Analysis of uncertainty-driven thermal histories in fused filament fabrication under material and process variability

Fused filament fabrication (FFF) produces a transient thermal field and shows layer-to-layer variability in the cooling response of deposited material in regions where similar behaviour would be expected. Since thermal history controls inter-layer bonding, residual stress development, and geometry stability, explaining the variability is tied to part quality. Understanding and quantifying these layer-to-layer differences requires a probabilistic mathematical model of the manufacturing process that propagates process-parameter variability and aleatoric material-property uncertainty through the deposition and cooling sequence, which deterministic approaches typically treat as fixed inputs. To develop such a framework, the in-situ fluctuations in deposition, chamber, and build-plate temperatures were measured using infrared thermography and thermocouples, and the resulting histories were incorporated as time-dependent initial and boundary conditions. The variability of temperature-dependent thermal conductivity and specific heat of ABS polymer, quantified through Bayesian modelling, was incorporated into the process model as a spatially correlated random field. The FFF transient heat-transfer problem was simulated in Abaqus/Standard using progressive element activation and a set of user-defined subroutines. The framework predicts a range of local thermal histories consistent with the experimentally measured profiles and supports interpretation of thermal fluctuations in FFF, providing a basis for process control and reliability assessment.