Improvement of control algorithms for the system of teafixing roller machine
                                                            YING Wenyuan, ZHAO Zhangfeng, ZHONG Jiang
                         （College of Mechanical Engineering, Zhejiang University of Technology, Hangzhou 310014, China）

Abstract: Aiming at the existence of bad leaves caused by the roller teafixing machine's dependence on humanexperience, and the low control accuracy of traditional PID algorithm, fuzzy algorithm and BP neural network PID selftuning technology were combined for improving control algorithm of the machine, based on the structure and process of the fixing machine, fuzzy algorithm, BP neural network PID control algorithm and STM32 ARM CortexM3 core singlechip microcomputer. In the closedloop control system based on that algorithm, tea leaves’ moisture content and leaftransport speed were taken as inputs, fuzzy algorithm was used to decide the corresponding roller temperature, and then BP neural network PID selftuning algorithm was used to achieve temperature control. Finally, Matlab was used to simulate and model the various parts of the system, and to compare the effect of fixation to verify the effect. The research results indicate that while the BP neural network PID control is applied to the teafixing control system, the adjustment time of the system step response is shorter than the conventional PID control algorithm, the overshoot is smaller, the control is more accurate. At the same time, the improved control system has a better performance in the quality of teafixing.

Key words: teafixing roller; fuzzy; BP neural network; PID selftuning