Advanced experimental tools designed for the assessment of the thermal load applied to the mixing tee and nozzle geometries in the PWR plant

Thermal fatigue studies have been started again, after the incident in the heat removal system of the Civaux NPP (may 1998). Thermal fatigue problem was suspected, the cracks occurred in the mixing tee were probably due to the fluctuation at large gap of temperature. It was clear that this domain needed data, the knowledge in the mixing areas should be improved. A wide experimental program supported by AREVA NP, CEA and EDF was launched (FATHER experiment 8 MW within a 160°C delta T ) based on endurance test and thermalhydraulic tests : the first test to restitute on mockup the damage occurred under realistic thermal hydraulic conditions and the thermal hydraulic tests to perform measurements for assessing the thermal load of the turbulent layer which impact the wall. In addition, the CEA commissioned an other experiment, smaller, called FATHERINO (delta T max. 75°C in water under a 4 m/s mixture velocity in a 50 mm diameter mixing tee mockup) where analytical tests would be conducted including tests for the instrumentation qualification. Temperature fluctuation amplitude can be various in the mixing area, and induced various loads in the wall. Thus, the thermal transfer to the wall, managed by the heat transfer coefficient can be various too, that depends mainly from the mixture velocity following the zone. The frequency of the fluctuation, the amplitude, the thermal transfer to the wall represent the basic parameters which have to be studied to assess the load, and consequently the damage initiation time after mechanical calculations. Data can be compared to the existing fatigue curve data. The physical parameters are complex to compute, because temperature signals are basically stationary but fluctuating (the heat transfer coefficient in fluctuating conditions). Following an original strategy, the objectives assigned in the FATHERINO experiment are to determine the areas where the fluctuation is large corresponding to a potential damage, and in a - - second time, to assess the thermal load. The CEA developed two types of experimental tools in order to investigate following the geometries, the zones of interest (high fluctuating temperature) and in a second stage, provide measurements within advanced specific instrumentation to assess the fluctuating temperature, the thermal flux, and the heat transfer coefficient. To investigate the areas of interest, an advanced mockup in brass material called "the skin of the fluid" were designed, particularly adapted to visualize the mean and fluctuating field of temperature using infra red measurements, and practically without attenuation. The Biot number reaches just 0.1, such as requires for a low attenuation of the temperature fluctuation in the fluid. The time histories of some pixels of interest were selected in order to complete the signal processing under the spectral representation. Using the same mockup geometry but in stainless material like the industrial case, the advanced sensors (in same material) designed by the CEA are installed in the locations defined by the infra red measurements. The specific sensors, Tf fluid, fluxmeter sensor, and coefh sensor, record in local, the fluctuation close to the wall in the fluid and in the wall. The heat flux sensor "fluxmeter" and the heat transfer sensor "coefh" are equipped with 3 micro thermocouples in their respective body, non intrusive and typically designed to catch the fluctuations within a low attenuation in the frequency range from ~ 0 to 20 Hz. By applying an inverse heat conduction method to the output data given by the fluxmeter, the wall temperature (mean and fluctuating values) at the internal surface can be accurately determined. The coefh sensor is like a fluxmeter sensor using the same technology but equipped with a thermocouple in the fluid to determine the heat transfer coefficient. In addition, the results from both experiments (brass and stainless steel mockups) are implemented as input data for the