The Journal of the Institute of Electrical Installation Engineers of Japan
Online ISSN : 2188-6946
Print ISSN : 0910-0350
ISSN-L : 0910-0350
The use of on-line learning and control Neural Network to DC/DC converter
Daisuke SoneKouki AyabeFUmiaki TakedaMototaka Sone
Author information
JOURNAL RESTRICTED ACCESS

2011 Volume 31 Issue 7 Pages 578-587

Details
Abstract
Currently,analog control is used for voltage regulation in the DC-DC converter. The control strategy is represented by PI,PD,and PID controllers. These control methods should achieve high performance output control during a linear load condition. However,when the load is constructed by nonlinear elements such as the choke coils and diodes,it is impossible to keep same performance because circuit conditions cannot be constant. Furthermore,a digital load with a high-speed and irregular change characteristic appeared in recent years.Therefore,we proposed a digital control system that is using a DSP as a control core in this paper. In the digital control system,we adopted the Neural Network(NN) method because it is possible to allow flexibly in various states of the load. In the proposed method,the DSP will learn the circuit condition in real time and realizes high performance output voltage control instead with analog control method
Content from these authors
© 2011 The Institute of Electrical Installation Engineers of Japan
Previous article
feedback
Top