Thanks for you replies.
You were right Xerf, it's worked well for me .
But I would like to ask an other question on this theme:
Performing an ALE, how do you fix the mesh in the space independently of the material flow ?
On the in- and outflows surfaces of my model I defined eulerian boundary surfaces whose nodes are fixed in the in- and outflow direction. It works well.
But on the rest of the part, the mesh moves with the material, causing distortion problems...I defined sliding boudary surfaces all around...but it is the same thing with lagrangian boundarys!