Blending biomass-based liquid biofuels for a circular economy: Measuring and predicting density for biodiesel and hydrocarbon mixtures at high pressures and temperatures by machine learning approach

Circular economy is an efficient approach to deal with the rapidly increasing demand for energy and its environmental impact. It involves switching to renewable and sustainable energy sources. By providing an alternative to traditional energy resources, biodiesel promotes a circular economy paradigm that embraces renewable energies rather than fossil fuels. As an alternative to fossil fuels, liquid biodiesel helps to improve energy efficiency and reduce pollution. Furthermore, blending biodiesel has proven to be economically and environmentally viable. Knowledge of the thermodynamic properties of biodiesel, such as densities and coefficients of expansivity and compressibility, plays an important role in understanding and determining the behavior of the fuel under different operating conditions. The blend densities of Soybean Oil Biodiesel (SOB) with hexane and Waste Cooking Oil Biodiesel (WCOB) with heptane in four volume ratios were experimentally studied. An Anton Paar vibrating tube densimeter HPM DMA has been used for accurate measurements of densities of the biofuels and their blends. Density measurements were performed at temperatures ranging from 298.15 K to 393.15 K, and pressures between 0.1 MPa and 140 MPa. This work also aims to develop accurate prediction models of the properties of biodiesel blends under various conditions, such as changes in temperature and pressure, which are essential for optimizing energy production processes and engine performance. In this context, the dataset was first investigated with Tait Equation of State (Eos) to determine thermodynamic parameters which including isothermal compressibility and isobaric expansivity. It showed generally low standard deviations. Furthermore, these new experimental densities were also modeled with Artificial Neural Network (ANN) as a machine learning method. This paper models proved that the machine learning method is a suitable tool to model the thermo-physical characteristics of liquid fuels. The findings of this study have the potential to provide valuable guidance to engine engineering applications developers regarding operating conditions and the selection of suitable biodiesel for their system.