Effects of various parameters on entropy generation and exergy destruction in a coil wire inserted heat exchanger by using deep learning neural network method

Abstract

In present study mainly the effect of inserting a coil wire type turbulator into a concentric type heat exchanger on its entropy generation (Ns), efficiency (epsilon) and exergy destruction (E*), is studied experimentally. Due to inserting coil wire inside the inner pipe of the exchanger presents a challenge, the coil wires are located through the pipe with various length (l/d), thickness (e/d) and pitch (p/d). The experiments are performed by also varying the Reynolds number from 30,000 to 80,000. The results showed the entropy generation number increases with increasing l/d and p/d of the coil wire turbulator, but decreases with increasing e/d. Exergy destruction is the exergy that is destroyed within the system boundary due to the effects of irreversibilities that generate the entropy. Therefore, it is proportional to the entropy generation. Hence similar trends are observed when calculating the exergy destruction. Finally increasing the thickness, pitch and length of the coil wire results with augmentation of the efficiency. The Deep Learning Neural Network Method, namely DNN is applied to the results of the experimental study for showing the relationship of the parameters to each other. Multiple Linear Regression (MLR) model is applied to the result for model verification. Comparison of the results of MLR and DNN methods show that DNN predictions are mostly reliable considering the relationships between the dependent and independent parameters.