Papers »

Residual performance assessment of steel frames through ambient strain-based model update: application to collapse test of steel frame using shake table

Recent advances in analytical modeling, building sensing system, and hazard simulation facilitate the development of a cyber-physical system which autonomously informs risk in the use of damaged buildings. Quantified damage information with good accuracy has a potential to drastically change the current process of making decisions on the continuous usage or re-occupancy of damaged buildings.

This paper presents a cyber-physical building system developed for post-earthquake damage assessment of steel frames. The location and extent of fracture damage are monitored by a wireless sensing network and an strain-based model updating technique. In the physical side of the system, the structural members suspected to severe damage during large earthquake are instrumented with Polyvinylidene fluoride (PVDF) sensors. PVDF sensors measure dynamic strains of the instrumented members under ambient vibration, and from which the inner force distribution, i.e., bending moment, at dominant modes is estimated. A comparative study of the inner force distribution enables the quantification of fracture damage with the application of a damage decoupling algorithm proposed by the authors. In the cyber-side of the system, a simple building model is developed in OpenSees platform with adjustable damage elements inserted to the members instrumented in the physical side. Then, the damage elements are updated accordingly as quantified fracture damage information being streamed after earthquake events.

The performance of the developed cyber-physical system is investigated through the application to the shake table testing of a steel frame specimen. The specimen is designed to accommodate fracture at beam-ends and column bases. The specimen is first damaged by an assumed mainshock ground motion and the analysis model is updated with damage information. Then, the performance against aftershocks is estimated. The test continues with sequential aftershocks and the model is updated accordingly until the specimen collapse. Finally, the paper discusses the needs for further research developments.

Author(s):

Masahiro Kurata    
Kyoto University
Japan

Akiko Suzuki    
Kyoto University
Japan

Xiaofua Li    
Kyoto University
Japan

Hiromichi Nishino    
Kyoto University
Japan

Kaede Minegishi    
Kyoto University
Japan