Abstract: Refractory multi-principal element alloys (RMPEAs) have favorable engineering application prospects due to their exemplary mechanical properties. However, there is a dearth of knowledge regarding the dynamic properties of RMPEAs, which constrains the material design of RMPEAs considering impact performances. To address this issue, in this study, the dynamic compression and spallation behavior of a single-phase body-centered cubic (BCC) TiZrNb RMPEA at impact velocities of 381-723 m s^-1 via single-stage gas gun plate impact experiments was investigated. The Hugoniot parameters were c₀=4.162 km s^-1 and s = 1.005, with a spall strength of 2.18-2.41 GPa. Microstructural analysis showed that spallation damage primarily involved a mix of intergranular and intragranular cracks. Dynamic deformation was mainly controlled by dislocation cross-slip and shear bands (SBs), with the Laves phase inducing localized stress concentrations that promoted void coalescence and reduced spall strength. Moreover, a quantitative relationship between the valence electron concentration, the atomic mass, the impact pressure and the shock bulk modulus was established, through the cold-energy mixture theory and a Particle Swarm Optimization-Back Propagation Neural Network (PSO-BPNN) model, which combined the theory of mechanics and the artificial intelligence algorithm, offering key insights into the materials design for the impact performance of MPEAs.

Key words: Shock compression, Refractory multi-principal element alloy, Dynamic deformation, Spallation mechanism, Artificial neural network