The study of this problem was the starting point of a three-year contract (1990)(1991)(1992)(1993) sponsored by the French Ministry of Research & Technology (MRT), and involving several laboratories specialized in NDT techniques. Since such an application was a fairly new NDE topic, the aim of this work was to study the basic fatigue behavior of several materials, as related to NDT techniques (directly taking into account industrial problems would have led to confusing results, since other

more »

... ior effects can influence the results). This communication will focus on the results IRSID obtained with the Barkhausen noise method, and will refer to other laboratOlies having used the same technique in this contract, such as INSA Lyon and EDF, the results of which have, at least partially, already been published [1] [2] . In a second step, the difficulties to overcome, especially with the instrumentation, in order to achieve a full-industrial testing, will be explored. FATIGUE TEST CONDmONS These tests concel1led two different kinds of steels: -a soft steel A48P2, containing 0.15 % C and displaying a felTito-peariitic structure, used for steam generator pipes. an alloyed steel 20CDV5 (0.20 % C+ Cr -Mo -V), with a martensitic structure, and mostly encountered in steam turbines clements (e.g. rotors or bolt~). The fatigue tests consisted in tension/compression cycling. Mainly low-cycle fatigue was studied, with a constant applied strain. Two levels of strain were tlied for the A48P2 steel: E = ± 0.2 % and E = ± 0.5 %. As for the alloyed 20CDV5 steel, only one high level was used: E = ± 1 %. Furthermore, some high-cycle fatigue was performed on the 2OCDV5 steel, applying a constant stress at levels lower than the yield strength (685 MPa for this material).