Evaluation for NDWCT Performance with Different Types of Packing Fills in Iraq Using ANN [dataset]

Iosr Journals, Qasim Saleh Mahdi, Muwafaq Rahi Al-Hachami
2015 Figshare  
Artificial Neural Network (ANN) is used to predicate experimental results for Natural Draft Wet Cooling Tower (NDWCT) rig using Levenberg-Marquardt back propagation algorithm in MATLAP. The experimental tests are done in hot and dry weather (Iraqi weather as an example). ANN results show good agreements with experimental results where average correlation coefficient (R) for all results is (0.994), average root mean square errors (RMSE) are (5.99, 0.91, 0.24, 0.51, 0.49, 0.2, and 5.46), and
more » ... ge of mean ratio between the errors and the network output values (MRE) are (1.72%, 1.32%, 3.93%, 1.78%, 3.77%, 8.4% , and 1.05%) for relative humidity change, tower range, water to air mass flow ratio, cooling capacity, heat rejected to air, effectiveness, and air enthalpy change respectively.
doi:10.6084/m9.figshare.1385162.v1 fatcat:vmgbh6wxwfaxvih2fkqrckzw54