APPLICATION OF NEURAL NETWORKS TO THE CONTROL OF BURGERS EQUATION WITH RANDOM FORCING

Yongseung Cho, Ramesh K. Agarwal, Yogesh Tupe

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

In this paper, the feasibility of using artificial neural network to obtain the control inputs for the flow governed by Burgers equation with random forcing is studied. The objective is to provide a real time active controller for the flow. The data obtained from the optimal (LQG) and robust (H) control are used to train the network. Due to the structure of a neural network, it is possible for the trained neural network to imitate the controller in real time. The network consists of a layer with 4 input neurons, 2 hidden layers, a layer with 1 output neuron, and a bias neuron. The network is trained using backpropagation method. Several data sets are used for training and testing the network. The trained network is tested for the generality and for fault tolerance. A simulation with the neural network controller is conducted for the flow modeled by the Burgers equation with random forcing input. The results show that the employment of neural network as a controller can be effective for flow control problems.

Original languageEnglish
Title of host publicationFluids Engineering
EditorsA.S. Lavine, U. Chandra, M.M. Chen, C.T. Crowe, U. Fritsching, al et al
PublisherAmerican Society of Mechanical Engineers (ASME)
Pages81-85
Number of pages5
ISBN (Electronic)9780791815502
DOIs
StatePublished - 1996
EventASME 1996 International Mechanical Engineering Congress and Exposition, IMECE 1996 - Atlanta, United States
Duration: Nov 17 1996Nov 22 1996

Publication series

NameASME International Mechanical Engineering Congress and Exposition, Proceedings (IMECE)
Volume1996-S

Conference

ConferenceASME 1996 International Mechanical Engineering Congress and Exposition, IMECE 1996
Country/TerritoryUnited States
CityAtlanta
Period11/17/9611/22/96

Fingerprint

Dive into the research topics of 'APPLICATION OF NEURAL NETWORKS TO THE CONTROL OF BURGERS EQUATION WITH RANDOM FORCING'. Together they form a unique fingerprint.

Cite this