Damage detection of truss structures via a combined SAP2000 OAPI and teaching-learning-based optimization

Sepil L. D., Özgan K., Şimşek S., Kahya V.

International Civil Engineering and Architecture Congress, Trabzon, Türkiye, 12 - 14 Ekim 2023, ss.1283-1298

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.31462/icearc.2023.sme344
  • Basıldığı Şehir: Trabzon
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.1283-1298
  • Karadeniz Teknik Üniversitesi Adresli: Evet

Özet

Purpose: Creating a numerical model of a structure is essential for assessing damage based on model updating. However, this task for complex engineering structures with many elements, such as trusses, requires significant effort. Using commercial software can provide significant convenience in this regard. To address this challenge, a finite element (FE) model updating approach based on the existing commercial software SAP2000 OAPI and TeachingLearning-Based Optimization (TLBO) for assessing damage in complex truss structures is proposed. Study design/methodology/approach: The FE model of the monitored structure is first created in SAP2000 software to simplify the complexities of FE simulation. Then, the damage assessment of the structure is defined as an unconstrained optimization problem. An objective function is constructed as a weighted linear combination of three modal parameters: frequency, Coordinate Modal Assurance Criterion (COMAC), and flexibility. To detect and quantify any stiffness degradation induced by damage, the optimization problem is solved by TLBO. The iterative optimization task is automatically performed by establishing a link between MATLAB and SAP2000 via the OAPI feature of SAP2000. Findings: The proposed approach is demonstrated using two numerical test examples by considering measurement noise and sparse measured data. Parametric studies for the two different truss structures subject to the assumed damage scenarios are presented. Originality/value: The proposed FE model updating technique successfully integrates commercial FE modeling software with custom research software, facilitating the utilization of modern technology for damage assessment in fullscale structures. This technique holds potential for further development and application in real-world structural health monitoring systems. Keywords: Vibration data; Damage detection; Finite element model (FEM); Teaching-Learning-Based optimization