An Advanced Physics-Informed Neural Operator for Dynamic, Zero-shot, and Near Real-time Simulation of Aerospace Composite Material Curing Process
Submission Track: Full Paper
Submission Category: AI-Guided Design
Keywords: neural operators; physics-informed DeepONet; aerospace composites processing; design optimization
TL;DR: We developed an advanced physics-informed DeepONet for design optimization of highly-nonlinear systems with high-dimensional design spaces in composites manufacturing.
Abstract: One of the key prerequisites of AI-guided design for manufacturing advanced materials is the availability of a dynamic, zero-shot and near real-time predictive model for quick and efficient design exploration. Deep Operator Networks (DeepONet) and their physics-informed variants have shown significant promise in imparting such generalization and inference power to the models. However, for highly nonlinear real-world applications like aerospace composites processing, existing models often fail to capture underlying solutions accurately and are typically limited to single input functions, constraining rapid process design development. This paper introduces an advanced physics-informed DeepONet tailored for such highly nonlinear systems with multiple input functions. Equipped with architectural enhancements like nonlinear decoders and effective training strategies such as curriculum learning and domain decomposition, the proposed model handles high-dimensional design spaces with significantly improved accuracy, outperforming the vanilla physics-informed DeepONet by two orders of magnitude. Its zero-shot prediction capability across a broad design space makes it a powerful tool for accelerating composites process design and optimization, with potential applications in other engineering fields characterized by strong nonlinearity.
Submission Number: 7
Loading