Combustion of Raw Biogas/Hot Air in a Porous Radiant Burner and Performance Prediction Using a Machine Learning Algorithm

Abstract

Combustion of raw biogas/hot air was performed in a porous radiant burner associated with a solar heater, and performance was predicted by a linear regression model using a machine learning algorithm. The test was conducted for the combustion of three different compositions of raw biogas mixtures having CO2 percentages of 25%, 30%, and 35% at the thermal load of 200–400 kW/m2. The hot air was supplied at an average temperature of 50 °C from the solar heater air supply system for proper combustion in lean mixture conditions. The porous radiant burner associated with a solar heater has offered radiation efficiency of 15.34–47.93%, NOX of 1–3.1 ppm, and CO of 25–87 ppm for three different compositions of raw biogas mixtures at the thermal load of 200–400 kW/m2 and equivalence ratio of 0.70–0.91. The increased radiation efficiency has indicated that the porous radiant burner can be an alternative for low-calorie fuel like raw biogas. Data analysis and processing have been performed using the machine learning algorithm, and the linear regression model has been developed using the python programming language. The error between predicted and experimentally calculated radiation efficiency is 1.67%.