FULL-SCALE TESTING AND NUMERICAL MODELING OF A MULTISTORY MASONRY STRUCTURE SUBJECTED TO INTERNAL BLAST LOADING
Analytics
43 views ◎16 downloads ⇓
Abstract
As military and diplomatic representatives of the United States are deployed throughout the world, they must frequently make use of local, existing facilities; it is inevitable that some of these will be load bearing unreinforced masonry (URM) structures. Although generally suitable for conventional design loads, load bearing URM presents a unique hazard, with respect to collapse, when exposed to blast loading. There is therefore a need to study the blast resistance of load bearing URM construction in order to better protect US citizens assigned to dangerous locales. To address this, the Department of Civil and Environmental Engineering at the University of North Carolina at Charlotte conducted three blast tests inside a decommissioned, coal-fired, power plant prior to its scheduled demolition. The power plant's walls were constructed of URM and provided an excellent opportunity to study the response of URM walls in-situ. Post-test analytical studies investigated the ability of existing blast load prediction methodologies to model the case of a cylindrical charge with a low height of burst. It was found that even for the relatively simple blast chamber geometries of these tests, simplified analysis methods predicted blast impulses with an average net error of 22%. The study suggested that existing simplified analysis methods would benefit from additional development to better predict blast loads from cylinders detonated near the ground's surface. A hydrocode, CTH, was also used to perform two and three-dimensional simulations of the blast events. In order to use the hydrocode, Jones Wilkins Lee (JWL) equation of state (EOS) coefficients were developed for the experiment's Unimax dynamite charges; a novel energy-scaling technique was developed which permits the derivation of new JWL coefficients from an existing coefficient set. The hydrocode simulations were able to simulate blast impulses with an average absolute error of 34.5%. Moreover, the hydrocode simulations provided highly resolved spatio-temporal blast loading data for subsequent structural simulations.Equivalent single-degree-of-freedom (ESDOF) structural response models were then used to predict the out-of-plane deflections of blast chamber walls. A new resistance function was developed which permits a URM wall to crack at any height; numerical methodologies were also developed to compute transformation factors required for use in the ESDOF method. When combined with the CTH derived blast loading predictions, the ESDOF models were able to predict out-of-plane deflections with reasonable accuracy. Further investigations were performed using finite element models constructed in LS-DYNA; the models used elastic elements combined with contacts possessing a tension/shear cutoff and the ability to simulate fracture energy release. Using the CTH predicted blast loads and carefully selected constitutive parameters, the LS-DYNA models were able to both qualitatively and quantitatively predict blast chamber wall deflections and damage patterns. Moreover, the finite element models suggested several modes of response which cannot be modeled by current ESDOF methods; the effect of these response modes on the accuracy of ESDOF predictions warrants further study.