Biomass higher heating value (HHV) modeling on the basis of proximate analysis using iterative network-based fuzzy partial least squares coupled with principle component analysis (PCA-INFPLS)

In this study, a novel iterative network-based fuzzy partial least squares coupled with principle component analysis (PCA-INFPLS) was proposed to predict the HHV of biomass fuels as a function of fixed carbon (FC), volatile matter (VM), and ash content. In this methodology, the PCA analysis was used to eliminate the co-linearity of experimental data for providing the required background to the INFPLS model. In the INFPLS structure, adaptive network-based fuzzy inference system (ANFIS) was applied to correlate the inputs and the outputs of iterative PLS score vectors. Furthermore, the capability of the PCA-INFPLS approach in estimating the biomass fuels HHV was compared with those of the PLS, ANFIS, NFPLS, and INFPLS models. Generally, the PCA-INFPLS approach was much more efficient than the other applied methods in modeling the biomass fuels HHV. More specifically, the developed model predicted the HHV of biomass fuels with an R-2 > 0.96, an MSE < 0.51, and an MAPE < 2.5%. Therefore, this approach could be utilized for reliable and accurate approximation of the HHV of biomass feedstocks based on the proximate analysis instead of lengthy laboratorial measurements. The PCA-INFPLS approach was then embedded into a simple and user-friendly software for estimating the biomass fuels HHV based on the proximate analysis.