Ballistic limit equation-constrained and data-driven residual protective performance assessment of Whipple shields with pitting damage

Micrometeoroid and orbital-debris (MMOD) impacts pose a critical hypervelocity threat to crewed spacecraft. Although Whipple shields disperse impact energy and generate pitting damage on the rear-wall to avoid perforation, repeated subcritical impacts on pitted rear-wall can aggravate distributed damage and cause degradation of residual protective performance (RPP). Online RPP quantification of Whipple shields with such latent damage is therefore essential for in-orbit inspection and risk-informed maintenance, yet remains challenging under high-dimensional complex type damage fields and small-sample availability. In this paper, a physics-data-driven RPP evaluation framework is proposed. RPP label fields are generated from a physics-based semi-analytical ballistic-limit equation (BLE), and 180 high-fidelity simulation-derived hypervelocity-impact (HVI) cases are constructed to represent representative debris-cloud induced pitting damage states. Using in-situ acoustic-emission (AE) signals as input data information, four architectures—CNN, U-Net, Transformer, and VQ VAE-based surrogate model (VQ-SM)—are benchmarked in terms of global-field fidelity, weak-zone localization, weak-zone area error, critical minimum-RPP error, extrapolation capability, and BLE-guided regularization. U-Net provides the most accurate global RPP reconstruction and minimum-RPP prediction, whereas VQ-SM better preserves weak-zone masks. BLE guidance yields regime-dependent, rather than universally positive, gains. This framework directly maps post-impact sensor data information to BLE-constrained residual-capacity fields, supporting risk-informed maintenance decisions for MMOD-induced pitting-damaged Whipple shields.