TY  - JOUR
T1  - Development of Hybrid Computational Intelligence Model for
Estimating Relative Cooling Power of Manganite-Based
Materials for Magnetic Refrigeration Enhancement
AU - O. Owolabi, Taoreed AU - Aldhafferi, Nahier AU - O. Akande, Kabiru AU - O. Olatunji, Sunday AU - Alqahtani, Abdullah 
JO  - Journal of Engineering and Applied Sciences
VL  - 13
IS  - 6
SP  - 1575
EP  - 1583
PY  - 2018
DA  - 2001/08/19
SN  - 1816-949x
DO  - jeasci.2018.1575.1583
UR  - https://makhillpublications.co/view-article.php?doi=jeasci.2018.1575.1583
KW  - Manganite-based materials
KW  -relative cooling power
KW  -sensitivity-based linear learning method
KW  -ionic radii
KW  -gravitational search algorithm
KW  -magnetic refrigeration
AB  - The significance of Relative Cooling Power (RCP) of manganite-based magnetic refrigerant in
Magnetic Refrigeration (MR) technology cannot be over-emphasized. Although, MR system overcomes the
setbacks of conventional gas compression technology with its better performance, low cost and little or no
environmental hazard. However, experimental determination of the refrigerant RCP is subjected to procedures
and routines that are not only challenging but also consume appreciable time and other valuable resources.
This necessitates for a simple and cost effective modeling technique that preserves the experimental precision
and accuracy. Therefore, this research develops Sensitivity-Based Linear Learning Method (SBLLM) of training
two-layer feedforward neural network for estimating RCP of manganite-based materials using ionic radii and
dopants concentration as inputs to the model. The number of epoch and hidden neurons of the network are
optimized using Gravitational Search Algorithm (GSA). The results of the developed GSA-SBLLM Model agree
well with the experimentally measured values. The strength and robustness of the developed GSA-SBLLM
Model include its ability to incorporate up to four different dopants and their respective concentrations to
manganite material for magnetic refrigerant RCP estimation. This ability coupled with the precision of its
estimates is of significant impact in magnetic refrigeration enhancement without experimental challenges.
ER  -