Optimizing preparation of compound modified Na-montmorillonite/rubber composites based on radial basis function neural network and its reinforcement-flame retardant properties

With Na-montmorillonite as the research object, mixture of cetyl trimethyl ammonium bromide and absolute ethyl alcohol, mixture of silane coupling agent KH550 and absolute ethyl alcohol as modifier to prepare compound modified Na-montmorillonite, compound modified Na-montmorillonite/rubber composites were prepared by accelerant, sulfur, zinc oxide, stearic acid, rubber and carbon black partially substituted by compound modified Na-montmorillonite. The preparation process parameters of compound modified Na-montmorillonite/rubber composites were optimized by radial basis function(RBF) neural network, compound modified Na-montmorillonite and compound modified Na-montmorillonite/rubber composites were tested and characterized. The results show that RBF neural network of reinforcement properties and the flame retardant property of compound modified Na-montmorillonite/rubber composites have the best approximation effect when "spread coeficient" is 0.50-0.65. Compound modified Na-montmorillonite/rubber composites show good reinforcement properties and flame retardant property (tensile strength of 19.1 MPa, tear strength of 43.5 kN/m and limit oxygen index of 32.83%) when the amount of deionized water is 1 074 g, amount of cetyl trimethyl ammonium bromide is 13.7 g, amount of absolute ethyl alcohol is 14.8 g, amount of silane coupling agent KH550 is 0.32 g and stirring rate is 2 890 r/min.