An extended physics-informed neural operator for accelerated design optimization in composites autoclave processing

Composite materials have become indispensable in aerospace, automotive, and marine industries due to their exceptional mechanical properties. Among these, thermoset composites manufactured in autoclaves require precise control over temperature and pressure profiles. Optimizing the cure cycle and equipment design parameters is crucial to attain the desired properties in the manufactured part. Traditional optimization methods require substantial computational time and effort due to their reliance on resource-intensive simulations, such as finite element analysis, and the complexity of rigorous optimization algorithms. Data-agnostic AI-based surrogate models, such as physics-informed neural operators, offer a promising alternative to conventional simulations, providing drastically reduced inference time, unparalleled data efficiency, and zero-shot super-resolution capability. However, the predictive accuracy of these models is often constrained to small, low-dimensional design spaces or systems with relatively simple dynamics. To address these challenges, we propose an accelerated gradient-based optimization framework powered by a novel neural operator called the eXtended Physics-Informed Deep Operator Network (XPIDON). The proposed architecture ensures accurate predictions across large, high-dimensional design spaces and nonlinear dynamical regimes. This is achieved through temporal domain decomposition, input coordinate normalization in subdomains to mitigate spectral bias and nonlinear decoding to better capture complex physical behaviors. As an efficient, differentiable surrogate, XPIDON enables near-real-time spatiotemporal predictions for arbitrary design conditions. Our end-to-end framework, which combines XPIDON with a gradient-based optimizer (Adam), improves the predictive performance by 50% compared to existing neural operators and yields a $3\times$ speedup over gradient-free approaches in obtaining optimal design variables for composites autoclave curing processes.