ANN: Prediction of an experimental heat transfer analysis of concentric tube heat exchanger with corrugated inner tubes. (25th June 2017)

• Record Type: Journal Article
• Authors: Verma, Tikendra Nath; Nashine, Prerana; Singh, Dheerendra Vikram; Singh, Thokchom Subhaschandra; Panwar, Deepali
• Is Part Of: Applied thermal engineering. Volume 120(2017)
• Page Start: 219
• Page End: 227
• Publication Date: 2017-06-25
• Subjects: Heat exchanger -- ANN -- Helical grooves -- Corrugated surfaces; Heat engineering -- Periodicals; Heating -- Equipment and supplies -- Periodicals; Periodicals; 621.40205
• DOI: 10.1016/j.applthermaleng.2017.03.126
• Languages: English
• ISSNs: 1359-4311

Abstract

Highlights: Full description of proposed fabricated heat exchanger with different corrugated and non-corrugated pies. Different pictorial configuration of pipes according to different pitch and depth. ANN modeling for predicting the heat transfer performance of proposed fabricated heat exchanger. Systematic procedure of formulating the mathematical expressions from validated ANN for predicting the same. Abstract: The main aim of this work to estimate the heat transfer performance of proposed fabricated heat exchanger using corrugated and non-corrugated pipes. Pitch and depths are varied in case of corrugated pipe. Authors have achieved the maximum heat transfer coefficient and Nusselt number with helical shaped ribs of 4 mm pitch and 1.5 mm depth with the variations in Reynolds number from 5000 to 17, 000, mass flow rates from 0.03 to 0.13 kg/s and 0.04 to 0.14 kg/s for cold and hot fluid respectively. Length and diameter of pipes are 25.4 mm and 2000 mm respectively. Authors have also modeled an artificial neural network (ANN) for predicting heat transfer coefficient, Nusselt number and Reynolds number. Value of coefficient of multiple determination are achieved. The R 2 (coefficient of multiple determination) - values is 0.99999, 0.999997 and 0.999993 for Reynolds number, heat transfer and Nusselt number respectively with unknown data set which is very satisfactory. The optimized architecture of ANN is 8-10-3.