| Summary: | In the current thesis, we attempt to understand the linear viscoelastic behaviour of S2 sea ice by studying its compliance function. Descriptions of the Power Law and the Generalised Kelvin models are provided. Starting with these analytic models, the compliance function was approximated with certain numerical techniques; truncated singular value decomposition (SVD), Tikhonov regularisation, and nonnegative least squares method were examined amongst all in this work. The approximate forms of the compliance function contain parameters that were determined by investigating strain response of ice to some given loading cases. Once the parameters were established, the numerical techniques that produced them were evaluated based on how well a resulting compliance function predicts experimental strain response. It was found that the truncated SVD gives the most reliable results. The test data for this project was extracted from creep tests and loading/unloading creep cycles at various stress rates and frequencies. Generally, the numerical methods were shown to determine the parameters of the approximate forms of compliance function rather well.
|
|---|