Study on Rotor Speed Identification of DTC System Based on PSO Algorithms of BP Neural Network

To solve the strong randomicity and slow convergence of the Particle Swarm Optimization (PSO) algorithms, two new particlepsilas position renewal formulas were analyzed on the basis of extrapolation in mathematics. A new modified PSO algorithm (called Leading PSO algorithms) was put forward. The direct torque control (DTC) system was built in the environment of Matlab (Simulink). The weight and threshold values of BP neural network were trained using the modified PSO algorithms. Some disadvantages such as slow convergence speed and easily plunging into the local solution were avoided effectively. The simulation result shows that the system works well, and the rotor speed identifier has great static and dynamic performance.