Sustainable composite from furfuryl alcohol and wood flour with outstanding fire resistance and its prediction using neural networks

Abstract

Novel composites were successfully produced using renewable green sources, furfuryl alcohol resin, commonly obtained from biomass, and wood flour. Compared with a conventional melt-blending technique used for the preparation of wood-plastic composites, this unique approach, utilizing low-viscosity thermoset resin with high affinity for wood, enables the avoidance of excessive treatment of wood flour. Four flame retardants possessing different flame-retardant mechanisms (expandable graphite (EG), ammonium dihydrogen phosphate (ADP), Exolit OP560 and dimethyl propane phosphonate) at two loading levels (7.5 and 15 wt%) were used to suppress the flammability of the composites evaluated by a cone calorimeter test, limiting oxygen index and UL 94. All investigated flame retardants significantly reduced maximum value of heat release rate (HRR) (EG and ADP approx. up to 75 %) and, moreover, ADP and EG significantly reduced the total smoke production (EG up to 25 % and ADP up to 96 %) confirming outstanding and unusual flammability suppression considering HRR reduction and a decrease in smoke production rate (SPR) at the same time. Besides that, the neural network prediction models for HRR and SPR from test time and mass loss rate were created and trained, giving the possibility to predict HRR and SPR values from simple and cheap tests, providing only mass loss rate at specific conditions.