Smart structure system for damage diagnosis of long span roof trusses

Abstract

Long span roof trusses, for their structural efficiency, are very popular in Hong Kong (e.g., the 90m by 90m roof span of Hong Kong Coliseum) and worldwide (eg., the 73m cantilever steel truss of the Shanghai Stadium). Collapse of this kind of structures is cataclysmic owing to the large area of coverage (e.g., the roof collapse of Terminal 2E at Charles de Gaulle Airport, France, on 23 May 2004; and the collapse of roof truss of the Sport Hall at City University of Hong Kong, Hong Kong SAR, on 20 May 2016). Many collapse reports of roof trusses revealed that symptoms (eg, cracking sounds and falling of concrete dust) were observed hours or even days before the collapse. The symptoms suggested that the structure may have damage accumulated on structural members and/or it may be overloaded with large defonnation. The main purpose of the proposed project is to develop a practical and cost-effective smart structure system for early detection of potential structural damage and abnormal deformation of long span roof trusses by proper sensing and intelligent diagnosis based on structural responses, such as accelerations, strains, modal parameters, under operational conditions. The proposed project put emphasis on the development of a damage diagnosis methodology for extracting information from various measurands and quantifying the damage status of the target structure. A Bayesian probabilistic approach is adopted as it fundamentally processes the statistical information from data for making inference and decision making. Major thrusts will be in the following areas. First, sensitive measurands shall be identified together with the corresponding optimal sensor configurations (including number of sensors and the measured locations and directions) to enhance the effectiveness of the proposed methodology. Second, the most “appropriate” model class of the target structure shall be selected conditional on the set of measured data to ensure infonnative damage detection results. Third, long-terrn monitoring database of measured data shall be developed for distinguishing damage-induced changes in measurands from their variation due to environmental effects (e.g., temperature and ambient excitation intensity). It is believed that the proposed methodology can provide valuable information about the structural damages and abnonnal deformation of long span roof trusses, so as to allow for early remediate measures and warning alarm for occupants to escape from the affected area. It is envisaged that the

basic methodology can also be applicable to other structural types with similar frequency ranges.