A Newton-Krylov method for solid mechanics

In this article a Newton-Krylov method is discussed, which is part of the inexact Newton's methods family. They combine iterative solution methods, especially Krylov methods (conjugate gradient, GMRES, etc.) with the Newton's method, whereas, in solid mechanics, direct solvers are often preferred for the solution of linearised systems. After having introduced a versatile and robust preconditioner, we discuss the efficiency of this approach on a highly non-linear industrial problem.