Spectroscopy-Based Machine Learning Approach to Predict Engine Fuel Properties of Biodiesel

Abstract

Various feedstocks can be employed for biodiesel production, leading to considerable variation in composition and engine fuel characteristics. Using biodiesels originating from diverse feedstocks introduces notable variations in engine characteristics. Therefore, it is imperative to scrutinize the composition and properties of biodiesel before deployment in engines, a task facilitated by predictive models. Additionally, the international commercialization of biodiesel fuel is contingent upon stringent regulations. The traditional experimental measurement of biodiesel properties is laborious and expensive, necessitating skilled personnel. Predictive models offer an alternative approach by estimating biodiesel properties without depending on experimental measurements. This research is centered on building models that correlate mid-infrared spectra of biodiesel and critical fuel properties, encompassing kinematic viscosity, cetane number, and calorific value. The novelty of this investigation lies in exploring the suitability of support vector machine (SVM) regression, a burgeoning machine learning algorithm, for developing these models. Hyperparameter optimization for the SVM models was conducted using the grid search method, Bayesian optimization, and gray wolf optimization algorithms. The resultant SVM models exhibited a noteworthy reduction in mean absolute percentage error (MAPE) for the prediction of biodiesel viscosity (3.1%), cetane number (3%), and calorific value (2.1%). SVM regression, thus, emerges as a proficient machine learning algorithm capable of establishing correlations between the mid-infrared spectra of biodiesel and its properties, facilitating the reliable prediction of biodiesel characteristics.