Forecast of the higher heating value based on proximate analysis by using support vector machines and multilayer perceptron in bioenergy resources

As biomass gets to be more relevant to energy feedstocks, the capacity to anticipate its Higher Heating Value (HHV) by more efficient algorithms from schedule information such as proximate analysis empowers quick choices around utilization in bioenergy. The present work studies a novel artificial smart model based on an interesting algorithm, relied on support vector machines (SVMs) jointly with the grid search (GS) optimizer, for characterization of HHV of raw biomass from parameters ascertained experimentally. Additionally, a multilayer perceptron (MLP) approach was built from the same experimental data for comparison objectives. The results of the current study are the relevance of each physico-chemical parameters on the raw biomass HHV through this novel model and forecasting the HHV of biomass. In this sense, when the novel model was applied to the observed dataset, a coefficient of determination and correlation coefficient equal to 0.8517 and 0.9229, were achieved for the HHV estimation, respectively. The importance of the use of learning machines is illustrated in the evaluation of the energy resources to energy systems to show an efficient algorithm for bioenergy purposes. The concordance between observed data and the GS/SVM-relied model indicated the good efficiency of the second.

Automated summary

This study developed an artificial smart model based on support vector machines and grid search optimizer for predicting and characterizing the Higher Heating Value (HHV) of raw biomass. The results showed that the model was accurate in predicting the HHV of biomass and highlighted the importance of physico-chemical parameters in determining the HHV.