Session: 17-01-01: Research Posters

Paper Number: 150704

150704 - Machine Learning (Ml) Approach to Predict the Viscosity of Ionic Liquids (Ils) Based Nanofluids 

Conventional methods of energy generation have been found to have a detrimental impact on the environment. As a result, researchers are actively investigating alternative, clean, and renewable sources of energy. Among these options, solar power has emerged as a highly efficient and sustainable means of generating electricity. One of the methods for employing solar power Is the utilization of concentrated solar power systems (CSPs) systems where heat energy is stored in heat transfer fluids (HTFs) from mirrors or lenses that are positioned to concentrate sunlight to a small area. Traditional HTFs that have been used in the past to produce steam for energy production include Therminol VP-1 (eutectic mixture of biphenyl and diphenyl oxide), thermal oil, and molten salt. The problems associated with traditional HTFs are high melting points and low decomposition temperatures which increase operation costs, reduce system efficiency, and diminish energy storage potential. Ionic liquids are a class of organic salts that have low melting points, low volatility, low flammability, and high thermal conductivity which make them ideal candidates for HTFs. Additionally, ionic liquids have been found to possess negligible vapor pressure and high thermal stability that simplify the heat transfer process. Ionic liquids (ILs)-based nanofluids are a new class of HTFs created by dispersing metal/metal oxide nanoparticles into ionic fluids. We have evidence showing that the heat transfer efficiency and thermal stability of ILs increase when embedded with nanoparticles (ILs-based nanofluids) even at high temperatures. Subsequently, metallic nanoparticles have high thermal conductivities and thereby increase the thermal properties of ILs-based nanofluids. Several groups have shown that ILs-based nanofluids exhibit increased thermophysical properties.

Although the thermophysical properties of ILs-based nanofluids have escalated, little research has been conducted to demonstrate their feasibility in existing CSPs. In order for ILs-based nanofluids to be utilized on a commercial scale, it is imperative that they can be employed without incurring expensive modifications to CSPs. A crucial aspect that necessitates investigation is the efficiency of pumping ILs-based nanofluids within the confines of pre-existing CSPs. Viscosity is the fluid property that is directly related to the pumping power. Here we attempt to predict the viscosity of ILs-based nanofluids using a machine learning (ML) approach.  A novel Gaussian process regression (GPR) model will be utilized for viscosity prediction. The proposed model will help to predict the optimized thermophysical properties which will eventually reduce the experimental cost for the design of advanced HTFs for next-generation solar thermal applications.

Presenting Author: Truman Brabham University of South Carolina Aiken

Presenting Author Biography: The presenter is an undergraduate student at the University of South Carolina Aiken.

Authors:

Turner Peeples University of South Carolina Aiken
Truman Brabham University of South Carolina Aiken
Jamil Khan University of South Carolina
Titan Paul University of South Carolina Aiken