Continuous, non-linear, optimal speed control of a Distributed Generation Power Pack using Artificial Neural Networks

Distributed Generation Power Packs with a combustion engine prime mover are still widely used to supply electric power in a variety of applications. These applications range from backup power supply systems to providing power in places where grid connection is either technically impractical or financially uneconomic. Due to the ever increasing cost of diesel fuel and the environmental issues associated with its use, the optimisation of these AC generators and the reduction of fuel consumption is vital. This paper presents how Artificial Neural Networks can be utilised in order to obtain a continuous function which relates variable load demand to optimal speed demand. The Artificial Neural Network toolbox within MATLAB is used to create, train and test the Artificial Neural Networks. This paper also shows the results of an experimental system used in order to emulate the Distributed Generation Power Pack. Overall it is shown that is possible to operate a variable speed system under optimal, non-linear, speed control using Artificial Neural Networks. © 2014 IEEE