Nuclear industry generates radioactive wastes of various nature such as solids, liquids but also sludges coming from effluent treatment facilities or historical residues stored in pool or tanks. The physico-chemical nature of those sludges leads to a complex flow behaviour making it difficult to handle and convey prior their immobilization in a conditioning matrix. In order to fluidize these suspensions of varying compositions, the mechanical action of power ultrasound is envisaged. It has recently been shown, thanks to a set-up coupling power ultrasound and rheology, that it is possible to significantly reduce the yield stress and viscosity of the slurry by applying ultrasound. The aim of this thesis is to pursue the studies already undertaken (physical chemistry, microstructure, ultrasound and rheology) on reconstituted sludge or simplified model suspensions, focusing more specifically on two aspects. The first, more fundamental, will aim to gain a better understanding of the interaction between power ultrasound and matter, with a particular focus on the origin of the effects observed (interfaces vs. volume). The second aspect will be more applied, with the development of original larger-scale experimental devices capable of generating flows closer to industrial situations. For this phD work, we are looking for a motivated, serious and curious candidate. Given the multidisciplinary character of the subject, mixing physics, physico-Chemistry and experimental development, the candidate could valorize his new skills in various industrial fields such as nuclear, civil engineering and depollution domain.