Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides an intelligent temperature control system and method for optical cable extrusion molding, which are used for solving the problems in the background art.
The invention provides the following technical scheme: an intelligent temperature control system for optical cable extrusion molding comprises a historical data storage module, a target data acquisition module, a theoretical temperature analysis module, an actual temperature analysis module and an intelligent temperature control module;
The target data acquisition module is used for acquiring target data and transmitting the target data to the theoretical temperature analysis module and the actual temperature analysis module; the target data acquisition module comprises a class-one data acquisition unit and a class-two data acquisition unit;
the theoretical analysis module builds a fully-connected neural network model based on the data of the historical data storage module, and analyzes one type of data of the one type of data acquisition unit to obtain a theoretical molding temperature value;
The actual temperature analysis module receives the theoretical forming temperature value of the theoretical temperature analysis module and the second class data of the second class data acquisition unit, calculates and analyzes the heat loss in the extrusion process, and combines the theoretical forming temperature value to obtain a final actual forming temperature value;
the temperature intelligent control module receives data of the actual temperature analysis module and intelligently regulates and controls the temperature;
The system comprises a first-class data acquisition unit, a second-class data acquisition unit, an actual temperature analysis module and a control unit, wherein the first-class data acquisition unit is used for acquiring first-class data and then transmitting the first-class data to the theoretical temperature analysis module, and the second-class data acquisition unit is used for acquiring second-class data and then transmitting the second-class data to the actual temperature analysis module, wherein the first-class data comprises the material melting temperature, the material curing temperature, the optical cable diameter and the production speed of an optical cable outer sheath; the second-class data comprise the surface area of an extrusion head of the extrusion die, the ambient temperature, the convective heat transfer coefficient of air, the heat radiation coefficient and the blackbody radiation coefficient;
The specific processing mode for constructing the fully-connected neural network model is as follows:
Initializing a fully connected neural network model basic architecture: the basic architecture comprises an input layer, a hidden layer, an activation function, a loss function, a learning rate, a regularizer and an output layer;
After initializing weights and biases, obtaining gradients of a loss function on each parameter through a back propagation algorithm, and updating each parameter by using a gradient descent method according to the gradients and the learning rate;
The activating function selects a Sigmoid function;
The loss function selects a mean square error loss function: Wherein y i is the true value of the ith sample, p i is the predicted value of the ith sample, and if the extrusion head works m times when the outer sheath is extruded in the extrusion molding of the optical cable, m samples exist, i=1, 2,3 … … m;
In the training process, the updating of the parameters is towards the gradient descending direction of the loss function, and the updating formula of the parameters is as follows: Wherein w h,r is an updated parameter of the r layer of the full-connection neural network model, w q,r is a current parameter of the r layer of the full-connection neural network model, alpha r is a learning rate of the r layer of the full-connection neural network model, and T Loss r is a gradient of a loss function of the r layer of the full-connection neural network model;
If the loss function is not changed in the continuous beta iteration process, ending the model convergence training to obtain a fully-connected neural network model after the training is completed;
The calculation formula of the heat loss is as follows: Wherein, Q all is the total heat loss value in the extrusion molding process, Q d is the heat dissipated by convection of air around the extrusion head in the extrusion molding process, and Q f is the heat lost by heat radiation of the extrusion head;
Wherein S is the surface area of an extrusion head of an extrusion die, gamma is the air convection heat transfer coefficient, C h is the ambient temperature, and ρ 0 is the theoretical molding temperature value;
Wherein epsilon is a heat radiation coefficient, epsilon is satisfied by [0.25,0.75], and H f is a blackbody radiation coefficient;
The actual forming temperature value is as follows: Wherein T s is the actual forming temperature value, M is the mass of the extrusion head, and c is the specific heat capacity of the material of the extrusion head.
Preferably, the intelligent temperature control module receives the data of the actual temperature analysis module, and the specific mode of intelligent temperature rising for the temperature is as follows:
calculating the time required for the extrusion head to rise from the current temperature to the actual forming temperature value: Wherein T 1 is the time required for the extrusion head to rise from the current temperature to the actual molding temperature value, T d is the current temperature of the extrusion head, and η 1 is the heating efficiency of the extrusion head.
Preferably, the intelligent temperature control module receives the data of the actual temperature analysis module, and the specific mode of intelligently reducing the temperature is as follows:
calculating the time required for the extrusion head to be cooled from the current temperature to the actual forming temperature value: Wherein T 2 is the time required for the extrusion head to cool down from the current temperature to the actual forming temperature value, η 2 is the cooling efficiency of the extrusion head, and T d is the current temperature of the extrusion head.
An intelligent control method for the temperature of optical cable extrusion molding comprises the following steps:
step S01: storing the historical data, and taking the stored historical data as a training set;
Step S02: constructing a fully connected neural network model based on the training set in the step S01;
Step S03: obtaining a type of data, inputting the data into a constructed fully-connected neural network model, and outputting a theoretical molding temperature value;
step S04: obtaining second-class data, calculating the total heat loss value in the extrusion molding process, and combining the theoretical molding temperature value output in the step S03 to obtain an actual molding temperature value;
Step S05: and (3) performing intelligent regulation and control on the temperature based on the actual molding temperature value in the step S04.
The invention has the technical effects and advantages that:
The invention is beneficial to constructing a fully-connected neural network model by arranging the theoretical temperature analysis module and the actual temperature analysis module, is easy to realize, has stronger representation capability, can well represent the relation between one type of data and theoretical extrusion forming temperature values, is added with a regularizer to control the complexity of the fully-connected neural network model and prevent overfitting, can control the temperature according to the requirements of different optical cable outer jackets, improves the forming quality of the outer jackets, and widens the universality; meanwhile, the heat loss existing in the extrusion process is analyzed, the situation that excessive heat loss occurs is prevented, the temperature cannot reach the required value, the forming quality of the outer sheath is improved, the forming effect of the optical cable outer sheath in the extrusion forming process is guaranteed, intelligent and accurate control is carried out on the temperature of the extrusion head, and the accuracy of temperature control of the extrusion head is improved.
Detailed Description
The embodiments of the present invention will be clearly and completely described below with reference to the drawings in the present invention, and the configurations of the structures described in the following embodiments are merely examples, and the system and method for intelligently controlling the temperature of extrusion molding of an optical cable according to the present invention are not limited to the structures described in the following embodiments, and all other embodiments obtained by a person skilled in the art without making any creative effort are within the scope of the present invention.
As shown in fig. 1, the invention provides an intelligent temperature control system for optical cable extrusion molding, which comprises a historical data storage module, a target data acquisition module, a theoretical temperature analysis module, an actual temperature analysis module and an intelligent temperature control module;
the system comprises a historical data storage module, a theoretical temperature analysis module, a target data acquisition module, a temperature intelligent control module, an actual temperature analysis module and an extrusion head, wherein the historical data storage module is used for storing historical data and transmitting the data to the theoretical temperature analysis module as model training data, the target data acquisition module is used for acquiring and transmitting the process parameters of forming the optical cable jacket of a manufacturing batch to the historical data storage module and simultaneously transmitting the data to the theoretical temperature analysis module for storage, and simultaneously transmitting the historical data to the theoretical temperature analysis module for carrying out theoretical temperature analysis on the optical cable jacket, the actual temperature analysis module is used for calculating and analyzing the heat loss of the optical cable jacket and combining the theoretical temperature analysis module to obtain a final actual forming temperature value, so that the temperature cannot reach a set temperature value and further influence the quality of the optical cable jacket.
In this embodiment, it should be specifically described that the historical data storage module is configured to store data, and transmit the stored historical data as a training set to the theoretical temperature analysis module;
The target data acquisition module is used for acquiring target data, comprising one type of data and two types of data, and transmitting the target data to the theoretical temperature analysis module and the actual temperature analysis module; the target data acquisition module comprises a class-one data acquisition unit and a class-two data acquisition unit, wherein the class-one data acquisition unit is used for acquiring class-one data and then transmitting the class-one data to the theoretical temperature analysis module, and the class-two data acquisition unit is used for acquiring class-two data and then transmitting the class-two data to the actual temperature analysis module; such a class of data includes, but is not limited to, the material melting temperature, material curing temperature, cable diameter, and production speed of the cable jacket; the second class of data includes, but is not limited to, extrusion die extrusion head surface area, ambient temperature, air convection heat transfer coefficient, heat emissivity coefficient, and blackbody emissivity coefficient.
In this embodiment, it should be specifically described that the theoretical analysis module constructs a fully connected neural network model based on the data of the historical data storage module, and analyzes one type of data of one type of data acquisition unit to obtain a theoretical forming temperature value, that is, a theoretical temperature value of an extrusion head when the outer sheath is extruded in extrusion forming of the optical cable;
the fully-connected neural network model is easy to realize, has strong representation capability, can well represent the relation between one type of data and theoretical extrusion molding temperature values, can carry out temperature adjustment and control according to the requirements of different optical cable outer jackets, improves the molding quality of the outer jackets, and widens the universality of the system;
The specific processing mode for constructing the fully-connected neural network model is as follows:
Initializing a fully connected neural network model basic architecture: the basic architecture comprises an input layer, a hidden layer, an activation function, a loss function, a learning rate, a regularizer and an output layer; the number of the input layer nodes is the same as the number of the feature vectors of the type of data, and the input layer organizes the type of data into a format which can be processed by the fully-connected neural network model; illustratively, extracting a characteristic vector of a material melting temperature of an outer jacket of the optical cable as a node of the input layer; in the embodiment, 1 output layer node is selected, and the output result is a theoretical molding temperature value; the number of neurons of the hidden layer is determined by a regularizer, and the regularizer adopts L2 norm regularization to control the complexity of the fully-connected neural network model and prevent overfitting;
After initializing weights and biases, obtaining gradients of a loss function on each parameter through a back propagation algorithm, and updating each parameter by using a gradient descent method according to the gradients and the learning rate; the gradient descent method can optimize the value of the parameters, so that the loss function of the model on training data is as small as possible; the parameters are final stable values after algorithm iteration and correction, namely weight and bias; the back propagation algorithm and the gradient descent method are both the prior art, and the embodiment is not repeated in detail;
The activating function selects a Sigmoid function;
The loss function selects a mean square error loss function: Wherein y i is the true value of the ith sample, and p i is the predicted value of the ith sample; if the extrusion head works m times when the outer sheath is extruded in the extrusion molding of the optical cable, m samples exist, i=1, 2 and 3 … … m, the actual temperature generated by the extrusion head is recorded as a true value, and the theoretical molding temperature value output by the output layer is recorded as a predicted value; the true value can be obtained by arranging a non-contact temperature sensor;
In the training process, the updating of the parameters is towards the gradient descending direction of the loss function, and the updating formula of the parameters is as follows: Wherein w h,r is an updated parameter of the r layer of the full-connection neural network model, w q,r is a current parameter of the r layer of the full-connection neural network model, alpha r is a learning rate of the r layer of the full-connection neural network model, and T Loss r is a gradient of a loss function of the r layer of the full-connection neural network model;
If the loss function is not reduced in the continuous beta iteration process, ending the model convergence training to obtain a fully connected neural network model after the training is completed;
And after the theoretical analysis module receives one type of data of the one type of data acquisition unit, outputting a theoretical molding temperature value rho 0 through the constructed fully-connected neural network model.
In this embodiment, it should be specifically described that the actual temperature analysis module receives the theoretical molding temperature value of the theoretical temperature analysis module and the second class data of the second class data acquisition unit, performs calculation and analysis on heat loss in the extrusion process, and combines the theoretical molding temperature value to obtain a final actual molding temperature value;
The heat loss existing in the extrusion process is analyzed, the situation that excessive heat loss occurs is prevented, the temperature cannot reach the required value, the forming quality of the outer sheath is improved, and the forming effect of the optical cable outer sheath in the extrusion forming process is ensured;
The calculation formula of the heat loss is as follows: Wherein, Q all is the total heat loss value in the extrusion molding process, Q d is the heat dissipated by convection of air around the extrusion head in the extrusion molding process, and Q f is the heat lost by heat radiation of the extrusion head;
Wherein S is the surface area of an extrusion head of the extrusion die, γ is the convective heat transfer coefficient of air, and γ=8w/m 2,Ch is taken as the ambient temperature in this embodiment;
Wherein epsilon is a heat radiation coefficient, and epsilon is satisfied with epsilon [0.25,0.75], and the specific numerical value is not specifically limited in the embodiment; h f is the blackbody radiation coefficient, the value of this example is 5.8W/(m 2·k4);
The actual forming temperature value is as follows: Wherein T s is the actual forming temperature value, M is the mass of the extrusion head, and c is the specific heat capacity of the material of the extrusion head.
In this embodiment, it needs to be specifically described that, the temperature intelligent control module receives the data of the actual temperature analysis module, and performs intelligent regulation and control on the temperature, and specific processing modes include, but are not limited to, performing intelligent regulation and control on the temperature by calculating the time required for heating up or the time required for cooling down:
calculating the time required for the extrusion head to rise from the current temperature to the actual forming temperature value, so as to intelligently raise the temperature of the extrusion head; Wherein T 1 is the time required for the extrusion head to rise from the current temperature to the actual forming temperature value, T d is the current temperature of the extrusion head, and the value of T d can be obtained by measurement of a non-contact temperature sensor; η 1 is the heating efficiency of the extrusion head, namely the value of the temperature rise of the extrusion head in a certain time t', which can be obtained through a plurality of actual tests;
Calculating the time required by the extrusion head to cool from the current temperature to the actual forming temperature value, thereby intelligently cooling the extrusion head; Wherein t 2 is the time required for the extrusion head to cool down from the current temperature to the actual forming temperature value, η 2 is the cooling efficiency of the extrusion head, that is, the value of the temperature decrease of the extrusion head in a certain time t', which can be obtained through a plurality of actual tests;
And stopping heating or cooling when the extrusion head reaches the actual forming temperature value.
As shown in fig. 2, the invention provides an intelligent control method for the temperature of optical cable extrusion molding, which comprises the following steps:
step S01: storing the historical data, and taking the stored historical data as a training set;
Step S02: constructing a fully connected neural network model based on the training set in the step S01;
Step S03: obtaining a type of data, inputting the data into a constructed fully-connected neural network model, and outputting a theoretical molding temperature value;
step S04: obtaining second-class data, calculating the total heat loss value in the extrusion molding process, and combining the theoretical molding temperature value output in the step S03 to obtain an actual molding temperature value;
step S05: and (3) performing intelligent regulation and control on the temperature of the extrusion head based on the actual forming temperature value in the step S04.
In this embodiment, it should be specifically explained that the difference between the present embodiment and the prior art is mainly that the present embodiment has a theoretical temperature analysis module and an actual temperature analysis module, and by constructing a fully connected neural network model, the present embodiment is easy to implement, has a relatively strong representation capability, and can well represent the relationship between one type of data and a theoretical extrusion molding temperature value, and adds a regularizer to control the complexity of the fully connected neural network model and prevent overfitting, so that temperature adjustment and control can be performed according to the requirements of different optical cable outer jackets, thereby improving the molding quality of the outer jackets, and widening universality; meanwhile, the heat loss existing in the extrusion process is analyzed, the situation that excessive heat loss occurs is prevented, the temperature cannot reach the required value, the forming quality of the outer sheath is improved, and the forming effect of the optical cable outer sheath in the extrusion forming process is guaranteed.
Finally: the foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
The foregoing is merely illustrative of the present application, and the present application is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.