Accurate prediction of the frequency response functions of tobacco composite panels using experimental modal analysis and model updating

Accurate vibration prediction of a Tobacco Composite Panel (TCP) using finite element (FE) analysis requires a physically accurate representation of both stiffness and damping. While stiffness governs natural frequency locations, Frequency Response Function (FRF) resonance peak amplitude and bandwidth are predominantly controlled by modal damping, and aligning resonance frequencies alone does not ensure accurate FRF prediction. Conventional FE model updating approaches largely focus on stiffness correction and introduce damping through proportional assumptions or nominal modal values, which frequently produces acceptable frequency agreement but persistent errors in FRF peak magnitude and bandwidth. This limitation is particularly pronounced in bio-based composites such as TCP, where damping is inherently high, variable between modes and specimens, and cannot be adequately represented by classical low-damping assumptions. The primary novelty of this study is the development of a staged FE model updating methodology that explicitly decouples stiffness updating from modal damping calibration, providing a physically traceable and reproducible framework for improving FRF prediction accuracy in composite structures with significant damping variability. The framework is developed and validated through experimental modal analysis (EMA) of tobacco composite strip specimens under approximate free-free boundary conditions, supported by quasi-static tensile testing for stiffness comparison and verified using steel dog-bone specimens as a controlled reference. Numerical models are constructed in PATRAN and solved using MSC NASTRAN for modal and FRF analyses, with model updating implemented in FEMTools. The updating sequence comprises four traceable stages, namely an initial FE model, a tensile-test-based stiffness model, a modalupdating-based stiffness model, and a final damping-updated model. In the stiffness stage, elastic parameters are updated to reduce natural frequency error to below 0.52% and improve mode shape correlation using the Modal Assurance Criterion. In the damping stage, EMA-derived modal damping ratios are systematically assigned and selectively refined mode by mode to improve FRF peak amplitude and bandwidth while preserving the aligned resonance locations. Applying this methodology, the final updated TCP FE model achieves Signature Assurance Criterion values of 0.88 to 0.90, representing a substantial improvement from the baseline value of 0.27 and confirming strong global FRF agreement. The proposed methodology resolves the stiffnessdamping coupling ambiguity that undermines conventional updating approaches, and establishes a technically defensible route for FRF-based FE model updating of biobased composite structures where damping variability is a defining characteristic.