摘要
DC-DC变换器在某些特殊参数下,其工作行为会发生很大的变化,由稳态发展成混沌态,输出特性呈恶化趋势.以PWM型BUCK变换器为研究对象对其在不同电路参数下的输出信号进行了分析研究,提出采用人工神经网络识别DC-DC变换器的混沌行为,首先对DC/DC变换器的输出信号进行特征值的提取,并提出了改进学习算法,加快收敛速度.最终通过仿真和实验验证了该方法的可行性.
A DC-DC converter can give rise to a great variety of behaviors and unstable output signals, depending on the values of the parameters of the circuit. The behaviors can become chaotic from stable work state and the output performance tends to be worse. In this paper, a PWM Buck converter is studied as a research object, the output signals of DC-DC converter in different parameters of the circuit are analyzed. A method of using neural network to identify chaotic behavior of DC-DC converter is put forward. The characteristic of the output signals of DC/DC converter is extracted firstly and the algorithm of BP is improved to quicken convergent speed. Finally, it validity is proven by simulation and experiment.
出处
《安徽机电学院学报》
2002年第4期1-6,共6页
Journal of Anhui Institute of Mechanical and Electrical Engineering
基金
国家自然科学基金资助项目(59677021)
