A combination of linear and nonlinear activation functions in neural networks for modeling a de-superheater

Morteza Mohammadzaheri, Lei Chen, Ali Ghaffari, John Willison

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

This paper deals with modeling a power plant component with mild nonlinear characteristics using a modified neural network structure. The hidden layer of the proposed neural network has a combination of neurons with linear and nonlinear activation functions. This approach is particularly suitable for nonlinear system with a low grade of nonlinearity, which can not be modeled satisfactorily by neural networks with purely nonlinear hidden layers or by the method of least square of errors (the ideal modeling method of linear systems). In this approach, two channels are installed in a hidden layer of the neural network to cover both linear and nonlinear behavior of systems. If the nonlinear characteristics of the system (i.e. de-superheater) are not negligible, then the nonlinear channel of the neural network is activated; that is, after training, the connections in nonlinear channel get considerable weights. The approach was applied to a de-superheater of a 325 MW power generating plant. The actual plant response, obtained from field experiments, is compared with the response of the proposed model and the responses of linear and neuro-fuzzy models as well as a neural network with purely nonlinear hidden layer. A better accuracy is observed using the proposed approach.

Original languageEnglish
Pages (from-to)398-407
Number of pages10
JournalSimulation Modelling Practice and Theory
Volume17
Issue number2
DOIs
Publication statusPublished - Feb 2009

Fingerprint

Superheaters
Activation Function
Nonlinear Function
Chemical activation
Neural Networks
Neural networks
Modeling
Power plants
Control nonlinearities
Field Experiment
Neuro-fuzzy
Power Plant
Neurons
Linear systems
Fuzzy Model
Nonlinear systems
Modeling Method
Network Structure
Least Squares
Neuron

Keywords

  • Artificial neural network
  • De-superheater
  • Modeling
  • Power plant

ASJC Scopus subject areas

  • Software
  • Modelling and Simulation
  • Hardware and Architecture

Cite this

A combination of linear and nonlinear activation functions in neural networks for modeling a de-superheater. / Mohammadzaheri, Morteza; Chen, Lei; Ghaffari, Ali; Willison, John.

In: Simulation Modelling Practice and Theory, Vol. 17, No. 2, 02.2009, p. 398-407.

Research output: Contribution to journalArticle

@article{54a7ab3210ae4f71a747a8c9eb198422,
title = "A combination of linear and nonlinear activation functions in neural networks for modeling a de-superheater",
abstract = "This paper deals with modeling a power plant component with mild nonlinear characteristics using a modified neural network structure. The hidden layer of the proposed neural network has a combination of neurons with linear and nonlinear activation functions. This approach is particularly suitable for nonlinear system with a low grade of nonlinearity, which can not be modeled satisfactorily by neural networks with purely nonlinear hidden layers or by the method of least square of errors (the ideal modeling method of linear systems). In this approach, two channels are installed in a hidden layer of the neural network to cover both linear and nonlinear behavior of systems. If the nonlinear characteristics of the system (i.e. de-superheater) are not negligible, then the nonlinear channel of the neural network is activated; that is, after training, the connections in nonlinear channel get considerable weights. The approach was applied to a de-superheater of a 325 MW power generating plant. The actual plant response, obtained from field experiments, is compared with the response of the proposed model and the responses of linear and neuro-fuzzy models as well as a neural network with purely nonlinear hidden layer. A better accuracy is observed using the proposed approach.",
keywords = "Artificial neural network, De-superheater, Modeling, Power plant",
author = "Morteza Mohammadzaheri and Lei Chen and Ali Ghaffari and John Willison",
year = "2009",
month = "2",
doi = "10.1016/j.simpat.2008.09.015",
language = "English",
volume = "17",
pages = "398--407",
journal = "Simulation Modelling Practice and Theory",
issn = "1569-190X",
publisher = "Elsevier",
number = "2",

}

TY - JOUR

T1 - A combination of linear and nonlinear activation functions in neural networks for modeling a de-superheater

AU - Mohammadzaheri, Morteza

AU - Chen, Lei

AU - Ghaffari, Ali

AU - Willison, John

PY - 2009/2

Y1 - 2009/2

N2 - This paper deals with modeling a power plant component with mild nonlinear characteristics using a modified neural network structure. The hidden layer of the proposed neural network has a combination of neurons with linear and nonlinear activation functions. This approach is particularly suitable for nonlinear system with a low grade of nonlinearity, which can not be modeled satisfactorily by neural networks with purely nonlinear hidden layers or by the method of least square of errors (the ideal modeling method of linear systems). In this approach, two channels are installed in a hidden layer of the neural network to cover both linear and nonlinear behavior of systems. If the nonlinear characteristics of the system (i.e. de-superheater) are not negligible, then the nonlinear channel of the neural network is activated; that is, after training, the connections in nonlinear channel get considerable weights. The approach was applied to a de-superheater of a 325 MW power generating plant. The actual plant response, obtained from field experiments, is compared with the response of the proposed model and the responses of linear and neuro-fuzzy models as well as a neural network with purely nonlinear hidden layer. A better accuracy is observed using the proposed approach.

AB - This paper deals with modeling a power plant component with mild nonlinear characteristics using a modified neural network structure. The hidden layer of the proposed neural network has a combination of neurons with linear and nonlinear activation functions. This approach is particularly suitable for nonlinear system with a low grade of nonlinearity, which can not be modeled satisfactorily by neural networks with purely nonlinear hidden layers or by the method of least square of errors (the ideal modeling method of linear systems). In this approach, two channels are installed in a hidden layer of the neural network to cover both linear and nonlinear behavior of systems. If the nonlinear characteristics of the system (i.e. de-superheater) are not negligible, then the nonlinear channel of the neural network is activated; that is, after training, the connections in nonlinear channel get considerable weights. The approach was applied to a de-superheater of a 325 MW power generating plant. The actual plant response, obtained from field experiments, is compared with the response of the proposed model and the responses of linear and neuro-fuzzy models as well as a neural network with purely nonlinear hidden layer. A better accuracy is observed using the proposed approach.

KW - Artificial neural network

KW - De-superheater

KW - Modeling

KW - Power plant

UR - http://www.scopus.com/inward/record.url?scp=57749111546&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=57749111546&partnerID=8YFLogxK

U2 - 10.1016/j.simpat.2008.09.015

DO - 10.1016/j.simpat.2008.09.015

M3 - Article

AN - SCOPUS:57749111546

VL - 17

SP - 398

EP - 407

JO - Simulation Modelling Practice and Theory

JF - Simulation Modelling Practice and Theory

SN - 1569-190X

IS - 2

ER -