Finite Element Model Updating and Damage Prediction for Historical Timber Buildings: Case Study of a Pavilion in Kai-Yuan Temple, China

Historical timber buildings have sustained various forms of damage over extended periods due to material degradation, wind loads, and seismic actions. Establishing an accurate finite element (FE) model and predicting the damage progression in such structures remain significant challenges. Therefore, this study proposed a model-updating method for historical timber buildings. The existing damage in historical timber buildings was captured using three-dimensional (3D) laser scanning technology. Subsequently, the initially established FE model was updated by response surface (RS) optimization and the multiobjective genetic algorithm (MOGA). Additionally, a structural damage prediction method for historical buildings was introduced based on the updated model. The proposed methodology was applied to a pavilion located in Kaiyuan Temple, Fujian Province, China. An FE model considering existing damage was established and updated in ANSYS, enabling an accurate assessment and prediction of damage progression. The model-updating process demonstrated significant improvement, with average discrepancies in mortise-tenon separation and structural inclination between the FE model and the actual structure reduced by 79% and 69.5%, respectively. Particularly, the maximum percentage discrepancy decreased from 25.5% to approximately 5%. Furthermore, the continuous increase in the uneven settlement of foundation has a great impact on the safety of the pavilion. Therefore, damage progression was predicted by simulating increased uneven foundation settlement. The results indicated that both mortise-tenon separation and structural inclination would exceed the critical limit value when the pillar settlement exceeds 2.5 mm. This study provides a robust and practical method for model updating and damage prediction in historical timber buildings, offering valuable insights for their preservation and structural assessment.