基于神经网络MRAS的速度辨识仿真研究

Research on Speed Identification Simulation Based on Neural Network MRAS

准确的速度辨识是无速度传感器矢量控制系统的关键。针对速度辨识,文中首先研究了基于MRAS的转速估计,并以此为基础,利用神经网络的自学习与自适应能力研究了一种基于BP神经网络MRAS的转速估计方法,实现了从低速到高速的宽范围调速。仿真结果表明,该方法可以实现低速到高速的速度辨识,速度辨识精度高,并且具有很好的抗干扰性和鲁棒性。

Accurate speed identification is the key of speed sensorless vector control system.For speed identification,the speed estimation based on MRAS was first studied in this paper.Based on this,a speed estimation method based on BP neural network MRAS was studied by using the self-learning and self-adaptive ability of neural network,and a wide range speed regulation from low speed to high speed was realized.The simulation results showed that this method can achieve low speed to high speed speed identification.The speed identification accuracy was high,and had good anti-interference and robustness.