A machine learning approach in drying of microalga Chlorella minutissima in a single rotary drum dryer for biodiesel production

Abstract

Drying of Chlorella minutissima slurry was carried out in a single rotary drum dryer at varied drum surface temperature and rotational speed to achieve higher lipid yield for biodiesel conversion. Residual moisture content and lipid yield of the corresponding drum speed and drum surface temperature were mapped. Application of the machine learning tool classified the range of residual moisture content to be <10% (wet biomass) for high lipid recovery with an accuracy of 97%. Based on the drying time, lipid recovery, and energy consumption, drying at 80°C drum surface temperature with 0.3 rpm resulted ˃90% lipid recovery as compared to the bone-dried biomass. The energy consumption of 7.328 kWh for 1 kg of dried biomass was recorded with profound reduction in drying time as compared to oven and tray drying of algal slurry. The fatty acid methyl ester profile of C. minutissima biodiesel constituted higher proportion of saturated fatty acids followed by monounsaturated fatty acid and polyunsaturated fatty acids. Various fuel characteristics, namely, viscosity, density, free glycerin, triglycerides, diglycerides, monoglycerides, copper strip corrosion, flash point, cetane index, iodine value, calorific value, acid value, saponification value, phosphorous, sulfur, water, and ash contents were analyzed, and showed values within the specified limits of national and international biodiesel standards. Thus, considering the drying time, energy consumption, and lipid recovery with suitable fuel properties, the drum dryer could be recommended for drying the microalgal slurry for large-scale operation.