作　　者： ; ; ;

机构地区： 华南理工大学电力学院

出　　处： 《继电器》 2004年第16期16-19,共4页

摘　　要： 神经网络控制是一种性能卓越的控制策略,神经网络具有出色的知识抽取与学习能力及较强的控制鲁棒性。将神经网络控制策略引入DC/DC变换器,基于BP神经网络,构造了一种DC/DC变换器的新型控制方法。以Buck变换器为例,为其设计了神经网络控制器,对其性能进行了仿真研究,并与传统的PI调节器的性能进行了比较。仿真结果表明,在输入电压或负载有快速波动的情况下,神经网络控制系统比PI调节器具有更好的动态响应特性。 The neural network control method is a novel type of control strategy. It has outstanding advantages in knowledge extracting and learning, and strong robustness in control. This paper is mainly to study the feasibility of introducing the neural network technique into the control of DC/DC converters. A new control strategy based on the BP neural network for DC/DC converter is proposed, and a neural network controller for the Buck type DC/DC converter is designed. The simulation results indicate that in the situation of input voltage and load with quick fluctuations , compared with the DC/DC converter which is based on the traditional PI control, the one based on the neural network control has a much better dynamic response characteristic.

关 键 词： 变换器 神经网络 控制策略 非线性函数 鲁棒性 容错性

领　　域： [电子电信] [自动化与计算机技术] [自动化与计算机技术]