Gradient-based multi-hazard optimization of MTMDs for tall buildings

Abstract

Past research has shown that, from a long-term damage perspective, a multi-hazard approach should often be considered for the design of tall buildings. This also stands for the design of supplemental dissipation devices such as multiple tuned mass dampers (MTMDs). Although MTMDs present several advantages over conventional TMDs, their design is a rather complex process. Thus, an efficient optimization algorithm is necessary to design large-scale MTMDs. This paper presents a powerful multi-hazard optimization-based methodology for the design of MTMDs attached to tall buildings. The total added mass is minimized, while the multi-hazard life-cycle cost (LCC) is taken as a constraint. To enable an efficient optimization of large-scale MTMDs, a gradient-based algorithm is developed, supported by a sensitivity analysis of the PEER-based multi-hazard LCC formulation. Material interpolation techniques are used to provide distinct design alternatives. The methodology is applied to a realistic tall building. First, a small-scale problem is used to compare the new methodology with classic evolutionary algorithms. The proposed framework required less than 2% of the function evaluations executed by the benchmark algorithms to deliver competitive solutions. Subsequently, large-scale problems using different formulations are presented. Solving problems with 133 TMDs, the optimized solutions allow to reduce the multi-hazard LCC by 35%, adding 0.69–0.86% of the building mass. The results show that the algorithm can provide economic designs at acceptable computational costs, and the methodology can capture consistently the multi-hazard behavior.