Automated identification of linear viscoelastic constitutive laws with EUCLID

We extend EUCLID, a computational strategy for automated material model discovery and identification, to linear viscoelasticity. For this case, we perform priori selection by adopting generalized Maxwell expressed Prony series, deploy EUCLID identification. The methodology is based on four ingredients: i. full-field displacement net force data; ii. very wide library — in our large number of terms the series; iii. momentum balance constraint; iv. sparsity constraint. devised comprises two stages. Stage 1 relies sparse regression; it enforces data exploits sparsity-promoting regularization drastically reduce series identify parameters. 2 k-means clustering; starting from reduced set stage 1, further reduces their grouping together elements with close relaxation times summing corresponding moduli. Automated procedures are proposed choice parameter clusters 2. overall demonstrated artificial numerical data, both without addition noise, shown efficiently accurately viscoelastic five across orders magnitude, out several hundreds spanning seven magnitude.