Distributed processing applied to neural networks and dual-frequency output control for multilevel inverters
Abstract
Work is described for controlling a multilevel inverter based dual frequency induction heating power supply using an artificial neural network. These types of power supplies are very useful for simplifying the process of induction heating for parts with uneven geometries. Neural networks can be computationally intensive and may require a lot of time to process, if a single computer is used to carry out the neural network. Distributed processing can reduce this time by distributing the calculations amongst several computers. This effort distributes an artificial neural network program among several homogenous computers, a computing cluster to be exact. To distribute the application Remote Procedure Calls are used, to provide fast and reliable communication between the computers. The results show that distributed processing is an efficient way to reduce processing times for neural network application.
Subject Area
Electrical engineering|Artificial intelligence
Recommended Citation
de la Cruz, Sergio, "Distributed processing applied to neural networks and dual-frequency output control for multilevel inverters" (2004). ETD Collection for University of Texas, El Paso. AAIEP10779.
https://scholarworks.utep.edu/dissertations/AAIEP10779