ABSTRACT

ABSTRACT: This paper describes a recursive algorithm termed the Modified Windowing Method (MWM) to fit Prony series models to creep data. The proposed method can generate discrete retardation spectra from test data without producing negative spectrum lines. The original Windowing Method (WM) assumes the equilibrium compliance of the Prony series representation to be equal to the largest data point. In that case, some errors may be introduced in calculation of spectrum lines from the data that do not cover the equilibrium region. In contrast to the WM, the MWM provides a more appropriate estimated value for the equilibrium compliance by fitting a straight line to the data in the last modelling window on the logarithmic time scale, thus resulting in a better fit. The effectiveness of the MWM was demonstrated through a numerical example in which the two algorithms were both applied to data read from a mathematical model for creep compliance. Finally, the MWM was applied to the Bending Beam Rheometer (BBR) data for determining the discrete retardation spectra of three asphalt binders.