0382 1 Department of Industrial Engineering, University of Girona, Spain Email: cahis@watt.udg.es 2 Department of Industrial Engineering, University of Girona, Spain 3 Department of Industrial Engineering, University of Girona, Spain AN INNOVATIVE ELASTO-PLASTIC ENERGY DISSIPATOR FOR THE STRUCTURAL AND NON-STRUCTURAL BUILDING PROTECTION Xavier CAHÍS 1 , Lluís TORRES 2 And Luis BOZZO 3 SUMMARY This paper presents a new steel shear link dissipador fabricated without welding in areas where yielding takes place, and making possible to design very small web thickness devices. The yielding part of the device is milled from one piece of rectangular shaped steel bars. Four shear link specimens have been tested. This particular set of dissipators were developed for the protection of nonstructural walls. The characterization tests consisted in cyclic loading under displacement control at quasi-static mode. The specimens start dissipating energy from 0.5 mm of displacement and from loads of 14.45 KN. All dissipators develop large deformations without web buckling, with a total accumulate displacement over 420 mm and an average shear strain from 0.11 to 0.16 rad before first damage in the web. Dissipated energy goes from 10 to 21 KJ, and specimens dissipate an additional amount of energy after first web damage because their thick and narrow flanges and stiffeners. Consequently, this new device has a two mode response which results in a very robust and safe dissipator. Proposed numerical models and simple mathematical expressions offer good results as compared to the experimentally obtained values. Values obtained with the equation presented by Kasai and Popov [1986] to determine buckling in shear stiffened links have been compared with numerically determined ones, with less than 6% of difference between them. Experimental analysis in a SDOF system has been performed with the shear link dissipator, in a 4x4 m shaking table (ISMES laboratories - Italy). Numerical most significant values obtained with a simple numerical model differ less than 10% from experimental data. INTRODUCTION From experimental and numerical studies, masonry partitions add significantly to the stiffness of a frame and alter its damping characteristics. Neglecting them in the analysis model is not a satisfactory design practice. However, presently it is very difficult to take advantage of their stiffness and strength since the masonry infill walls usually cannot resist large frame deformations resulting that buffeting from the frame destroys the partitions. Limiting the drift is the most wide spread solution to protect building equipment and nonstructural components. In general, codes specify drift limits such as a percentage of a story height. A proper earthquake resistant design of a building should guarantee that, at each story, the drift limit is not exceeded, protecting the infill walls from excessive deformations and brittle failures. Another proposed solution for masonry wall protection is the uncoupling of the resistant structure from the infill walls providing a gap filled with foam rubber or other flexible inexpensive material. This avoids the introduction of uncertainties in the numeric model used for the design of the frame. The technique, however, does not take advantage of the stiffness and strength of the walls. Besides, if the bare frame is flexible and the ground motion is severe, buffeting may not be avoided destroying