Accelerated Lifetime Testing of Thermal Insulation Elements

Abstract

Thermal insulation materials play an important role in the area of energy technology. Thermal insulation elements (TIEs) are used in fields where high quality and high convenience insulation is required as e.g. in buildings. The TIE manufacturing sector is evolving, but not completely mature. Producers and users have an urgent demand for quality control techniques. Hitherto, quality control and service lifetime prediction for TIEs have mainly been considered from a physical viewpoint with strong emphasis on measurement issues. Rigorous statistical approaches are still missing. From a review of the physical models for TIE degradation over time we build a mixed nonlinear regression model of degradation as a function of time and ambient temperature. The model accounts for measurement-to-measurement and for unit-to-unit variation. We investigate inferential techniques for model parameter estimation and lifetime prediction, and we study the design of accelerated experiments on TIEs.