基于神经网络的异步电机直接转矩控制研究-电力电子与电力传动专业论文.docxVIP

Abstract The direct torque contorl(DTC) of induction motors is one of the high performance control strategies， which was proposed after the vector coltrol scheme. Its basic idea is： if the amplitude of magnetic flux of motor’s gap is constant,we can adjust rotating speed of stator’s magnetic flux to control torque and rotating speed of rotor.Compared with vector control technique,it has many adventages such as fast torque response, simple control structure,easy to be digitalized ， etc.Athough this kind of technique has received some successful application recently,there are also a great deal of problems in the theories and pratices which deserved to be researched . This thesis aims at studying of neural-network-based DTC system,based on neural networks theory and the problems in practice. The main work of the thesis are as follows: The disadvantages of classical stator flux observing areanalysed; then according to the principle of back-propaganda networks, a simple, robust method of observing stator flux is implemented by neural network. With one neural network in DTC system, to carry out the voltage vector selection function. DTC algorithm is implemented by neural network. Compa red with the classical DTC, the neural-work-based scheme incursshorter execution times and then the performance is improved .Hence, it i s benefit to practice in medium and pint-sized power system. Stator flux observation and voltage vector selector in this thesis are all based on the feedforawrd neural network.The learning algothrim is LM algothrim,a modified back backprogapation algothrim,which has faster computation speed. Computer simulation results are obtained and analyzed,and the effectiveness of the proposed schemes is tested. Keywords: Directtorque control; Induction motor; Neural network; Stator-flux observation; Voltage vector selector II 目 录 第 1 章 绪论 1 1.1 课 题 背 景 1 1.2 直接 转 矩 控 制 的 研 究 现 状 1 1.3 应 用 神 经 网 络 的 可 行 性 与 优 越 性 4 1.4 论 文 的 主 要 工 作 和 结 构 安 排 4 1.5 本 章