Composite laminates reliability assessment using diffusion process backed up by perspective forms of non-parametric kernel estimators

Abstract

Numerous technical systems are modeled to identify dependability and safety measures. The modeling approach here is a real data driven approach. We work with the results obtained by assessing a progressive material—natural fibers reinforced laminate composites. The knowledge of certain properties, namely reliability performance, has not been sufficiently studied. Our approaches include the Wiener diffusion process, parametric and non-parametric estimations and mainly a hazard function which is estimated parametrically and non-parametrically by kernel estimators. We also demonstrate the significance between parametric and non-parametric estimations, the way they influence the selection of a resulting model and prospective practical applicability.