The Research And Design Of Direct Torque Control System Key Parameter Observer

Once the speed sensorless direct torque control technology has been brought up, it is widely concerned and obtains mushroom growth with its novel control antilogy, succinct system organization, superior static & dynamic characteristics. In order to realize more high-powered speed sensorless direct torque control, we must improve the identification precision of key parameter among others. For this reason, this article contrapose two important key parameters: The observation of revolution and flux linkage, carry out a series of research &design .With matlab /simulink, this article has built the simulation model of direct torque control system, and then inserts the observer of its neural network based on mutated ant colony into system model. To import advanced intelligent control antilogy into conventional direct torque, then join intelligence optimistic algorithm into conventional neural network control organically, through the network observer built to carry out online observation of electric machine parameter. In order to enhance the system operation effectiveness, this article has imported mutated evolutionary algorithm of genetic algorithm into basic ant colony optimization, Then use the mutated ant colony optimization gained to optimize the connection weight in neural network, effectively solve the problem that neural network is easy to involve in minimal point, meanwhile the speed of convergence will be highly enhanced than only adopt single acting basic ant colony optimization. From the simulation result we can see, this kind of neural network based on mutated ant colony identifier can accurately identify the variation of motor speed online, Moreover, its import can observably improve the static & dynamic characteristics of original system.Besides, contrapose the problem which performance of direct torque control system is bad at low speed, this article has brought forward a solution which constitute full order staobserver instead of traditional integrator. The simulation results indicate that this solution not only can greatly improve the low-speed performance of system, but also has favorable dynamic tracking performance.Finally, from the spot debugging result of electric machine control using intelligence observe and control comprehensive test-bed, it is observed that this article's design plan is good practice and efficient.