| This AI temperature controller is the outcome of my three years graduate studies in research on process control. The AI temperature controller uses the LPC2119 as the core, equips the general-purpose thermal sensor interface circuits with programmable gain, drive circuit for thyristors (including triac) and solid-state relay (SSR), keyboard and LED display circuit,4-20mA current loop circuit and RS485 communications circuit making it can be both the core of a single-loop system controller and the node controller in a large-scale DCS system, each node controller can communicate with the host computer via RS485 bus. Embedded software is the soul of this controller, in particular, is one of the control algorithms, they are implemented by embedded C language. The model of the control objects will always be different from one occasion to another, as well as the requirement for the control accuracy. Therefore, I integrate four control algorithms in the controller, namely ON-OFF control algorithms, conventional PID control algorithm, fuzzy control algorithm and fuzzy PID algorithm to adapt to different requirements and on-site operators' habits. At the same time, I also design the SCADA software which is accomplished by the MATLAB GUI to support the AI temperature controller.In this paper, the design for the hardware, software, control strategies, and several other aspects of this AI temperature controller will be explained in details. In order to test the performance of the controller, I used an oven in the lab as the control object and obtained several groups of system's response curves after doing a lot of experiments by different control algorithms (see Chapter 5) and finally made some suggestions for the occasions to use these control algorithms. The article also contains the theorical analysis and research on the neural network in the application of temperature control and the MATLAB simulation results show that the temperature control of the neural network has great application prospect. |
PDF Full Text Request