Thermal shock analysis of the BNL/AGS mercury target experiments

The Spallation Neutron Source is a 1-MW accelerator-based pulsed neutron source facility being designed by several national laboratories and led by Oak Ridge National Laboratory. Liquid mercury will be used as the target material. Several important issues must be resolved to ensure the feasibility of mercury as a target material. One of the key issues is related to the thermal shock loads caused by rapid energy deposition of the proton beam to the mercury target. The rate of temperature rise is enormous (∼107 C/s) during the very brief beam pulse (∼0.5 micros). The pressure waves will interact with the target structure walls and the bulk flow field. Understanding and predicting propagation of pressure pulses are critical for establishing the feasibility of construction and safe operation of the facility. To resolve such issues, a benchmarking database needs to be developed to validate computational models calculating the pressure wave behavior in the mercury target. Collaboration was initiated to conduct experiments wit h a close- to full-scale cylindrical vessel filled with mercury. Specific experiments were conducted at the Brookhaven National Laboratory's Alternating Gradient Synchrotron (AGS) in June of 1997. In these experiments, a high-energy (24-GeV) proton beam deposited ∼61% of its energy in the mercury target over a time period of ∼0.1 micros. Several optical strain gauges were attached to the surface of the steel target wall. The proton pulse shape was roughly parabolic and was estimated to be of ∼0.05 m in radius. Computational models were developed and validated against the AGS data. The ANSYS/Multiphysics code system modeled the test target in Ref. 3, using acoustic elements for mercury fluid and solid elements for the target wall. The authors discuss a two-dimensional model of the ANSYS/Multiphysics code for the same test target. In this effort, solid elements were used to model the liquid mercury without allowing shear between these mercury elements. The ANSYS code uses the implicit Newmark time integration method, coupled with Newton-Raphson solution techniques to obtain the time histories. Materials were assumed to be homogeneous and isotropic, and viscous dissipation was neglected in the calculations. Figure 1 shows a comparison between the calculation and the data. Also, a comparison against the CTH results is included in the same figure