Effect of fuel composition uncertainty on grate firing biomass combustor performance: a Bayesian model averaging approach

Abstract

Biomass has great potential to meet greenhouse gas reduction and fuel supply security in the future. Although the grate biomass combustors are increasingly deployed worldwide to generate energy from solid biomass, challenges in understanding the system operation to some extent have remained. This paper analyzes the effects of fuel composition uncertainty on the biomass grate combustor’s performance, which have not been solved so far. A 1D transient numerical model of the biomass fuel bed combustion is developed. A set of thermal gravimetric analysis (TGA) experiments on randomly selected biomass particles from the same fuel supplier are conducted to achieve the proximate analysis of the particles. The Bayesian model averaging (BMA) method was exercised to deliver the fuel uncertainty into the CFD model of the fuel bed. Results revealed that the fuel composition variability can significantly affect the solid fuel conversion so that ignoring them can result in incomplete combustion. In three various scenarios proposed, combustor is analyzed: (I) using primary fuel composition given by the producer, (II) mean value of fuel composition obtained from the BMA model, and lastly (III) fuel composition under fully uncertainty conditions. Results revealed that overlooking the fuel uncertainty results in overestimating system energy output by 8.3% and also can waste 1611-kg feed annually which is roughly 5% of whole consumed fuel. Meanwhile, owing to uncertainty associated with fuel composition, flame temperature can fluctuate up to 15 °C. According to the uncertainty analysis, char content of wood pellets has dominating role in fuel quality. Finally, a life cycle analysis (LCA) is conducted for the first, second, and coal-fueled system scenarios.