Numerical Simulation of Macro Impact Laser Peening and the Effect of Peening on Fatigue Life and Weld Residual Stresses

Author:  Said Taheri, Emricka Julan, Eric Lorentz
Source:  ICSP-13
Doc ID:  2017133
Year of Publication:  2017
Abstract:  
Introduction: Stress corrosion cracking (SCC) in nickel-based alloy 600 is one of the significant ageing degradations in major components of pressurised water reactors. Components such as steam generator tubes or bottom mounted instrumentation (BMI) in reactor pressure vessels have experienced SCC. Some other components made of austenitic stainless steels subjected to high cycle thermal fluctuations have exhibited thermal fatigue crazing. One common detrimental parameter is frequently the tensile weld residual stress (WRS). Laser peening (LP) is a surface mitigation technique, as is shot peening (SP), for improving the life of metallic components by generating a compressive surface residual stress (RS) field induced by high-power laser pulses. LP has been applied in Japan on several BMI. Numerical simulations of LP is performed by 3D FEM using a high-speed explicit dynamic code named Europlexus. Numerical validation of Europlexus code for laser peening is performed by comparison of RS obtained by Europlexus and several other codes [1,2]. These RS are also compared with measured RS at stabilised states [1]. However simulated RS did not represent a stabilised state due to considerable CPU times which would be needed to obtain it. A Johnson-Cook (JC) law is used for all simulations. Parameters of this law for an In600 are identified [3] at a small strain rate 3 10 on the stress strain curve, at 103 to 5*103 by the Hopkinson bar test and at 6 10 , which is strain rate of an LP operation, by the VISAR velocimetry technique using laser shock and Doppler effects [4]. In the literature, a characteristic representation of RS after impact is given by the stress plot (Sxx or Syy, figure 1a) parallel to the impacted surface versus the depth at just one point (this plane is perpendicular to the direction of impact Z). In fact a likelihood assumption under LP or SP operation is that with a large number of impacts the RS field is homogeneous in the X,Y plane (except near the edge of the treated area). The RS field is thus only dependent on the Z coordinate. Almost total absence of the effect of WRS on RS after LP has been shown in [1] where WRS are approximated by a thermo-visco-plastic simulation. The thermal loading for this simulation is obtained using measured WRS on the surface. In this case however we have no knowledge about the validity of the simulated WRS at depth, so an axisymmetric simulation of welding is performed hier. Fatigue life may be impacted by strain hardening due to peening, for alloys with memory effects [9]. The beneficial effect of peening on the fatigue life of aluminium alloys (aeronautical industry) and ferritic steel (car industry) has been reported extensively in the literature. However the situation is different for In600 and austenitic stainless steels used in nuclear power plants, as these latter alloys have a memory effect of maximum monotonic (or cyclic) strain hardening. Indeed, it has been shown that [7] for 304 and 316 stainless steel, cyclic pre-hardening of 10 cycles at ±2% or 14% monotone strain hardening significantly increases the fatigue lifetime in stress-controlled tests but reduces it in strain-controlled tests (high cycle thermal fatigue). Thus in stress control, the beneficial effect of prehardening due to peening is added to the beneficial effect of compressive stress for fatigue life, while in strain control, the detrimental effect of pre-hardening reduces the beneficial effect of compressive stress. Numerical simulations show [8] that SP impacts may create a plastic strain of about 20% while LP creates a plastic strain of about 2%. This may be due to a higher strain rate for LP of 6 10 against approximately 5 10 for SP. Indeed, in [9] it is shown that, SP has a beneficial effect for 304 stainless steel in strain control at zero mean stress but not at 60 MPa mean stress.