Gas flow regulation method for ducted rocket based on particle swarm optimization

In order to improve the control performance of gas flow control system in the ducted rocket， the velocity factor corresponding to different pressure and free volume was obtained by using the neural network weight optimized by bare bones particle swarm under the framework of the TD-PID （PID improved by tracking differentiator）algorithm， based on which a TD-PID controller with neural network optimized by bare bones particle swarm （BBPSO-NN-TD-PID） was designed. The simulation results show that compared with the TD-PID controller with velocity factor optimized by bare bones particle swarm （BBPSO-TD-PID）， the algorithm can effectively reduce the response time， but the pressure overshoot is pretty much constant while the flow negative regulation increases slightly in the whole regulation process. After the PID parameters were optimized together with the speed factor， a TD-PID controller with neural network and PID parameters optimized by bare bones particle swarm （BBPSO-NN-TD-PID*） was designed， and the corresponding result of pressure overshoot as well as plus and minus flow negative regulation are improved， and only response time increases slightly， which further improves the performance of the gas flow control system. By comparing the influence of different time lag slopes and step disturbance amplitudes on the response curves， it was found that although there are situations where the actuation of slide valve under high pressure is slow or the particle deposition causes a sudden change of the throat angle， the controller optimized by BBPSO-NN-TD-PID* algorithm can still maintain good control performance.