Nonlinear dynamic analysis of heavy timber frame structures including passive damping devices

Graduation date: 2003Engineering analysis of heavy timber structures under dynamic loads\ud is generally lacking. Some literature suggests that heavy timber structures\ud are stiffer than expected, but popular engineering opinion is to the contrary.\ud In addition, prior research has shown that passive friction dampers can be\ud used in timber structures for the intended purpose of increasing energy\ud dissipation under seismic loadings. This study was conducted based on the\ud hypothesis that heavy timber structures require bracing for seismic design\ud and further that heavy timber performance under seismic loads will benefit\ud from the addition of passive dampers.\ud A computational study was designed to assess the performance of a\ud two-dimensional model heavy timber structure when it is subjected to a\ud seismic ground motion. Five building systems were included in the\ud investigation; all were two story, single bay timber frames with 1-percent\ud inherent material damping. The frames were differentiated by connection\ud characteristics, the presence or absence of bracing, and the presence or\ud absence of friction damper devices. The five building systems were: (1) moment-resisting frame (rigid connections, no bracing); (2) semi-rigid\ud frame (semi-rigid connections, no bracing); (3) pinned braced-frame (pinned\ud connections, concentric braces); (4) braced damped pinned-frame (pinned\ud connections, cross braces with dampers); (5) semi-rigid damped frame\ud (semi-rigid connections, concentric bracing with dampers). By subjecting\ud these five structural configurations to the same ground motion, the effects\ud of the connection rigidity, bracing, and dampers can be resolved as\ud measured by story drift and member forces.\ud The ground motion used was a corrected, unscaled, USGS Loma\ud Prieta record with maximum acceleration of 1.05 g. The study frame was\ud designed with access floor loading with computer use for a geographic\ud location near San Francisco following the Uniform Building Code and the\ud National Design Specification for Wood Construction. Beams and columns\ud were glulam. The semi-rigid connections were designed as a dowel circle\ud connection following DIN 1052, which is widely used in Europe.\ud The software used for the investigation was DRAIN-2DX. Semi-rigid\ud connection behavior was implemented using the Florence Model. The\ud DRAIN-2DX software does not have a dedicated damper element. A\ud damper element would produce a rectangular displacement-force\ud hysteresis when subjected to a fully reversed cyclic load; the key features of\ud this hysteresis behavior are brace stiffness followed by nearly zero stiffness\ud at a defined force. A brace-damper superelement was created by using two\ud compression/tension link elements in parallel with an inelastic bar element.\ud This combination of elements produced the necessary behaviors, and the\ud performance was verified in a single story, single bay frame subjected to a\ud piecewise forcing function and demonstrated in a two story, one bay timber\ud structure.\ud The results of the analyses for the five building systems are specific\ud to the Loma Prieta seismic ground motions. The semi-rigid moment-resisting\ud frame remained within allowable drift limits, which suggests that the dowel circle connection produces satisfactory moment resistance. The\ud braced pinned-frame and the braced-damped, pinned and semi-rigid frames\ud responded similarly to each other because the large cross-section timber\ud brace in the low rise, low mass system results in an extremely stiff\ud structure. These structures are so stiff that the damper does not function.\ud Evidence suggests that a rotational damper may be functional in semi-rigid\ud connections of unbraced timber frames, but computational development of\ud the rotational damper is needed