基于BP神经网络的异步电机转速估计

The Speed Estimation of Induction Motor Based on BP Neural Network

无速度传感器矢量控制已经在异步电机上取得广泛应用,转速估计的精确程度是该控制方法实现的重要因素,而电机转速受复杂的强耦合多变量共同作用。基于此,文中尝试采用具有自学习和自适应能力的神经网络控制方法来估计转速,仿真证明,采用神经网络预测的电机转速误差可以控制在±1转之内,证实该方案的合理性。

The speed sensorless vector control of induction motor has been widely used. Accuracy of speed estimation is an important factor to realize the control method. The motor speed was affected by the complex coupling of multiple variables. Based on this,the neural network was used as a self-learning and adaptive ability to estimate the speed. The simulation results showed that the motor speed error predicted by the neural network can be controlled within ± 1 rpm,and the rationality of the scheme was confirmed.