Neural network based performance prediction of medium to high ash Indian coal in Fluidized Bed Gasifier

The environmental awareness has pushed the Indian coal industries to make a transition towards strengthening and emphasizing of energy generation from coal through clean technologies, which has resulted in introduction of gasification as an alternate but promising source of clean energy. The feasibility of a gasification process is driven by two major factors i.e. gas yield and gas composition. Conventionally, these factors are tested, analysed and interpreted for a limited number of feed coal samples for respective gasifiers, where the gasifier performance highly depends on the its controlled environment as well as feed coal’s petrographic (maturity) and chemical (grade) properties. The Fluidized Bed gasifier has an overall advantage of handling high-ash feed coal at homogeneous temperature profile maintained below its ash melting point. Through present study, an attempt has been made to use an Artificial Neural Network (ANN) for learning from feed coal properties and thereafter prediction of gasifier performance in terms of carbon conversion and concentration of gases produced (CO, CO2 and H2). The gas composition and carbon conversion efficiency predicted from the trained MLFN helped in prediction of gasifier performance, which can be very much useful for gasifier design for Indian feed coal at a preliminary stage.