Study of MHD Corrosion and Transport of Corrosion Products of Ferritic/Martensitic Steels in the Flowing PbLi and its Application to Fusion Blanket
Two important components of a liquid breeder blanket of a fusion power reactor are the liquid breeder/coolant and the steel structure that the liquid is enclosed in. One candidate combination for such components is Lead-Lithium (PbLi) eutectic alloy and advanced Reduced Activation Ferritic/Martensitic (RAFM) steel. Implementation of RAFM steel and PbLi in blanket applications still requires material compatibility studies as many questions related to physical/chemical interactions in the RAFM/PbLi system remain unanswered. First of all, the mass loss caused by the flow-induced corrosion of the steel walls at temperatures in the range 450 C -550 C needs to be better characterized. Second, another serious practical concern is the transport of activated corrosion products and their precipitation in the cold section of the loop. Third, an important modeling parameter, the saturation concentration of iron in PbLi, needs further evaluations as the existing correlations demonstrate scattering of several orders of magnitude. Besides, the existing experimental data on corrosion are often contradictive and the underlying physics is not well understood, especially if the PbLi flow is turbulent and strongly affected by the applied magnetic field due to magnetohydrodynamic (MHD) effects in the flowing liquid metal.
The research performed here is aimed at: (1) better understanding of corrosion processes in the system including RAFM steel and flowing PbLi in the presence of a strong magnetic field and (2) prediction of corrosion losses in conditions of a Dual Coolant Lead Lithium (DCLL) blanket, which is at present the key liquid metal blanket concept in the US. To do this, numerical and analytical tools have been developed and then applied to the analysis of corrosion processes.
First, efforts were taken to develop a computational suite called TRANSMAG (Transport phenomena in Magnetohydrodynamic Flows) as an analysis tool for corrosion processes in the PbLi/RAFM system, including transport of corrosion products in MHD laminar and turbulent flows. The computational approach in TRANSMAG is based on simultaneous solution of flow, energy and mass transfer equations with or without a magnetic field, assuming mass transfer controlled corrosion and uniform dissolution of iron in the flowing PbLi. Then, the new computational tool was used to solve an inverse mass transfer problem where the saturation concentration of iron in PbLi was reconstructed from the experimental data resulting in the following correlation: , where T is the temperature of PbLi in K and is in wppm. The new correlation for saturation concentration was then used in the analysis of corrosion processes in laminar flows in a rectangular duct in the presence of a strong transverse magnetic field. As shown in this study, the mass loss increases with the magnetic field such that the corrosion rate in the presence of a magnetic field can be a few times higher compared to purely hydrodynamic flows. In addition, the corrosion behavior was found to be different between the side wall of the duct (parallel to the magnetic field) and the Hartmann wall (perpendicular to the magnetic field) due to formation of high-velocity jets at the side walls. In the blanket conditions, the side walls experience a stronger corrosion attack demonstrating a mass loss up to 2-3 times higher compared to the Hartmann walls. The analysis for a case of a strong magnetic field suggests scaling laws for the mass loss ML in rectangular ducts, which include the effects of the temperature T, mean bulk velocity Um and the applying magnetic field B0: for the side wall, and for the Hartmann wall, where q, s ~ 0.5. As seen from these laws, the mass loss at the Hartmann wall is not affected by a magnetic field providing the magnetic field is high.
Further analysis was performed for corrosion in the Hartmann flow, which is the MHD analog of the hydrodynamic Poiseuille flow. The main goal of the analysis is to elucidate the effect of a magnetic field on the corrosion mass loss in the case when the applied magnetic field is perpendicular to the flow-confining wall. It was found that the corrosion rate is strongly dependent of the ratio between the thickness of the concentration boundary layer and that of the magnetohydrodynamic Hartmann boundary layer. Once the concentration boundary layer becomes thicker than the Hartmann layer, further increase in the magnetic field does not affect the corrosion rate. A self-similar solution for the concentration field was derived for two particular cases: (i) the thickness of the concentration boundary layer is much smaller than the thickness of the Hartmann layer and (ii) the Hartmann layer is much thinner than the concentration boundary layer. The derived solutions comply very well with the numerical data and thus can be recommended for calculations of the corrosion mass loss in fusion applications and also to analyze experimental data.
Analysis of the effect of a magnetic field on corrosion of RAFM steel in a turbulent PbLi flow is performed using numerical simulations. The impact of the magnetic field strength and its direction for this mass transfer problem is analyzed with the aid of a mass transfer equation for dissolved products coupled with the MHD equations. This approach utilizes a special form of the "K-ε" model of turbulence, which takes into account the effect of turbulence suppression by a magnetic field. Computations are performed for three orientations of the magnetic field, with respect to the main flow (streamwise, spanwise and wall-normal B-field) in the temperature range from 400 C to 550 C, which is of particular interest for fusion cooling applications. Changes in the corrosion rate caused by MHD effects have been analyzed with regard to turbulence modification by a magnetic field and to formation of the Hartmann boundary layer at the walls perpendicular to the magnetic field. As demonstrated, for all three magnetic field orientations, decrease of the corrosion rate occurs as the magnetic field increases. However, a wall-normal magnetic field has a stronger effect on the reduction of the corrosion rate compared to the other two magnetic field orientations due to more intensive turbulence suppression. For the case of a wall-normal magnetic field, a correlation for the turbulent dimensionless mass transfer coefficient (Sherwood number, Sh) has been constructed based on the numerical data, which shows the effect of the flow velocity via the Reynolds number (Re) and that of the applied magnetic field via the Hartmann number (Ha): , where Sherwood number in a purely hydrodynamic flow Sh0 is a function of Re.
The developed analytical and computational tools have been used in the calculations of the corrosion mass loss in the poloidal ducts of the DCLL blanket under conditions of the so-called US DEMO reactor. The present analysis is limited to the outboard region of the reactor where the magnetic field is ~ 4 T. One of the goals of the analysis is to establish conditions when a high PbLi temperature at the blanket exit of ~700C needed for high thermal efficiency of the power conversion cycle can be achieved, while the corrosion mass loss is maintained within the allowable limits. At present, the suggested maximum for the corrosion wall thinning is 20 m/yr. The analysis includes parametric studies, using the electrical conductivity of the insulating flow channel insert (FCI) and the PbLi temperature as parameters. Also, more detailed computations have been performed using computed temperature distributions from the 3D MHD/thermofluid analysis. The obtained corrosion data suggest that the most corrosion losses occur in the thin gap between the First Wall and the FCI (side-wall section of the gap), while the corrosion losses in the Hartmann-wall section of the gap are almost negligible due to very low velocities there. Also, the maximum temperature at the interface between the RAFM wall and the flowing PbLi (which guaranties the average wall thinning < 20 m/yr) was estimated at about 470 C. This is consistent with the estimate from a more conservative analysis in the past.