Prediction of the Dynamics of a Fluidized Bed Reactor using Artificial Neural Networks

Publication Year:
2007
Usage 2947
Downloads 2735
Abstract Views 212
Repository URL:
http://dc.engconfintl.org/fluidization_xii/88
Author(s):
Karimipour, Shayan; Mostoufi, Navid; Sotudeh -Gharebagh, Rahmat
Tags:
Fluidized Bed; Chaos; k-Nearest Neighbor; Artificial Neural Network; CFD; Chemical Engineering; Engineering
article description
The dynamic behavior of fluidized bed has been studied based on the chaos theory. The experiments were done in a fluidized bed of 0.15 m diameter using an optical fiber probe. The interval between successive clusters in the fluidized bed were calculated from the time series signals and proved to be chaotic by calculating the correlation dimension. An artificial neural network (ANN) was adapted and trained to predict the generated time series. The ANN results were compared with the predictions of the k-Nearest Neighbor (kNN) method to show the superiority of ANN in chaotic time series prediction.