Research On Temperature Control System Of Thermal Analyzer Based On Neural Network

Thermal analysis makes use of the different exothermic effects under heating and cooling states to determine the physical properties of materials,and its main work is the measurement and control of the temperature.The temperature control system of the thermal analyzer and heating furnace is a complex system with the characteristics of nonlinearity,dynamic variability,and lagging.Therefore,it is difficult to establish the an exact mathematical model for the controlled object by using the PID control method.This project leverages the advantages of the neural network—self-learning,self-organization,and approaching any nonlinear function—to solve the above difficulty.Integrating the neutral network with PID control method,this work designs a thermal management system for controlling the temperature of the thermal analyzer and heating furnace,and monitoring the temperature parameters in real time via linking the Android mobile phone with the control system by the Bluetooth technology.The set-points of the neurons in input layer,the neurons in hidden layer,and the neurons in output layer are 3,5,3,respectively.A neural network-based PID control model of 3-5-3 structure is designed to achieve three functions of proportion,integration,and differentiation functions.In the MATLAB environment,the simulation experiment of the thermal analyzer and heating furnace is carried out.The structure of the control model of the thermal analyzer and heating furnace is determined.Meanwhile,the effectiveness and rationality of the neural network-based PID control method is verified.To further validate the feasibility of this method and the control effect of this system in practice,we design a system of thermal analysis and heating furnace temperature control using the CC2650 microcontroller.The hardware consists of the power circuit,the Bluetooth communication circuit,the interface debug circuit,the temperature detection circuit,the temperature control circuit,and the LCD circuit.The software is composed by the temperature collecting program,the pulse width modulation program,the neural network-based PID control program,etc.The Android mobile phone system comprises of the Bluetooth scan interface,the main interface,and the file interface.In particularly,the system test and performance evaluation are carried out by the extensive experiments.A large number of experiments are complemented on the thermal analyzer and heating furnace temperature control system.In addition,the temperature change of the thermal analyzer and heating furnace is dynamically monitored in real time by the Android mobile phone.The experimental results indicate that the system realizes the functions of the temperature collecting,the AD conversion,the parameter calculation of the neural network-based PID,the Bluetooth communication,and the real-time display.Besides,it is proved that the precision of this system meets the requirement of the thermal analyzer and hearting furnace control.Moreover,it is concluded that this system has good reliability and stability in the real environment.