3-dimensional (3D) equivalent static wind loads (ESWLs), including two swaying and one twisting load components, are estimated based on spatiotemporally varying aerodynamic wind loads measured in wind tunnel tests. A prominent problem in the determination of ESWLs according to such random wind loadings is the combination issue for resulting load effects with low out-crossing rate in order to reflect actual maximum responses of the building. In view of correlations among these three load components, a scalar combination rule is advanced and the derived load cases in each incident wind direction in wind tunnel tests needs to be integrated as a set of ESWL cases provided for design purpose.

An optimization-based framework is proposed by firstly assuming that the probability function of...[ Read more ]

3-dimensional (3D) equivalent static wind loads (ESWLs), including two swaying and one twisting load components, are estimated based on spatiotemporally varying aerodynamic wind loads measured in wind tunnel tests. A prominent problem in the determination of ESWLs according to such random wind loadings is the combination issue for resulting load effects with low out-crossing rate in order to reflect actual maximum responses of the building. In view of correlations among these three load components, a scalar combination rule is advanced and the derived load cases in each incident wind direction in wind tunnel tests needs to be integrated as a set of ESWL cases provided for design purpose.

An optimization-based framework is proposed by firstly assuming that the probability function of three load components in one incident wind direction is expressed by a multivariate normal distribution so that the equivalence surface of probability corresponding to a specified value of the peak factor becomes an ellipsoidal one. The optimization algorithm is applied in searching for a convex hull which serves as a design envelope covering ellipsoidal thresholds in all incident wind directions. Individual load cases could be expressed in terms of the coordinates of the vertexes on the optimized design envelope. Two main objective formulations are well developed: 1) To achieve more precise prediction of wind load combinations by minimizing the volume of the design envelope in consideration that the difference between polyhedral surfaces and ellipsoidal threshold should be as small as possible; 2) To minimize the risk of overestimation of wind-induced maximum top displacement by providing constraint functions for load combination factors in critical load cases. The Pareto optimal theory is also integrated with this new approach allowing for predicting the appropriate number of load cases. Not only is this comprehensive approach able to demonstrate a systematic way of determining wind load combinations more accurately, but also a benefit of predicting a set of Pareto-optimal number of load cases is achieved.