STRUCTURAL DESIGN OPTIMIZATION FOR VIBRATION CONTROL OF DIAGRID STRUCTURES USING A NEURAL DYNAMIC MODEL

Diagrid systems are a unique and attractive structural system with a geometric configuration that has high structural efficiency and aesthetic flexibility, signature building structures with diagrid systems include the Hearst Tower in New York, USA, the former 30 St. Mary Axe in London, UK and the ModeGakuen Cocoon Tower in Tokyo, JP. However, there is a lack of studies about the seismic performance and vibration control of diagrid building structures. A building is equipped with damping devices to mitigate the structural dynamic response against high-intensity earthquakes. Control systems are usually designed after the structural design is determined, there are few studies that perform Control Co-Design, which consists of an integrated design where the building and the control system are established simultaneously. In this paper, the optimal weight of a building with an integrated control system is solved using a soft-computing optimization algorithm denominated hybrid Counter Propagation Neural Dynamic model (CPND) that determines member sizes on design specifications from AISC-15 and ASCE7-16 when the building is subjected to gravity loads and a series of high-intensity synthetic earthquakes. This integrated design approach is executed for three different scenarios: structure without control, structure with passive control, and structure with semi-active control. The proposed methodology is evaluated using diagrid building structures of different heights with an integrated semi-active or passive control system. The reduction of material in the diagrid elements of each building shows the efficiency of the proposed method for the integrated design of structures.