In order to validate a cross section, the balancing theory is commonly used [1]. Such a processing is limitated by the fact that the balancing theory is valid only if the cross section orientation is in the same direction than the main constraint. Such a limitation would disappear in a 3D balancing process.