Constellation, le dépôt institutionnel de l'Université du Québec à Chicoutimi

Résumé

Heightening the water repellency of surfaces can serve anti-icing purposes by removing water drops before they freeze and adhere to a surface. Here we study the impact dynamics of water droplets on silicone rubber surfaces—ranging from hydrophobic to superhydrophobic—at −20, −10, and 25 °C. We evaluate the influence of static contact angle, contact angle hysteresis, surface roughness, temperature, impacting velocity, and droplet diameter on droplet behavior (e.g., deposition, bouncing, splash). Minor effect of temperature on droplet dynamics on microstructured surfaces for a wide range of Weber and Reynolds numbers is observed. Experimental observations show that full bouncing only occurs on superhydrophobic surfaces with a CA > 160° and a CAH < 2° at temperatures above 0 °C for We < 110 and Re < 5000. Increasing the impact velocity of the droplet on rough surfaces heightens the probability of splashing. This experimental data is then coupled with machine-learning techniques (logistic regression, decision tree, and random forest) to comprehensively investigate droplet impact behavior on hydrophobic and superhydrophobic surfaces at various temperatures. We predict the behavior probability of impacting droplets on surfaces as a function of Weber number, Reynolds number and surface features (static contact angle, contact angle hysteresis, temperature, and surface roughness). Our experimental results and machine learning–based predictions are highly consistent, demonstrating that machine learning can effectively predict droplet motion on hydrophobic and superhydrophobic silicone rubber surfaces at different temperatures.