Tuesday  |  Salon 9  |  09:20 AM–09:40 AM

#15673, Identification and Monitoring of Damage Scenarios under Cyclic Loading of a Coated 300M Steel by Thermography Measurements

Determining a full S/N curve is time consuming and expensive. Understanding and modelling of fatigue mechanisms are the first steps to build a fatigue lifespan model and speed up the fatigue properties characterization. The aim of this paper is to implement and validate a test and post-processing protocol to detect and monitor crack propagation of a coated steel. 300M steel, shot peened and WC-CoCr coated, is chosen as model material for the study.
With high temporal and spatial resolution temperature fields, Infra-red (IR) measurements provide quantitative information on steel integrity under the coating. Under a sufficient cyclic stress level, a network of regularly spaced parallel crack rings can be detected after few hundred cycles. Fracture surfaces show a multi-initiation of the steel from the interface, a multi propagation through the shot peened area and a quick rupture when a crack reaches the bulk.
Cracks can be qualitatively studied but because of the non-adiabatic test condition, the thermal signal on the surface corresponds to a volumetric average of the coating and substrate thermo-elastic coupling. Two indicators are chosen to detect cracks (the first harmonic amplitude and the skewness) due to their sensitivity. To quantitatively interpret the signal, a non-linear thermo-mechanical finite (FE) elements model is needed. The first step is to calibrate the FE model to correlate the IR measurements. With the large number of unknown parameters (materials parameters and configuration characterization), only a good qualitative correlation is obtained as both indicators are sensitive to every FE model parameter. Indeed, the FE model shows a significant effect of the crack depth on indicators, meaning a potentially feasible crack monitoring.
Fatigue tests are performed and monitored by IR camera. Films are recorded regularly, so indicators can be computed with the number of cycles. Evolution of indicators is confirmed during tests near the rupture location. To correlate the temperature signals with crack depths, more specific indicators are built. The FE model is calibrated with those new indicators on well-known configurations. For other configurations, the FE model links the indicators to the crack depth. Combined with the test indicators, the crack depth might be plotted according to the number of cycles.

Pierrick Lepitre   ENSTA Bretagne
Sylvain Calloch   ENSTA Bretagne
Matthieu Dhondt   ENSTA Bretagne
Martin Surand   Safran Landing Systems
Cédric Doudard   ENSTA Bretagne