The tangent matrix T does not have to be exact; an approximate T can also provide the solution but it may take more iterations. In any case, the residual vector will be computed using the definition R...

The tangent matrix T does not have to be exact; an approximate T can also provide the solution but it may take more iterations. In any case, the residual vector will be computed using the definition R = KU − F and convergence is declared only when the residual vector R or the solution increment ∆U is sufficiently small as indicated in Eqs.