Property Prediction & Digital Twins
Supervised learning models and hybrid fuzzy logic networks forecast multi-property responses from print parameters, feeding digital twins that provide real-time energy and time insights for sustainable manufacturing.
Hassan et al., Composites Part C (2024) - Section 7: Predictive Modeling
Key Impact Metrics
2.21-3.29%
Prediction error for tensile strength using ANN + fuzzy logic
R² ≈ 0.994
Feed-forward perceptron accuracy across multiple targets
Real-time
Energy/time estimators for cyber-physical twins
Highlighted Studies
| Study | Highlights |
|---|---|
|
ANN + Fuzzy Logic Models
Hybrid Neuro-Fuzzy Systems
|
Predicted tensile strength with 2.21–3.29% error, revealing infill density as the dominant variable over speed or temperature. |
|
Feed-Forward Perceptrons
Multi-Output Neural Networks
|
Achieved R² ≈ 0.994 across tensile strength, surface roughness, build time, and material consumption targets using sigmoid activations. |
|
Digital Twin Estimators
Cyber-Physical Systems
|
Energy and print-time estimators based on multilayer perceptrons and gradient boosting provide live inputs for twins managing sustainability KPIs. |