Download or read book Simulation Methods for the Phase Behavior of Complex Fluids written by Hyuntae Jung and published by . This book was released on 2020 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: Computer simulations have been widely used in the the study of complex fluids to provide good consistency of macroscopic properties with experiments and understandings in the microscopic view. In the last 10 years, the rise in computing power with graphical processing units and modern data science techniques such as machine learning promote enormous improvements on molecular simulations for studies of complex phenomena which can be challenging, in particular, the phase behavior of complex mixtures. In this thesis, we propose several methods for obtaining phase coexistence of complex fluid mixtures and study phase behaviors of polymer solutions utilizing molecular simulations and a neural network technique. We develop a simulation method for getting phase coexistence from single cell simulations, which uses a spatial concentration autocorrelation function to spatially align instantaneous concentration profiles from different snapshots. Except in the neighborhood of the critical point, the Interface method shows excellent agreement with the phase diagrams of the Widom-Rowlinson model and the symmetric blends of freely jointed polymer molecules available from conventional methods. To obtain coexistence points even near the critical point, we propose a supervised machine learning framework fed by single cell simulation data to optimize convolutional neural network (CNN) filter size as one of hyperparameters. The CNN predictions with simulation data show good agreement with the phase boundary of references close to the critical point. Also, we found that understanding intermediate structures during a phase transition is important to consider them as training sets. Lastly, we study phase behaviors of poly(ethylene oxide) in imidazolium-based ionic liquids which have the unique low critical solubility temperature phase behavior. We use a hybrid simulation method composed of the Interface method and a neural network model to draw phase diagrams. As a result of prediction of critical points for various systems with substitution of H with CH3 on C2 position and different N-alkyl chain length of cation, our results agree with experimental results. Interestingly, a polymer chain near single C2 methylated cation is likely to form not only wrapping but also crown conformation which implies the entropic-driven phase separation.