Automatic control system and method for fiber cement board making

Abstract

The present application provides an automatic control system and method for fiber cement board making, which relates to the field of industrial automatic control. It monitors and collects real-time temperature values ​​and real-time stirring speed values, and then introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the back end to perform time series collaborative analysis, so as to capture and characterize the query response optimization feature representation between the real-time stirring speed time series characteristics and the real-time temperature local time series characteristics, so as to reflect the key query feature response and interaction between the stirring speed time series semantics and the temperature time series semantics, thereby automatically generating the stirring speed value at the next time point, thereby automatically performing process control of the heating stirring reaction, and thereby improving the intelligence and adaptability of the fiber cement board making process control.

Description

Technical Field

[0001] The present application relates to the field of industrial automatic control, and in particular to an automatic control system and method for fiber cement board making. Background Art

[0002] As an important building material, fiber cement board is widely used in the construction industry and is favored for its high strength, good weather resistance and fire resistance. However, the fiber cement board produced by the existing preparation process has the disadvantages of weak stress resistance and easy breakage, which limits its application in the field of building materials.

[0003] In response to the above technical problems, Chinese patent CN114230263B discloses a fiber cement board and its preparation method, which first prepares composite nano-microspheres with a specific morphology through chemical reaction, then mixes them with other main components such as cement binder, aggregate and reinforcing fiber, and finally prepares the fiber cement board through a series of process flows such as heating and stirring reaction, water filtration and wet material stacking and cutting. This method aims to improve the performance of the board, such as waterproofness, mechanical properties and flame retardant properties, by introducing nanomaterials, thereby improving the product quality of the final fiber cement board.

[0004] However, in the above-mentioned preparation method of fiber cement board, although the process of preparing nano hollow microspheres and mixing them with other components to prepare fiber cement board is described in detail, during the material mixing and heating and stirring reaction process, the traditional control method usually adopts a fixed parameter setting method, that is, maintaining a constant stirring speed and temperature during the stirring process. Although this method is simple, it ignores the dynamic synergistic relationship between temperature and stirring speed, which may lead to the reaction conditions of the mixture at different stages not being optimally matched, thereby affecting the quality and performance of the final product. For example, if the stirring speed is not enough at a certain temperature, it may cause uneven mixing of materials, thereby affecting the dispersion effect of the fibers and the strength of the final board; on the contrary, if the stirring speed is too fast and the temperature is not properly controlled, it may cause local overheating, affecting the performance of the material or causing the material to decompose, thereby affecting the final preparation quality and performance of the fiber cement board.

[0005] Therefore, an optimized automatic control solution for fiber cement board making is desired. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The present application provides an automatic control system and method for fiber cement board making, which monitors and collects real-time temperature values ​​and real-time stirring speed values, and then introduces data processing and analysis algorithms based on artificial intelligence and deep learning in the back end to perform time series collaborative analysis, so as to capture and characterize the query response optimization feature representation between the real-time stirring speed time series characteristics and the real-time temperature local time series characteristics, so as to reflect the key query feature response and interaction between the stirring speed time series semantics and the temperature time series semantics, so as to automatically generate the stirring speed value at the next time point, so as to automatically perform the process control of the heating stirring reaction, thereby improving the intelligence and adaptability of the fiber cement board making process control.

[0007] According to one aspect of the present application, there is provided an automatic control method for fiber cement board making, which comprises:

[0008] The water, cement binder and aggregate are mixed and stirred to obtain a mixed material; reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres are added to the mixed material, and the mixture is heated and stirred for reaction, and then the water is filtered to obtain a solid wet material; the solid wet material is stacked and cut to form a fiber cement wet blank; the fiber cement wet blank is pressurized, statically cured, and subjected to high temperature and high pressure steam curing to obtain a fiber cement board;

[0009] The method comprises adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material, heating and stirring the mixture for reaction, and filtering the water to obtain a solid wet material, including:

[0010] Get the time queue of the real-time temperature value collected by the temperature sensor;

[0011] Obtaining a time queue of real-time stirring speed values ​​collected by a speed sensor;

[0012] Sequentially encoding the time queue of the real-time temperature value and the time queue of the real-time stirring speed value to obtain a sequence of local time series characteristics of the real-time temperature and a time series characteristic of the real-time stirring speed;

[0013] Performing a time series feature query response optimization process on the sequence of the real-time temperature local time series features and the real-time stirring speed time series features to obtain a real-time stirring speed-real-time temperature query response optimization feature;

[0014] The stirring speed optimization control instruction is generated based on the real-time stirring speed-real-time temperature query response optimization feature.

[0015] Further, the time queue of the real-time temperature value and the time queue of the real-time stirring speed value are sequence-encoded to obtain a sequence of real-time temperature local time series features and a real-time stirring speed time series feature, including:

[0016] Inputting the time queue of the real-time temperature value into a sequence encoder based on a 1D-CNN model to obtain a sequence of real-time temperature local time series implicit correlation feature vectors as a sequence of real-time temperature local time series features;

[0017] The time queue of the real-time stirring speed value is input into a sequence encoder based on a forward LSTM model to obtain a real-time stirring speed time series implicit correlation feature vector as the real-time stirring speed time series feature.

[0018] Further, the sequence of the real-time temperature local time series features and the real-time stirring speed time series features are subjected to time series feature query response optimization processing to obtain a real-time stirring speed-real-time temperature query response optimization feature, including:

[0019] Performing feature aggregation processing based on linear embedding coding on the sequence of implicitly associated feature vectors of the real-time temperature local time series to obtain a prompt template;

[0020] Based on the prompt template, a sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector is cross-domain optimized query encoded to obtain a real-time stirring speed-real-time temperature query response optimized encoding vector as the real-time stirring speed-real-time temperature query response optimized feature.

[0021] Further, the sequence of the real-time temperature local time series implicitly associated feature vectors is subjected to feature aggregation processing based on linear embedding coding to obtain a prompt template, including:

[0022] Using a key embedding matrix, linear embedding coding is performed on each real-time temperature local time series implicitly associated feature vector in the sequence of real-time temperature local time series implicitly associated feature vectors to obtain a sequence of real-time temperature local time series implicitly associated feature vectors after linear transformation;

[0023] Arrange the sequence of the real-time temperature local time series implicitly associated characteristic vectors after the linear transformation into a matrix to obtain a real-time temperature local time series key matrix;

[0024] The maximum value of the real-time temperature local time sequence implicit correlation feature vector after each linear transformation in the real-time temperature local time sequence key matrix is ​​extracted to obtain the real-time temperature local time sequence key matrix significant feature vector as the prompt template.

[0025] Furthermore, a key embedding matrix is ​​used to perform linear embedding coding on each real-time temperature local time series implicit association feature vector in the sequence of the real-time temperature local time series implicit association feature vector to obtain a sequence of real-time temperature local time series implicit association feature vectors after linear transformation, including: multiplying the real-time temperature local time series implicit association feature vector by the key embedding matrix and then adding the key embedding bias vector by position to obtain the real-time temperature local time series implicit association feature vector after linear transformation.

[0026] Further, based on the prompt template, a sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector is cross-domain optimized query encoded to obtain a real-time stirring speed-real-time temperature query response optimized encoding vector as the real-time stirring speed-real-time temperature query response optimized feature, including:

[0027] Performing linear embedding coding on the real-time stirring speed time series implicitly associated feature vector using a query embedding matrix and a value embedding matrix to obtain a real-time stirring speed time series query vector and a real-time stirring speed time series value vector;

[0028] Input the real-time stirring speed time series query vector, the real-time stirring speed time series value vector, the real-time temperature local time series implicit association feature vectors after each linear transformation in the real-time temperature local time series key matrix, and the prompt template into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vectors;

[0029] The positional mean vector of the sequence of the real-time stirring speed-real-time temperature time series semantic cross-domain optimized query encoding vector is calculated to obtain the real-time stirring speed-real-time temperature query response optimized encoding vector.

[0030] Further, the real-time stirring speed time series query vector, the real-time stirring speed time series value vector, the real-time temperature local time series implicit association feature vector after each linear transformation in the real-time temperature local time series key matrix, and the prompt template are input into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vectors, including:

[0031] After calculating the multiplication between the real-time stirring speed time series query vector and the transposed vector of the real-time temperature local time series implicit association feature vector after the linear transformation, the obtained real-time stirring speed-real-time temperature time series semantic association feature matrix is ​​divided by the bi-norm of the significant feature vector of the real-time temperature local time series key matrix according to the position to obtain the real-time stirring speed-real-time temperature time series semantic association representation matrix;

[0032] The real-time stirring speed-real-time temperature time series semantic association representation matrix is ​​input The function is processed to obtain the real-time stirring speed-real-time temperature time series semantic association weight matrix;

[0033] After multiplying the real-time stirring speed-real-time temperature time series semantic association weight matrix with the significant feature vector of the real-time temperature local time series key matrix, the obtained feature vector is point-multiplied with the real-time stirring speed time series value vector to obtain the real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vector.

[0034] Furthermore, a stirring speed optimization control instruction is generated based on the real-time stirring speed-real-time temperature query response optimization feature, including: inputting the real-time stirring speed-real-time temperature query response optimization encoding vector into a decoder-based stirring speed optimization control module to obtain the stirring speed optimization control instruction, and the stirring speed optimization control instruction includes a recommended value of the stirring speed value at the next time point.

[0035] According to another aspect of the present application, there is provided an automatic control system for fiber cement board making, comprising:

[0036] A real-time temperature value acquisition unit, used to acquire a time queue of real-time temperature values ​​collected by a temperature sensor;

[0037] A real-time stirring speed value acquisition unit, used to acquire a time queue of a real-time stirring speed value collected by a speed sensor;

[0038] A sequence encoding unit, used for performing sequence encoding on the time queue of the real-time temperature value and the time queue of the real-time stirring speed value to obtain a sequence of local time series characteristics of the real-time temperature and a time series characteristic of the real-time stirring speed;

[0039] A time series feature query response optimization processing unit, used for performing time series feature query response optimization processing on the sequence of the real-time temperature local time series features and the real-time stirring speed time series features to obtain a real-time stirring speed-real-time temperature query response optimization feature;

[0040] The stirring speed optimization unit is used to generate stirring speed optimization control instructions based on the real-time stirring speed-real-time temperature query response optimization characteristics.

[0041] Furthermore, the sequence encoding unit is also used for:

[0042] Inputting the time queue of the real-time temperature value into a sequence encoder based on a 1D-CNN model to obtain a sequence of real-time temperature local time series implicit correlation feature vectors as a sequence of real-time temperature local time series features;

[0043] The time queue of the real-time stirring speed value is input into a sequence encoder based on a forward LSTM model to obtain a real-time stirring speed time series implicit correlation feature vector as the real-time stirring speed time series feature.

[0044] Compared with the prior art, the present application provides an automatic control system and method for fiber cement board making, which monitors and collects real-time temperature values ​​and real-time stirring speed values, and then introduces data processing and analysis algorithms based on artificial intelligence and deep learning in the back end to perform timing collaborative analysis, so as to capture and characterize the query response optimization feature representation between the real-time stirring speed timing characteristics and the real-time temperature local timing characteristics, so as to reflect the key query feature response and interaction between the stirring speed timing semantics and the temperature timing semantics, thereby automatically generating the stirring speed value at the next time point, thereby automatically performing process control of the heating stirring reaction, and thereby improving the intelligence and adaptability of the fiber cement board making process control. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0046] Figure 1 The figure is a flow chart of an automatic control method for fiber cement board making according to an embodiment of the present application.

[0047] Figure 2 The present invention is a flow chart of adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material in the automatic control method for making fiber cement board according to an embodiment of the present application, heating and stirring the mixture to react, and then filtering the water to obtain a solid wet material.

[0048] Figure 3 This is a data flow diagram of adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material in the automatic control method for fiber cement board making according to an embodiment of the present application, heating and stirring the mixture to react, and then filtering the water to obtain a solid wet material.

[0049] Figure 4 A flowchart of performing timing feature query response optimization processing on the sequence of the real-time temperature local timing features and the real-time stirring speed timing features in the automatic control method for fiber cement board making according to an embodiment of the present application to obtain the real-time stirring speed-real-time temperature query response optimization features.

[0050] Figure 5 The present invention is a flowchart of performing feature aggregation processing based on linear embedding coding on the sequence of the real-time temperature local time series implicit correlation feature vectors in the automatic control method for fiber cement board making according to an embodiment of the present application to obtain a prompt template.

[0051] Figure 6 The present invention is a flowchart of performing cross-domain optimization query encoding on the sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector based on the prompt template in the automatic control method for fiber cement board making according to an embodiment of the present application to obtain the real-time stirring speed-real-time temperature query response optimization encoding vector as the real-time stirring speed-real-time temperature query response optimization feature.

[0052] Figure 7 The system block diagram of an automatic control system for fiber cement board making according to an embodiment of the present application. DETAILED DESCRIPTION

[0053] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0054] The word “exemplary” is used exclusively herein to mean “serving as an example, example, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other embodiments.

[0055] In addition, in order to better illustrate the present application, numerous specific details are given in the following specific embodiments. It should be understood by those skilled in the art that the present application can also be implemented without certain specific details. In some examples, methods, means, components and circuits well known to those skilled in the art are not described in detail in order to highlight the subject matter of the present application.

[0056] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0057] Fiber cement board is an important building material, which is widely used in the construction industry due to its high strength, good weather resistance and fire resistance. However, there are some problems in the existing production process, such as weak stress resistance and fragility, which limit its application in the field of building materials.

[0058] In a fiber cement board and a preparation method thereof disclosed in Chinese patent CN114230263B, during the material mixing and heating and stirring reaction process, the traditional material mixing and heating and stirring reaction process control method usually adopts a fixed parameter setting method, that is, maintaining a constant stirring speed and temperature during the stirring process. This method ignores the dynamic synergistic relationship between temperature and stirring speed, which may result in the reaction conditions of the mixture at different stages not being optimally matched, thereby affecting the quality and performance of the final product. Specifically, the appropriate temperature can accelerate chemical reactions, while the appropriate stirring speed contributes to the uniform distribution of substances. If the two do not work well together, uneven mixing or incomplete reaction may result, which in turn affects the final performance of the product.

[0059] In view of the above technical problems, the present application proposes an automatic control method for fiber cement board making. Figure 1 FIG. 1 is a flow chart of an automatic control method for fiber cement board manufacturing according to an embodiment of the present application. Figure 1 As shown, according to the embodiment of the present application, the automatic control method for fiber cement board making includes: S110, mixing and stirring water, cement binder and aggregate to obtain a mixed material; S120, adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-like polyaniline nano hollow microspheres to the mixed material, heating and stirring the mixture, and then filtering the water to obtain a solid wet material; S130, stacking and cutting the solid wet material to form a fiber cement wet blank; S140, pressurizing, curing, and high-temperature and high-pressure steaming the fiber cement wet blank to obtain a fiber cement board.

[0060] In the above-mentioned automatic control method for fiber cement board making, the step S110 is to mix and stir water, cement binder and aggregate to obtain a mixed material. It should be understood that in construction projects, mixing and stirring water, cement binder and aggregate is a key step in producing concrete. Among them, water plays an important role in chemical reactions in this process. After cement is mixed with water, it undergoes a hydration reaction to form a hydrated cement paste, which gradually hardens and solidifies the concrete to eventually form a solid structure. Aggregates provide a skeleton for concrete and bear the main bearing capacity. Specifically, during the mixing process, attention should be paid to observing the consistency and uniformity of the mixture to ensure that the cement slurry and aggregate are evenly distributed to avoid drying or excessive moisture. By reasonably mixing these three materials, not only can the strength and durability of the concrete be ensured, but also its workability and impermeability and other properties can be affected.

[0061] In the above-mentioned automatic control method for fiber cement board making, the step S120 adds reinforcing fiber and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material, and after heating and stirring the reaction, the water is filtered to obtain a solid wet material. It should be understood that reinforcing fiber is a fiber material used to improve the mechanical properties of materials, and is widely used in composite materials to improve their strength, rigidity and toughness. Common reinforcing fibers include carbon fiber, glass fiber, aramid fiber and natural fiber. Modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres are an innovative functional material that combines the excellent properties of silicon dioxide, titanium dioxide and aluminum oxide and the conductive properties of polyaniline. This composite microsphere has a unique sea thorn-shaped structure, provides a large specific surface area, and enhances the adsorption and reactivity of the material. Specifically, in the technical solution of the present application, ultrasonic mixing is used to ensure the uniform dispersion of each component. Then, after adding the reinforcing fiber and microspheres, heating and stirring reaction are performed. Then, after the heating and stirring reaction is completed, the system will produce a certain amount of water and unreacted substances. In order to obtain the desired solid wet material, a filtering operation is required. The purpose of filtering is to remove moisture and impurities in the mixture to obtain a relatively dry solid wet material. Among them, the filtering process adopts a vacuum filtration method. Finally, a solid wet material is obtained. Here, the addition of reinforcing fibers and modified microspheres not only improves the mechanical properties of the composite material, but also gives it better conductivity and heat resistance.

[0062] In particular, in the automatic control method of fiber cement board making, it is crucial to add reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material, and to carry out the process of heating and stirring the reaction. Specifically, the addition of the modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres can significantly enhance the mechanical properties of the fiber cement board, such as tensile strength, compressive strength and toughness. These nanomaterials can fill the pores in the cement matrix, reduce internal defects, and thus improve the overall strength of the material. In addition, the uniform distribution of reinforcing fibers and nano hollow microspheres is the key to ensuring the stable performance of the fiber cement board. Through the heating and stirring reaction, these reinforcing materials can be better dispersed in the cement matrix to avoid local aggregation, thereby improving the uniformity of the final product. In addition, during the heating and stirring reaction, the synergistic interaction of temperature and stirring speed can affect the hydration process of the cement matrix, thereby regulating the microstructure of the final product. Appropriate reaction conditions are conducive to the formation of a dense and uniform microstructure, which is very important for improving the durability and performance of fiber cement boards.

[0063] Based on this, during the heating and stirring reaction process, the technical concept of the present application is to monitor and collect the real-time temperature value in real time through the temperature sensor, and monitor and collect the real-time stirring speed value in real time through the speed sensor, and then introduce the data processing and analysis algorithm based on artificial intelligence and deep learning in the back end to perform time series collaborative analysis on the real-time temperature value and the stirring speed value, so as to capture and characterize the query response optimization feature representation between the real-time stirring speed time series feature and the real-time temperature local time series feature. This feature representation can reflect the key query feature response and interaction between the stirring speed time series semantics and the temperature time series semantics, and is used for subsequent stirring speed optimization control. In this way, the stirring speed value at the next time point can be automatically generated according to the actual temperature data and stirring speed data, so as to automatically control the process of heating and stirring reaction, thereby improving the intelligence and adaptability of the fiber cement board making process control, and helping to ensure the preparation quality and performance of the fiber cement board.

[0064] Figure 2 The present invention is a flow chart of adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material in the automatic control method for making fiber cement board according to an embodiment of the present application, heating and stirring the mixture to react, and then filtering the water to obtain a solid wet material. Figure 3 This is a data flow diagram of adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material in the automatic control method for fiber cement board making according to an embodiment of the present application, heating and stirring the mixture, and filtering the water to obtain a solid wet material. Figure 2 and Figure 3 As shown, in the step S120, reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres are added to the mixed material, and after heating and stirring the reaction, water is filtered to obtain a solid wet material, including: S121, obtaining a time queue of real-time temperature values ​​collected by a temperature sensor; S122, obtaining a time queue of real-time stirring speed values ​​collected by a speed sensor; S123, sequence encoding the time queue of real-time temperature values ​​and the time queue of real-time stirring speed values ​​to obtain a sequence of real-time temperature local timing characteristics and a real-time stirring speed timing characteristic; S124, performing timing characteristic query response optimization processing on the sequence of real-time temperature local timing characteristics and the real-time stirring speed timing characteristic to obtain a real-time stirring speed-real-time temperature query response optimization characteristic; S125, generating a stirring speed optimization control instruction based on the real-time stirring speed-real-time temperature query response optimization characteristic.

[0065] Specifically, in the steps S121 and S122, a time queue of real-time temperature values ​​collected by the temperature sensor is obtained, and a time queue of real-time stirring speed values ​​collected by the speed sensor is obtained. It should be understood that in the technical solution of the present application, the change of temperature directly affects the physical and chemical properties of the material, especially in cement and other chemical reaction processes. Through the temperature sensor, the temperature value obtained in real time can form a time queue, which reflects the temperature change at each moment during the stirring process. The temperature data is usually output in the form of a digital signal, processed by the data acquisition module, and the temperature value at each time point is recorded. This real-time monitoring can detect abnormal temperature fluctuations in time to avoid material performance failures or process failures caused by improper temperature. Secondly, the change in stirring speed will directly affect the uniformity of the mixed material, and then affect the performance of the final product. Therefore, the speed sensor will also collect real-time speed data based on time to form a speed time queue. This process is similar to temperature monitoring. The speed sensor converts the speed information into a digital signal, which is processed and stored by the data acquisition system. The acquired speed time queue can adjust the stirring speed in real time to ensure that the mixing process meets the set process requirements.

[0066] In an embodiment of the present application, the step S123, sequence encoding the time queue of the real-time temperature values ​​and the time queue of the real-time stirring speed values ​​to obtain a sequence of real-time temperature local timing features and a real-time stirring speed timing feature, includes: S1231, inputting the time queue of the real-time temperature values ​​into a sequence encoder based on a 1D-CNN model to obtain a sequence of real-time temperature local timing implicitly associated feature vectors as the sequence of the real-time temperature local timing features; S1232, inputting the time queue of the real-time stirring speed values ​​into a sequence encoder based on a forward LSTM model to obtain a real-time stirring speed timing implicitly associated feature vector as the real-time stirring speed timing feature.

[0067] Specifically, in step S1231, the time queue of the real-time temperature value is input into a sequence encoder based on the 1D-CNN model to obtain a sequence of implicitly associated feature vectors of the local time series of the real-time temperature. It should be understood that the time queue of the real-time temperature value is an ordered sequence of numerical values, which usually contains periodic changes, emergencies and possible noise, reflecting the change of temperature over time. Directly processing the original data may be difficult to effectively capture these time series features, so a sequence encoder based on the 1D-CNN model is used for processing. Among them, 1D-CNN can extract local features of the input time series data by sliding on the time dimension by applying a convolution kernel (filter). The convolution operation can extract the features of the local area and emphasize certain important temperature change patterns, such as rising, falling or stable stages. In specific implementation, the time queue of the real-time temperature value is first converted into a format suitable for inputting the 1D-CNN model. This usually includes standardizing the temperature value to adapt it to the input requirements of the network. Then, the time queue is divided into multiple overlapping windows (sliding windows), each of which contains a certain number of continuous temperature values. These windows are used as input data and feature extraction is performed by 1D-CNN. In the convolution layer, multiple convolution kernels process each window in parallel to generate multiple feature maps. These feature maps not only retain the temporal characteristics of the time series, but also extract the local correlation of the temperature data. Then, after processing with convolution and activation functions (such as ReLU), the feature map is downsampled through the pooling layer to reduce the dimension of the data and retain the main features. The pooling operation can help reduce computational complexity, enhance the robustness of the model, and reduce sensitivity to noise. Finally, after multiple layers of convolution and pooling, the extracted feature vectors are flattened and input into the fully connected layer to form the final sequence of local temporal implicit correlation feature vectors, thereby capturing the dynamic characteristics of temperature changes over time and identifying the laws of temperature changes.

[0068] Specifically, in step S1232, the time queue of the real-time stirring speed value is input into the sequence encoder based on the forward LSTM model to obtain the real-time stirring speed time series implicit correlation feature vector. It should be understood that the LSTM network is suitable for processing time series data and can effectively capture the long-term dependencies and time series dynamic changes in the time series. Among them, LSTM overcomes the gradient vanishing problem of the traditional RNN (recurrent neural network) when processing long sequences by introducing a gating mechanism, so that it performs well in capturing the characteristics of the time series. In the technical solution of the present application, the change in speed may be affected by many factors, such as the properties of the material, the load of the agitator, etc. Therefore, understanding the time series characteristics of the speed is very important for optimizing the stirring process and improving product quality. Inputting these real-time speed values ​​into the forward LSTM model can help the model learn the pattern and law of the speed change over time, thereby obtaining meaningful implicit features. Specifically, the time queue of the real-time stirring speed value needs to be preprocessed first. This process includes data standardization to ensure that the mean and variance of the input data meet the requirements of the model. Then, the speed time queue is divided into multiple overlapping time windows. Each window will contain a certain number of continuous speed values, which will be used as the input of the LSTM model. Among them, the input of the LSTM model is usually three-dimensional, with the shape of (number of samples, time step, number of features), where the number of features is 1, representing the speed value. Then, these time windows are input to the LSTM layer for processing. LSTM can effectively retain and update long-term and short-term memory through its unique gating structure (input gate, forget gate and output gate), so that the model can pay attention to important historical information when learning time series features. This mechanism ensures that the model not only pays attention to the current speed value, but also comprehensively considers the previous speed changes and captures the trend and law of the speed in a specific time period. Finally, after being processed by the LSTM layer, the obtained implicit state vector can be regarded as the time series implicit correlation feature vector of the real-time stirring speed. The obtained real-time stirring speed time series implicit correlation feature vector summarizes the dynamic information of the speed change over time and can reflect the important patterns in the stirring process. In this way, it is helpful to extract the time series features in the stirring process and improve the system's understanding of the change of stirring speed.

[0069] Specifically, in step S124, the sequence of the local time series features of the real-time temperature and the time series features of the real-time stirring speed are subjected to time series feature query response optimization processing to obtain the real-time stirring speed-real-time temperature query response optimization feature. It should be understood that the sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector respectively contain the time series association features of the real-time stirring speed and each local time series feature information of the real-time temperature, and since the sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector in the original time series feature space contains a large amount of information that is irrelevant to the subsequent control task, this will make it difficult to efficiently focus on the truly useful stirring speed time series semantic part. In order to extract the most relevant information from the two key parameter timing features and to express the key part of the stirring speed timing semantics, and to guide the subsequent predictive control of stirring speed by optimizing the coding, so as to ensure the best coordination between temperature and stirring speed during the mixing process, in the technical solution of the present application, the sequence of the real-time temperature local timing features and the real-time stirring speed timing features are further optimized to obtain the real-time stirring speed-real-time temperature query response optimization features. Through the optimization of the timing feature query response, each real-time temperature local timing feature can be used to match and query the real-time stirring speed timing semantics that interact with the key features, so as to filter out the noise and highlight the key timing semantic information, thereby alleviating the problem of information sparsity and redundancy. At the same time, in this way, weights can be dynamically allocated within the entire sequence range, and the temporal context association between the real-time stirring speed and the real-time temperature can be better understood, thereby improving the learning effect of the long-distance dependency and providing a basis for the subsequent predictive control of stirring speed.

[0070] Figure 4 The present invention is a flowchart of performing timing feature query response optimization processing on the sequence of the real-time temperature local timing feature and the real-time stirring speed timing feature in the automatic control method for fiber cement board making according to the embodiment of the present application to obtain the real-time stirring speed-real-time temperature query response optimization feature. Figure 4As shown, in an embodiment of the present application, in the step S124, the sequence of the real-time temperature local time series features and the real-time stirring speed time series features are subjected to time series feature query response optimization processing to obtain the real-time stirring speed-real-time temperature query response optimization feature, including: S1241, the sequence of the real-time temperature local time series implicit association feature vectors is subjected to feature aggregation processing based on linear embedding coding to obtain a prompt template; S1242, based on the prompt template, the sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector is subjected to cross-domain optimization query coding to obtain the real-time stirring speed-real-time temperature query response optimization coding vector as the real-time stirring speed-real-time temperature query response optimization feature.

[0071] Figure 5 The present invention is a flowchart of performing linear embedding coding-based feature aggregation processing on the sequence of the real-time temperature local time series implicit correlation feature vectors in the automatic control method for fiber cement board manufacturing according to the embodiment of the present application to obtain a prompt template. Figure 5 As shown, in an embodiment of the present application, in the step S1241, the sequence of the real-time temperature local time series implicitly associated feature vectors is subjected to feature aggregation processing based on linear embedding coding to obtain a prompt template, including: S1241-1, using a key embedding matrix to perform linear embedding coding on each real-time temperature local time series implicitly associated feature vector in the sequence of the real-time temperature local time series implicitly associated feature vectors to obtain a sequence of real-time temperature local time series implicitly associated feature vectors after linear transformation; S1241-2, arranging the sequence of real-time temperature local time series implicitly associated feature vectors after linear transformation into a matrix to obtain a real-time temperature local time series key matrix; S1241-3, extracting the maximum value of each real-time temperature local time series implicitly associated feature vector after linear transformation in the real-time temperature local time series key matrix to obtain a significant feature vector of the real-time temperature local time series key matrix as the prompt template.

[0072] Specifically, in step S1241-1, a key embedding matrix is ​​used to perform linear embedding coding on each real-time temperature local time series implicitly associated feature vector in the sequence of the real-time temperature local time series implicitly associated feature vector to obtain a sequence of real-time temperature local time series implicitly associated feature vectors after linear transformation. It should be understood that the original sequence of the real-time temperature local time series implicitly associated feature vectors is embedded and mapped to a new potential space in order to capture the relationship between different dimensions and enhance the model's understanding of the complex temporal association pattern between temperature and stirring speed. After such linear embedding coding, the sequence of real-time temperature local time series implicitly associated feature vectors after linear transformation becomes more suitable for participating in subsequent attention mechanism calculations, thereby improving the model's expressiveness and enabling the model to more effectively identify important temporal feature associations.

[0073] In an embodiment of the present application, the step S1241-1 uses a key embedding matrix to perform linear embedding coding on each real-time temperature local timing implicit association feature vector in the sequence of the real-time temperature local timing implicit association feature vector to obtain a sequence of real-time temperature local timing implicit association feature vectors after linear transformation, including: multiplying the real-time temperature local timing implicit association feature vector by the key embedding matrix and then adding the key embedding bias vector by position to obtain the real-time temperature local timing implicit association feature vector after linear transformation.

[0074] Specifically, in step S1241-2, the sequence of the real-time temperature local time series implicitly associated feature vectors after the linear transformation is arranged in a matrix to obtain a real-time temperature local time series key matrix. It should be understood that the real-time temperature local time series implicitly associated feature vectors after the linear transformation are arranged in sequence to form a new data structure called a key matrix, so as to facilitate parallel processing, thereby improving calculation efficiency.

[0075] Specifically, in the step S1241-3, the maximum value of the real-time temperature local time sequence implicitly associated feature vectors after each linear transformation in the real-time temperature local time sequence key matrix is ​​extracted to obtain the real-time temperature local time sequence key matrix significant feature vector as the prompt template. It should be understood that the maximum value of each column (i.e., the real-time temperature local time sequence implicitly associated feature vector after each linear transformation) is selected from the key matrix to obtain the real-time temperature local time sequence key matrix significant feature vector. Then, the real-time temperature local time sequence key matrix significant feature vector is used as a prompt template, representing the most prominent or representative part of the real-time temperature local time sequence feature. The real-time temperature local time sequence key matrix significant feature vector is introduced as an additional guidance signal, which helps to reduce noise interference and strengthen the model's learning of important patterns. At the same time, it can also help the attention mechanism focus more on the key parts related to the subsequent tasks in the real-time temperature local time sequence feature, thereby improving the model's attention to the important temperature local time sequence detail semantics and the information with high relevance to the subsequent stirring speed optimization control task.

[0076] Figure 6 The present invention is a flowchart of performing cross-domain optimization query coding on the sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector based on the prompt template in the automatic control method for fiber cement board making according to the embodiment of the present application to obtain the real-time stirring speed-real-time temperature query response optimization coding vector as the real-time stirring speed-real-time temperature query response optimization feature. Figure 6As shown, in an embodiment of the present application, in the step S1242, based on the prompt template, the sequence of the real-time stirring speed time series implicitly associated feature vector and the real-time temperature local time series implicitly associated feature vector is cross-domain optimized query encoded to obtain a real-time stirring speed-real-time temperature query response optimized encoding vector as the real-time stirring speed-real-time temperature query response optimized feature, including: S1242-1, using the query embedding matrix and the value embedding matrix to linearly embed the real-time stirring speed time series implicitly associated feature vector to obtain a real-time stirring speed time series query vector and a real-time stirring speed time series value vector; S1242-2, input the real-time stirring speed timing query vector, the real-time stirring speed timing value vector, the real-time temperature local timing implicit association feature vector after each linear transformation in the real-time temperature local timing key matrix, and the prompt template into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of real-time stirring speed-real-time temperature timing semantic cross-domain optimization query encoding vectors; S1242-3, calculate the positional mean vector of the sequence of real-time stirring speed-real-time temperature timing semantic cross-domain optimization query encoding vectors to obtain the real-time stirring speed-real-time temperature query response optimization encoding vector.

[0077] Specifically, in step S1242-1, the query embedding matrix and the value embedding matrix are used to perform linear embedding coding on the real-time stirring speed time series implicitly associated feature vector to obtain the real-time stirring speed time series query vector and the real-time stirring speed time series value vector. It should be understood that the query embedding matrix and the value embedding matrix are used to act on the real-time stirring speed time series implicitly associated feature vector respectively to generate the corresponding real-time stirring speed time series query vector and the real-time stirring speed time series value vector. These two vectors each play different but complementary roles: the real-time stirring speed time series query vector is used to determine which real-time stirring speed time series semantic information needs to be paid attention to, while the real-time stirring speed time series value vector carries the actual time series semantic feature content. The combination of the two can accurately select feature information related to subsequent control tasks while maintaining the context, thereby enhancing the generalization ability and accuracy of the processing.

[0078] Specifically, in step S1242-2, the real-time stirring speed time series query vector, the real-time stirring speed time series value vector, the real-time temperature local time series implicitly associated feature vectors after each linear transformation in the real-time temperature local time series key matrix, and the prompt template are input into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of real-time stirring speed-real-time temperature time series semantic cross-domain optimized query encoding vectors. It should be understood that the real-time stirring speed time series query vector, the real-time stirring speed time series value vector, the real-time temperature local time series implicitly associated feature vectors after each linear transformation in the real-time temperature local time series key matrix, and the prompt template are jointly input into a specially designed heterogeneous converter structure, which usually includes a heterogeneous multi-head attention mechanism component, which can automatically adjust the weight allocation strategy according to the provided prompt template, realize the effective optimization of the stirring speed time series query feature, and enhance the ability to handle complex stirring speed and temperature time series query responses and stirring speed control tasks.

[0079] In an embodiment of the present application, the step S1242-2, the real-time stirring speed timing query vector, the real-time stirring speed timing value vector, the real-time temperature local timing implicitly associated feature vector after each linear transformation in the real-time temperature local timing key matrix, and the prompt template are input into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of real-time stirring speed-real-time temperature timing semantic cross-domain optimized query encoding vectors, including: calculating the multiplication between the real-time stirring speed timing query vector and the transposed vector of the real-time temperature local timing implicitly associated feature vector after the linear transformation, and then dividing the obtained real-time stirring speed-real-time temperature timing semantics association feature matrix by the two norms of the significant feature vectors of the real-time temperature local timing key matrix by position to obtain a real-time stirring speed-real-time temperature timing semantics association representation matrix; inputting the real-time stirring speed-real-time temperature timing semantics association representation matrix The real-time stirring speed-real-time temperature time series semantic association weight matrix is ​​processed in the function; after the real-time stirring speed-real-time temperature time series semantic association weight matrix is ​​multiplied by the significant feature vector of the real-time temperature local time series key matrix, the obtained feature vector is multiplied by the real-time stirring speed time series value vector according to the position point to obtain the real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vector.

[0080] Specifically, in the step S1242-3, the position mean vector of the sequence of the real-time stirring speed-real-time temperature time series semantic cross-domain optimization query coding vector is calculated to obtain the real-time stirring speed-real-time temperature query response optimization coding vector. It should be understood that the sequence of the real-time stirring speed-real-time temperature time series semantic cross-domain optimization query coding vector is averaged according to the position, and finally the real-time stirring speed-real-time temperature query response optimization coding vector is obtained. This operation not only simplifies the expression of the result, which is convenient for further decoding prediction generation tasks, but also retains the comprehensive characteristics of all position information in the sequence of the real-time stirring speed-real-time temperature time series semantic cross-domain optimization query coding vector, ensuring that the model output can reflect the cross-domain optimization query coding core attributes between the entire real-time stirring speed and real-time temperature time series semantics. This method helps the model better understand the long-term time series interaction response characteristics and change trends between stirring speed and temperature, reduce noise interference, and provide important and simple feature representations for subsequent stirring speed optimization control tasks.

[0081] Specifically, the sequence of the real-time temperature local time series features and the real-time stirring speed time series features are subjected to time series feature query response optimization processing using the following feature response interaction formula to obtain the real-time stirring speed-real-time temperature query response optimization feature;

[0082] Among them, the characteristic response interaction formula is: ; ; ; ; ; ; ; ;in, is the sequence of implicitly associated feature vectors of the local time series of real-time temperature, are the first, second and third in the sequence of the real-time temperature local time series implicit correlation feature vectors. The real-time temperature local time series implicit correlation feature vector, is the first in the sequence of implicitly associated feature vectors of the real-time temperature local time series The real-time temperature local time series implicit correlation feature vector, and are the key embedding matrix and key embedding bias vector respectively, is the sequence of implicitly associated eigenvectors of the local time series of real-time temperature after linear transformation. The real-time temperature local time series implicit correlation feature vector after linear transformation, is the local timing key matrix of real-time temperature, To extract the maximum value in a vector, is the significant eigenvector of the local time-series key matrix of real-time temperature, is the implicitly associated eigenvector of the real-time stirring speed time series, and are the query embedding matrix and the query bias vector respectively, and are the value embedding matrix and value bias vector respectively, is the real-time stirring speed time series query vector, is the real-time stirring speed time series value vector, is the two-norm of the vector, for function, is matrix multiplication, For point-by-position multiplication, The first in the sequence of real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vector A real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vector, The number of vectors in the sequence of the real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vector, The encoding vector is optimized for the real-time stirring speed-real-time temperature query response.

[0083] In an embodiment of the present application, the step S125 generates a stirring speed optimization control instruction based on the real-time stirring speed-real-time temperature query response optimization feature, including: inputting the real-time stirring speed-real-time temperature query response optimization coding vector into a stirring speed optimization control module based on a decoder to obtain the stirring speed optimization control instruction, and the stirring speed optimization control instruction includes a recommended value of the stirring speed value at the next time point. In other words, the query response optimization feature between the stirring speed and the temperature is used to perform decoding regression, so as to perform predictive optimization control of the stirring speed. In this way, the stirring speed value at the next time point can be automatically generated according to the actual temperature data and stirring speed data, so as to automatically perform process control of the heating stirring reaction, thereby improving the intelligence and adaptability of the fiber cement board making process control, and helping to ensure the preparation quality and performance of the fiber cement board.

[0084] Preferably, the sequences of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector respectively represent the short-range-long-range bidirectional time series association features of the real-time stirring speed value and the local time domain time series association features of the real-time temperature. In this way, when performing query coding optimization with the key matrix significant features as the prompt template, the key matrix significant prompt differences under the time series distribution pattern differences will lead to different query coding aggregation modes, causing the real-time stirring speed-real-time temperature query response optimization coding vector to have a complex full time domain space aggregation structure. Therefore, when the real-time stirring speed-real-time temperature query response optimization coding vector is input into the stirring speed optimization control module based on the decoder, it is expected to improve the decoding regression convergence and generalization effect of the real-time stirring speed-real-time temperature query response optimization coding vector under the complex time domain space aggregation structure.

[0085] Therefore, when the applicant of the present application inputs the real-time stirring speed-real-time temperature query response optimization coding vector into the decoding-based stirring speed optimization control module to obtain the stirring speed optimization control instruction, the real-time stirring speed-real-time temperature query response optimization coding vector is optimized. The optimization process includes: calculating the absolute value sum of each eigenvalue of the real-time stirring speed-real-time temperature query response optimization coding vector to obtain a first real-time stirring speed-real-time temperature query response optimization coding space structure value, and calculating the square root of the sum of the squares of each eigenvalue of the real-time stirring speed-real-time temperature query response optimization coding vector to obtain a second real-time stirring speed-real-time temperature query response optimization coding space structure value: ; ;in, represents each eigenvalue of the real-time stirring speed-real-time temperature query response optimization encoding vector, , represents the summation function, represents the absolute value function, represents the set of real numbers, represents the first real-time stirring speed-real-time temperature query response optimized encoding space structure value, represents the second real-time stirring speed-real-time temperature query response optimized encoding space structure value, represents the set of real numbers, Indicates the length of a vector;

[0086] Each eigenvalue of the real-time stirring speed-real-time temperature query response optimized coding vector is multiplied by the first real-time stirring speed-real-time temperature query response optimized coding space structure value and the second real-time stirring speed-real-time temperature query response optimized coding space structure value to obtain a first real-time stirring speed-real-time temperature query response optimized coding structure reference value and a second real-time stirring speed-real-time temperature query response optimized coding structure reference value corresponding to each eigenvalue; each eigenvalue of the real-time stirring speed-real-time temperature query response optimized coding vector is multiplied by the length of the real-time stirring speed-real-time temperature query response optimized coding vector and the square root of the length to obtain a first real-time stirring speed-real-time temperature query response optimized coding scale transformation value and a second real-time stirring speed-real-time temperature query response optimized coding scale transformation value corresponding to each eigenvalue; the first real-time stirring speed-real-time temperature query response optimized coding structure is multiplied by the length of the real-time stirring speed-real-time temperature query response optimized coding vector and the square root of the length to obtain a first real-time stirring speed-real-time temperature query response optimized coding scale transformation value and a second real-time stirring speed-real-time temperature query response optimized coding scale transformation value corresponding to each eigenvalue; The first real-time stirring speed-real-time temperature query response optimization coding space structure value is divided by the difference between the first real-time stirring speed-real-time temperature query response optimization coding space structure value and the first real-time stirring speed-real-time temperature query response optimization coding scale transformation value to obtain a first real-time stirring speed-real-time temperature query response optimization coding transformation adjustment value; the second real-time stirring speed-real-time temperature query response optimization coding structure reference value is divided by the difference between the second real-time stirring speed-real-time temperature query response optimization coding space structure value and the second real-time stirring speed-real-time temperature query response optimization coding scale transformation value to obtain a second real-time stirring speed-real-time temperature query response optimization coding transformation adjustment value; and the weighted sum of the first real-time stirring speed-real-time temperature query response optimization coding transformation adjustment value and the second real-time stirring speed-real-time temperature query response optimization coding transformation adjustment value is calculated to obtain each eigenvalue of the optimized real-time stirring speed-real-time temperature query response optimization coding vector.

[0087] Here, the optimized real-time stirring speed-real-time temperature query response optimization encoding vector It is expressed as: exist ,and In the case of: ; in, Represents each eigenvalue of the real-time stirring speed-real-time temperature query response optimization encoding vector, represents the first real-time stirring speed-real-time temperature query response optimized encoding space structure value, represents the second real-time stirring speed-real-time temperature query response optimized encoding space structure value, Indicates the first real-time stirring speed-real-time temperature query response optimization code conversion adjustment value, Indicates the second real-time stirring speed-real-time temperature query response optimization code conversion adjustment value, represents a first real-time stirring speed-real-time temperature query response optimized coding transformation adjustment feature vector obtained by vectorizing the plurality of first real-time stirring speed-real-time temperature query response optimized coding transformation adjustment values, represents a second real-time stirring speed-real-time temperature query response optimized coding transformation adjustment feature vector obtained by vectorizing a plurality of second real-time stirring speed-real-time temperature query response optimized coding transformation adjustment values, represents the optimized real-time stirring speed-real-time temperature query response optimization encoding vector, It means adding by position. Indicates the query response optimization encoding adjustment weighted value, represents positional multiplication, represents the set of real numbers, Indicates the length of a vector.

[0088] That is, for the spatial structure information of the feature set of the real-time stirring speed-real-time temperature query response optimization coding vector in the high-dimensional space, the scale-based frame transformation of each eigenvalue of the real-time stirring speed-real-time temperature query response optimization coding vector is performed by using the class norm space structured representation of the real-time stirring speed-real-time temperature query response optimization coding vector as a reference window, and the spatial structure-based frame attention weight adjustment of each eigenvalue of the real-time stirring speed-real-time temperature query response optimization coding vector is realized to ensure the spatial transformation invariance of the real-time stirring speed-real-time temperature query response optimization coding vector under the interaction of the feature space, thereby improving the convergence and generalization effect of the decoding regression of the feature set of the real-time stirring speed-real-time temperature query response optimization coding vector under the complex time domain spatial structure representation, and improving the accuracy of the obtained stirring speed optimization control instruction. In this way, the stirring speed value at the next time point can be automatically generated according to the actual temperature data and stirring speed data, so as to automatically perform the process control of the heating stirring reaction, thereby improving the intelligence and adaptability of the fiber cement board making process control, which helps to ensure the preparation quality and performance of the fiber cement board.

[0089] In the above-mentioned automatic control method for fiber cement board making, the step S130 is to stack and cut the solid wet material to form a fiber cement wet blank. It should be understood that the fiber cement wet blank is a new type of building material, which is mainly composed of cement, fiber and other modified materials, and has good strength and toughness. The fibers (such as glass fiber, aramid fiber or natural fiber) added to its structure effectively enhance the tensile and bending properties of the material, reduce the occurrence of cracks, and improve durability. Specifically, in the technical solution of the present application, the solid wet material is evenly layered and placed in the forming mold by stacking. The stacking process must ensure that each layer is well bonded to prevent problems such as delamination or bubbles during subsequent molding. In order to improve the bonding strength between the layers, slight compaction can be appropriately performed during the stacking process to ensure full contact between the layers. Then, the stacked wet blank is cut into the required size and shape by laser cutting to achieve uniform distribution and appropriate shape of the material, and obtain a high-quality wet blank product.

[0090] In the above-mentioned automatic control method for fiber cement board making, the step S140 is to pressurize, rest, and steam-cured the fiber cement wet board at high temperature and high pressure to obtain the fiber cement board. It should be understood that the pressurization step is to apply a certain pressure to the fiber cement wet board by using a press machine, the purpose of which is to improve the compactness of the board. In this process, the pores and bubbles of the wet board are compressed, and the combination between the fiber and the cement matrix is ​​tighter, thereby enhancing the strength of its overall structure. The pressurization process not only helps to remove excess water, but also makes the fiber and the cement matrix form a better dispersion and interweaving, thereby improving the mechanical properties of the board. Next, the resting stage is to let the pressurized wet board stand for a certain period of time so that the board can naturally solidify and hydrate. During the resting process, the moisture in the cement will gradually react with the cement particles to form hydration products (such as hydrated calcium silicate), so that the cement structure is gradually strengthened. This process helps to alleviate the local stress concentration caused by pressurization and avoid deformation or cracks in the board. At the same time, resting can also make the combination of fiber and cement more firm, further improving the overall performance of the material. Finally, it enters the high-temperature and high-pressure steam curing stage. By placing the wet blank board in a high-temperature and high-pressure steam environment (usually carried out in a steam curing kettle), the cement hydration reaction is accelerated under temperature and pressure conditions, and the strength and hardness of the fiber cement board are significantly improved. During the high-temperature and high-pressure steam curing process, the rate of cement hydration reaction is greatly accelerated, the internal structure of the board becomes denser, the moisture is further removed, and the water resistance and frost resistance of the cement board are enhanced. Among them, the high-temperature and high-pressure steam curing treatment needs to be carried out at a temperature between 120°C and 180°C and a pressure of 2 to 3 atmospheres, and the duration ranges from a few hours to more than ten hours. In this way, the hydration reaction of the cement matrix of the wet blank board can be promoted, and its hardness and density can be enhanced, thereby obtaining a fiber cement board with better physical properties.

[0091] In summary, an automatic control method for fiber cement board making based on the embodiment of the present application is clarified, which monitors and collects real-time temperature values ​​and real-time stirring speed values, and then introduces data processing and analysis algorithms based on artificial intelligence and deep learning in the back end to perform timing collaborative analysis, so as to capture and characterize the query response optimization feature representation between the real-time stirring speed timing characteristics and the real-time temperature local timing characteristics, so as to reflect the key query feature response and interaction between the stirring speed timing semantics and the temperature timing semantics, thereby automatically generating the stirring speed value at the next time point, thereby automatically performing process control of the heating and stirring reaction, thereby improving the intelligence and adaptability of the fiber cement board making process control, and helping to ensure the preparation quality and performance of the fiber cement board.

[0092] Figure 7 FIG. 1 is a system block diagram of an automatic control system for fiber cement board making according to an embodiment of the present application. Figure 7As shown, according to an embodiment of the present application, an automatic control system 100 for fiber cement board making includes: a real-time temperature value acquisition unit 110, used to acquire a time queue of real-time temperature values ​​collected by a temperature sensor; a real-time stirring speed value acquisition unit 120, used to acquire a time queue of real-time stirring speed values ​​collected by a speed sensor; a sequence encoding unit 130, used to perform sequence encoding on the time queue of the real-time temperature values ​​and the time queue of the real-time stirring speed values ​​to obtain a sequence of real-time temperature local time series characteristics and a real-time stirring speed time series characteristic; a time series characteristic query response optimization processing unit 140, used to perform time series characteristic query response optimization processing on the sequence of real-time temperature local time series characteristics and the real-time stirring speed time series characteristics to obtain a real-time stirring speed-real-time temperature query response optimization characteristic; a stirring speed optimization unit 150, used to generate a stirring speed optimization control instruction based on the real-time stirring speed-real-time temperature query response optimization characteristic.

[0093] Here, those skilled in the art will appreciate that the specific operations of each step in the above-mentioned automatic control system for fiber cement board making have been described in detail above. Figures 1 to 6 The automatic control method for fiber cement board making has been described in detail in the description of the automatic control method for fiber cement board making, and therefore, its repeated description will be omitted.

[0094] In summary, an automatic control system for fiber cement board making based on the embodiment of the present application is clarified, which monitors and collects real-time temperature values ​​and real-time stirring speed values, and then introduces data processing and analysis algorithms based on artificial intelligence and deep learning at the back end to perform timing collaborative analysis, so as to capture and characterize the query response optimization feature representation between the real-time stirring speed timing characteristics and the real-time temperature local timing characteristics, so as to reflect the key query feature response and interaction between the stirring speed timing semantics and the temperature timing semantics, thereby automatically generating the stirring speed value at the next time point, thereby automatically performing process control of the heating and stirring reaction, thereby improving the intelligence and adaptability of the fiber cement board making process control, and helping to ensure the preparation quality and performance of the fiber cement board.

Claims

1. An automatic control method for fiber cement board making, characterized in that: include: Mixing and stirring water, cement binder and aggregate to obtain a mixed material; Adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material, heating and stirring the mixture for reaction, and then filtering the water to obtain a solid wet material; The solid wet material is stacked and cut to form a fiber cement wet board; The fiber cement wet board is subjected to pressurization, static curing, and high-temperature and high-pressure steam curing to obtain a fiber cement board; The method comprises adding reinforcing fibers and modified SiO2 / TiO2 / Al2O3 composite sea thorn-shaped polyaniline nano hollow microspheres to the mixed material, heating and stirring the mixture for reaction, and filtering the water to obtain a solid wet material, including: Get the time queue of the real-time temperature value collected by the temperature sensor; Obtaining a time queue of real-time stirring speed values ​​collected by a speed sensor; Sequentially encoding the time queue of the real-time temperature value and the time queue of the real-time stirring speed value to obtain a sequence of local time series characteristics of the real-time temperature and a time series characteristic of the real-time stirring speed; Performing a time series feature query response optimization process on the sequence of the real-time temperature local time series features and the real-time stirring speed time series features to obtain a real-time stirring speed-real-time temperature query response optimization feature; Generating a stirring speed optimization control instruction based on the real-time stirring speed-real-time temperature query response optimization feature; The time queue of the real-time temperature value and the time queue of the real-time stirring speed value are sequentially encoded to obtain a sequence of local time series features of the real-time temperature and a time series feature of the real-time stirring speed, including: Inputting the time queue of the real-time temperature value into a sequence encoder based on a 1D-CNN model to obtain a sequence of real-time temperature local time series implicit correlation feature vectors as a sequence of real-time temperature local time series features; The time queue of the real-time stirring speed value is input into a sequence encoder based on a forward LSTM model to obtain a real-time stirring speed time series implicit correlation feature vector as the real-time stirring speed time series feature.

2. The automatic control method for fiber cement board making according to claim 1, characterized in that: The sequence of the real-time temperature local time series features and the real-time stirring speed time series features are subjected to time series feature query response optimization processing to obtain a real-time stirring speed-real-time temperature query response optimization feature, including: Performing feature aggregation processing based on linear embedding coding on the sequence of implicitly associated feature vectors of the real-time temperature local time series to obtain a prompt template; Based on the prompt template, a sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector is cross-domain optimized query encoded to obtain a real-time stirring speed-real-time temperature query response optimized encoding vector as the real-time stirring speed-real-time temperature query response optimized feature.

3. The automatic control method for fiber cement board making according to claim 2, characterized in that: Performing feature aggregation processing based on linear embedding coding on the sequence of the real-time temperature local time series implicitly associated feature vectors to obtain a prompt template includes: Using a key embedding matrix, linear embedding coding is performed on each real-time temperature local time series implicitly associated feature vector in the sequence of real-time temperature local time series implicitly associated feature vectors to obtain a sequence of real-time temperature local time series implicitly associated feature vectors after linear transformation; Arrange the sequence of the real-time temperature local time series implicitly associated characteristic vectors after the linear transformation into a matrix to obtain a real-time temperature local time series key matrix; The maximum value of the real-time temperature local time sequence implicit correlation feature vector after each linear transformation in the real-time temperature local time sequence key matrix is ​​extracted to obtain the real-time temperature local time sequence key matrix significant feature vector as the prompt template.

4. The automatic control method for fiber cement board making according to claim 3, characterized in that: A key embedding matrix is ​​used to perform linear embedding coding on each real-time temperature local time series implicit association feature vector in the sequence of the real-time temperature local time series implicit association feature vector to obtain a sequence of real-time temperature local time series implicit association feature vectors after linear transformation, including: multiplying the real-time temperature local time series implicit association feature vector by the key embedding matrix and then adding the key embedding bias vector by position to obtain the real-time temperature local time series implicit association feature vector after linear transformation.

5. The automatic control method for fiber cement board making according to claim 4, characterized in that: Based on the prompt template, a sequence of the real-time stirring speed time series implicit association feature vector and the real-time temperature local time series implicit association feature vector is cross-domain optimized query encoded to obtain a real-time stirring speed-real-time temperature query response optimized encoding vector as the real-time stirring speed-real-time temperature query response optimized feature, including: Performing linear embedding coding on the real-time stirring speed time series implicitly associated feature vector using a query embedding matrix and a value embedding matrix to obtain a real-time stirring speed time series query vector and a real-time stirring speed time series value vector; Input the real-time stirring speed time series query vector, the real-time stirring speed time series value vector, the real-time temperature local time series implicit association feature vectors after each linear transformation in the real-time temperature local time series key matrix, and the prompt template into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vectors; The positional mean vector of the sequence of the real-time stirring speed-real-time temperature time series semantic cross-domain optimized query encoding vector is calculated to obtain the real-time stirring speed-real-time temperature query response optimized encoding vector.

6. The automatic control method for fiber cement board making according to claim 5, characterized in that: The real-time stirring speed time series query vector, the real-time stirring speed time series value vector, the real-time temperature local time series implicit association feature vector after each linear transformation in the real-time temperature local time series key matrix, and the prompt template are input into a heterogeneous converter structure based on template prompt optimization to obtain a sequence of real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vectors, including: After calculating the multiplication between the real-time stirring speed time series query vector and the transposed vector of the real-time temperature local time series implicit association feature vector after the linear transformation, the obtained real-time stirring speed-real-time temperature time series semantic association feature matrix is ​​divided by the bi-norm of the significant feature vector of the real-time temperature local time series key matrix according to the position to obtain the real-time stirring speed-real-time temperature time series semantic association representation matrix; The real-time stirring speed-real-time temperature time series semantic association representation matrix is ​​input The function is processed to obtain the real-time stirring speed-real-time temperature time series semantic association weight matrix; After multiplying the real-time stirring speed-real-time temperature time series semantic association weight matrix with the significant feature vector of the real-time temperature local time series key matrix, the obtained feature vector is point-multiplied with the real-time stirring speed time series value vector to obtain the real-time stirring speed-real-time temperature time series semantic cross-domain optimization query encoding vector.

7. The automatic control method for fiber cement board making according to claim 6, characterized in that: Generating a stirring speed optimization control instruction based on the real-time stirring speed-real-time temperature query response optimization feature includes: inputting the real-time stirring speed-real-time temperature query response optimization coding vector into a decoder-based stirring speed optimization control module to obtain the stirring speed optimization control instruction, wherein the stirring speed optimization control instruction includes a recommended value of the stirring speed value at the next time point.

8. An automatic control system for fiber cement board making, characterized in that: include: A real-time temperature value acquisition unit, used to acquire a time queue of real-time temperature values ​​collected by a temperature sensor; A real-time stirring speed value acquisition unit, used to acquire a time queue of a real-time stirring speed value collected by a speed sensor; A sequence encoding unit, used for performing sequence encoding on the time queue of the real-time temperature value and the time queue of the real-time stirring speed value to obtain a sequence of local time series characteristics of the real-time temperature and a time series characteristic of the real-time stirring speed; A time series feature query response optimization processing unit, used for performing time series feature query response optimization processing on the sequence of the real-time temperature local time series features and the real-time stirring speed time series features to obtain a real-time stirring speed-real-time temperature query response optimization feature; A stirring speed optimization unit, used for generating a stirring speed optimization control instruction based on the real-time stirring speed-real-time temperature query response optimization feature; Wherein, the sequence encoding unit is used to: Inputting the time queue of the real-time temperature value into a sequence encoder based on a 1D-CNN model to obtain a sequence of real-time temperature local time series implicit correlation feature vectors as a sequence of real-time temperature local time series features; The time queue of the real-time stirring speed value is input into a sequence encoder based on a forward LSTM model to obtain a real-time stirring speed time series implicit correlation feature vector as the real-time stirring speed time series feature.

Images