SMARTCLAVE: A High-Fidelity Digital Twin for Predicting Process-Induced Local Defects and Distortion of Autoclave-Cured Composite Parts

Fiber-reinforced composite components for high-performance applications are typically manufactured using autoclave processing. However, achieving optimal process conditions for high-quality autoclave-cured composite parts remains a challenge due to the complex interplay of thermal, chemical, and mechanical phenomena within the resin-fiber system. Existing tools are limited to global material responses and can only identify defect-prone zones, lacking the resolution to characterize local defects. Moreover, thermal boundary conditions are often oversimplified, relying on region-specific constant heat transfer coefficients (HTCs) or calibrations from limited experimental data. This work introduces SMARTCLAVE, a high-fidelity digital twin equipped with integrated multi-scale, multi-physics sub-models to predict local process-induced material responses (defects and distortion) during autoclave curing. A key feature of SMARTCLAVE is its advanced thermal CFD module, which generates spatially and temporally varying physics-based HTC maps. This drive coupled thermochemical, consolidation, and stress models for downstream finite element analysis. SMARTCLAVE accurately captures 3D evolution of temperature, resin pressure, and degree of cure, as well as local defects including voids, resin-rich zones, and ply waviness, along with mesoscale part distortion. We validate SMARTCLAVE against experimental data obtained from autoclave curing of a tapered beam composite conducted at NIAR, following industry-standard procedures. Additionally, we present the graphical user interface (GUI) of SMARTCLAVE and demonstrate that the tool can effectively complement existing commercial software for global-scale defect prediction, such as COMPRO.