In a PWR the reactor coolant flow that goes through the reactor internals and the fuel assemblies is characterized by high turbulence and this flow is able to induce some structural vibration. A few years ago, some nuclear power plants were obliged to shut down for many months, due to the heavy damage caused by vibration. The design of reactors must be carefully checked taking into account the possible interaction between hydraulic excitation and reactor structure response. The reactor assembly of a PWR consists of: (1) a reactor vessel which withstands the internal pressure of the primary fluid and maintains the reactor core; (2) reactor internals which maintain fuel assemblies, guide the control rods and wear a thermal shield in order to reduce the fast neutron exposure of the reactor vessel wall; and (3) fuel assemblies and control rods. The SAFRAN test loop consists of a reduced-scale ( 1 8 ) model of a reactor vessel, reactor internals, dummies representing fuel assemblies and a system of three loops including pumps and damping tanks connected to the reactor vessel, the purpose of which is to simulate the flow distribution of a three-loop PWR. The scaling laws for designing the model and the test loop are: same geometry and attachment conditions; same flow velocity: V model = V reactor; same Cauchy number, i.e. same ratio of inertia forces to stiffness forces; and same Euler number, i.e. same ratio of inertia forces to pressure forces. Nevertheless, it is not possible to use the same Reynolds number. The ratio between the Reynolds number of the reactor and the Reynolds number of the model, for the same fluid velocity, is 70. This is mainly due to scale ratio and to the viscosity of the fluid in the hot condition. But in most cases, we are above the critical values of Reynolds number where there is a variation of the Strouhal number S = ƒD/V. The measured frequencies in the model will be eight times the frequencies occurring in the reactor. In general, the construction technology used for the model is the same as that used for the reactor. All the structures in contact with the fluid are made of stainless steel. The instrumentation used on the SAFRAN test loop consists of accelerometers, pressure sensors and relative displacement sensors. Vibration phenomena are studied using two different approaches. In the first approach, the vibration properties of the structure are measured by means of tests performed in air and water to obtain, in both cases, frequencies, modes, damping and stiffness values. The hydraulic excitation sources are measured by tests on the loop: frequencies, Δp values, direct- and cross-correlation lengths. During these tests, structures are stiffened in order to prevent their motion. By means of a computer program based on the POWELL method, the structural response can be calculated according to the density of Δp distributed around the structure. The second approach consists of measuring directly the structural response to hydraulic excitations. Comparison of the results given by these two approaches shows: (a) the system non-linearities and (b) the coupling between the fluid and the structure. By using two different approaches a better knowledge of complex phenomena can be gained.