The biomass-based hydrogen production yield prediction model based on PSO-BPNN

Huge energy demand of the society is mainly derived from fossil fuels. However, a large amount of fossil fuel combustion emits enormous amounts of greenhouse gases. To reduce greenhouse gas emissions, carbon-free energy has become a hot research topic. Hydrogen energy is widely regarded as one of the most promising carbon-free energy sources due to its high-grade energy and complete clean combustion. Among the established hydrogen production technologies, biomass-based hydrogen production technology is considered to be one of the most promising process routes due to its abundant raw material sources, low price, less energy consumption, mild reaction conditions, etc. However, the biomass-based hydrogen production process is unstable due to the numerous factors that affect the hydrogen production yield and the unclear mechanism of biomass-based hydrogen production. To better meet the industrial production demand, the biomass-based hydrogen production yield prediction model has been proposed. Predicting biomass-based hydrogen production yield is beneficial for the industry to adjust the key production process parameters, such as material dosage and environmental temperature, to avoid material wastage and achieve online control. Thus, this study proposes a biomass-based hydrogen production yield prediction model based on the particle swarm optimization-backpropagation neural network hybrid algorithm. The proposed model is verified by the data collected from the biomass-based hydrogen production process with a single substrate type and that with multiple substrate types. The prediction results show that the accuracy of the proposed model is very high with an R 2 of 0.99.