Towards a Digital Twin for Mitigating Void Formation in Autoclave Composite Parts

Abstract High-performance polymeric composites are increasingly used in the design of aircraft structural components; however, susceptibility to manufacturing irregularities, including porosity/voids, remains a primary challenge delaying the implementation of advanced composites in modern aircraft. Voids may be precursors to structural damage significantly affecting structural integrity and remaining useful life of the aircraft. High-performance aerospace composite parts are commonly manufactured from resin-saturated pre-impregnated plies of uncured material laid-up over a rigid tool and consolidated and cured in an autoclave. In many applications, especially in thick and curved composite sections common in aerospace designs, autoclave consolidation alone is not sufficient to remove the air, or bulk, that might have been entrapped within the laminate during the layup process. Vacuum consolidation, or “debulking”, has been a standard practice extensively used by manufacturers to reduce the amount of bulk prior to autoclave curing. The debulking process typically takes long time and requires intensive manual operations. The underlying physical principles governing the formation and evolution of voids during these early stages of the manufacturing process are not yet well understood, and controlling parameters remain largely determined empirically or based on prior experience. Driven by the need to improve such understanding, there has been a recent interest in using newly available high-fidelity non-destructive inspection (NDI) techniques, such as X-ray Computed Tomography, for in situ observations of composites internal structure during the early stages of manufacturing. In situ observations are allowing researchers to identify the driving mechanisms involved during defect formation, and develop improved predictive models. The objective of this work is to show that high-fidelity NDI data and new developments in numerical modeling might be combined into the creation of a Digital Twin for mitigation of void formation in composite parts. In particular, X-ray CT data is used to extract bulk content and distribution in uncured carbon-epoxy curved-beam specimens after manual layup, and such information is transferred into a finite element (FE) model for simulation of debulking. The FE model uses a fracture-based approach that relies on pore-pressure cohesive zone modeling recently proposed for the discrete representation of entrapped air pockets in uncured resin-saturated prepregs. Preliminary results support a promising ability of the digital twin concept for optimizing the debulking process of autoclave composites towards mitigating void formation.