Flow boiling through helical coils is an effective heat transfer enhancement technique where the centripetal force distributes the liquid film on the wall resulting in thinner liquid film and higher critical heat flux. Although there are several empirical correlations in the literature, most of these correlations are applicable for specific operating conditions. Recently, Artificial Neural Networks (ANNs) technique has been used for performance prediction in various thermal engineering topics. This paper presents the application of feed forward neural network with Levenberg-Marquardt training algorithm to predict the heat transfer coefficients of flow boiling inside 2.8mm diameter helically coiled tube. The Normalized Prandtl, Dean, Convective and Boiling numbers were utilized as network input while the two-phase to liquid only heat transfer coefficients ratio was used as output. This network was trained using 353 data points from published literature and validated with data from current experimental measurements with deviation of ± 30%.
展开▼
