一种异步电机自适应线性神经元速度观测器  

Speed observer of ADALINE for induction motors

作　　者：苗轶如[1] 刘和平[1] 杜俊秀 彭东林[2] MIAO Yi-ru;LIU He-ping;DU Jun-xiu;PENG Dong-lin(State Key Laboratory of Power Transmission Equipment&System Security and New Technology,Chongqing University, Chongqing 400044,China;Engineering Research Center of Mechanical Testing Technology and Equipment of Ministry of Education,Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]重庆大学输配电装备及系统安全与新技术国家重点实验室,重庆400044 [2]重庆理工大学机械检测技术与装备教育部工程研究中心,重庆400054

出　　处：《电机与控制学报》2018年第11期52-57,66,共7页Electric Machines and Control

基　　金：国家自然科学基金(51127001)

摘　　要：针对电动汽车驱动用交流异步电机的无速度传感器矢量控制系统,提出一种自适应线性神经元(ADALINE)速度观测器。首先建立三相异步电机的数学模型,然后在电压模型与电流模型所构成参考自适应(MRAS)速度观测器的基础上,以转子磁链的广义误差作为速度观测器的输入,利用ADALINE算法作为系统自适应律,推导出神经网络权重的自动修正方法,保证在宽调速范围内均能获得准确的电机速度观测值。最后通过建立仿真模型与搭建实验平台进行验证,结果一致表明在高速和低速情况下速度观测值均能准确跟随实际值,动态性能良好,证明所提出的理论分析以及观测器实现方法的可行性与正确性。A new speed observer based on adaptive linear element(ADALINE)for vector control system of induction motors without speed transducer used in Electric Vehicles was presented.The math model of induction motor was established.The generalized error of rotor flux was taken as the input of speed observer,which was designed according to the theory of ADALINE based on model reference adaptive system(MRAS)consist of current model and voltage model.Automatic correction method of neural network weights was derived,which ensured that accurate speed observation could be obtained over a wide speed range.By establishing the simulation model and experimental platform to verify the system,the simulation and experimental results show that speed observation accurately follow the actual value without steady error at both high and low speeds,and speed observer has good dynamic performance.Thus,the correctness and feasibility of proposed theoretical analysis and observer implementation are proved.

关 键 词：矢量控制 异步电机 无速度传感器 速度观测器 自适应线性神经元 

分 类 号：TM301.2[电气工程—电机]