Experimental study and machine learning based prediction of heat transfer dynamics during droplet impingement on high temperature substrates

Droplet dynamics on heated surface is a complex phenomenon influenced by surface characteristics, and temperature gradients making it challenging to quantify heat transfer accurately through traditional methods. This study presents a novel approach for analysing heat transfer from high temperature microstructured surfaces during droplet impingement using image processing. This is used to analyze the evaporation mechanism and estimation of droplet area to calculate heat transfer coefficient. Distilled water at room temperature corresponding to atmospheric pressure of volume 10 µl is impinged on three different substrates namely polished, channelled and micropillared. The surface temperature is varied from 80°C to 115°C. The temporal analysis of droplets on high temperature surface is studied using non-intrusive thermal imaging technique and droplet dynamics is visualized using high-speed camera. The relative uncertainty in repeatability and reproducibility of the estimated values of evaporation rate is ±8.65% and ±7.2% respectively. The evaporation rate at different wettability is discussed. The instantaneous droplet area and instantaneous heat transfer coefficient is calculated for the substrates at different temperatures. It is inferred that for polished surface, channeled surface and micropillared surface, initial variation of instantaneous heat transfer coefficient is negligible for substrate temperature of 80 ℃ and 85 ℃. For the surface temperature range of 90 ℃ to 115 ℃, the instantaneous heat transfer coefficient is observed to be higher for micropillared surface on an overall pattern. Analysis of Nusselt number shows that there is nearly 15% enhancement in evaporative heat transfer from microplillared surface as compared to others. A deep multi-layer perceptron (MLP) based prediction has been performed to predict the non-dimensional heat transfer.