An Efficient Dual Response Surface-Based Convex Robust Optimization of Multi-storied RC Buildings Subjected to Surface Blast Loading Characterized by Interval Uncertainty

This work deals with robust design optimization (RDO) of structures when information about the input parameters is not adequate to treat them probabilistically. This case often arises while designing a structure under surface blast loading. Neither the mean, standard deviation, nor the probability distribution function is available for surface blast parameters, like charge weight, stand-off distance, overpressure, duration of the pulse, shock front velocity, drag coefficient, etc. In such cases, the parameters can be treated as interval types, where only the ranges of variations of the uncertain parameters are sufficient to model them in a hyper-ellipsoidal convex domain. Subsequently, a convex programming (CP) approach can be used to optimize the system. This approach works fine for simple deterministic systems where loading is equivalent to static in nature. But, the surface blast is a dynamic impulse type load with significant stochastic characteristics. The time-history of blast loading varies significantly even with the same hazard parameter setup due to the inherent uncertainty associated with blast load modeling. A new dual response surface method (RSM)-based RDO procedure is developed in the present paper. The dual RSM captures the stochastic nature of load time-history which in-turn is used to develop a new formulation of RDO. The effectiveness of the proposed approach is explained using a numerical problem of an RC 20-storied building. The RDO is solved by sequential quadratic programming. The results of RDO are compared with the conventional IS: 4991- 1968 based approach to indicate the need for the present approach.