COMPARATIVE PREDICTIVE MODELLING AND ANALYSIS OF THE PROCESS OF A HEAT EXCHANGER USING RESPONSE SURFACE METHODOLOGY AND ARTIFICIAL NEURAL NETWORK
UMUDIKE JOURNAL OF ENGINEERING AND TECHNOLOGY · 04.06.2026 · 27
DOI: 10.33922/j.ujet_v12i2_4 ↗Kurz gefasst
Fachartikel in der Zeitschrift UMUDIKE JOURNAL OF ENGINEERING AND TECHNOLOGY, online erschienen am 04.06.2026. Verfasst von N. U. Unwana-Obong.
Abstract (Originalsprache)
This study performed a comparative analysis of the Response Surface Methodology (RSM) and the Artificial Neural Network (ANN) models employed to predicts the thermo-hydraulic performance of a Gas Turbine Heat Exchanger Unit 16 (HEU-16) at Transcorp Power Limited where the operating parameters was obtained from Mark V Speed control system. The three independent operating parameters were: oil inlet temperature, water inlet temperature, and mass flow rate and the responses evaluated were heat transfer coefficient, pressure drop, and thermal effectiveness of the heat exchanger system. Experimental data obtained were analyzed using Design-Expert software for RSM modeling, while ANN modeling was carried out using multilayer feed-forward, back-propagation (B.P.) algorithms, Levenberg-Marquardt training algorithm was the best at predicting the performance of the heat exchanger. The results revealed that the operating parameters significantly affected the thermal and hydraulic behavior of the exchanger. Increase in inlet temperatures and mass flow rate enhanced the heat transfer coefficient and thermal effectiveness, although pressure drop also increased correspondingly. The developed quadratic RSM models were statistically significant with high coefficients of determination, indicating good agreement between predicted and experimental values. However, ANN demonstrated superior predictive accuracy with higher R² values and lower mean square error compared to RSM, due to its enhanced capability in handling nonlinear process relationships. The predicted results established optimum operating conditions at an oil inlet temperature of 69.47 °C, water inlet temperature of 54.02 °C, and mass flow rate of 269.42 m3/s. Under these conditions, the predicted heat transfer coefficient, pressure drop, and thermal effectiveness were 691.08 W/m²·K, 1.88 kPa, and 0.718 respectively, with a desirability value of 1.0. The study demonstrated that ANN and RSM are reliable tools for thermal system prediction and performance enhancement in industrial heat exchanger applications.
Zitieren
APA
Unwana-Obong, N. U. (2026). COMPARATIVE PREDICTIVE MODELLING AND ANALYSIS OF THE PROCESS OF A HEAT EXCHANGER USING RESPONSE SURFACE METHODOLOGY AND ARTIFICIAL NEURAL NETWORK. UMUDIKE JOURNAL OF ENGINEERING AND TECHNOLOGY, 27. https://doi.org/10.33922/j.ujet_v12i2_4
Vancouver
Unwana-Obong NU. COMPARATIVE PREDICTIVE MODELLING AND ANALYSIS OF THE PROCESS OF A HEAT EXCHANGER USING RESPONSE SURFACE METHODOLOGY AND ARTIFICIAL NEURAL NETWORK. UMUDIKE JOURNAL OF ENGINEERING AND TECHNOLOGY. 2026:27. doi:10.33922/j.ujet_v12i2_4
BibTeX
@article{UnwanaObong2026,
author = {Unwana-Obong, N. U.},
title = {COMPARATIVE PREDICTIVE MODELLING AND ANALYSIS OF THE PROCESS OF A HEAT EXCHANGER USING RESPONSE SURFACE METHODOLOGY AND ARTIFICIAL NEURAL NETWORK},
journal = {UMUDIKE JOURNAL OF ENGINEERING AND TECHNOLOGY},
year = {2026},
pages = {27},
doi = {10.33922/j.ujet_v12i2_4}
}
EUVIMED