Applying a supervised ANN (artificial neural network) approach to the prognostication of driven wheel energy efficiency indices

H. Taghavifar, A. Mardani

Research output: Contribution to journalArticle

37 Citations (Scopus)

Abstract

This paper examines the prediction of energy efficiency indices of driven wheels (i.e. traction coefficient and tractive power efficiency) as affected by wheel load, slippage and forward velocity at three different levels with three replicates to form a total of 162 data points. The pertinent experiments were carried out in the soil bin testing facility. A feed-forward ANN (artificial neural network) with standard BP (back propagation) algorithm was practiced to construct a supervised representation to predict the energy efficiency indices of driven wheels. It was deduced, in view of the statistical performance criteria (i.e. MSE (mean squared error) and R2), that a supervised ANN with 3-8-10-2 topology and Levenberg–Marquardt training algorithm represented the optimal model. Modeling implementations indicated that ANN is a powerful technique to prognosticate the stochastic energy efficiency indices as affected by soil-wheel interactions with MSE of 0.001194 and R2 of 0.987 and 0.9772 for traction coefficient and tractive power efficiency. It was found that traction coefficient and tractive power efficiency increase with increased slippage. A similar trend is valid for the influence of wheel load on the objective parameters. Wherein increase of velocity led to an increment of tractive power efficiency, velocity had no significant effect on traction coefficient.
Original languageEnglish
Pages (from-to)651-657
Number of pages7
JournalEnergy
Volume68
Early online date7 Feb 2014
DOIs
Publication statusPublished - 15 Apr 2014
Externally publishedYes

Fingerprint

Energy efficiency
Wheels
Neural networks
Soils
Backpropagation algorithms
Bins
Topology
Testing
Experiments

Keywords

  • ANN
  • Energy efficiency
  • Traction coefficient
  • Tractive power efficiency

Cite this

Applying a supervised ANN (artificial neural network) approach to the prognostication of driven wheel energy efficiency indices. / Taghavifar, H.; Mardani, A.

In: Energy, Vol. 68, 15.04.2014, p. 651-657.

Research output: Contribution to journalArticle

@article{664f6c44465b403db44ac4a5b6433e8b,
title = "Applying a supervised ANN (artificial neural network) approach to the prognostication of driven wheel energy efficiency indices",
abstract = "This paper examines the prediction of energy efficiency indices of driven wheels (i.e. traction coefficient and tractive power efficiency) as affected by wheel load, slippage and forward velocity at three different levels with three replicates to form a total of 162 data points. The pertinent experiments were carried out in the soil bin testing facility. A feed-forward ANN (artificial neural network) with standard BP (back propagation) algorithm was practiced to construct a supervised representation to predict the energy efficiency indices of driven wheels. It was deduced, in view of the statistical performance criteria (i.e. MSE (mean squared error) and R2), that a supervised ANN with 3-8-10-2 topology and Levenberg–Marquardt training algorithm represented the optimal model. Modeling implementations indicated that ANN is a powerful technique to prognosticate the stochastic energy efficiency indices as affected by soil-wheel interactions with MSE of 0.001194 and R2 of 0.987 and 0.9772 for traction coefficient and tractive power efficiency. It was found that traction coefficient and tractive power efficiency increase with increased slippage. A similar trend is valid for the influence of wheel load on the objective parameters. Wherein increase of velocity led to an increment of tractive power efficiency, velocity had no significant effect on traction coefficient.",
keywords = "ANN, Energy efficiency, Traction coefficient, Tractive power efficiency",
author = "H. Taghavifar and A. Mardani",
year = "2014",
month = "4",
day = "15",
doi = "10.1016/j.energy.2014.01.048",
language = "English",
volume = "68",
pages = "651--657",
journal = "Energy",
issn = "0360-5442",
publisher = "Elsevier",

}

TY - JOUR

T1 - Applying a supervised ANN (artificial neural network) approach to the prognostication of driven wheel energy efficiency indices

AU - Taghavifar, H.

AU - Mardani, A.

PY - 2014/4/15

Y1 - 2014/4/15

N2 - This paper examines the prediction of energy efficiency indices of driven wheels (i.e. traction coefficient and tractive power efficiency) as affected by wheel load, slippage and forward velocity at three different levels with three replicates to form a total of 162 data points. The pertinent experiments were carried out in the soil bin testing facility. A feed-forward ANN (artificial neural network) with standard BP (back propagation) algorithm was practiced to construct a supervised representation to predict the energy efficiency indices of driven wheels. It was deduced, in view of the statistical performance criteria (i.e. MSE (mean squared error) and R2), that a supervised ANN with 3-8-10-2 topology and Levenberg–Marquardt training algorithm represented the optimal model. Modeling implementations indicated that ANN is a powerful technique to prognosticate the stochastic energy efficiency indices as affected by soil-wheel interactions with MSE of 0.001194 and R2 of 0.987 and 0.9772 for traction coefficient and tractive power efficiency. It was found that traction coefficient and tractive power efficiency increase with increased slippage. A similar trend is valid for the influence of wheel load on the objective parameters. Wherein increase of velocity led to an increment of tractive power efficiency, velocity had no significant effect on traction coefficient.

AB - This paper examines the prediction of energy efficiency indices of driven wheels (i.e. traction coefficient and tractive power efficiency) as affected by wheel load, slippage and forward velocity at three different levels with three replicates to form a total of 162 data points. The pertinent experiments were carried out in the soil bin testing facility. A feed-forward ANN (artificial neural network) with standard BP (back propagation) algorithm was practiced to construct a supervised representation to predict the energy efficiency indices of driven wheels. It was deduced, in view of the statistical performance criteria (i.e. MSE (mean squared error) and R2), that a supervised ANN with 3-8-10-2 topology and Levenberg–Marquardt training algorithm represented the optimal model. Modeling implementations indicated that ANN is a powerful technique to prognosticate the stochastic energy efficiency indices as affected by soil-wheel interactions with MSE of 0.001194 and R2 of 0.987 and 0.9772 for traction coefficient and tractive power efficiency. It was found that traction coefficient and tractive power efficiency increase with increased slippage. A similar trend is valid for the influence of wheel load on the objective parameters. Wherein increase of velocity led to an increment of tractive power efficiency, velocity had no significant effect on traction coefficient.

KW - ANN

KW - Energy efficiency

KW - Traction coefficient

KW - Tractive power efficiency

UR - http://www.scopus.com/inward/record.url?eid=2-s2.0-84898029886&partnerID=MN8TOARS

U2 - 10.1016/j.energy.2014.01.048

DO - 10.1016/j.energy.2014.01.048

M3 - Article

VL - 68

SP - 651

EP - 657

JO - Energy

JF - Energy

SN - 0360-5442

ER -