Abstrait

Statistical inferential models for flexible polyurethane foam properties

O.O.Ogunleye, F.A.Oyawale

Models of foamproperties in terms of rawmaterialmix that forms the thrust of this paper is uncommon in the foamindustry. Eighty sets of data on foam formulation and physical propertieswere collected fromfour certified companies using central composite design (CCD). To verify data reliability, ten of the mixes were randomly selected, physical properties value determined and compared to the existing formulation using t-statistics. Density (P1); compression set (P2); elongation (P3); hardness-index (P4) and tensile strength (P5) were formulated as functions of the rawmaterials (toluene-diisocynate, water, amine, silicone-oil, stannous - octoate and ethylene-chloride.) using regression analysis. The mean density, compression-set, elongation, hardness-index and tensile strength of the data verification sample were 23.40kgm-3, 8.46%, 170.01%, 150.13Nand 117.54kNm-2, respectively, while 23.26kgm-3, 8.49%, 169.26%, 150.14Nand 117.48kNm-2 for the existing foams. Both setswere not significantly different (p<0.05).The functions P1, P2, P3, P4, P5 with respective standard errors of 0.539, 0.097, 0.989, 0.987 and 0.513 were not significant (p<0.05) while the coefficients of determination were 0.983, 0.898, 0.997, 0.976 and 0.896 respectively. These models are useful for making optimal decisions under various economic conditions for foam production.