Machine learning approach to predict the biofuel production via biomass gasification and natural gas integrating to develop a low-carbon and environmental-friendly design: Thermodynamic-conceptual assessment.

ABSTRACT

A hybrid energy cycle (HEC) based on biomass gasification can be suggested as an efficient, modern and low-carbon energy power plant. In the current article, a thermodynamic-conceptual design of a HEC based on biomass and solar energies has been developed in order to generate electric power, heat and hydrogen energy. The planned HEC consists of six main units two electric energy production units, a heat recovery unit (HRU), a hydrogen energy generation cycle based on water electrolysis, a thermal power generation unit (based on LFR field), and a biofuel production unit (based on biomass gasification process). Conceptual analysis is based on the development of energy, exergy and exergoeconomic assessments. Besides that, the reduction rate of pollutant emission through the planned HEC compared to conventional power plants is presented. In the planned HEC, when hydrogen energy is not needed, excess hydrogen is feed into the combustion chamber to improve system performance and reduce the need for natural gas. Accordingly, the rate of polluting gases emitted from the cycle can be mitigated due to the reduction of fossil fuels consumption. Further, based on the machine learning technique (MLT), the level of biofuel produced from the mentioned process is estimated. In this regard, two algorithms (i.e., Support vector machine and Gaussian process regression) have been employed to develop the prediction model. The findings indicated that the considered HEC can produce about 10.2 MW of electricity, 153 kW of thermal power, and 71.8 kmol/h of hydrogen energy. In both training and testing sets, the Support vector machine model exhibits better behavior compared the two Gaussian process regression model. Based on machine learning technique, with increasing gasification pressure, the level of biofuel obtained from the process does not increase significantly.