Application of Genetic Algorithm for Increasing the Utilization of Poor-quality Biodiesel through Blending

Abstract

Biodiesel is a viable alternative to petroleum diesel. The properties of the biodiesel depend on the feedstock used to produce it. A significant difference in properties exists between different biodiesels. Therefore, standards for biodiesel fuel had been developed considering many factors such as safe handling, corrosion, ignition quality, stability, cold flow property, and performance. For using biodiesel as fuel, the properties of the biodiesel should be within the limits specified in the standard. Unfortunately, biodiesel produced from many feedstocks does not comply with the specifications for all the properties. To utilize biodiesel with poor quality, biodiesels can be blended so that the properties of the blend comply with the specifications. Determining the optimal biodiesel blend ratio experimentally requires a lot of effort particularly when the number of parameters to be optimized is more and when the number of constituent biodiesels in the blend is more. In this work, the application of non-dominated sorting genetic algorithm to predict optimal blends for different scenarios is demonstrated.