PSO Based Emotional BPN and RBF Neural Network Models for Wind Speed Prediction

The present research focuses on developing certain proposed machine learning neural network
architectures along with certain mathematical criterion and stochastic population based swarm
intelligence technique particle swarm optimization inspired by nature behavior to carry out wind speed
prediction in renewable energy systems with real time wind farm datasets. In the developed machine
learning model, the work concentrated on developing emotional neural network architecture models
that are optimized employing the particle swarm optimization approach and the optimized emotional
models are employed to carry out effective wind speed prediction for the given real time wind farm
data. Four neural network models are proposed – PSO – EBPN (Emotional Back Propagation Neural
Network) model, PSO – ERBFNN (Emotional Radial Basis Function Neural Network) model, PSO –
EBPN model with hidden neuron criterion and PSO – ERBFNN model with hidden neuron criterion
and as well all these four network models are employed to compute the predicted wind speed output.
The developed models for wind speed prediction has performed in a better manner avoiding local and
global minima problem and as well had a reasonable better convergence rate.