculating the sustained-release time deviation, we can intuitively understand the gap between the actual sustained-release time and the target sustained-release time, and then guide the optimization and adjustment of the coating parameters.

[0038] The gypsum particle raw material is transported to the coating equipment, and the coating equipment is controlled by the optimal coating parameters to perform multi-layer coating adjustment on the gypsum particle raw material.

[0039] In one embodiment, during the production of gypsum particles, the system terminal transports the gypsum particle raw materials to the coating equipment. These devices operate according to the determined optimal coating parameters. Through the precise control of these parameters, the coating equipment can adjust the gypsum particle raw materials to coat multiple layers to ensure that each layer of coating is uniform, stable, and meets the expected performance requirements. This process can improve the quality of gypsum particles and improve their physical and chemical properties to meet the needs of different applications.

[0040] In summary, the embodiments of the present application have at least the following technical effects: In the embodiment of the present application, gypsum particle raw materials are prepared and their particle size and acidity characteristics are collected, and then a coating solution is prepared and its solution properties and composition information are collected. Based on this information, a gypsum preparation twin model is established and connected to the coating process optimization model. The coating process optimization model is also embedded with a sustained-release effect prediction network. Subsequently, this network is used to optimize the sustained-release effect, and the optimal coating parameters including the number of coating layers and the thickness of each coating layer are obtained. During the coating process, the gypsum particle raw materials are transported to the coating equipment, and multi-layer coating adjustment is performed according to the optimal coating parameters. In the coating solution and raw material fitness analysis stage, if the fitness detection index meets the preset threshold, the coating process optimization model is activated; if not, the solution is prompted to be replaced. In addition, in order to train the sustained-release effect prediction network, the gypsum preparation twin model is used to obtain sustained-release data samples without environmental interference and under environmental interference, and the network is trained to improve the prediction accuracy. In the optimization process, the deviation of the slow-release effect of the initial optimized coating parameters is predicted and minimized, and finally the optimized coating parameters that meet the preset slow-release effect are obtained. These technical effects jointly solve the technical problem that the slow-release effect of the coating layer is difficult to accurately control due to the characteristics of many impurities and small grains in the process of generating gypsum with citric acid gypsum, and realize the harmless treatment of citric acid gypsum through intelligent process adjustment, improve the slow-release performance of building gypsum powder, and ensure the efficiency of gypsum preparation.

[0041] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. The processes depicted in the accompanying drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0042] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

[0043] This specification and drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for adjusting the preparation process of gypsum, characterized in that: The method comprises: preparing gypsum particle raw materials, and collecting particle size characteristics and acidity characteristics of the gypsum particle raw materials; preparing a coating solution for coating the gypsum particle raw material, and collecting solution property information and solution composition information of the coating solution; Establishing a gypsum preparation twin model according to the particle size characteristics, the acidity characteristics, the solution property information and the solution composition information; Connecting the gypsum preparation twin model with the coating process optimization model, wherein the coating process optimization model is embedded with a sustained-release effect prediction network; Optimizing the sustained-release effect according to the sustained-release effect prediction network embedded in the coating process optimization model, and obtaining the optimal coating parameters that meet the preset sustained-release effect, wherein the optimal coating parameters include the number of coating layers and the thickness of each coating layer; The gypsum particle raw material is transported to the coating equipment, and the coating equipment is controlled by the optimal coating parameters to perform multi-layer coating adjustment on the gypsum particle raw material.

2. The method according to claim 1, characterized in that After collecting the solution property information and solution composition information of the coating solution, it also includes: Performing a compatibility analysis on the coating solution and the gypsum particle raw material, and outputting a compatibility detection index, wherein the compatibility analysis includes a chemical reaction rate, solution adhesion, acid tolerance rate, and impurity precipitation rate; When the fitness detection index is greater than a preset fitness threshold, activating the coating process optimization model; When the fitness detection index is less than the preset fitness threshold, a solution replacement reminder is output.

3. The method according to claim 1, characterized in that Training the sustained-release effect prediction network, including: Obtaining initialization parameters of the sustained-release effect prediction network; Prepare a twin model according to the gypsum, and obtain a first sustained-release data sample, wherein the first sustained-release data sample is monitoring data of a sustained-release process without environmental interference; Performing sample prediction on the first sustained-release data sample according to the sustained-release effect prediction network to obtain a first sustained-release stability rate sample and a first sustained-release duration sample; Training is performed according to the first sustained-release stability rate sample and the first sustained-release duration sample until a preset accuracy rate is reached, and the parameters of the sustained-release effect prediction network are updated.

4. The method according to claim 3, characterized in that Obtain the first sustained-release stability rate sample, including: Decomposing the first sustained-release data sample, and outputting a first initial sustained-release data sample and a first final sustained-release data sample; Outputting a first initial sustained-release stability rate sample and a first final sustained-release stability rate sample for the first initial sustained-release data sample and the first final sustained-release data sample; A weight calculation is performed on the first initial sustained-release stability rate sample and the first final sustained-release stability rate sample, and a first sustained-release stability rate sample is output.

5. The method according to claim 3, characterized in that Training the sustained-release effect prediction network also includes: Prepare a twin model according to the gypsum, and obtain a second sustained-release data sample, wherein the second sustained-release data sample is monitoring data of a sustained-release process under environmental interference; Performing sample prediction on the second sustained-release data sample according to the sustained-release effect prediction network to obtain a second sustained-release stability rate sample and a second sustained-release duration sample; Training is performed according to the second sustained-release stability rate sample and the second sustained-release duration sample until the preset accuracy rate is reached, and the parameters of the sustained-release effect prediction network are updated.

6. The method according to claim 1, characterized in that The sustained-release effect is optimized according to the sustained-release effect prediction network embedded in the coating process optimization model to obtain the optimal coating parameters that meet the preset sustained-release effect, including: Analyze the first set of input data and the second set of input data according to the coating process optimization model to obtain initial optimization coating parameters; The slow-release effect prediction network is used to predict the initial optimal coating parameters and output a slow-release effect deviation, wherein the slow-release effect deviation includes a slow-release stability rate deviation. and sustained release duration deviation ; By minimizing the sustained-release effect deviation, the optimal coating parameters that meet the preset sustained-release effect are obtained.

7. The method according to claim 6, characterized in that The sustained release stability rate deviation The calculation formula is as follows: ; in, is the number of coating layers, shows the sustained release time of the i-th coating layer, is the average value of the sustained-release time of each layer, The standard deviation of the sustained-release stability rate was calculated for all the coating layers to evaluate the dispersion of the sustained-release time.

8. The method according to claim 6, characterized in that The sustained release time deviation The calculation formula is as follows: ; in, Indicates the actual average sustained-release time and the target sustained-release time The absolute deviation between It represents the average sustained release time of all coating layers. represents the sustained release time of the i-th coating layer, is the thickness of the i-th layer, is the sustained release rate constant of the i-th layer, is the expected release time, is the number of coating layers.

Images