Research On Direct Thrust Control Of Micro Turbojet Engine

Traditional control system for Micro Turbojet Engine (MTE) can not accurately control thrust inreal time, holds large design margin, can not make full use of MTE’s performance. Therefore directthrust control of MTE is researched.With a commercial MTE as research object, a thrust estimator based on neural network isdesigned. A large quantity of sample data is acquired to train the neural network through rig tests. Threeparameters (engine speed, exhaust gas temperature and fuel supply duty ratio) are selected as the inputswith correlation analysis. And time delayed feed forward neural network is chosen to design the estimator.Test results show that the estimator has good performance.MTE is difficult to control with ordinary PID method because of nonlinearity. For this reason,fuzzy PID control method is studied. PID parameters are adjusted by fuzzy mathematical method.Furthermore, a feed-forward fuzzy self-tuning PID controller is designed. Thus closed-loop control ofMTE thrust is realized. At last, rig test of the controller is carried out and the test results show that it hasgood control performance.The research demonstrates the availability of the thrust estimator design method based onartificial neural network and the availability of the closed-loop thrust control of MTE.