The Augmented Lagrangian method is used to iterate several times over a single time step, to correct contact forces (augmentation) to get a gap $g$ smaller than a given tolerance (defined by the parameters PRECONNORMALGAP and PRECONTANGENTGAP of the ElementProperties). Contact forces at the iteration $i$ are written :
$$ F = L_i + k_N\,g $$
where $k_N$ is the penalty and $L_i$ the Lagrangian. For the next augmentation, the predicted Lagrangian is
$$ L_{i+1} = L_i + k_N\,g $$
or it is extrapolated. The gap tolerance is defined by the properties of the contact element. This accuracy must be given to activate the method.
Parameters are set by functions associated to metafor.getAugmentedLagrangianManager() :
alm = metafor.getAugmentedLagrangianManager() alm.setMaxNbOfIterations(itma)
where
itma | maximal number of augmentations for a given time step (by default equal to 10) |