Fiber cement board production line information big data system and method

Abstract

The invention discloses a fiber cement board production line information big data system and method, and the method employs a data processing and analysis algorithm based on artificial intelligence and deep learning to carry out the time sequence feature analysis and fine-grained time sequence feature response correlation coding of real-time temperature, real-time humidity, real-time viscosity and real-time stirring speed. In this way, time sequence fine-grained response interaction characteristics between the temperature, humidity and viscosity and the stirring speed in the stirring process are learned and captured, and time sequence reasoning of stirring speed control is carried out through Bayesian posterior probability calculation among the time sequence fine-grained response interaction characteristics of the temperature, humidity and viscosity in the stirring process, so that the stirring speed value of the next time point is determined. And generating an adjusting instruction. In this way, intelligent control over the stirring process of fiber cement board production can be achieved, and the stability and uniformity of the stirring process can be guaranteed.

Description

Technical Field

[0001] The present application relates to the field of big data, and more specifically, to a fiber cement board production line information big data system and method. Background Art

[0002] Fiber cement board is a building board made of cement as the main binder, with an appropriate amount of reinforcing fibers (such as organic fibers, synthetic fibers or inorganic fibers) added, and a series of processes such as mixing, molding, and curing. Fiber cement board has the characteristics of fire resistance, waterproofness, high strength, and good durability, and is widely used in the construction industry.

[0003] The production process of fiber cement board mainly includes multiple steps such as raw material proportioning, mixing, molding, and curing. In the mixing process, the mixing uniformity of raw materials is crucial to the performance of the material. In other words, the uniformity and stability of mixing directly affect the quality indicators such as strength, durability and appearance of the final product. Therefore, how to effectively monitor and adjust the mixing process of fiber cement board production to ensure the stability and uniformity of the mixing process is an important issue.

[0004] However, traditional mixing control methods rely on manual experience and timing adjustments, lack real-time data support, and are difficult to predictively control the mixing process, which can easily lead to instability in the production process. In addition, the traditional mixing control method has a low level of intelligence, which is not only inefficient, but also may lead to inconsistent product quality, increasing production costs and wasting resources.

[0005] Therefore, a fiber cement board production line information big data system is desired, which can drive the intelligent control of the fiber cement board production mixing process based on big data. Summary of the invention

[0006] In order to solve the above technical problems, the present application is proposed. The embodiment of the present application provides a fiber cement board production line information big data system and method, which uses data processing and analysis algorithms based on artificial intelligence and deep learning to perform time series feature analysis and fine-grained time series feature response association coding for real-time temperature, real-time humidity, real-time viscosity and real-time stirring speed, so as to learn and capture the time series fine-grained response interaction characteristics between the temperature, humidity and viscosity in the stirring process and the stirring speed, and perform time series reasoning of stirring speed control by calculating the Bayesian posterior probability between the three time series fine-grained response interaction characteristics, so as to determine the stirring speed value at the next time point to generate adjustment instructions. In this way, intelligent control of the stirring process of fiber cement board production can be achieved, which helps to ensure the stability and uniformity of the stirring process.

[0007] According to one aspect of the present application, a fiber cement board production line information big data system is provided, which includes: a first information collector, a second information collector, a big data platform and a user end, including: the first information collector includes a temperature sensor, a humidity sensor and a viscosity sensor, wherein the temperature sensor, the humidity sensor and the viscosity sensor are respectively installed in a mixer, and are used to collect real-time temperature, real-time humidity and real-time viscosity;

[0008] The second information collector is a stirring speed monitor, which is used to collect real-time stirring speed;

[0009] The big data platform is used to receive and store the real-time temperature, real-time humidity, real-time viscosity and real-time stirring speed collected by the first information collector and the second information collector, and generate an adjustment instruction based on the real-time temperature, the real-time humidity, the real-time viscosity and the real-time stirring speed;

[0010] The user end is used to receive the adjustment instruction and generate a fiber cement board production control prompt signal based on the adjustment instruction.

[0011] According to another aspect of the present application, a fiber cement board production line information big data method is provided, which includes:

[0012] Acquire a time queue of real-time temperature, a time queue of real-time humidity and a time queue of real-time viscosity collected by a temperature sensor, a humidity sensor and a viscosity sensor installed in the mixer;

[0013] Obtaining a time queue of real-time stirring speed collected by the stirring speed monitor;

[0014] Performing feature extraction on the time queue of the real-time temperature, the time queue of the real-time humidity, the time queue of the real-time viscosity and the time queue of the real-time stirring speed respectively to obtain a real-time temperature time series correlation feature vector, a real-time humidity time series correlation feature vector, a real-time viscosity time series correlation feature vector and a real-time stirring speed time series correlation feature vector;

[0015] Input the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector, the real-time humidity time series association feature vector and the real-time stirring speed time series association feature vector, and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector into a fine-grained local decoupled sequence feature response network to obtain a real-time temperature-stirring speed time series interaction representation vector, a real-time humidity-stirring speed time series interaction representation vector, and a real-time viscosity-stirring speed time series interaction representation vector;

[0016] Performing a posteriori reasoning on stirring control based on the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector, and the real-time viscosity-stirring speed time series interaction representation vector to obtain a stirring control multimodal time series interaction reasoning representation vector;

[0017] Based on the stirring control multimodal temporal interactive reasoning representation vector, the adjustment instruction is determined, and the adjustment instruction is used to indicate that the stirring speed value at the next time point should increase, decrease or remain unchanged.

[0018] Compared with the prior art, the fiber cement board production line information big data system and method provided by the present application adopts data processing and analysis algorithms based on artificial intelligence and deep learning to perform time series feature analysis and fine-grained time series feature response association coding for real-time temperature, real-time humidity, real-time viscosity and real-time stirring speed, thereby learning and capturing the time series fine-grained response interaction features between the temperature, humidity and viscosity and the stirring speed during the stirring process, and performing time series reasoning of stirring speed control by calculating the Bayesian posterior probability between the three time series fine-grained response interaction features, thereby determining the stirring speed value at the next time point to generate adjustment instructions. In this way, intelligent control of the stirring process of fiber cement board production can be achieved, which helps to ensure the stability and uniformity of the stirring process. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] 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.

[0020] Figure 1 is a block diagram of a fiber cement board production line information big data system according to an embodiment of the present application;

[0021] Figure 2 A schematic diagram of data flow of a fiber cement board production line information big data system according to an embodiment of the present application;

[0022] Figure 3 A block diagram of a big data platform in a fiber cement board production line information big data system according to an embodiment of the present application;

[0023] Figure 4 It is a flow chart of the fiber cement board production line information big data method according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0025] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0026] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0027] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0028] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0029] Traditional mixing control methods mostly rely on manual experience and timing adjustments, lack real-time data support, and are difficult to predictively control the mixing process, which can easily lead to instability in the production process. In addition, the intelligence level of traditional mixing control methods is low, which is not only inefficient, but may also lead to inconsistent product quality, increase production costs and waste resources. Therefore, a fiber cement board production line information big data system is desired, which can drive the intelligent control of the fiber cement board production mixing process based on big data.

[0030] In the technical solution of the present application, a fiber cement board production line information big data system is proposed. Figure 1 It is a block diagram of a fiber cement board production line information big data system according to an embodiment of the present application. Figure 2 FIG. 1 is a schematic diagram of data flow in a fiber cement board production line information big data system according to an embodiment of the present application. Figure 1 and Figure 2As shown, according to the fiber cement board production line information big data system 300 of the embodiment of the present application, it includes: a first information collector 310; a second information collector 320; a big data platform 330 and a user terminal 340.

[0031] In particular, the first information collector 310 includes a temperature sensor, a humidity sensor and a viscosity sensor, wherein the temperature sensor, the humidity sensor and the viscosity sensor are respectively installed in the mixer to collect real-time temperature, real-time humidity and real-time viscosity.

[0032] In particular, the second information collector 320 is a stirring speed monitor for collecting real-time stirring speed.

[0033] In particular, the big data platform 330. Wherein, the big data platform is used to receive and store the real-time temperature, real-time humidity, real-time viscosity and real-time stirring speed collected by the first information collector and the second information collector, and generate adjustment instructions based on the real-time temperature, the real-time humidity, the real-time viscosity and the real-time stirring speed. In particular, in a specific example of the present application, Figure 3As shown, the big data platform 330 includes: a stirring state data acquisition module 331, which is used to obtain the time queue of real-time temperature, the time queue of real-time humidity and the time queue of real-time viscosity collected by the temperature sensor, humidity sensor and viscosity sensor installed in the mixer; a stirring speed acquisition module 332, which is used to obtain the time queue of real-time stirring speed collected by the stirring speed monitor; a stirring parameter time series feature extraction module 333, which is used to extract the time queue of real-time temperature, the time queue of real-time humidity, the time queue of real-time viscosity and the time queue of real-time stirring speed respectively to obtain the real-time temperature time series correlation feature vector, the real-time humidity time series correlation feature vector, the real-time viscosity time series correlation feature vector and the real-time stirring speed time series correlation feature vector; a stirring parameter multimodal time series semantic fine-grained response interaction module 334, which is used to extract the real-time temperature time series correlation feature vector and the real-time stirring speed time series correlation feature vector, the real-time humidity time series correlation feature vector, the real-time viscosity time series correlation feature vector and the real-time stirring speed time series correlation feature vector. The real-time humidity time series association feature vector and the real-time stirring speed time series association feature vector and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector are respectively input into the fine-grained local decoupled sequence feature response network to obtain the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector; a stirring control multimodal time series semantic interaction reasoning module 335 is used to perform stirring control posterior reasoning based on the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector to obtain the stirring control multimodal time series interaction reasoning representation vector; a stirring speed control module 336 is used to determine the adjustment instruction based on the stirring control multimodal time series interaction reasoning representation vector, and the adjustment instruction is used to indicate that the stirring speed value at the next time point should increase, decrease or remain unchanged.

[0034] Specifically, the stirring state data acquisition module 331 and the stirring speed acquisition module 332 are used to obtain the time queue of real-time temperature, real-time humidity and real-time viscosity collected by the temperature sensor, humidity sensor and viscosity sensor installed in the mixer; and to obtain the time queue of real-time stirring speed collected by the stirring speed monitor. It should be understood that in the stirring production process of fiber cement board, factors such as temperature, humidity and viscosity will affect the fluidity and stability of the mixture. There is a complex nonlinear relationship between these factors and the stirring speed. Therefore, it is necessary to achieve the best stirring effect through real-time monitoring and intelligent control to ensure the stability and uniformity of the stirring process.

[0035] Specifically, the stirring parameter time series feature extraction module 333 is used to extract the time series of the real-time temperature, the time series of the real-time humidity, the time series of the real-time viscosity and the time series of the real-time stirring speed to obtain the real-time temperature time series correlation feature vector, the real-time humidity time series correlation feature vector, the real-time viscosity time series correlation feature vector and the real-time stirring speed time series correlation feature vector. In a specific example of the present application, the time series of the real-time temperature, the time series of the real-time humidity, the time series of the real-time viscosity and the time series of the real-time stirring speed are input into the sequence encoder based on the RNN model to obtain the real-time temperature time series correlation feature vector, the real-time humidity time series correlation feature vector, the real-time viscosity time series correlation feature vector and the real-time stirring speed time series correlation feature vector. That is, the sequence encoder of the RNN model is used to extract the time series correlation feature information of the real-time temperature, the real-time humidity, the real-time viscosity and the real-time stirring speed in the time dimension, so as to obtain the real-time temperature time series correlation feature vector, the real-time humidity time series correlation feature vector, the real-time viscosity time series correlation feature vector and the real-time stirring speed time series correlation feature vector.

[0036] It is worth mentioning that a recurrent neural network (RNN) is a special type of neural network designed specifically for processing sequence data, such as text, speech, and time series. Unlike feedforward neural networks, RNNs have feedback connections that allow information to flow cyclically in the network.

[0037] Specifically, the stirring parameter multimodal time series semantic fine-grained response interaction module 334 is used to input the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector, the real-time humidity time series association feature vector and the real-time stirring speed time series association feature vector, and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector into the fine-grained local decoupled sequence feature response network to obtain the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector, and the real-time viscosity-stirring speed time series interaction representation vector. It should be understood that the time series characteristics of the real-time temperature, real-time humidity, and real-time viscosity in the time dimension have a time series response correlation relationship and complex interaction influence with the stirring speed time series characteristics. Therefore, in order to explore and characterize the characteristic response interaction information and correlation relationship between the real-time temperature time series correlation characteristics, the real-time humidity time series correlation characteristics and the real-time viscosity time series correlation characteristics and the stirring speed time series correlation characteristics, respectively, in the technical solution of the present application, the real-time temperature time series correlation feature vector and the real-time stirring speed time series correlation feature vector, the real-time humidity time series correlation feature vector and the real-time stirring speed time series correlation feature vector and the real-time viscosity time series correlation feature vector and the real-time stirring speed time series correlation feature vector are further input into the fine-grained local decoupled sequence feature response network respectively to obtain the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector.

[0038] In an embodiment of the present application, the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector, the real-time humidity time series association feature vector and the real-time stirring speed time series association feature vector, and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector are respectively input into a fine-grained local decoupled sequence feature response network to obtain a real-time temperature-stirring speed time series interaction representation vector, a real-time humidity-stirring speed time series interaction representation vector, and a real-time viscosity-stirring speed time series interaction representation vector, including: first, feature decoupling the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector to obtain a sequence of real-time temperature time series association local semantic representation vectors and a sequence of real-time stirring speed time series association local semantic representation vectors; through feature decoupling, the original temperature and stirring speed time series feature vectors are decomposed into multiple time series association local semantic representation vectors, so that the model can independently analyze and process the details of each local time series feature of the real-time temperature and stirring speed in the time dimension, so that the fine-grained time series correlation characteristics between temperature and stirring speed in different time periods can be analyzed more carefully. This decoupling can enhance the model's ability to capture feature dynamics and help it better understand and express the intrinsic properties of the data. Next, the local semantic association interaction vectors between each corresponding set of real-time temperature time series associated local semantic representation vectors and real-time stirring speed time series associated local semantic representation vectors in the sequence of the real-time temperature time series associated local semantic representation vectors and the sequence of the real-time stirring speed time series associated local semantic representation vectors are calculated to obtain a sequence of real-time temperature-stirring speed local time series semantic association interaction vectors; wherein the real-time temperature-stirring speed local time series semantic association interaction vectors reveals the potential correlation between the real-time temperature local time series feature semantics and the stirring speed local time series feature semantics, providing a quantitative perspective for understanding how the two features interact at the local level, which is crucial for capturing the local pattern and structure of the interactive information between real-time temperature and stirring speed. Then, the fine-grained attention weights of each real-time temperature-stirring speed local time series semantic association interaction vector in the sequence of the real-time temperature-stirring speed local time series semantic association interaction vector are calculated to obtain a sequence of real-time temperature-stirring speed local time series fine-grained attention weights; that is, the fine-grained attention weights are calculated for each real-time temperature-stirring speed local time series semantic association interaction vector, and the fine-grained attention weights highlight the key elements in this interaction vector, namely, the real-time temperature and stirring speed local time series semantic association features, so that the model can focus on the most important feature interaction information. It should be understood that by giving more important elements a greater weight, the fine-grained attention weights enable the model to pay more attention to the feature parts that are most critical to the subsequent stirring speed control task, thereby improving the quality of feature representation and the prediction accuracy of the model.Further, the sequence of the real-time temperature-stirring speed local time series fine-grained attention weights is input into the random deactivation module to obtain the sequence of the real-time temperature-stirring speed local time series fine-grained attention weights after the deactivation process; the random deactivation module is used to achieve the sparseness of the weights and enhance the generalization ability of the model. The random deactivation process reduces the risk of overfitting of the model by randomly "closing" a part of the weights, while retaining the most important local time series feature interaction information between the real-time temperature and the stirring speed. This enables the model to better generalize and adapt to different production environments and conditions when facing new data. Subsequently, the sequence of the real-time temperature-stirring speed local time series fine-grained attention weights after the deactivation process is used as the weight distribution, and the weighted sum of the sequence of the real-time temperature-stirring speed local time series semantic association interaction vectors is calculated to obtain the real-time temperature-stirring speed time series interaction representation vector. Here, the sequence of the real-time temperature-stirring speed local time series fine-grained attention weights after the deactivation process is used as the weight distribution, and the sequence of the real-time temperature-stirring speed local time series semantic association interaction vectors is weighted summed. The weighted sum operation combines local interaction information and attention weights to generate a real-time temperature-stirring speed timing interaction representation vector. This global timing interaction representation integrates the interaction information and attention weights between each real-time temperature-stirring speed local timing semantic feature, enabling the model to understand the relationship between features more comprehensively, thereby improving the accuracy and reliability of subsequent predictions of the stirring speed control timing.

[0039] The process of calculating the local semantic association interaction vector between each group of corresponding real-time temperature time series associated local semantic representation vectors and real-time stirring speed time series associated local semantic representation vectors in the sequence of the real-time temperature time series associated local semantic representation vectors and the sequence of the real-time stirring speed time series associated local semantic representation vectors to obtain the sequence of real-time temperature-stirring speed local time series semantic association interaction vectors includes: performing position difference, position dot multiplication and position addition processing on each group of corresponding real-time temperature time series associated local semantic representation vectors and real-time stirring speed time series associated local semantic representation vectors to obtain the real-time temperature-stirring speed local time series semantic difference. Representation vector, real-time temperature-stirring speed local temporal semantic dot product representation vector and real-time temperature-stirring speed local temporal semantic sum representation vector; cascade the real-time temperature-stirring speed local temporal semantic differential representation vector, the real-time temperature-stirring speed local temporal semantic dot product representation vector and the real-time temperature-stirring speed local temporal semantic sum representation vector to obtain a real-time temperature-stirring speed local temporal semantic associated multidimensional representation fusion vector; perform one-dimensional convolution encoding on the real-time temperature-stirring speed local temporal semantic associated multidimensional representation fusion vector and then pass it through a maximum pooling layer to obtain a real-time temperature-stirring speed local temporal semantic associated interaction vector.

[0040] More specifically, the process of calculating the fine-grained attention weights of each real-time temperature-stirring speed local temporal sequence semantic association interaction vector in the sequence of the real-time temperature-stirring speed local temporal sequence semantic association interaction vector to obtain a sequence of real-time temperature-stirring speed local temporal sequence fine-grained attention weights includes: matrix multiplying the real-time temperature-stirring speed local temporal sequence semantic association interaction vector and the weight matrix, and then adding them positionally with the bias vector to obtain a linearly transformed real-time temperature-stirring speed local temporal sequence semantic association interaction vector; using the tanh function to activate the linearly transformed real-time temperature-stirring speed local temporal sequence semantic association interaction vector, and then vector multiplying it with the weight coefficient vector to obtain the real-time temperature-stirring speed local temporal sequence fine-grained attention weight.

[0041] In summary, in the above embodiment, the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector, the real-time humidity time series association feature vector and the real-time stirring speed time series association feature vector, and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector are respectively input into the fine-grained local decoupled sequence feature response network to obtain the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector, including: inputting the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector into the fine-grained local decoupled sequence feature response network to process them with the following fine-grained interaction response aggregation formula to obtain the real-time temperature-stirring speed time series interaction representation vector; wherein the fine-grained interaction response aggregation formula is:

[0042] DecoupLe(f1)={f 11 , f 12 , ..., f 1i , ..., f 1n}

[0043] Decouple(f2)={f 21 , f 22 , ..., f 2i , ..., f 2n}

[0044]

[0045] E i =v T tanh(W r f int (f 1i , f 2i )+b r )

[0046] Wi = dropout(E i )

[0047] V int =∑ i f int (f 1I , f 2i )·W i

[0048] Wherein, f1 is the real-time temperature time series associated feature vector, f2 is the real-time stirring speed time series associated feature vector, Decouple(·) is the feature decoupling operation, and f 11 , f 12 , ..., f 1i , ..., f 1n are the 1st, 2nd, ..., ith, ..., nth real-time temperature time series associated local semantic representation vectors in the sequence of real-time temperature time series associated local semantic representation vectors, f 21 , f 22 , ..., f 2i , ..., f 2n are the 1st, 2nd, ..., ith, ..., nth real-time stirring speed time series associated local semantic representation vectors in the sequence of real-time stirring speed time series associated local semantic representation vectors, [·, ·, ·] is a cascade operation, conv1D is a one-dimensional convolution operation, is positional subtraction, ⊙ is positional multiplication, is added by position, MaxPool(·) is the maximum value pooling process, and f int (f 1i , f 2i ) is the real-time temperature-stirring speed local temporal semantic association interaction vector, W r and b r are the weight matrix and bias vector respectively, tanh(·) is the tanh function, v T is the weight coefficient vector, E i is the real-time temperature-stirring speed local temporal fine-grained attention weight, dropout(·) is the inactivation process, W i is the local time series fine-grained attention weight of real-time temperature-stirring speed after inactivation, V int is a vector representing the real-time temperature-stirring speed time series interaction.

[0049] Specifically, the stirring control multimodal temporal semantic interaction reasoning module 335 is used to perform stirring control posterior reasoning based on the real-time temperature-stirring speed temporal interaction representation vector, the real-time humidity-stirring speed temporal interaction representation vector and the real-time viscosity-stirring speed temporal interaction representation vector to obtain the stirring control multimodal temporal interaction reasoning representation vector. In a specific example of the present application, the real-time temperature-stirring speed temporal interaction representation vector, the real-time humidity-stirring speed temporal interaction representation vector and the real-time viscosity-stirring speed temporal interaction representation vector are input into the stirring control posterior reasoner based on the Bayesian probability network to obtain the stirring control multimodal temporal interaction reasoning representation vector. Because the real-time temperature-stirring speed temporal interaction representation vector, the real-time humidity-stirring speed temporal interaction representation vector and the real-time viscosity-stirring speed temporal interaction representation vector respectively contain the temporal semantic fine-grained interactive association relationship and feature representation information between the real-time temperature, real-time humidity and real-time viscosity and the stirring speed. Therefore, in order to effectively integrate these multi-modal data timing response interaction information, so as to comprehensively consider the influence of multiple stirring variables on the stirring process, so as to obtain a more comprehensive understanding of the stirring process, and provide a basis for the subsequent predictive control of stirring speed, in the technical solution of the present application, the real-time temperature-stirring speed timing interaction representation vector, the real-time humidity-stirring speed timing interaction representation vector and the real-time viscosity-stirring speed timing interaction representation vector are further input into the stirring control posterior inference device based on the Bayesian probability network to obtain the stirring control multi-modal timing interaction reasoning representation vector. Through the processing of the stirring control posterior inference device based on the Bayesian probability network, it is possible to learn how temperature, humidity and viscosity affect the stirring speed, that is, to depict the interaction and influence between the temperature-stirring speed timing response interaction characteristics, the humidity-stirring speed timing response interaction characteristics and the viscosity-stirring speed timing response interaction characteristics, and use this as prior knowledge to infer the timing changes and trends of the stirring speed, so as to provide a basis for the subsequent predictive control of the stirring speed.

[0050] Specifically, the stirring speed control module 336 is used to determine the adjustment instruction based on the stirring control multimodal temporal interactive reasoning representation vector, and the adjustment instruction is used to indicate that the stirring speed value at the next time point should increase, decrease or remain unchanged. In a specific example of the present application, the stirring control multimodal temporal interactive reasoning representation vector is input into a classifier-based refrigerant flow control module to obtain an adjustment instruction, and the adjustment instruction is used to indicate that the stirring speed value at the next time point should increase, decrease or remain unchanged. That is, the multimodal temporal interactive reasoning semantics of stirring control is used to perform classification processing to determine the stirring speed value at the next time point to generate an adjustment instruction. In this way, intelligent control of the stirring process in the production of fiber cement boards can be achieved, which helps to ensure the stability and uniformity of the stirring process.

[0051] In the technical solution of the present application, the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector respectively represent the time series interaction response aggregation characteristics between the real-time temperature and the stirring speed, the time series interaction response aggregation characteristics between the real-time humidity and the stirring speed, and the time series interaction response aggregation characteristics between the real-time viscosity and the stirring speed. In this way, when the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector are input into the stirring control posterior inference device based on the Bayesian probability network, the obtained stirring control multimodal time series interaction reasoning representation vector will cause the distribution regression mode shift due to the difference in the eigenvalue fine-grained modal correspondence between the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector, that is, it causes the overflow of the adjustment instruction obtained by the stirring control multimodal time series interaction reasoning representation vector input into the stirring speed big data intelligent control module based on the classifier, affecting its accuracy.

[0052] Based on this, the stirring control multimodal temporal interactive reasoning representation vector is input into the classifier-based stirring speed big data intelligent control module to obtain the adjustment instructions, which specifically includes:

[0053] Calculate the absolute value sum of each eigenvalue of the stirring control multimodal temporal interactive reasoning representation vector to obtain a first stirring control multimodal temporal interactive reasoning representation and a modulation value, and calculate the square root of the sum of the squares of each eigenvalue of the stirring control multimodal temporal interactive reasoning representation vector to obtain a second stirring control multimodal temporal interactive reasoning representation and a modulation value;

[0054] After performing dot-wise subtraction on the stirring control multimodal temporal interactive reasoning representation vector and the second stirring control multimodal temporal interactive reasoning representation and modulation value, perform dot-wise multiplication on the number of eigenvalues ​​of the stirring control multimodal temporal interactive reasoning representation vector and the reciprocal of the first stirring control multimodal temporal interactive reasoning representation and modulation value, and take the reciprocal of each eigenvalue to obtain the first stirring control multimodal temporal interactive reasoning representation phase conversion vector;

[0055] After performing dot-wise subtraction on the stirring control multimodal temporal interactive reasoning representation vector and the first stirring control multimodal temporal interactive reasoning representation and modulation value, perform dot-wise multiplication on the square root of the number of eigenvalues ​​of the stirring control multimodal temporal interactive reasoning representation vector and the reciprocal of the second stirring control multimodal temporal interactive reasoning representation and modulation value, and take the reciprocal of each eigenvalue to obtain the second stirring control multimodal temporal interactive reasoning representation phase conversion vector;

[0056] Subtract the dot product vector of the second stirring control multimodal temporal interactive reasoning representation phase transition vector from the first stirring control multimodal temporal interactive reasoning representation phase transition vector and the weighted hyperparameter to obtain an optimized stirring control multimodal temporal interactive reasoning representation vector;

[0057] The optimized stirring control multimodal temporal interactive reasoning representation vector is passed through a classifier-based stirring speed big data intelligent control module to obtain adjustment instructions.

[0058] Here, the optimization representation of the stirring control multimodal temporal interactive reasoning representation vector V is:

[0059]

[0060] v i ∈V∈R n

[0061] Wherein, V is the representation vector of the multi-modal temporal interactive reasoning of the stirring control, R represents a real number set, and v i represents the eigenvalue of the ith position of the stirring control multimodal temporal interactive reasoning representation vector, n represents the number of eigenvalues ​​of the stirring control multimodal temporal interactive reasoning representation vector, α represents the first stirring control multimodal temporal interactive reasoning representation and modulation value, β represents the second stirring control multimodal temporal interactive reasoning representation and modulation value, ⊙ represents the positional point multiplication, represents subtraction by position, (·)⊙-1 represents the inverse of each eigenvalue of the calculated eigenvector, V1 represents the first stirring control multimodal temporal interactive reasoning representation phase transition vector, V2 represents the second stirring control multimodal temporal interactive reasoning representation phase transition vector, ω represents a weighted hyperparameter, and V′ represents the optimized stirring control multimodal temporal interactive reasoning representation vector.

[0062] Therefore, in the preferred example, the difference of the feature value of the stirring control multimodal temporal interactive reasoning representation vector relative to the feature set of the vector as a whole of the stirring control multimodal temporal interactive reasoning representation vector and the difference of the modulation representation are used as the semantic change intensity information, and the phase-like conversion corresponding to the position-based intensity modulation is performed through different and modulation representation forms, so that the aggregation enhancement of the semantic change phase perception can improve the axial aggregation receptive field along the feature aggregation direction by performing a spatial translation operation based on alternating stacking under the vector set scale balance of the stirring control multimodal temporal interactive reasoning representation vector, thereby improving the perception effect of the aggregation semantics of the stirring control multimodal temporal interactive reasoning representation vector on the detail semantic change, so as to improve the expression effect of the stirring control multimodal temporal interactive reasoning representation vector, and in this way, avoid the invalid overflow of the classification result caused by the deviation of the distribution mode of the stirring control multimodal temporal interactive reasoning representation vector, thereby improving the accuracy of the adjustment instruction obtained by the intelligent control module of the stirring speed big data based on the classifier. In this way, more accurate stirring speed adaptive control can be achieved according to the real-time stirring state, which helps to ensure the stability and uniformity of the stirring process.

[0063] In particular, the user terminal 340 is used to receive the adjustment instruction and generate a fiber cement board control prompt signal based on the adjustment instruction.

[0064] As described above, the fiber cement board production line information big data system 300 according to the embodiment of the present application can be implemented in various wireless terminals, such as a server with a fiber cement board production line information big data algorithm. In a possible implementation, the fiber cement board production line information big data system 300 according to the embodiment of the present application can be integrated into the wireless terminal as a software module and / or a hardware module. For example, the fiber cement board production line information big data system 300 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the fiber cement board production line information big data system 300 can also be one of the many hardware modules of the wireless terminal.

[0065] Alternatively, in another example, the fiber cement board production line information big data system 300 and the wireless terminal may also be separate devices, and the fiber cement board production line information big data system 300 may be connected to the wireless terminal via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0066] Furthermore, a fiber cement board production line information big data method is also provided.

[0067] Figure 4 FIG. 1 is a flowchart of a fiber cement board production line information big data method according to an embodiment of the present application. Figure 4 As shown, according to the fiber cement board production line information big data method of the embodiment of the present application, the steps include: S1, obtaining the time queue of real-time temperature, the time queue of real-time humidity and the time queue of real-time viscosity collected by the temperature sensor, humidity sensor and viscosity sensor installed in the mixer; S2, obtaining the time queue of real-time stirring speed collected by the stirring speed monitor; S3, respectively extracting features of the time queue of real-time temperature, the time queue of real-time humidity, the time queue of real-time viscosity and the time queue of real-time stirring speed to obtain real-time temperature time series correlation feature vector, real-time humidity time series correlation feature vector, real-time viscosity time series correlation feature vector and real-time stirring speed time series correlation feature vector; S4, extracting the real-time temperature time series correlation feature vector, the real-time stirring speed time series correlation feature vector, the real-time humidity time series correlation feature vector, the real-time viscosity time series correlation feature vector and the real-time stirring speed time series correlation feature vector. The feature vector and the real-time stirring speed time series association feature vector and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector are respectively input into the fine-grained local decoupled sequence feature response network to obtain the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector; S5, based on the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector and the real-time viscosity-stirring speed time series interaction representation vector, a posteriori reasoning of stirring control is performed to obtain a stirring control multimodal time series interaction reasoning representation vector; S6, based on the stirring control multimodal time series interaction reasoning representation vector, the adjustment instruction is determined, and the adjustment instruction is used to indicate that the stirring speed value at the next time point should increase, decrease or remain unchanged.

[0068] In summary, the fiber cement board production line information big data method according to the embodiment of the present application is explained, which uses data processing and analysis algorithms based on artificial intelligence and deep learning to perform time series feature analysis and fine-grained time series feature response association coding for real-time temperature, real-time humidity, real-time viscosity and real-time stirring speed, thereby learning and capturing the time series fine-grained response interaction features between the temperature, humidity and viscosity and the stirring speed during the stirring process, and performing time series reasoning of stirring speed control by calculating the Bayesian posterior probability between the three time series fine-grained response interaction features, thereby determining the stirring speed value at the next time point to generate adjustment instructions. In this way, intelligent control of the stirring process of fiber cement board production can be achieved, which helps to ensure the stability and uniformity of the stirring process.

[0069] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of terms used herein is intended to best explain the principles of the embodiments, practical applications, or improvements to the technology in the market, or to enable other persons of ordinary skill in the art to understand the embodiments disclosed herein.

Claims

1. A fiber cement board production line information big data system, comprising a first information collector, a second information collector, a big data platform and a user terminal, characterized in that: include: The first information collector includes a temperature sensor, a humidity sensor and a viscosity sensor, wherein the temperature sensor, the humidity sensor and the viscosity sensor are respectively installed in the mixer to collect real-time temperature, real-time humidity and real-time viscosity; The second information collector is a stirring speed monitor, which is used to collect real-time stirring speed; The big data platform is used to receive and store the real-time temperature, real-time humidity, real-time viscosity and real-time stirring speed collected by the first information collector and the second information collector, and generate an adjustment instruction based on the real-time temperature, the real-time humidity, the real-time viscosity and the real-time stirring speed; The user end is used to receive the adjustment instruction and generate a fiber cement board production control prompt signal based on the adjustment instruction.

2. The fiber cement board production line information big data system according to claim 1 is characterized in that: The big data platform includes: A stirring state data acquisition module, used to obtain a time queue of real-time temperature, a time queue of real-time humidity and a time queue of real-time viscosity collected by a temperature sensor, a humidity sensor and a viscosity sensor installed in the stirrer; A stirring speed acquisition module, used to obtain a time queue of real-time stirring speed acquired by the stirring speed monitor; A stirring parameter time series feature extraction module is used to extract features from the time queue of the real-time temperature, the time queue of the real-time humidity, the time queue of the real-time viscosity and the time queue of the real-time stirring speed to obtain a real-time temperature time series correlation feature vector, a real-time humidity time series correlation feature vector, a real-time viscosity time series correlation feature vector and a real-time stirring speed time series correlation feature vector; A stirring parameter multimodal time series semantic fine-grained response interaction module, used for inputting the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector, the real-time humidity time series association feature vector and the real-time stirring speed time series association feature vector, and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector into a fine-grained local decoupled sequence feature response network to obtain a real-time temperature-stirring speed time series interaction representation vector, a real-time humidity-stirring speed time series interaction representation vector, and a real-time viscosity-stirring speed time series interaction representation vector; A stirring control multimodal temporal semantic interactive reasoning module, used for performing stirring control posterior reasoning based on the real-time temperature-stirring speed temporal interactive representation vector, the real-time humidity-stirring speed temporal interactive representation vector and the real-time viscosity-stirring speed temporal interactive representation vector to obtain a stirring control multimodal temporal interactive reasoning representation vector; The stirring speed control module is used to determine the adjustment instruction based on the stirring control multimodal temporal interactive reasoning representation vector, and the adjustment instruction is used to indicate whether the stirring speed value at the next time point should increase, decrease or remain unchanged.

3. The fiber cement board production line information big data system according to claim 2 is characterized in that: The stirring parameter timing feature extraction module is used to: input the real-time temperature time queue, the real-time humidity time queue, the real-time viscosity time queue and the real-time stirring speed time queue into a sequence encoder based on the RNN model to obtain the real-time temperature timing association feature vector, the real-time humidity timing association feature vector, the real-time viscosity timing association feature vector and the real-time stirring speed timing association feature vector.

4. The fiber cement board production line information big data system according to claim 3 is characterized in that: The stirring parameter multimodal temporal semantic fine-grained response interaction module includes: A feature decoupling unit, used for performing feature decoupling on the real-time temperature time series associated feature vector and the real-time stirring speed time series associated feature vector to obtain a sequence of real-time temperature time series associated local semantic representation vectors and a sequence of real-time stirring speed time series associated local semantic representation vectors; A local semantic association interaction calculation unit, used for calculating the local semantic association interaction vector between each corresponding group of real-time temperature time series association local semantic representation vectors and real-time stirring speed time series association local semantic representation vectors in the sequence of the real-time temperature time series association local semantic representation vectors and the sequence of the real-time stirring speed time series association local semantic representation vectors to obtain a sequence of real-time temperature-stirring speed local time series semantic association interaction vectors; A fine-grained attention weight calculation unit, used to calculate the fine-grained attention weight of each real-time temperature-stirring speed local temporal semantic association interaction vector in the sequence of real-time temperature-stirring speed local temporal semantic association interaction vectors to obtain a sequence of real-time temperature-stirring speed local temporal fine-grained attention weights; A deactivation unit, used for inputting the sequence of real-time temperature-stirring speed local time series fine-grained attention weights into a random deactivation module to obtain a sequence of real-time temperature-stirring speed local time series fine-grained attention weights after deactivation processing; The real-time temperature-stirring speed time series interaction representation unit is used to calculate the weighted sum of the sequence of the real-time temperature-stirring speed local time series fine-grained attention weights after the inactivation process as the weight distribution to obtain the real-time temperature-stirring speed time series interaction representation vector.

5. The fiber cement board production line information big data system according to claim 4, characterized in that: The local semantic association interaction calculation unit is used to: Perform position difference, position dot multiplication and position addition processing on each group of corresponding real-time temperature time series associated local semantic representation vectors and real-time stirring speed time series associated local semantic representation vectors to obtain real-time temperature-stirring speed local time series semantic difference representation vectors, real-time temperature-stirring speed local time series semantic dot product representation vectors and real-time temperature-stirring speed local time series semantic addition representation vectors; Cascading the real-time temperature-stirring speed local time series semantic difference representation vector, the real-time temperature-stirring speed local time series semantic dot product representation vector and the real-time temperature-stirring speed local time series semantic addition representation vector to obtain a real-time temperature-stirring speed local time series semantic association multi-dimensional representation fusion vector; The real-time temperature-stirring speed local temporal semantic association multidimensional representation fusion vector is subjected to one-dimensional convolution encoding and then passed through a maximum pooling layer to obtain a real-time temperature-stirring speed local temporal semantic association interaction vector.

6. The fiber cement board production line information big data system according to claim 5, characterized in that: The fine-grained attention weight calculation unit is used to: The real-time temperature-stirring speed local time series semantic association interaction vector and the weight matrix are matrix-multiplied and then added to the bias vector by position to obtain a linear transformation real-time temperature-stirring speed local time series semantic association interaction vector; The linear transformation real-time temperature-stirring speed local temporal sequence semantic association interaction vector is activated using a tanh function and then vector-multiplied with a weight coefficient vector to obtain a real-time temperature-stirring speed local temporal sequence fine-grained attention weight.

7. The fiber cement board production line information big data system according to claim 6, characterized in that: The stirring control multimodal temporal semantic interactive reasoning module is used to: input the real-time temperature-stirring speed temporal interaction representation vector, the real-time humidity-stirring speed temporal interaction representation vector and the real-time viscosity-stirring speed temporal interaction representation vector into the stirring control posterior reasoner based on the Bayesian probability network to obtain the stirring control multimodal temporal interactive reasoning representation vector.

8. The fiber cement board production line information big data system according to claim 7, characterized in that: The stirring speed control module is used to: input the stirring control multimodal temporal interactive reasoning representation vector into a classifier-based refrigerant flow control module to obtain an adjustment instruction, wherein the adjustment instruction is used to indicate whether the stirring speed value at the next time point should increase, decrease or remain unchanged.

9. A fiber cement board production line information big data method, characterized in that: include: Acquire a time queue of real-time temperature, a time queue of real-time humidity and a time queue of real-time viscosity collected by a temperature sensor, a humidity sensor and a viscosity sensor installed in the mixer; Obtaining a time queue of real-time stirring speed collected by the stirring speed monitor; Performing feature extraction on the time queue of the real-time temperature, the time queue of the real-time humidity, the time queue of the real-time viscosity and the time queue of the real-time stirring speed respectively to obtain a real-time temperature time series correlation feature vector, a real-time humidity time series correlation feature vector, a real-time viscosity time series correlation feature vector and a real-time stirring speed time series correlation feature vector; Input the real-time temperature time series association feature vector and the real-time stirring speed time series association feature vector, the real-time humidity time series association feature vector and the real-time stirring speed time series association feature vector, and the real-time viscosity time series association feature vector and the real-time stirring speed time series association feature vector into a fine-grained local decoupled sequence feature response network to obtain a real-time temperature-stirring speed time series interaction representation vector, a real-time humidity-stirring speed time series interaction representation vector, and a real-time viscosity-stirring speed time series interaction representation vector; Performing a posteriori reasoning on stirring control based on the real-time temperature-stirring speed time series interaction representation vector, the real-time humidity-stirring speed time series interaction representation vector, and the real-time viscosity-stirring speed time series interaction representation vector to obtain a stirring control multimodal time series interaction reasoning representation vector; Based on the stirring control multimodal temporal interactive reasoning representation vector, the adjustment instruction is determined, and the adjustment instruction is used to indicate that the stirring speed value at the next time point should increase, decrease or remain unchanged.

Images