Self-learning and cloud computing fusion method and waterproof roll tension control method
Through the PIO-BP neural network model integrating self-learning and cloud computing, the multivariable, coupling and nonlinear problems of waterproof coil tension adjustment are solved, efficient tension control is achieved, and the production efficiency and product quality of waterproof coils are improved. .
Patent Information
- Application Number
- CN202411125988.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-10
- Publication Date
- 2024-11-12
AI Technical Summary
The existing waterproof coil tension control method relies on manual experience and traditional PLC+PID control mode, which is difficult to meet the tension adjustment needs of multivariable, coupled, and nonlinear characteristics, resulting in poor control effect and equipment damage.
The method of self-learning and cloud computing is adopted to process and analyze tension data through the cloud-trained PIO-BP neural network model, and combine multiple statistical methods to achieve real-time prediction and adjustment.
It improves the control accuracy and stability of the tension of waterproof coils, reduces the defective rate, and improves production efficiency.
Smart Images

Figure CN118938798A_ABST
Abstract
Description
[0001] This application is a divisional application with an application date of June 10, 2022, application number 2022106565467, and name “Waterproof membrane tension control system and method based on the integration of self-learning and cloud computing”. Technical Field
[0002] The present invention belongs to the technical field of asphalt waterproofing coiled material forming, and specifically relates to a self-learning and cloud computing fusion method, and also relates to an asphalt waterproofing coiled material tension control method based on the fusion of self-learning and cloud computing. Background Art
[0003] At present, modified asphalt waterproofing membrane is a basic industry for my country's national economy and social development, and an important guarantee for the development of strategic emerging industries. It can also improve living conditions, manage the ecological environment, and is also an important support for the development of a circular economy.
[0004] Conventional modified asphalt waterproofing membrane production equipment is mainly composed of modified asphalt preparation system, molding production line system and production auxiliary system. The molding production line system is the most critical system. In the actual production process, the quality and stability of tension control directly affects the normal production and product quality of the membrane molding production line.
[0005] Specifically, the tension fluctuation of the molding production line system will directly lead to problems such as wrinkling and deviation of the roll body. Therefore, it not only affects the production of the molding line, but also easily damages the equipment during long-term operation. At the same time, the application research of modern advanced technologies such as artificial intelligence, cloud storage and cloud computing technology in the field of intelligent control of asphalt waterproofing roll tension has just started.
[0006] However, most of the existing waterproofing membrane tension control methods are based on manual adjustment based on human experience. Of course, there are also a few that use traditional PLC as the control hardware for waterproofing membrane tension, or use PID algorithm programs in tension control strategies. However, the tension adjustment in the waterproofing membrane production process itself has multi-variable, coupled, nonlinear and other characteristics. Whether the traditional method relies on manual adjustment by workers or on the PLC+PID control mode, it actually relies on the past experience of technicians in the waterproofing membrane production field. Therefore, the control effect is poor. At the same time, the traditional waterproofing membrane tension control module is subject to the limitations of the hardware itself, which makes it difficult to meet the requirements of a large amount of real-time data processing and analysis related to tension control during the production process. Technologies such as the new generation of self-learning and cloud computing have high computing speed and efficiency, large data throughput and other properties. If the two can be integrated and applied to the tension control of waterproofing membranes, it can promote it to achieve a qualitative leap. Summary of the invention
[0007] The purpose of the present invention is to overcome the deficiencies of the prior art and provide a new asphalt waterproofing membrane tension control method based on the fusion of self-learning and cloud computing.
[0008] At the same time, the present invention also relates to a method for integrating self-learning and cloud computing.
[0009] In order to achieve the above object, the technical solution adopted by the present invention is:
[0010] A method for integrating self-learning and cloud computing, wherein the cloud training adopted by the cloud computing module is a self-learning algorithm PIO-BP neural network model integrating a pigeon swarm optimization algorithm and a BP neural network, and the implementation process of the model is as follows:
[0011]
[0012] Model X i , V i are the initial position and velocity of the virtual pigeon during the simulated navigation process, and in the multidimensional search space, their respective positions and velocities are updated in each iteration. The update process is as follows:
[0013]
[0014] Model X gbest After Nc-1 iterations, by comparing the positions of all pigeons, when the number of iterations reaches the required number, the map and compass operators are stopped and the surface operator continues to work. At the same time, N Nc To record the number of half of the pigeons, those pigeons far away from the destination are unfamiliar with the surface, they will no longer be able to distinguish the path, and for For all the respective center positions of the Nc-1 generation, if each pigeon flies a straight-line distance to reach the destination, the following formula will be used:
[0015]
[0016] When the iteration loop reaches the maximum number of iterations, the surface operator stops working and the pigeon's position X i and speed V i As the initial weight and threshold of the BP neural network, and start the training of the BP neural network algorithm model to obtain the control model;
[0017] At the same time, when cloud computing is used to process the parameter data in the cloud, the data is processed using multivariate statistical methods to obtain self-learning algorithm training data that better meets the actual situation on site.
[0018] Another technical solution of the present invention is: a waterproof coiled material tension control method based on the integration of self-learning and cloud computing, which comprises the following steps:
[0019] Step 1: The PLC parameter acquisition module collects the tension data, displacement sensor, frequency input value and feedback value of the frequency converter, and input parameters of the human-machine interface of each tension acquisition point of the asphalt waterproofing membrane production equipment in real time. The field data server is used to store the data of the PLC parameter acquisition module, and the field data server can transmit the data to the data storage service module;
[0020] Step 2: The state values of each parameter collected in step 1 are stored in the on-site data server and transmitted to the cloud data server;
[0021] Step 3: The state parameters obtained in step 2 are processed by self-learning in the cloud to remove noise and data anomalies in the signal, and the corresponding normalization and dimension removal are performed to obtain tension data history information;
[0022] Step 4: Use the learning algorithm PIO-BP neural network model to train the algorithm to obtain the asphalt waterproofing membrane tension control model, and store the above model in the cloud data server. The cloud computing module uses the self-learning algorithm PIO-BP neural network model that integrates the pigeon flock optimization algorithm and the BP neural network, and the model implementation process is as follows:
[0023]
[0024] Model X i , V i are the initial position and velocity of the virtual pigeon during the simulated navigation process, and in the multidimensional search space, their respective positions and velocities are updated in each iteration. The update process is as follows:
[0025]
[0026]
[0027] Model X gbest After Nc-1 iterations, by comparing the positions of all pigeons, when the number of iterations reaches the required number, the map and compass operators are stopped and the surface operator continues to work. At the same time, N Nc To record the number of half of the pigeons, those pigeons far away from the destination are unfamiliar with the surface, they will no longer be able to distinguish the path, and for All respective central positions in the Nc-1 generation;
[0028] At the same time, when the parameter data is processed in the cloud by cloud computing, the data is processed by multivariate statistical methods to obtain self-learning algorithm training data that better meets the actual on-site conditions of asphalt waterproofing membranes;
[0029] Step 5: The asphalt waterproofing membrane tension control model stored in the cloud data server in step 4 is used as the cloud tension intelligent control model, and the on-site tension controller automatically downloads the cloud tension intelligent control model that matches the actual operating parameters of the production line;
[0030] Step 6: The on-site tension control module uses the feedback point data of each tension collection point as the input signal of the cloud-based tension intelligent control model obtained in the tension controller to perform real-time adjustment of the tension at each tension collection point.
[0031] According to a specific implementation and preferred aspect of the present invention, if each pigeon reaches the destination by flying a straight-line distance, there will be the following formula:
[0032]
[0033] When the iteration loop reaches the maximum number of iterations, the surface operator stops working and the pigeon's position X i and speed V i As the initial weights and thresholds of the BP neural network, the training of the BP neural network algorithm model is started to obtain the tension control model of the asphalt waterproofing membrane.
[0034] In some specific implementations, in the cloud computing module of step 4, a 3-layer BP neural network is selected for cloud training, and the structure adopts 8, 7 and 1 nodes respectively. The BP neural network includes an input layer, a hidden layer and an output layer, wherein the input layer nodes include 1# tension collection point, 2# tension collection point, 3# tension collection point, displacement, inverter input value, product mode, ambient temperature and humidity; the hidden layer activation function uses relu.
[0035] According to another specific implementation and preferred aspect of the present invention, in step 1, it is also necessary to collect state parameters of the tension control execution module, and at the same time, when the tension control module is in action, the difference between the theoretical displacement and the actual displacement is also used as a historical data set of the self-learning algorithm. Preferably, in step 1, it is also necessary to use the temperature and humidity information of the on-site production process as a historical data set of the self-learning algorithm.
[0036] According to another specific implementation and preferred aspect of the present invention, the tension controller used in step five can automatically download the cloud tension intelligent control model that matches it in the data storage service module according to the actual operating parameters of the production line, and the tension controller can also interact with the PLC parameter acquisition module; the tension control execution module uses the feedback point data of each tension acquisition point as the input signal of the cloud tension intelligent control model obtained in the tension controller to execute the adjustment of the tension at each tension acquisition point. It is prevented that the cloud computing (PIO-BP neural network model) module cannot be adjusted after the network is disconnected. In this way, it can also be explained that the data storage at the field end is not only used to store the historical data of each state parameter of the tension control actuator, but also to prevent the loss of historical data required by the cloud computing (PIO-BP neural network model) module when the network is disconnected, resulting in data fault problems.
[0037] Preferably, the tension control execution module includes an adjustment power device and a synchronous tension adjustment frame arranged at each tension collection point, wherein the synchronous tension adjustment frame is installed together with the hardware of the production line, and when multiple programs are run on the tension controller, no coupling interference will be generated between the tension collection points.
[0038] In addition, in step five, the tension controller installed at the asphalt waterproofing membrane production line obtains the cloud-based tension intelligent control model obtained in step four via FTP, and at the same time serves as an intelligent prediction model for on-site tension control and is combined with step six to make real-time predictions and adjust the tension at each tension collection point; and / or, the cloud-based tension intelligent control model of step five serves as an intelligent prediction model for on-site tension control to make real-time predictions and adjust the tension at each tension collection point.
[0039] Preferably, in step 4, the model training accuracy value is set to 0.0001, and the historical training data set after mean filtering in step 3 is imported, and the position and speed of the optimal solution for pigeon flock navigation are used as the initial weights and thresholds of the BP neural network.
[0040] Another technical solution of the present invention is:
[0041] Due to the application of the above technical solution, the present invention has the following advantages compared with the prior art:
[0042] The present invention integrates self-learning and cloud computing technology to make real-time prediction and adjustment of the tension control of waterproof membranes based on tension data, displacement sensor, inverter frequency input value and feedback value, and human-machine interface input parameters. At the same time, the algorithm based on self-learning can effectively solve the shortcomings of tension regulation itself, such as multivariable, coupling, and nonlinearity. Cloud computing can better meet the needs of large amounts of real-time data processing and analysis in the production process of asphalt waterproof membranes. Therefore, it effectively improves the control accuracy and stability of the tension of waterproof membranes, greatly reduces the defective rate of waterproof membranes, and improves the production efficiency of waterproof membranes. BRIEF DESCRIPTION OF THE DRAWINGS
[0043] Figure 1 It is a schematic diagram of the structure of the asphalt waterproofing membrane tension control method of the present invention;
[0044] Figure 2 It is a schematic diagram of the architecture of the PIO-BP neural network model of the present invention;
[0045] Figure 3 It is a structural schematic diagram of the asphalt waterproofing coiled material tension control system of the present invention;
[0046] Among them: 1. Cloud service unit; 10. Cloud server; 100. Cloud algorithm service module; 101. Data storage service module; 11. Cloud computing module; A. Asphalt waterproofing membrane tension control model; A', Cloud tension intelligent control model;
[0047] 2. Data acquisition unit; 20. PLC parameter acquisition module; 21. Field data server;
[0048] 3. Tension control execution unit; 30. Tension controller; 31. Tension control execution module; 310. Adjustment power device; 311. Synchronous tension adjustment frame;
[0049] B1, B2, B3, status parameters. DETAILED DESCRIPTION
[0050] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0051] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents. The present invention will be further described in detail below with reference to the embodiments given in the accompanying drawings.
[0052] like Figures 1 to 3 As shown, the asphalt waterproofing membrane tension control system based on the integration of self-learning and cloud computing in this embodiment includes a cloud service unit 1, a data acquisition unit 2, and a tension control execution unit 3.
[0053] Specifically, the cloud service unit 1 includes a cloud server 10 and a cloud computing module 11, wherein the cloud server 10 includes a cloud algorithm service module 100 and a data storage service module 101, and the cloud computing module 11 uses the cloud algorithm service module 100 and the data storage service module 101 to complete cloud training to obtain the asphalt waterproofing membrane tension control model A.
[0054] In this example, the cloud training adopted by the cloud computing module 11 is a self-learning algorithm PIO-BP neural network model that integrates the pigeon swarm optimization algorithm and the BP neural network, and the implementation process of the model is as follows:
[0055]
[0056] Model X i , V i The initial position and velocity of the pigeon in the virtual pigeon simulation navigation process, and in the multidimensional search space, the respective position and velocity will be updated in each iteration. The update process is as follows:
[0057]
[0058] Model X gbest It is the global optimal position obtained by comparing the positions of all pigeons after Nc-1 iterations. When the number of iterations reaches the required number, the map and compass operators stop working and continue working in the surface operator. Nc To record the number of half of the pigeons, those pigeons far away from the destination are not familiar with the surface, they will no longer be able to distinguish the path. For all the respective center positions of the Nc-1 generation, if each pigeon flies a straight-line distance to reach the destination, the following formula will be used:
[0059]
[0060] When the iteration loop reaches the maximum number of iterations, the surface operator stops working.i and speed V i As the initial weight and threshold of the BP neural network, and start the training of the BP neural network algorithm model, obtain the modified asphalt waterproofing membrane tension control model A, and store the algorithm model that meets the training accuracy in the cloud server 10;
[0061] The data acquisition unit 2 includes a PLC parameter acquisition module 20 and a field data server 21, wherein the PLC parameter acquisition module 20 is used to collect tension data, displacement sensors, inverter frequency input values and feedback values, and human-machine interface input parameters (state parameters B1) of each tension collection point of the asphalt waterproofing membrane production equipment in real time.
[0062] At the same time, the PLC parameter acquisition module also requires temperature and humidity information (state parameter B1) during the on-site production process.
[0063] The field data server 21 is used to store the data of the PLC parameter acquisition module 20 , and the field data server 21 can transmit the data to the data storage service module 101 via industrial Ethernet.
[0064] The tension control execution unit 3 includes a tension controller 30 and a tension control execution module 31, wherein the tension controller 30 can automatically download a cloud-based tension intelligent control model A' that matches the actual operating parameters of the production line in the data storage service module 101, and the tension control execution module 31 uses the feedback point data of each tension collection point as the input signal of the cloud-based tension intelligent control model A' obtained in the tension controller 30 to adjust the tension at each tension collection point.
[0065] The tension controller 30 interacts with the PLC parameter acquisition module 20 to prevent the cloud computing (PIO-BP neural network model) module from being disconnected from the network and being unable to make adjustments. This also shows that the data storage at the field end is not only used to store the historical data of each state parameter of the tension control actuator, but also to prevent the historical data required by the cloud computing (PIO-BP neural network model) module from being lost when the network is disconnected, resulting in data gaps.
[0066] The tension control execution module 31 includes an adjustment power device 310 (tension control execution cylinder) and a synchronous tension adjustment frame 311 set at each tension collection point, wherein the synchronous tension adjustment frame 311 is installed together with the hardware of the production line. When multiple programs are run on the tension controller 30, there will be no coupling interference between the tension collection points.
[0067] The PLC parameter acquisition module 20 can also acquire the tension control execution module state parameter B3.
[0068] In summary, the control method of the asphalt waterproofing coiled material tension control system implemented in this embodiment includes the following steps:
[0069] Step 1: The PLC parameter acquisition module collects the tension data, displacement sensor, frequency input value and feedback value of the frequency converter, and input parameters of the human-machine interface at each tension acquisition point of the asphalt waterproofing membrane production equipment; collects the state parameters of the tension control execution module and the temperature and humidity information during the on-site production process;
[0070] Step 2: The state values of each parameter collected in step 1 are stored in the on-site data server and transmitted to the cloud data server;
[0071] Step 3: The state parameters obtained in step 2 are processed by self-learning in the cloud to remove noise and data anomalies in the signal, and the corresponding normalization and dimension removal are performed to obtain tension data history information;
[0072] Step 4: Use the learning algorithm PIO-BP neural network model to train the algorithm, where the self-learning algorithm uses a 3-layer BP neural network with an 8-7-1 structure. The input layer nodes include 1# tension collection point, 2# tension collection point, 3# tension collection point, displacement, inverter input value, product mode, ambient temperature and humidity; the hidden layer activation function uses relu, and the model training accuracy value is set to 0.0001. At the same time, the historical training data set after mean filtering in step 3 is imported, and the optimal position and speed of the pigeon group navigation are used as the initial weight and threshold of the BP neural network. Then, the BP neural network algorithm model training is started to obtain the asphalt waterproofing membrane tension control model, and the algorithm model that meets the training accuracy is stored in the cloud server;
[0073] Step 5: The tension controller installed on the asphalt waterproofing membrane production line uses the asphalt waterproofing membrane tension control model stored in the cloud data server in step 4 as the cloud tension intelligent control model through FTP. The on-site tension controller automatically downloads the cloud tension intelligent control model that matches the actual operating parameters of the production line, and uses it as the intelligent prediction model for on-site tension control.
[0074] Step 6: The on-site tension control module uses the feedback data of each tension collection point as the input signal of the cloud-based tension intelligent control model obtained in the tension controller to make real-time predictions and adjust the tension at each tension collection point.
[0075] In addition, in step 6, the tension control module at the field end does not directly collect field data, but uses PLC to collect the signals of each sensor through network communication for the tension control module to read.
[0076] Therefore, the method for controlling the tension of the asphalt waterproofing membrane of the present embodiment, through the settings of steps one to six, can integrate self-learning and cloud computing technology to control the tension of the modified asphalt waterproofing membrane, make real-time prediction and adjustment, and can also effectively solve the shortcomings of tension regulation itself, such as multivariable, coupled, and nonlinear. At the same time, cloud computing can better meet the needs of large amounts of real-time data processing and analysis in the production process of asphalt waterproofing membranes. Therefore, it effectively improves the control accuracy and stability of the tension of the waterproofing membrane, greatly reduces the defective rate of the waterproofing membrane, and improves the production efficiency of the waterproofing membrane.
[0077] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with the technology to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.
Claims
1. A method for integrating self-learning and cloud computing, characterized in that: The cloud training adopted by its cloud computing module is a self-learning algorithm PIO-BP neural network model that integrates the pigeon swarm optimization algorithm and the BP neural network. The implementation process of the model is as follows: Model X i , V i are the initial position and velocity of the virtual pigeon during the simulated navigation process, and in the multidimensional search space, their respective positions and velocities are updated in each iteration. The update process is as follows: Model X gbest After Nc-1 iterations, by comparing the positions of all pigeons, when the number of iterations reaches the required number, the map and compass operators are stopped and the surface operator continues to work. At the same time, N Nc To record the number of half of the pigeons, those pigeons far away from the destination are unfamiliar with the surface, they will no longer be able to distinguish the path, and for For all the respective center positions of the Nc-1 generation, if each pigeon flies a straight-line distance to reach the destination, the following formula will be used: When the iteration loop reaches the maximum number of iterations, the surface operator stops working and the pigeon's position X i and speed V i As the initial weight and threshold of the BP neural network, and start the training of the BP neural network algorithm model to obtain the control model; At the same time, when cloud computing is used to process the parameter data in the cloud, the data is processed using multivariate statistical methods to obtain self-learning algorithm training data that better meets the actual situation on site.
2. A waterproof coil tension control method based on the integration of self-learning and cloud computing, characterized in that: It includes the following steps: Step 1: The PLC parameter acquisition module collects the tension data, displacement sensor, frequency input value and feedback value of the frequency converter, and input parameters of the human-machine interface of each tension acquisition point of the asphalt waterproofing membrane production equipment in real time. The field data server is used to store the data of the PLC parameter acquisition module, and the field data server can transmit the data to the data storage service module; Step 2: The state values of each parameter collected in step 1 are stored in the on-site data server and transmitted to the cloud data server; Step 3: The state parameters obtained in step 2 are processed by self-learning in the cloud to remove noise and data anomalies in the signal, and the corresponding normalization and dimension removal are performed to obtain tension data history information; Step 4: Use the learning algorithm PIO-BP neural network model to train the algorithm to obtain the asphalt waterproofing membrane tension control model, and store the above model in the cloud data server. The cloud computing module uses the self-learning algorithm PIO-BP neural network model that integrates the pigeon flock optimization algorithm and the BP neural network, and the model implementation process is as follows: Model X i , V i are the initial position and velocity of the virtual pigeon during the simulated navigation process, and in the multidimensional search space, their respective positions and velocities are updated in each iteration. The update process is as follows: Model X gbest After Nc-1 iterations, by comparing the positions of all pigeons, when the number of iterations reaches the required number, the map and compass operators are stopped and the surface operator continues to work. At the same time, N Nc To record the number of half of the pigeons, those pigeons far away from the destination are unfamiliar with the surface, they will no longer be able to distinguish the path, and for All respective central positions in the Nc-1 generation; At the same time, when the parameter data is processed in the cloud by cloud computing, the data is processed by multivariate statistical methods to obtain self-learning algorithm training data that better meets the actual on-site conditions of asphalt waterproofing membranes; Step 5: The asphalt waterproofing membrane tension control model stored in the cloud data server in step 4 is used as the cloud tension intelligent control model, and the on-site tension controller automatically downloads the cloud tension intelligent control model that matches the actual operating parameters of the production line; Step 6: The on-site tension control module uses the feedback point data of each tension collection point as the input signal of the cloud-based tension intelligent control model obtained in the tension controller to perform real-time adjustment of the tension at each tension collection point.
3. The waterproof coiled material tension control method based on the integration of self-learning and cloud computing according to claim 2 is characterized in that: If each pigeon flies a straight-line distance to reach its destination, the formula will be as follows: When the iteration loop reaches the maximum number of iterations, the surface operator stops working and the pigeon's position X i and speed V i As the initial weights and thresholds of the BP neural network, the training of the BP neural network algorithm model is started to obtain the tension control model of the asphalt waterproofing membrane.
4. The waterproof coiled material tension control method based on the integration of self-learning and cloud computing according to claim 2 is characterized in that: In the cloud computing module of step 4, a 3-layer BP neural network is selected for cloud training, and the structure adopts 8, 7 and 1 nodes respectively.
5. The waterproof coiled material tension control method based on the integration of self-learning and cloud computing according to claim 4 is characterized in that: The BP neural network includes an input layer, a hidden layer, and an output layer. The input layer nodes include 1# tension collection point, 2# tension collection point, 3# tension collection point, displacement, inverter input value, product mode, ambient temperature and humidity; the hidden layer activation function uses relu.
6. The waterproof coiled material tension control method based on the integration of self-learning and cloud computing according to claim 2 is characterized in that: In step one, it is also necessary to collect the state parameters of the tension control execution module. At the same time, when the tension control module is in action, the difference between the theoretical displacement and the actual displacement is also used as a historical data set of the self-learning algorithm; and / or, in step one, the temperature and humidity information of the on-site production process is also needed as a historical data set of the self-learning algorithm.
7. The asphalt waterproofing coiled material tension control method based on the integration of self-learning and cloud computing according to claim 2 is characterized by: The tension controller used in step 5 can automatically download the cloud tension intelligent control model that matches the actual operating parameters of the production line in the data storage service module, and the tension controller can also interact with the PLC parameter acquisition module; The tension control execution module uses the feedback point data of each tension collection point as the input signal of the cloud-based tension intelligent control model obtained in the tension controller to adjust the tension at each tension collection point.
8. The asphalt waterproofing coiled material tension control method based on the integration of self-learning and cloud computing according to claim 7 is characterized by: The tension control execution module includes an adjustment power device and a synchronous tension adjustment frame set at each tension collection point, wherein the synchronous tension adjustment frame is installed together with the hardware of the production line. When multiple programs are run on the tension controller, no coupling interference will be generated between the tension collection points.
9. The asphalt waterproofing coiled material tension control method based on the integration of self-learning and cloud computing according to claim 2 is characterized by: In step five, the tension controller installed at the asphalt waterproofing membrane production line obtains the cloud-based tension intelligent control model obtained in step four via FTP, and at the same time serves as an intelligent prediction model for on-site tension control and is combined with step six to make real-time predictions and adjust the tension at each tension collection point; and / or, the cloud-based tension intelligent control model of step five serves as an intelligent prediction model for on-site tension control to make real-time predictions and adjust the tension at each tension collection point.
10. The asphalt waterproofing coiled material tension control method based on the integration of self-learning and cloud computing according to claim 2 is characterized in that: In step 4, the model training accuracy value is set to 0.0001, and the historical training data set after mean filtering in step 3 is imported, and the position and speed of the optimal solution for pigeon flock navigation are used as the initial weights and thresholds of the BP neural network.
