General inverse neural current control for buck converter

Abstract

The neural networks have been used in different applications of automatic control systems, many of them in converters of continuous voltage. Different works have been presented where neuronal controllers are used in Buck converters for voltage control, but works where this technique is used for the current control of these devices to a rated load, remain unknown; It was made a search to know the state-of-the-art of this thematic, there was not found applications regarding to this issue. In this simulation, a generalized inverse neuronal control is used, in order to control de current in a Buck Converter. Afterwards, to improve the performance of the controller, it is used a controller of neuronal modules in parallel, trained and activated by unit load regions, this is proposed in order to evaluate their viability to be used this type of neural control, in a posterior implementation of Hardware. In the present article a controller of these characteristics in a Buck converter has been simulated, with output current of 8A, and a nominal charge of 1.5 Ω and lower impedances, using inverse neural controller. © Springer Science+Business Media B.V. 2008.