Research On Intelligent Control System Of Microwave Heating Based On Improved BP Neural Network Algorithm

As a new type of heating technology, microwave, helping people heat uniformly and efficiently without pollution and save time, reflects the tremendous advantages in industrial, agricultural, medical fields and so on with its good characteristics. So nowadays microwave has a wide practical prospect with the innovative development of the precise and intelligently control of the microwave output power which will be the key technology.This paper shows an automatic control system of industrial microwave based on improved BP neural network PID conjugate gradient algorithm, and mainly describes the basic principles of the magnetron as the microwave source, the basic characteristics of industrial microwave heating, the hardware circuit of measuring and controlling the temperature and he corresponding intelligent control algorithms. The main work is as follows:1. Firstly, this paper describes the introduction of the heating theory and the basic characteristics of microwave, the structure of this technology system, detailed description of the operating principle and basic structure of magnetron and the research and design for microwave transmission means and the heating chamber conducted.2. Secondly, analyzes the advantages and disadvantages of several types of temperature sensors and especially improves the thermocouple temperature sensor to suit the microwave heating system. This paper plans to complete a design of hardware circuit which use DSP as the core processor, including acquisition, monitoring and control. Temperature acquisition is make up of K-thermocouple and MAX6675integrated chips. Temperature monitoring shows us the real-time monitoring of temperature on the LCD. And the upper computer software through MAX232serial interface to communicate with the temperature monitoring. Temperature control uses PWM (pulse modulation techniques) to adjust the magnetron to achieve the intelligent control of the temperature.3. For the temperature nonlinearity, hysteresis and time-varying characteristics, this paper will adopt an improved conjugate gradient BP neural network PID algorithm, having fast response speed, high control accuracy and anti-jamming features, which will have been theoretically improved that it is feasible in the intelligent system of microwave heating after the simulation experiment of using MATLAB separately for conventional PID algorithm, algorithm neural network and improved BP neural network PID conjugate gradient algorithm. Finally, the practicality and effectiveness based on DSP and improved conjugate gradient BP neural network PID algorithm is tested by heating activated carbon.