Identification of natural rubber samples for high-voltage insulation applications

Abstract

Latex presents high variability due to inherent differences among varieties from different countries, producers or crop seasonality. Natural rubber formulations from natural latex, to be used in insulating materials intended for high-voltage applications, require a wide variety of compounding and multitude of industrial processes. These aspects make it very difficult ensuring the same dielectric properties of the final product. At manufacturing level, it is very important to apply strict control processes to ensure that the final product fulfills all quality specifications. In this paper, a promising approach was applied to automatically identify natural rubber samples with suitable dielectric behavior from those with unsuitable dielectric behavior. This approach is based on the study of FTIR spectral data by applying suitable multivariable methods, such as principal component analysis, canonical variate analysis and k-nearest neighbors. The accurate and fast results reported in this work prove the suitability and potential of the proposed approach.