Developing a Prediction Model for the Blending Level and Properties of Biodiesel-Diesel Blends Using Near-Infrared Spectroscopy Method

Abstract

Biodiesel is an oxygenated, sulfur-free, biodegradable, and non-toxic alternative fuel that can be derived from renewable resources, such as vegetable oils, animal fat and waste cooking oil. However, biodiesel made from different source materials usually have different physical and chemical properties, and the concentration of biodiesel in biodiesel-diesel blends varies from pumps to pumps and from users to users, which all have significant impacts on engine performance and efficiency, and are a challenge to biodiesel quality control as well. In this study, the Near-Infrared Spectroscopy method was used for relatively inexpensive and rapid on-line measurement of the concentration and properties of biodiesel in biodiesel-diesel blends. Methyl esters of five different oils were tested in this study, the soybean oil, canola oil, palm oil, waste cooling oil, and coconut oil. Two different brands of commercial grade No. 2 on highway diesel and one brand of off-road No.2 diesel fuels were used to blend with each biodiesel fuel at various concentrations to prepare calibration samples. Specific gravity of each sample was measured using ASTM standard methods. An ASD spectrometer equipped with a Multi-Use Fiber-optic Fixture was employed to scan the fuel samples in the whole NIR range. Two different regression methods were applied to develop a model for estimating the percentage blending level and specific gravity of biodiesel-diesel blends. The models developed have a RMSEP of 2.73 % for biodiesel concentration and a RMSEP of 0.2 % in specific gravity prediction.