Automatic control system and method for fiber cement board polishing equipment

Abstract

The present application provides an automatic control system and method for a fiber cement board polishing device, which relates to the field of intelligent production control. The automatic control system and method collect the time series of the vibration signal and pressure data of the polishing head of the polishing device, and use data analysis and processing technology based on artificial intelligence to perform local time series correlation analysis and core principal component time series dynamic aggregation on the polishing head pressure data, and perform waveform semantic extraction on the vibration signal of the polishing head to automatically recommend the polishing head pressure data value at the next time point. In this way, the parameter changes during the polishing process can be monitored in real time, and the pressure of the polishing head can be dynamically adjusted, thereby reducing the workload of the operator and improving the accuracy of control. It can also adapt to the subtle differences of different materials to perform adaptive pressure adjustment, thereby realizing intelligent control of the polishing process.

Description

Technical Field

[0001] The present application relates to the field of intelligent production control, and more particularly, in an embodiment of the present application, to an automatic control system and method for a fiber cement board polishing device. Background Art

[0002] Fiber cement board is a composite material composed of cement, sand, water and fiber (such as glass fiber, polyester fiber, etc.). This board has the characteristics of high strength, good weather resistance and strong fire resistance, and is widely used in the construction industry. During the production process, fiber cement board needs to be polished to achieve the requirements of smooth surface, improve aesthetics and service life. By controlling the polishing process, it is possible to ensure the uniformity of polishing, improve the flatness and glossiness of the board, reduce surface cracks, and ensure the long-term durability and visual effect of the board in exterior wall decoration.

[0003] However, the traditional method mainly relies on preset programs or basic sensor technology when controlling the polishing process, which has some obvious limitations. Specifically, this control method cannot monitor the temporal dynamic changes of parameters during the polishing process in real time, so that the pressure or speed of the polishing head cannot be adjusted according to real-time data, resulting in insufficient control accuracy. In addition, the fixedness of the preset program cannot adapt to the subtle differences between different batches of materials, further limiting the optimization of the polishing effect. In addition, relying on preset programs requires manual regular or irregular setting and adjustment of parameters, which is not only time-consuming and labor-intensive, but also prone to inconsistent parameter settings due to differences in operator experience, which in turn affects the stability of product quality and reduces production efficiency.

[0004] Therefore, an automatic control solution for a fiber cement board polishing device is desired. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed.

[0006] According to one aspect of the present application, there is provided an automatic control method for a fiber cement board polishing device, comprising:

[0007] The fiber cement board to be processed is placed in a designated processing area of ​​a polishing device by a manipulator, and the fiber cement board to be processed is fixed; the fiber cement board to be processed is polished by the polishing device to obtain a polished fiber cement board; the fiber cement board after polishing is quality inspected to obtain a quality inspection result; the fiber cement board after polishing is screened based on the quality inspection result to obtain a qualified fiber cement board;

[0008] The method of polishing the fiber cement board to be processed by the polishing device to obtain the polished fiber cement board comprises: during the polishing process, the pressure of the polishing head of the polishing device is controlled in real time, including:

[0009] During the polishing process, collecting a vibration signal of a polishing head of the polishing device;

[0010] Acquire a time series of pressure data of a polishing head of the polishing device;

[0011] Performing a dynamic aggregation analysis of core principal component features based on time series features on the time series of the polishing head pressure data to obtain an aggregated representation of the polishing head pressure time series features;

[0012] After extracting the semantic features of the vibration signal waveform of the polishing head to obtain the semantic features of the polishing head vibration waveform, interactively fusion is performed on the semantic features of the polishing head vibration waveform and the aggregated representation of the polishing head pressure timing features to obtain the polishing head pressure timing-vibration waveform semantic fusion features;

[0013] Based on the polishing head pressure time series-vibration waveform semantic fusion features, a recommended decoding value of the polishing head pressure data at the next time point is obtained.

[0014] Based on the automatic control method of the fiber cement board polishing equipment, the time series of the polishing head pressure data is subjected to a dynamic aggregation analysis of the core principal component characteristics based on the time series characteristics to obtain an aggregation representation of the polishing head pressure time series characteristics, including:

[0015] The time series of the polishing head pressure data is passed through a pressure time series encoder based on a 1D-CNN model to obtain a sequence of local time series correlation feature vectors of the polishing head pressure;

[0016] The core principal component time series feature dynamic aggregation is performed on the sequence of the local time series correlation feature vectors of the polishing head pressure to obtain the aggregated representation of the polishing head pressure time series feature.

[0017] Based on the above-mentioned automatic control method for the fiber cement board polishing equipment, the sequence of the local time series correlation feature vectors of the polishing head pressure is dynamically aggregated with the core principal component time series features to obtain the aggregated representation of the polishing head pressure time series features, including:

[0018] Extracting a current polishing head pressure local time series correlation feature vector from the sequence of the polishing head pressure local time series correlation feature vectors, and performing a kernel principal component analysis on the current polishing head pressure local time series correlation feature vector to obtain a current polishing head pressure local time series kernel principal component representation vector;

[0019] defining other polishing head pressure local time series associated feature vectors in the sequence of polishing head pressure local time series associated feature vectors as historical polishing head pressure local time series associated feature vectors to obtain a sequence of historical polishing head pressure local time series associated feature vectors, and performing a kernel principal component analysis on each historical polishing head pressure local time series associated feature vector in the sequence of historical polishing head pressure local time series associated feature vectors to obtain a sequence of kernel principal component representation vectors of historical polishing head pressure local time series;

[0020] Calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector, and time modulating the obtained sequence of polishing head pressure time series propagation attenuation factors to obtain a sequence of polishing head pressure time series span modulation propagation attenuation factors;

[0021] Based on the sequence of the polishing head pressure timing span modulation propagation attenuation factors, the sequence of the current polishing head pressure local timing kernel principal component representation vector and the historical polishing head pressure local timing kernel principal component representation vector are subjected to timing gradient attenuation dynamic propagation aggregation to obtain the polishing head pressure timing feature aggregation representation vector as the polishing head pressure timing feature aggregation representation.

[0022] Based on the above-mentioned automatic control method for the fiber cement board polishing equipment, the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector is calculated, and the obtained sequence of polishing head pressure time series propagation attenuation factor is time-modulated to obtain a sequence of polishing head pressure time series span modulation propagation attenuation factor, including:

[0023] Calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector relative to the current polishing head pressure local time series kernel principal component representation vector to obtain the sequence of the polishing head pressure time series propagation attenuation factors;

[0024] Based on the time span between each historical polishing head pressure local timing kernel principal component representation vector in the sequence of historical polishing head pressure local timing kernel principal component representation vectors and the current polishing head pressure local timing kernel principal component representation vector, each polishing head pressure timing propagation attenuation factor in the sequence of polishing head pressure timing propagation attenuation factors is modulated in the time dimension to obtain the sequence of polishing head pressure timing span modulated propagation attenuation factors.

[0025] Based on the above-mentioned automatic control method for the fiber cement board polishing device, the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector relative to the current polishing head pressure local time series kernel principal component representation vector is calculated to obtain the sequence of the polishing head pressure time series propagation attenuation factors, including:

[0026] Dividing the characteristic value of the corresponding position of the historical polishing head pressure local time series core principal component representation vector and the current polishing head pressure local time series core principal component representation vector to obtain a polishing head pressure time series core principal component characteristic attenuation vector;

[0027] Calculating the base-two logarithmic function value of the absolute value of each eigenvalue of the polishing head pressure time series core principal component characteristic attenuation vector to obtain the polishing head pressure time series core principal component characteristic attenuation logarithmic vector;

[0028] Calculate the position-by-position dot product of the historical polishing head pressure local time series kernel principal component representation vector and the polishing head pressure time series kernel principal component characteristic attenuation logarithm vector, and perform position-by-position multiplication on the obtained dot product vector to obtain the polishing head pressure time series propagation attenuation value;

[0029] An exponential function of the polishing head pressure time series propagation attenuation value with the natural constant e as the base is calculated to obtain the polishing head pressure time series propagation attenuation factor.

[0030] Based on the above-mentioned automatic control method for the fiber cement board polishing device, each polishing head pressure timing propagation attenuation factor in the sequence of the polishing head pressure timing propagation attenuation factor is modulated in the time dimension based on the time span between each historical polishing head pressure local timing kernel principal component representation vector in the sequence of the historical polishing head pressure local timing kernel principal component representation vector and the current polishing head pressure local timing kernel principal component representation vector to obtain the sequence of the polishing head pressure timing span modulation propagation attenuation factor, including:

[0031] Subtract the timestamp of the current polishing head pressure local time series kernel principal component representation vector from the timestamp of the historical polishing head pressure local time series kernel principal component representation vector and round down to obtain a polishing head pressure time span value;

[0032] Calculate an exponential function with the natural constant e as the base and the polishing head pressure time span value as the exponent to obtain the polishing head pressure time span modulation value;

[0033] The polishing head pressure time series propagation attenuation factor corresponding to the historical polishing head pressure local time series kernel principal component representation vector is divided by the polishing head pressure time span modulation value to obtain the polishing head pressure time series span modulation propagation attenuation factor.

[0034] Based on the above-mentioned automatic control method of the fiber cement board polishing equipment, based on the sequence of the polishing head pressure time series span modulation propagation attenuation factor, the sequence of the current polishing head pressure local time series kernel principal component representation vector and the sequence of the historical polishing head pressure local time series kernel principal component representation vector are subjected to time series gradient attenuation dynamic propagation aggregation to obtain the polishing head pressure time series feature aggregation representation vector as the polishing head pressure time series feature aggregation representation, including:

[0035] Inputting the sequence of the polishing head pressure timing span modulation propagation attenuation factors into the gated function screening module to obtain a sequence of the polishing head pressure timing span modulation propagation attenuation weights;

[0036] Based on the sequence of the polishing head pressure time series span modulation propagation attenuation weights, calculating the weighted sum of the sequence of the historical polishing head pressure local time series kernel principal component representation vectors to obtain the historical polishing head pressure kernel principal component significant propagation aggregation representation vector;

[0037] Calculate the positional sum of the historical polishing head pressure core principal component significant propagation aggregate representation vector and the current polishing head pressure local time series core principal component representation vector to obtain the polishing head pressure time series feature aggregate representation vector;

[0038] The sequence of the polishing head pressure timing span modulation propagation attenuation factors is input into the gate function screening module to obtain the sequence of the polishing head pressure timing span modulation propagation attenuation weights, including:

[0039] Comparing each polishing head pressure timing span modulation propagation attenuation factor in the sequence of polishing head pressure timing span modulation propagation attenuation factors with a predetermined threshold value to obtain a sequence of polishing head pressure timing span modulation propagation attenuation weights;

[0040] Among them, in response to the polishing head pressure timing span modulation propagation attenuation factor being greater than the predetermined threshold, the polishing head pressure timing span modulation propagation attenuation factor greater than the predetermined threshold is input into the sigmoid function; in response to the polishing head pressure timing span modulation propagation attenuation factor being less than or equal to the predetermined threshold, the polishing head pressure timing span modulation propagation attenuation factor less than or equal to the predetermined threshold is set to zero.

[0041] Based on the above-mentioned automatic control method of the fiber cement board polishing equipment, after extracting the semantic features of the vibration signal waveform of the polishing head to obtain the semantic features of the polishing head vibration waveform, the semantic features of the polishing head vibration waveform and the aggregated representation of the polishing head pressure timing feature are interactively fused to obtain the polishing head pressure timing-vibration waveform semantic fusion features, including:

[0042] The vibration signal of the polishing head is passed through a vibration signal waveform semantic feature extractor based on a deep separable convolutional neural network model to obtain a polishing head vibration waveform semantic feature vector as the polishing head vibration waveform semantic feature;

[0043] The polishing head pressure timing feature aggregation representation vector and the polishing head vibration waveform semantic feature vector are cascaded to obtain a polishing head pressure timing-vibration waveform semantic fusion feature vector as the polishing head pressure timing-vibration waveform semantic fusion feature.

[0044] Based on the above-mentioned automatic control method of the fiber cement board polishing equipment, based on the semantic fusion features of the polishing head pressure timing-vibration waveform, a recommended decoding value of the polishing head pressure data at the next time point is obtained, including: passing the polishing head pressure timing-vibration waveform semantic fusion feature vector through a polishing head pressure controller based on a decoder to obtain the recommended decoding value of the polishing head pressure data at the next time point.

[0045] According to another aspect of the present application, an automatic control system for a fiber cement board polishing device is provided, comprising:

[0046] A polishing head vibration signal acquisition module is used to collect the vibration signal of the polishing head of the polishing equipment during the polishing process;

[0047] A polishing head pressure data time series acquisition module, used to obtain the time series of polishing head pressure data of the polishing device;

[0048] A core principal component feature dynamic aggregation analysis module is used to perform core principal component feature dynamic aggregation analysis based on time series features on the time series of the polishing head pressure data to obtain a polishing head pressure time series feature aggregation representation;

[0049] A vibration signal waveform semantic feature processing module is used to extract the vibration signal waveform semantic features of the vibration signal of the polishing head to obtain the polishing head vibration waveform semantic features, and then interactively fuse the polishing head vibration waveform semantic features and the polishing head pressure timing feature aggregation representation to obtain the polishing head pressure timing-vibration waveform semantic fusion feature;

[0050] The polishing head pressure data recommendation decoding module is used to obtain the recommended decoding value of the polishing head pressure data at the next time point based on the polishing head pressure time series-vibration waveform semantic fusion feature.

[0051] Compared with the prior art, the present application provides an automatic control system and method for a fiber cement board polishing device, which collects the time series of the vibration signal of the polishing head of the polishing device and the pressure data of the polishing head, and uses artificial intelligence-based data analysis and processing technology to perform local time series correlation analysis and core principal component time series dynamic aggregation on the polishing head pressure data, and extracts waveform semantics from the vibration signal of the polishing head to automatically recommend the polishing head pressure data value at the next time point. In this way, the parameter changes during the polishing process can be monitored in real time, and the pressure of the polishing head can be dynamically adjusted, thereby reducing the workload of the operator and improving the accuracy of control. It can also adapt to the subtle differences of different materials to perform adaptive pressure adjustments, thereby realizing intelligent control of the polishing process. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] 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.

[0053] Figure 1 The present invention is a flowchart of an automatic control method of a fiber cement board polishing device according to an embodiment of the present application.

[0054] Figure 2 The present invention is a flow chart of the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application, in which the fiber cement board to be processed is polished by the polishing equipment to obtain the polished fiber cement board.

[0055] Figure 3 This is a data flow diagram of the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application, in which the fiber cement board to be processed is polished by the polishing equipment to obtain the polished fiber cement board.

[0056] Figure 4 The present invention is a flowchart for performing a dynamic aggregation analysis of the core principal component characteristics of the time series of the polishing head pressure data based on the time series characteristics in the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application to obtain an aggregated representation of the polishing head pressure time series characteristics.

[0057] Figure 5 The present invention is a flowchart of dynamically aggregating the core principal component timing characteristics of the sequence of local timing correlation feature vectors of the polishing head pressure in the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application to obtain the aggregated representation of the polishing head pressure timing characteristics.

[0058] Figure 6 A flowchart for calculating the timing propagation attenuation factor of each historical polishing head pressure local timing kernel principal component representation vector in the sequence of historical polishing head pressure local timing kernel principal component representation vectors in the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application, and time-modulating the obtained sequence of polishing head pressure timing propagation attenuation factors to obtain a sequence of polishing head pressure timing span modulation propagation attenuation factors.

[0059] Figure 7 It is a system block diagram of an automatic control system of a fiber cement board polishing device according to an embodiment of the present application. DETAILED DESCRIPTION

[0060] 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.

[0061] 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.

[0062] 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.

[0063] 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.

[0064] Fiber cement board is a building material made of cement, cellulose fiber and other additives, with high strength, fire resistance, waterproof, corrosion resistance and other characteristics. In the traditional control method of fiber cement board polishing process, it usually relies on preset programs and basic sensor technology, which has problems such as inability to monitor and adjust parameters in real time, poor adaptability, and reliance on manual operation, which affects the polishing effect and production efficiency.

[0065] In view of the above technical problems, the present application proposes an automatic control method for a fiber cement board polishing device. Figure 1FIG. 1 is a flow chart of an automatic control method of a fiber cement board polishing device according to an embodiment of the present application. Figure 1 As shown, according to the automatic control method of the fiber cement board polishing equipment of the embodiment of the present application, the method includes: step S110, placing the fiber cement board to be processed in the designated processing area of ​​the polishing equipment by a robot, and fixing the fiber cement board to be processed; step S120, polishing the fiber cement board to be processed by the polishing equipment to obtain a polished fiber cement board; step S130, performing quality inspection on the polished fiber cement board to obtain a quality inspection result; step S140, screening the polished fiber cement board based on the quality inspection result to obtain a qualified fiber cement board.

[0066] In the automatic control method of the fiber cement board polishing equipment, the step S110 is to place the fiber cement board to be processed in the designated processing area of ​​the polishing equipment by the robot arm, and fix the fiber cement board to be processed. It should be understood that the robot arm has high-precision positioning capabilities and can accurately move the fiber cement board to a preset position to ensure that each board can be processed under the same standard. In this way, the need for manual intervention can be reduced, the possibility of human error can be reduced, and the work efficiency of the production line can be improved.

[0067] In the automatic control method of the fiber cement board polishing equipment, the step S120 is to polish the fiber cement board to be processed by the polishing equipment to obtain the polished fiber cement board. It should be understood that fiber cement board is widely used in the fields of construction and decoration due to its excellent weather resistance and strength, wherein the polishing process can make the surface of the fiber cement board smooth and flat, and improve the aesthetics and durability. Specifically, the polishing process generally includes multiple stages, the first is rough polishing, using a coarse polishing material to remove larger surface flaws and unevenness, and then enters the medium polishing stage to further smooth the surface and reduce the roughness. Finally, fine polishing is performed, and a high gloss effect is achieved through a delicate polishing material. In this way, the fiber cement board to be processed will be effectively transformed into a smooth and polished fiber cement board with good visual effects, significantly improving its market competitiveness and application value.

[0068] Wherein, the step S120, polishing the fiber cement board to be processed by the polishing device to obtain the polished fiber cement board, includes: controlling the pressure of the polishing head of the polishing device in real time during the polishing process.

[0069] Accordingly, in the real-time control of the polishing head pressure of the polishing equipment during the polishing process, the technical concept of the present application is to collect the time series of the vibration signal of the polishing head of the polishing equipment and the polishing head pressure data, and use data analysis and processing technology based on artificial intelligence to perform local time series correlation analysis and core principal component time series dynamic aggregation on the polishing head pressure data, and perform waveform semantic extraction on the vibration signal of the polishing head, so as to automatically recommend the polishing head pressure data value at the next time point based on the aggregation representation of the polishing head pressure time series features and the fusion representation between the semantic features of the polishing head vibration waveform. In this way, the parameter changes in the polishing process can be monitored in real time to dynamically adjust the pressure of the polishing head, thereby reducing the workload of the operator, ensuring that the polishing process is always in the best state, and improving the accuracy of control. And it can adapt to the subtle differences of different materials to perform adaptive pressure adjustment, so as to achieve intelligent control of the polishing process.

[0070] Figure 2 The present invention is a flow chart of the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application, in which the fiber cement board to be processed is polished by the polishing equipment to obtain the polished fiber cement board. Figure 3 The following is a data flow diagram of the automatic control method of the fiber cement board polishing device according to the embodiment of the present application, in which the fiber cement board to be processed is polished by the polishing device to obtain the polished fiber cement board. Figure 2 and Figure 3 As shown, the step S120, polishing the fiber cement board to be processed by the polishing equipment to obtain the polished fiber cement board, includes: step S121, collecting the vibration signal of the polishing head of the polishing equipment during the polishing process; step S122, obtaining the time series of the polishing head pressure data of the polishing equipment; step S123, performing a dynamic aggregation analysis of the core principal component features based on the time series features on the time series of the polishing head pressure data to obtain the aggregated representation of the polishing head pressure time series features; step S124, extracting the semantic features of the vibration signal waveform of the vibration signal of the polishing head to obtain the semantic features of the polishing head vibration waveform, and then interactively fusing the semantic features of the polishing head vibration waveform and the aggregated representation of the polishing head pressure time series features to obtain the polishing head pressure time series-vibration waveform semantic fusion features; step S125, based on the polishing head pressure time series-vibration waveform semantic fusion features, obtaining the recommended decoding value of the polishing head pressure data at the next time point.

[0071] Figure 4The present invention is a flowchart of performing a dynamic aggregation analysis of the core principal component characteristics of the polishing head pressure data based on the time series characteristics in the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application to obtain the aggregation representation of the polishing head pressure time series characteristics. Figure 4 As shown, in an embodiment of the present application, the step S123, performing a dynamic aggregation analysis of the core principal component features based on the timing characteristics on the time series of the polishing head pressure data to obtain an aggregated representation of the polishing head pressure timing characteristics, includes: S1231, passing the time series of the polishing head pressure data through a pressure timing encoder based on a 1D-CNN model to obtain a sequence of local timing-related feature vectors of the polishing head pressure; S1232, performing a dynamic aggregation of the core principal component timing characteristics on the sequence of local timing-related feature vectors of the polishing head pressure to obtain an aggregated representation of the polishing head pressure timing characteristics.

[0072] Specifically, in step S1231, considering that the time series of the polishing head pressure data has temporal correlation features within the local time interval, the 1D-CNN model is a deep learning architecture suitable for processing one-dimensional data (such as time series), which can automatically extract local features and capture patterns and trends in the data through convolution operations. Based on this, in the technical solution of the present application, the time series of the polishing head pressure data is passed through a pressure time series encoder based on the 1D-CNN model to capture and refine the temporal features within each local time segment, and obtain a sequence of local temporal correlation feature vectors of the polishing head pressure. Specifically, when processing the polishing head pressure data, the 1D-CNN (one-dimensional convolutional neural network) first inputs the time series data into the 1D-CNN model to perform feature extraction using multiple convolutional layers and activation functions. These convolutional layers can identify important patterns and trends in the time series, such as the peaks and valleys of pressure and their rate of change. Then, after convolution processing, the model will generate a series of local feature vectors that reflect the pressure changes in the time series in different time windows. Next, the feature vector is reduced in dimension through the pooling layer to retain the most representative features to reduce the computational complexity and improve the efficiency of subsequent analysis. Finally, a sequence of local temporal correlation feature vectors of polishing head pressure is obtained to provide support for further analysis and decision-making.

[0073] In particular, in the step S1232, it is taken into account that the key information in different time periods in the sequence of the local timing-related feature vector of the polishing head pressure has different contributions in the entire time domain, that is, the influence of past information on current information will gradually weaken with the passage of time. In order to better capture and represent the key features in the polishing head pressure data, so as to better understand and analyze the changes of data over time, thereby achieving more precise control of the polishing process, in the technical solution of the present application, the sequence of the local timing-related feature vector of the polishing head pressure is dynamically aggregated with the core principal component timing features to obtain the polishing head pressure timing feature aggregation representation vector as the polishing head pressure timing feature aggregation representation.

[0074] Figure 5 The present invention is a flowchart of dynamically aggregating the core principal component time series characteristics of the sequence of the local time series correlation characteristic vectors of the polishing head pressure in the automatic control method of the fiber cement board polishing equipment according to the embodiment of the present application to obtain the aggregated representation of the polishing head pressure time series characteristics. Figure 5 As shown, in an embodiment of the present application, the step S1232, dynamically aggregates the core principal component time series features of the sequence of the polishing head pressure local time series associated feature vectors to obtain the aggregated representation of the polishing head pressure time series features, including: S1232-1, extracting the current polishing head pressure local time series associated feature vector from the sequence of the polishing head pressure local time series associated feature vectors, and performing core principal component analysis on the current polishing head pressure local time series associated feature vector to obtain the current polishing head pressure local time series core principal component representation vector; S1232-2, defining other polishing head pressure local time series associated feature vectors in the sequence of the polishing head pressure local time series associated feature vectors as historical polishing head pressure local time series associated feature vectors to obtain a sequence of historical polishing head pressure local time series associated feature vectors, and performing core principal component analysis on the sequence of historical polishing head pressure local time series associated feature vectors. Perform kernel principal component analysis on each historical polishing head pressure local timing associated feature vector to obtain a sequence of historical polishing head pressure local timing kernel principal component representation vectors; S1232-3, calculate the timing propagation attenuation factor of each historical polishing head pressure local timing kernel principal component representation vector in the sequence of historical polishing head pressure local timing kernel principal component representation vectors, and time modulate the obtained sequence of polishing head pressure timing propagation attenuation factors to obtain a sequence of polishing head pressure timing span modulation propagation attenuation factors; S1232-4, based on the sequence of polishing head pressure timing span modulation propagation attenuation factors, perform timing gradient attenuation dynamic propagation aggregation on the current polishing head pressure local timing kernel principal component representation vector and the sequence of historical polishing head pressure local timing kernel principal component representation vectors to obtain a polishing head pressure timing feature aggregation representation vector as the polishing head pressure timing feature aggregation representation.

[0075] Specifically, in the step S1232-1 and the step S1232-2, the sequence of the local time series correlation feature vector of the current polishing head pressure and the sequence of the local time series correlation feature vector of the historical polishing head pressure are extracted from the sequence of the local time series correlation feature vector of the polishing head pressure, and the kernel principal component analysis is performed on them to obtain the sequence of the local time series kernel principal component representation vector of the current polishing head pressure and the local time series kernel principal component representation vector of the historical polishing head pressure. It should be understood that the kernel principal component analysis (KPCA) is a nonlinear feature extraction technology, which extends the traditional principal component analysis (PCA) and can be used for data dimensionality reduction. Among them, by using KPCA to perform nonlinear mapping on the current and historical polishing head pressure feature vectors, the main direction of the data can be found in the high-dimensional space, thereby realizing the dimensionality reduction of nonlinear data. Compared with the traditional PCA, KPCA is more suitable for processing data sets with complex structures. In this way, the main change trends and characteristics in the current and historical polishing head pressure data can be better understood and represented, thereby providing more effective feature representation for the next step of dynamic aggregation and prediction.

[0076] Specifically, in step S1232-3, the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector is calculated, and the obtained sequence of polishing head pressure time series propagation attenuation factor is time modulated to obtain a sequence of polishing head pressure time series span modulation propagation attenuation factor. It should be understood that in order to capture the trend of feature changes over time and the information loss, the randomness and uncertainty in the data are measured to reveal the dynamic behavior of the data in the time series and the potential information attenuation loss. The sequence of polishing head pressure time series propagation attenuation factor is obtained by calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector. In this way, different weights can be assigned to the historical polishing head pressure features of different time periods. That is, the more recent historical data has a greater impact on the current state, while the impact of the more distant historical data gradually weakens, so as to more reasonably reflect the importance of historical data. Then, based on the time span between the current polishing head pressure local timing kernel principal component representation vector and each historical polishing head pressure local timing kernel principal component representation vector, the time dimension of each polishing head pressure timing propagation attenuation factor is modulated so as to more effectively capture and consider the timing correlation between different time span distances, ensuring that even significant changes between nodes that are far apart in time are regarded as part of normal fluctuations, and obtaining a sequence of polishing head pressure timing span modulated propagation attenuation factors.

[0077] Figure 6The present invention is a flowchart for calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector in the automatic control method of the fiber cement board polishing device according to the embodiment of the present application, and performing time modulation on the obtained sequence of polishing head pressure time series propagation attenuation factors to obtain a sequence of polishing head pressure time series span modulation propagation attenuation factors. Figure 6 As shown, in the embodiment of the present application, the step S1232-3, calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector, and performing time modulation on the obtained sequence of the polishing head pressure time series propagation attenuation factor to obtain a sequence of the polishing head pressure time series span modulation propagation attenuation factor, includes: S1232-31, calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector relative to the current polishing head pressure local time series kernel principal component representation vector. The timing propagation attenuation factor of the local timing kernel principal component representation vector of the polishing head pressure is obtained to obtain the sequence of the timing propagation attenuation factors of the polishing head pressure; S1232-32, based on the time span between each historical local timing kernel principal component representation vector of the polishing head pressure in the sequence of the historical local timing kernel principal component representation vector of the polishing head pressure and the current local timing kernel principal component representation vector of the polishing head pressure, each polishing head pressure timing propagation attenuation factor in the sequence of the polishing head pressure timing propagation attenuation factor is modulated in the time dimension to obtain the sequence of the polishing head pressure timing span modulated propagation attenuation factors.

[0078] In an embodiment of the present application, the step S1232-31, calculating the timing propagation attenuation factor of each historical polishing head pressure local timing core principal component representation vector in the sequence of the historical polishing head pressure local timing core principal component representation vector relative to the current polishing head pressure local timing core principal component representation vector to obtain the sequence of the polishing head pressure timing propagation attenuation factors, includes: dividing the characteristic value of the corresponding position of the historical polishing head pressure local timing core principal component representation vector and the current polishing head pressure local timing core principal component representation vector to obtain the polishing head pressure timing core principal component characteristic attenuation factor. vector; calculate the base-two logarithmic function value of the absolute value of each eigenvalue of the polishing head pressure timing core principal component characteristic attenuation vector to obtain the polishing head pressure timing core principal component characteristic attenuation logarithmic vector; calculate the position-by-position dot product of the historical polishing head pressure local timing core principal component representation vector and the polishing head pressure timing core principal component characteristic attenuation logarithmic vector, and perform position-by-position multiplication on the obtained dot product vector to obtain the polishing head pressure timing propagation attenuation value; calculate the exponential function of the polishing head pressure timing propagation attenuation value with the natural constant e as the base to obtain the polishing head pressure timing propagation attenuation factor.

[0079] In an embodiment of the present application, the step S1232-32 performs time dimension modulation on each polishing head pressure timing propagation attenuation factor in the sequence of the polishing head pressure timing propagation attenuation factor based on the time span between each historical polishing head pressure local timing core principal component representation vector in the sequence of the historical polishing head pressure local timing core principal component representation vector and the current polishing head pressure local timing core principal component representation vector to obtain the sequence of the polishing head pressure timing span modulation propagation attenuation factor, including: subtracting the timestamp of the current polishing head pressure local timing core principal component representation vector from the timestamp of the historical polishing head pressure local timing core principal component representation vector and rounding down to obtain the polishing head pressure time span value; calculating an exponential function with the natural constant e as the base and the polishing head pressure time span value as the exponent to obtain the polishing head pressure time span modulation value; dividing the polishing head pressure timing propagation attenuation factor corresponding to the historical polishing head pressure local timing core principal component representation vector by the polishing head pressure time span modulation value to obtain the polishing head pressure timing span modulation propagation attenuation factor.

[0080] Specifically, in step S1232-4, based on the sequence of the polishing head pressure time span modulation propagation attenuation factors, the sequence of the current polishing head pressure local time series core principal component representation vector and the sequence of the historical polishing head pressure local time series core principal component representation vector are subjected to time gradient attenuation dynamic propagation aggregation to obtain the polishing head pressure time series feature aggregation representation vector as the polishing head pressure time series feature aggregation representation. It should be understood that in order to dynamically adjust the weights of each historical data point, so as to more flexibly process information of different time periods, the modulated factors are input into the gate function screening module to obtain the sequence of the polishing head pressure time series span modulation propagation attenuation weights. That is, the gating mechanism can help determine which historical data is more important and critical and which can be ignored. Afterwards, the weighted sum of the sequence of the historical polishing head pressure local time series core principal component representation vector is calculated by using the respective polishing head pressure time series span modulation propagation attenuation weights in the sequence of the respective polishing head pressure time series span modulation propagation attenuation weights as weights to integrate the time series dynamic change trend of the entire historical feature, and obtain the historical polishing head pressure core principal component significant propagation aggregation representation vector. Finally, the position-wise sum of the historical polishing head pressure core principal component significant propagation aggregation representation vector and the current polishing head pressure local timing core principal component representation vector is calculated to obtain the final polishing head pressure timing feature aggregation representation vector.

[0081] In an embodiment of the present application, the step S1232-4, based on the sequence of the polishing head pressure timing span modulation propagation attenuation factor, performs time gradient attenuation dynamic propagation aggregation on the sequence of the current polishing head pressure local timing core principal component representation vector and the sequence of the historical polishing head pressure local timing core principal component representation vector to obtain the polishing head pressure timing feature aggregation representation vector as the polishing head pressure timing feature aggregation representation, including: inputting the sequence of the polishing head pressure timing span modulation propagation attenuation factor into the gated function screening module to obtain the sequence of the polishing head pressure timing span modulation propagation attenuation weights; based on the sequence of the polishing head pressure timing span modulation propagation attenuation weights, calculating the weighted sum of the sequence of the historical polishing head pressure local timing core principal component representation vectors to obtain the historical polishing head pressure core principal component significant propagation aggregation representation vector; calculating the positional sum of the historical polishing head pressure core principal component significant propagation aggregation representation vector and the current polishing head pressure local timing core principal component representation vector to obtain the polishing head pressure timing feature aggregation representation vector;

[0082] The step of inputting the sequence of the polishing head pressure timing span modulation propagation attenuation factors into the gated function screening module to obtain the sequence of the polishing head pressure timing span modulation propagation attenuation weights comprises: comparing each polishing head pressure timing span modulation propagation attenuation factor in the sequence of the polishing head pressure timing span modulation propagation attenuation factors with a predetermined threshold value to obtain the sequence of the polishing head pressure timing span modulation propagation attenuation weights;

[0083] Among them, in response to the polishing head pressure timing span modulation propagation attenuation factor being greater than the predetermined threshold, the polishing head pressure timing span modulation propagation attenuation factor greater than the predetermined threshold is input into the sigmoid function; in response to the polishing head pressure timing span modulation propagation attenuation factor being less than or equal to the predetermined threshold, the polishing head pressure timing span modulation propagation attenuation factor less than or equal to the predetermined threshold is set to zero.

[0084] Specifically, the sequence of the local time series correlation feature vectors of the polishing head pressure is dynamically aggregated with the core principal component time series features to obtain the polishing head pressure time series feature aggregation representation vector, which can be expressed by the formula: ; ; ; ; ; ; ;in, is the sequence of local temporal correlation feature vectors of the polishing head pressure, is the first, second, ..., first in the sequence of the local temporal correlation characteristic vector of the polishing head pressure ,..., ,..., The local time series correlation feature vector of polishing head pressure, Yes Perform kernel principal component analysis. for The corresponding polishing head pressure local time series kernel principal component representation vector, To calculate the norm of a vector, is the standard deviation of the local temporal correlation eigenvector of the polishing head pressure, represents an exponential function with the natural constant e as the base, It is calculated and The similarity between yes The corresponding polishing head pressure centralization matrix, is the identity matrix, is the scaling parameter, is a column vector of all ones, yes The transposed vector of yes The corresponding polishing head pressure core matrix composed of multiple similarities is: yes The corresponding polishing head pressure centralization core matrix, is the first in the sequence of the principal component vector of the polishing head pressure core The eigenvalues ​​of the principal component vectors of the polishing head pressure core, for The corresponding polishing head pressure core principal component vector, is the first eigenvalues, To return the maximum value value, It is The largest eigenvalue, For cascade processing, is the polishing head pressure cascade vector composed of the polishing head pressure core principal component vectors, , , , is the first, second, and third principal component vectors of the polishing head pressure core sequence. , No. polishing head pressure kernel principal component vector, yes The corresponding polishing head pressure kernel matrix row vector, is the The transposed vector of is the sequence of the principal component representation vectors of the local time series kernel of the polishing head pressure, are the first, second, ..., first, second, third, fourth, fifth, fifth, sixth, sixth, sixth, seventh ... ,..., The local time series kernel principal component representation vector of the polishing head pressure is for The eigenvalues ​​at each position in , for The eigenvalues ​​at each position in , for The number of eigenvalues ​​in , for and The polishing head pressure timing propagation attenuation factor between and Respectively represent and The timestamp of the principal component representation vector of the local time series kernel of the polishing head pressure, To round down, yes and The polishing head pressure timing span modulation propagation attenuation factor between, For masking, yes function, is a predetermined threshold, is the pressure timing span modulation propagation attenuation weight of each polishing head in the sequence of the pressure timing span modulation propagation attenuation weight of the polishing head, is the number of vectors in the sequence of the principal component representation vectors of the local time series kernel of the polishing head pressure, is the aggregate representation vector of the polishing head pressure timing characteristics.

[0085] In an embodiment of the present application, the step S124, after performing vibration signal waveform semantic feature extraction on the vibration signal of the polishing head to obtain the polishing head vibration waveform semantic feature, the polishing head vibration waveform semantic feature and the polishing head pressure timing feature aggregate representation are interactively fused to obtain the polishing head pressure timing-vibration waveform semantic fusion feature, including: passing the vibration signal of the polishing head through a vibration signal waveform semantic feature extractor based on a deep separable convolutional neural network model to obtain a polishing head vibration waveform semantic feature vector as the polishing head vibration waveform semantic feature; cascading the polishing head pressure timing feature aggregate representation vector and the polishing head vibration waveform semantic feature vector to obtain a polishing head pressure timing-vibration waveform semantic fusion feature vector as the polishing head pressure timing-vibration waveform semantic fusion feature.

[0086] Specifically, the vibration signal of the polishing head is passed through a vibration signal waveform semantic feature extractor based on a deep separable convolutional neural network model to obtain a polishing head vibration waveform semantic feature vector as the polishing head vibration waveform semantic feature. It should be understood that, considering that the vibration signal of the polishing head contains key waveform semantic features of the vibration signal such as frequency components, amplitude changes, etc., and in order to extract meaningful features from the vibration signal, so as to better understand and analyze the pattern in the vibration signal, in the technical solution of the present application, the vibration signal of the polishing head is passed through a vibration signal waveform semantic feature extractor based on a deep separable convolutional neural network model to obtain a polishing head vibration waveform semantic feature vector. That is, the deep separable convolutional neural network model can achieve deep learning of complex vibration signals at a lower computational cost and extract potential semantic features through its efficient feature extraction mechanism. Among them, the deep separable convolutional neural network reduces the computational complexity by decomposing the standard convolution into deep convolution and point-by-point convolution. In this way, the deep separable convolutional neural network can more effectively extract local features and spatial relationships in the signal, thereby improving the efficiency and accuracy of feature extraction. In the specific operation, each waveform of the vibration signal is first passed through the deep convolution layer to generate a series of intermediate feature maps to reflect the multi-dimensional information of the signal. Next, the point-by-point convolution layer combines these feature maps to form a more representative feature vector. The final vibration waveform semantic feature vector can effectively express the semantic information of the vibration signal.

[0087] Specifically, the polishing head pressure time series feature aggregation representation vector and the polishing head vibration waveform semantic feature vector are cascaded to obtain the polishing head pressure time series-vibration waveform semantic fusion feature vector as the polishing head pressure time series-vibration waveform semantic fusion feature. It should be understood that considering that the polishing head pressure time series feature aggregation representation vector reflects the pressure change trend during the polishing process, this is helpful to understand the pressure distribution during the polishing process. The polishing head vibration waveform semantic feature vector reflects the vibration mode during the polishing process, which helps to detect whether there is abnormal vibration or instability during the polishing process. If pressure data or vibration data are relied on alone, misjudgment may occur due to the limitations of the data itself. For example, pressure data may not fully reflect the vibration changes caused by changes in material hardness, and vibration data may not fully reflect the quality problems caused by improper pressure regulation. Therefore, in order to integrate two different types of data features to obtain a more comprehensive and representative feature representation to improve the accuracy of polishing process control, in the technical solution of the present application, the polishing head pressure time series feature aggregation representation vector and the polishing head vibration waveform semantic feature vector are interactively fused to obtain the polishing head pressure time series-vibration waveform semantic fusion feature vector.

[0088] In an embodiment of the present application, the step S125, based on the polishing head pressure timing-vibration waveform semantic fusion feature, obtains the recommended decoding value of the polishing head pressure data at the next time point, including: passing the polishing head pressure timing-vibration waveform semantic fusion feature vector through a polishing head pressure controller based on a decoder to obtain the recommended decoding value of the polishing head pressure data at the next time point. That is, the polishing head pressure timing-vibration waveform semantic fusion feature obtained by the interaction of the polishing head pressure timing feature aggregation representation vector and the polishing head vibration waveform semantic feature vector is decoded and processed to automatically recommend the polishing head pressure data value at the next time point. In this way, the parameter changes in the polishing process can be monitored in real time to dynamically adjust the pressure of the polishing head, thereby reducing the workload of the operator, ensuring that the polishing process is always in the best state, and improving the accuracy of control. And it can adapt to the subtle differences of different materials to perform pressure adaptive adjustment, thereby realizing intelligent control of the polishing process.

[0089] Preferably, the polishing head pressure timing-vibration waveform semantic fusion feature vector is passed through a decoder-based polishing head pressure controller to obtain a recommended decoding value of the polishing head pressure data at the next time point, including: respectively calculating the first feature sum value of all eigenvalues ​​of the polishing head pressure timing-vibration waveform semantic fusion feature vector, the second feature sum value of all eigenvalues ​​of the polishing head pressure timing feature aggregation representation vector, and the third feature sum value of all eigenvalues ​​of the polishing head vibration waveform semantic feature vector; after dividing the first feature sum value by the second feature sum value and the third feature sum value, respectively, calculating the square root of the sum of the squares of the two quotients to obtain the polishing head pressure timing-vibration waveform semantic fusion distribution probability value ; Calculate the product of the first polishing head pressure timing-vibration waveform semantic fusion feature value of the polishing head pressure timing-vibration waveform semantic fusion feature vector and the polishing head pressure timing-vibration waveform semantic fusion distribution probability value to obtain the first polishing head pressure timing-vibration waveform semantic fusion feature distribution probability product; Calculate the product of the second polishing head pressure timing-vibration waveform semantic fusion feature value of the polishing head pressure timing-vibration waveform semantic fusion feature vector and the absolute value of the difference between one and the polishing head pressure timing-vibration waveform semantic fusion distribution probability value to obtain the second polishing head pressure timing-vibration waveform semantic fusion feature inverse distribution probability product; Calculate the first polishing head pressure timing-vibration waveform semantic fusion The first polishing head pressure timing-vibration waveform semantic fusion feature inverse distribution probability quotient is obtained by dividing the combined characteristic value by the quotient of one minus the absolute value of the difference between the polishing head pressure timing-vibration waveform semantic fusion distribution probability value and the polishing head pressure timing-vibration waveform semantic fusion distribution probability value; the second polishing head pressure timing-vibration waveform semantic fusion feature value and the polishing head pressure timing-vibration waveform semantic fusion distribution probability value are calculated to obtain the second polishing head pressure timing-vibration waveform semantic fusion feature distribution probability quotient; the first polishing head pressure timing-vibration waveform semantic fusion feature distribution probability product and the second polishing head pressure timing-vibration waveform semantic fusion feature inverse distribution probability product are added and the first polishing head pressure timing-vibration waveform semantic fusion feature inverse distribution probability product is subtracted. The difference between the distribution probability quotient and the distribution probability quotient of the second polishing head pressure timing-vibration waveform semantic fusion feature is obtained to obtain the polishing head pressure timing-vibration waveform semantic fusion correction feature value; for each first polishing head pressure timing-vibration waveform semantic fusion feature value and second polishing head pressure timing-vibration waveform semantic fusion feature value of the polishing head pressure timing-vibration waveform semantic fusion feature vector, the polishing head pressure timing-vibration waveform semantic fusion correction feature matrix composed of the polishing head pressure timing-vibration waveform semantic fusion correction feature value is matrix multiplied with the polishing head pressure timing-vibration waveform semantic fusion feature vector to obtain the corrected polishing head pressure timing-vibration waveform semantic fusion feature vector;The corrected polishing head pressure time series-vibration waveform semantic fusion feature vector is passed through a polishing head pressure controller based on a decoder to obtain a recommended decoding value of the polishing head pressure data at the next time point. ;

[0090] Among them, the polishing head pressure time series-vibration waveform semantic fusion feature vector The correction process is expressed as: ; ; ;in, The first character represents the semantic fusion feature vector of the polishing head pressure time series and vibration waveform The eigenvalues ​​at the positions, represents the probability value of the semantic fusion distribution of the polishing head pressure time series and vibration waveform, The first character represents the semantic fusion feature vector of the polishing head pressure time series and vibration waveform The eigenvalues ​​at the positions, represents the absolute value function, represents the polishing head pressure timing-vibration waveform semantic fusion correction feature value at each position in the polishing head pressure timing-vibration waveform semantic fusion correction feature matrix, Represents the polishing head pressure timing-vibration waveform semantic fusion correction feature matrix, Represents the polishing head pressure timing-vibration waveform semantic fusion feature vector, Represents the corrected polishing head pressure timing-vibration waveform semantic fusion feature vector;

[0091] Among them, the probability value of polishing head pressure time series-vibration waveform semantic fusion distribution The calculation formula is as follows: Among them, ∑ represents the summation function, represents the polishing head pressure timing-vibration waveform semantic fusion feature vector, ' represents the first character of the semantic fusion feature vector of the polishing head pressure time series and vibration waveform The eigenvalues ​​at the positions, represents the polishing head pressure time series feature aggregation representation vector, The characteristic values ​​of each position in the vector representing the polishing head pressure time series characteristic aggregation are represented, represents the semantic feature vector of the polishing head vibration waveform, represents the feature value of each position in the semantic feature vector of the polishing head vibration waveform, Represents the probability value of the semantic fusion distribution of the polishing head pressure timing-vibration waveform.

[0092] That is, when the polishing head pressure timing feature aggregation representation vector and the polishing head vibration waveform semantic feature vector respectively represent the timing node aggregation timing association features of the polishing head pressure data and the image semantic features of the vibration signal of the polishing head, when they are interactively fused, the polishing head pressure timing-vibration waveform semantic fusion vector will also cause interactive fusion diversity due to the distribution differences of different modal data in their respective time-space dimensions, thereby affecting the decoding mapping iteration consistency and reducing the accuracy of the decoding results.

[0093] Therefore, the present application aims at the region-boundary integral relationship of the high-dimensional feature manifold of the polishing head pressure timing-vibration waveform semantic fusion feature vector, and defines the feature set of the polishing head pressure timing-vibration waveform semantic fusion feature vector as a manifold constraint boundary relative to the spatial distribution probability of the polishing head pressure timing feature aggregation representation vector and the polishing head vibration waveform semantic feature vector, so as to approximate the high-dimensional manifold of the feature set of the polishing head pressure timing-vibration waveform semantic fusion feature vector, which is composed of a single connected region representation of eigenvalue pairs, so as to avoid the mapping ambiguity of the local manifold representation in the high-dimensional feature decoding regression convergence space from the diversified feature representation of the vector of the polishing head pressure timing-vibration waveform semantic fusion, improve the execution iteration consistency of each local feature distribution in the mapping task, improve the feature decoding regression convergence effect, and improve the numerical accuracy of the recommended decoding value of the polishing head pressure data at the next time point obtained by the polishing head pressure controller based on the decoder. In this way, the parameter changes in the polishing process can be monitored in real time to dynamically adjust the pressure of the polishing head, thereby reducing the workload of the operator, ensuring that the polishing process is always in the best state, and improving the accuracy of control. It can also adaptively adjust the pressure to suit the subtle differences in different materials, thereby achieving intelligent control of the polishing process.

[0094] In the automatic control method of the fiber cement board polishing equipment described above, the step S130 performs a quality inspection on the fiber cement board after the polishing process to obtain a quality inspection result. It should be understood that the polishing process will affect the physical and chemical properties of the fiber cement board, such as compressive strength, water resistance, and wear resistance. Therefore, it is necessary to conduct a comprehensive quality inspection on the fiber cement board to promptly discover possible defects in the processing process, such as surface scratches, color difference, uneven thickness, etc., which will affect the performance and appearance of the final product. In addition, the quality inspection can also evaluate the impact of the polishing effect on the overall structure of the board to ensure that the basic properties of the material are not damaged while improving the appearance.

[0095] In the automatic control method of the fiber cement board polishing equipment, the step S140 is to screen the polished fiber cement board based on the quality inspection results to obtain the qualified fiber cement board. It should be understood that quality inspection can provide detailed data and information. By screening and analyzing the inspection results, it can be clear which boards meet the prescribed quality standards and which have defects or are unqualified. Among them, the screening process usually adopts a grading method to divide the quality inspection results into three categories: qualified, marginally qualified and unqualified. Qualified products can be directly put on the market or processed later; marginally qualified products need to further evaluate their application scenarios to ensure that they are used in a suitable environment; unqualified products must be eliminated to avoid circulation in the market and prevent potential risks and losses to customers. In this way, qualified fiber cement boards that meet the standards can be screened out, thereby improving production efficiency and reducing defective product rates.

[0096] In summary, the automatic control method of the fiber cement board polishing equipment based on the embodiment of the present application is explained, which collects the time series of the vibration signal of the polishing head of the polishing equipment and the pressure data of the polishing head, and adopts data analysis and processing technology based on artificial intelligence to perform local time series correlation analysis and core principal component time series dynamic aggregation on the polishing head pressure data, and extracts waveform semantics from the vibration signal of the polishing head to automatically recommend the polishing head pressure data value at the next time point. In this way, the parameter changes during the polishing process can be monitored in real time, and the pressure of the polishing head can be dynamically adjusted, thereby reducing the workload of the operator and improving the accuracy of control. And it can adapt to the subtle differences of different materials to perform adaptive pressure adjustment, so as to achieve intelligent control of the polishing process.

[0097] Figure 7 FIG. 1 is a system block diagram of an automatic control system for a fiber cement board polishing device according to an embodiment of the present application. Figure 7As shown, according to the automatic control system 100 of the fiber cement board polishing equipment of the embodiment of the present application, it includes: a polishing head vibration signal acquisition module 110, which is used to collect the vibration signal of the polishing head of the polishing equipment during the polishing process; a polishing head pressure data time series acquisition module 120, which is used to obtain the time series of the polishing head pressure data of the polishing equipment; a core principal component feature dynamic aggregation analysis module 130, which is used to perform core principal component feature dynamic aggregation analysis based on time series features on the time series of the polishing head pressure data to obtain a polishing head pressure time series feature aggregation representation; a vibration signal waveform semantic feature processing module 140, which is used to extract the vibration signal waveform semantic features of the vibration signal of the polishing head to obtain the polishing head vibration waveform semantic features, and then interactively fuse the polishing head vibration waveform semantic features and the polishing head pressure time series feature aggregation representation to obtain the polishing head pressure time series-vibration waveform semantic fusion features; a polishing head pressure data recommendation decoding module 150, which is used to obtain the polishing head pressure data recommended decoding value at the next time point based on the polishing head pressure time series-vibration waveform semantic fusion features.

[0098] Here, those skilled in the art will appreciate that the specific operations of each step in the automatic control system of the above-mentioned fiber cement board polishing equipment have been described in detail above. Figures 1 to 6 The automatic control method of the fiber cement board polishing equipment has been described in detail, and therefore, its repeated description will be omitted.

[0099] As described above, the automatic control system 100 of the fiber cement board polishing device according to the embodiment of the present application can be implemented in various terminal devices. In one example, the automatic control system 100 of the fiber cement board polishing device can be integrated into the terminal device as a software module and / or a hardware module. For example, the automatic control system 100 of the fiber cement board polishing device can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the automatic control system 100 of the fiber cement board polishing device can also be one of the many hardware modules of the terminal device.

[0100] Alternatively, in another example, the automatic control system 100 of the fiber cement board polishing device and the terminal device may also be separate devices, and the automatic control system 100 of the fiber cement board polishing device may be connected to the terminal device via a wired and / or wireless network and transmit interactive information in accordance with an agreed data format.

[0101] In summary, the automatic control system of the fiber cement board polishing equipment based on the embodiment of the present application is explained, which collects the time series of the vibration signal of the polishing head of the polishing equipment and the pressure data of the polishing head, and adopts data analysis and processing technology based on artificial intelligence to perform local time series correlation analysis and core principal component time series dynamic aggregation on the polishing head pressure data, and extracts waveform semantics from the vibration signal of the polishing head to automatically recommend the polishing head pressure data value at the next time point. In this way, the parameter changes during the polishing process can be monitored in real time, and the pressure of the polishing head can be dynamically adjusted, thereby reducing the workload of the operator and improving the accuracy of control. And it can adapt to the subtle differences of different materials to perform adaptive pressure adjustment, so as to achieve intelligent control of the polishing process.

Claims

1. An automatic control method for a fiber cement board polishing device, characterized in that: include: The fiber cement board to be processed is placed in a designated processing area of ​​a polishing device by a manipulator, and the fiber cement board to be processed is fixed; the fiber cement board to be processed is polished by the polishing device to obtain a polished fiber cement board; Performing quality inspection on the polished fiber cement board to obtain a quality inspection result; screening the polished fiber cement board based on the quality inspection result to obtain a qualified fiber cement board; The method of polishing the fiber cement board to be processed by the polishing device to obtain the polished fiber cement board includes: controlling the pressure of the polishing head of the polishing device in real time during the polishing process, including: During the polishing process, collecting a vibration signal of a polishing head of the polishing device; Acquire a time series of pressure data of a polishing head of the polishing device; Performing a dynamic aggregation analysis of core principal component features based on time series features on the time series of the polishing head pressure data to obtain an aggregated representation of the polishing head pressure time series features; After extracting the semantic features of the vibration signal waveform of the polishing head to obtain the semantic features of the polishing head vibration waveform, interactively fusion is performed on the semantic features of the polishing head vibration waveform and the aggregated representation of the polishing head pressure timing features to obtain the polishing head pressure timing-vibration waveform semantic fusion features; Based on the polishing head pressure time series-vibration waveform semantic fusion feature, a recommended decoding value of the polishing head pressure data at the next time point is obtained; Among them, after extracting the semantic features of the vibration signal waveform of the polishing head to obtain the semantic features of the polishing head vibration waveform, the semantic features of the polishing head vibration waveform and the aggregated representation of the polishing head pressure timing features are interactively fused to obtain the polishing head pressure timing-vibration waveform semantic fusion features, including: The vibration signal of the polishing head is passed through a vibration signal waveform semantic feature extractor based on a deep separable convolutional neural network model to obtain a polishing head vibration waveform semantic feature vector as the polishing head vibration waveform semantic feature; The polishing head pressure timing feature aggregation representation vector and the polishing head vibration waveform semantic feature vector are cascaded to obtain the polishing head pressure timing-vibration waveform semantic fusion feature vector as the polishing head pressure timing-vibration waveform semantic fusion feature.

2. The automatic control method of the fiber cement board polishing equipment according to claim 1, characterized in that: Performing a dynamic aggregation analysis of the core principal component features based on the time series features on the time series data of the polishing head pressure to obtain an aggregated representation of the time series features of the polishing head pressure, including: The time series of the polishing head pressure data is passed through a pressure time series encoder based on a 1D-CNN model to obtain a sequence of local time series correlation feature vectors of the polishing head pressure; The core principal component time series feature dynamic aggregation is performed on the sequence of the local time series correlation feature vectors of the polishing head pressure to obtain the aggregated representation of the polishing head pressure time series feature.

3. The automatic control method of the fiber cement board polishing equipment according to claim 2, characterized in that: Performing dynamic aggregation of core principal component time series features on the sequence of local time series correlation feature vectors of the polishing head pressure to obtain an aggregated representation of the polishing head pressure time series features includes: Extracting a current polishing head pressure local time series correlation feature vector from the sequence of the polishing head pressure local time series correlation feature vectors, and performing a kernel principal component analysis on the current polishing head pressure local time series correlation feature vector to obtain a current polishing head pressure local time series kernel principal component representation vector; defining other polishing head pressure local time series associated feature vectors in the sequence of polishing head pressure local time series associated feature vectors as historical polishing head pressure local time series associated feature vectors to obtain a sequence of historical polishing head pressure local time series associated feature vectors, and performing a kernel principal component analysis on each historical polishing head pressure local time series associated feature vector in the sequence of historical polishing head pressure local time series associated feature vectors to obtain a sequence of kernel principal component representation vectors of historical polishing head pressure local time series; Calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector, and time modulating the obtained sequence of polishing head pressure time series propagation attenuation factors to obtain a sequence of polishing head pressure time series span modulation propagation attenuation factors; Based on the sequence of the polishing head pressure timing span modulation propagation attenuation factors, the sequence of the current polishing head pressure local timing kernel principal component representation vector and the historical polishing head pressure local timing kernel principal component representation vector are subjected to timing gradient attenuation dynamic propagation aggregation to obtain the polishing head pressure timing feature aggregation representation vector as the polishing head pressure timing feature aggregation representation.

4. The automatic control method of the fiber cement board polishing equipment according to claim 3, characterized in that: Calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector, and time modulating the obtained sequence of polishing head pressure time series propagation attenuation factors to obtain a sequence of polishing head pressure time series span modulation propagation attenuation factors, including: Calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector relative to the current polishing head pressure local time series kernel principal component representation vector to obtain the sequence of the polishing head pressure time series propagation attenuation factors; Based on the time span between each historical polishing head pressure local timing kernel principal component representation vector in the sequence of historical polishing head pressure local timing kernel principal component representation vectors and the current polishing head pressure local timing kernel principal component representation vector, each polishing head pressure timing propagation attenuation factor in the sequence of polishing head pressure timing propagation attenuation factors is modulated in the time dimension to obtain the sequence of polishing head pressure timing span modulated propagation attenuation factors.

5. The automatic control method of the fiber cement board polishing equipment according to claim 4, characterized in that: Calculating the time series propagation attenuation factor of each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector relative to the current polishing head pressure local time series kernel principal component representation vector to obtain the sequence of the polishing head pressure time series propagation attenuation factors, including: Dividing the characteristic value of the corresponding position of the historical polishing head pressure local time series core principal component representation vector and the current polishing head pressure local time series core principal component representation vector to obtain a polishing head pressure time series core principal component characteristic attenuation vector; Calculating the base-two logarithmic function value of the absolute value of each eigenvalue of the polishing head pressure time series core principal component characteristic attenuation vector to obtain the polishing head pressure time series core principal component characteristic attenuation logarithmic vector; Calculate the position-by-position dot product of the historical polishing head pressure local time series kernel principal component representation vector and the polishing head pressure time series kernel principal component characteristic attenuation logarithm vector, and perform position-by-position multiplication on the obtained dot product vector to obtain the polishing head pressure time series propagation attenuation value; An exponential function of the polishing head pressure time series propagation attenuation value with the natural constant e as the base is calculated to obtain the polishing head pressure time series propagation attenuation factor.

6. The automatic control method of the fiber cement board polishing equipment according to claim 5, characterized in that: Based on the time span between each historical polishing head pressure local time series kernel principal component representation vector in the sequence of the historical polishing head pressure local time series kernel principal component representation vector and the current polishing head pressure local time series kernel principal component representation vector, each polishing head pressure time series propagation attenuation factor in the sequence of the polishing head pressure time series propagation attenuation factor is modulated in the time dimension to obtain the sequence of the polishing head pressure time series span modulation propagation attenuation factor, including: Subtract the timestamp of the current polishing head pressure local time series kernel principal component representation vector from the timestamp of the historical polishing head pressure local time series kernel principal component representation vector and round down to obtain a polishing head pressure time span value; Calculate an exponential function with the natural constant e as the base and the polishing head pressure time span value as the exponent to obtain the polishing head pressure time span modulation value; The polishing head pressure time series propagation attenuation factor corresponding to the historical polishing head pressure local time series kernel principal component representation vector is divided by the polishing head pressure time span modulation value to obtain the polishing head pressure time series span modulation propagation attenuation factor.

7. The automatic control method of the fiber cement board polishing equipment according to claim 6, characterized in that: Based on the sequence of the polishing head pressure time series span modulation propagation attenuation factors, the sequence of the current polishing head pressure local time series core principal component representation vector and the sequence of the historical polishing head pressure local time series core principal component representation vector are subjected to time series gradient attenuation dynamic propagation aggregation to obtain the polishing head pressure time series feature aggregation representation vector as the polishing head pressure time series feature aggregation representation, including: Inputting the sequence of the polishing head pressure timing span modulation propagation attenuation factors into the gated function screening module to obtain a sequence of the polishing head pressure timing span modulation propagation attenuation weights; Based on the sequence of the polishing head pressure time series span modulation propagation attenuation weights, calculating the weighted sum of the sequence of the historical polishing head pressure local time series kernel principal component representation vectors to obtain the historical polishing head pressure kernel principal component significant propagation aggregation representation vector; Calculate the positional sum of the historical polishing head pressure core principal component significant propagation aggregate representation vector and the current polishing head pressure local time series core principal component representation vector to obtain the polishing head pressure time series feature aggregate representation vector; The sequence of the polishing head pressure timing span modulation propagation attenuation factors is input into the gate function screening module to obtain the sequence of the polishing head pressure timing span modulation propagation attenuation weights, including: Comparing each polishing head pressure timing span modulation propagation attenuation factor in the sequence of polishing head pressure timing span modulation propagation attenuation factors with a predetermined threshold value to obtain a sequence of polishing head pressure timing span modulation propagation attenuation weights; Among them, in response to the polishing head pressure timing span modulation propagation attenuation factor being greater than the predetermined threshold, the polishing head pressure timing span modulation propagation attenuation factor greater than the predetermined threshold is input into the sigmoid function; in response to the polishing head pressure timing span modulation propagation attenuation factor being less than or equal to the predetermined threshold, the polishing head pressure timing span modulation propagation attenuation factor less than or equal to the predetermined threshold is set to zero.

8. The automatic control method of the fiber cement board polishing equipment according to claim 7, characterized in that: Based on the semantic fusion features of the polishing head pressure timing-vibration waveform, a recommended decoding value of the polishing head pressure data at the next time point is obtained, including: passing the polishing head pressure timing-vibration waveform semantic fusion feature vector through a decoder-based polishing head pressure controller to obtain the recommended decoding value of the polishing head pressure data at the next time point.

9. An automatic control system for a fiber cement board polishing device, using the automatic control method for a fiber cement board polishing device according to any one of claims 1 to 8, characterized in that: include: A polishing head vibration signal acquisition module is used to collect the vibration signal of the polishing head of the polishing equipment during the polishing process; A polishing head pressure data time series acquisition module, used to obtain the time series of polishing head pressure data of the polishing device; A core principal component feature dynamic aggregation analysis module is used to perform core principal component feature dynamic aggregation analysis based on time series features on the time series of the polishing head pressure data to obtain a polishing head pressure time series feature aggregation representation; A vibration signal waveform semantic feature processing module is used to extract the vibration signal waveform semantic features of the vibration signal of the polishing head to obtain the polishing head vibration waveform semantic features, and then interactively fuse the polishing head vibration waveform semantic features and the polishing head pressure timing feature aggregation representation to obtain the polishing head pressure timing-vibration waveform semantic fusion feature; The polishing head pressure data recommendation decoding module is used to obtain the recommended decoding value of the polishing head pressure data at the next time point based on the polishing head pressure time series-vibration waveform semantic fusion feature.

Images