EXPERIMENTAL STUDY OF SEISMIC STRENGTHENING AND RETROFITTING MEASURES IN MASONRY BUILDING

Abstract

The brick and stone masonry buildings have proved to be the most vulnerable to earthquake forces and have suffered maximum damage. There is a lot of research conducted in the development of earthquake resistant and retrofitting measures in masonry buildings. Some of these methods have been introduced in the code of practices. Moreover, the buildings constructed by employing existing methods also suffer some damage. All these methods have not been adequately verified by experimental tests for their effectiveness. Therefore, it is necessary to assess the existing methods of strengthening and retrofitting as well as to suggest certain modifications in order to further reduce the damages in masonry buildings. The majority of testing so far done for evaluating the seismic resistance is on small scale brick masonry models. Very few tests have been conducted on stone masonry models. Advances in servo-hydraulic technology and computer simulation are making experimental testing more feasible in earthquake engineering, but fundamentally, such researches are being concentrated principally on steel and concrete structures whereas majority of population in India live in low strength masonry houses constructed with stone, bricks, mud, adobe etc. The effectiveness of various strengthening and retrofitting techniques of stone and brick masonry buildings are evaluated respectively through shake table testing and quasistatic testing of realistic masonry structure. It has also been felt that the earthquake resistance of masonry buildings should be studied through analytical methods so that the strengthening of buildings could be designed on a rational basis. Six full scale models of one storeyed stone masonry houses with different strengthening measures have been tested under progressively increasing intensity of shock on shock table facility. After the damage of models, these are retrofitted by existing techniques prescribed in the code and tested again. The behaviour of each model including the pattern of cracking, identification of weak zones, damage with progressively increasing shocks have been studied. The natural frequency, equivalent viscous damping factor, acceleration at base of table and at roof of models has been measured from experiments. The roof acceleration is found to amplify before cracking and then it is reduced after cracking which demonstrates a typical base isolation effect. A correlation of shock table motion with actual earthquake ground motion has been conducted on the basis of Arya & Saragoni's destructive potential factor. It considers the effect of peak accelerations, frequency content and duration of motions simultaneously. The parameter for shock 3 (W-21) is 0.927 while in Uttarkashi earthquake it is 0.816. Thus the pattern of damage in shock 3 can be quite similar to Uttarkashi earthquake. The models without any earthquake resistance measures has very little resistance to shock, as a result complete collapse has been observed in shock test. The codal provision such as lintel band, roof band, comer and jamb reinforcement are effective in improving the behaviour of stone masonry models. The provision of additional band at sill level in addition to existing provisions has significantly improved the behaviour of stone masonry model in shock test. The cracking and bulging is reduced in walls and damage to walls in severe shock is also considerably reduced. The complete failure of model does not occur. The injection of cementitous grout on localized damaged areas can restore the original strength and stiffness. The introduction of external horizontal tie bar to a wall can increase the strength and ductility of the model. Moreover, welded wire mesh in damaged region not only increases the lateral resistance of the wall but also prevents shear and flexure failure of the models. A three-dimensional (3D) elastic analysis of the tested structure has been carried out by finite element method. The COSMOS/M programme has been used with 8 noded brick element for walls and the roof has been modeled with 3D truss element. The strengthening measures have been represented by increase of Young's modulus (E) in respective region. The natural frequencies and mode shapes have been determined from VI finite element model. The majority of modes of vibration shows shear deformations in piers ofwall and bending at lintel level. The vertical vibrations and torsional behaviour of model is also observed in higher modes. The elastic dynamic response has been obtained by response spectrum method for signatures of shock obtained from experiment. The shear stress contours are plotted for both the shear walls. It can be observed that in, unstrengthened model the maximum shear stress are concentrated in the regions above lintel level, mid portion in piers ofshear wall and around the opening. In strengthened model (lintel band) the maximum shear stresses are concentrated at the lintel level. This shows that cracking will occur in these regions. There is a good agreement in the region of cracking in the shock table tested model with the results ofFEM analysis Seismic behaviour of three brick masonry models, half-full scale size, with different strengthening measures has been studied under cyclic loading in quasi-static test facility with an aim to evaluate its effectiveness in enhancing the seismic performance. The models have been tested upto ultimate failure and their strength, deformabilty, energy dissipation capacity, hysteretic behaviour, damping, crack pattern has been studied. The models have been constructed and fixed on the strong floor. The hydraulic actuators attached to the reaction wall through a load cell are fixed on the top of the model. The model is subjected to alternate cyclic loading in the form of sine sweep waves. After substantial damage in cyclic testing the model 2 and 3 were retrofitted by two different techniques. The model 2 have been retrofitted by epoxy sand mortar while model 3 is retrofitted by grout injection and welded wire mesh. The retrofit models have been tested up to ultimate failure and their strength, deformabilty, energy dissipation capacity, hysteretic behaviour, damping, crack patterns have been compared with the original models. Elastic and strength properties of masonry i.e. compressive strength, tensile and shear strength , modulus of elasticity, and damping, have been obtained by testing the specially prepared model specimens. Experimental tests reveal that a good shear connection between the roof slab and masonry walls is necessary all through the length of the slab which would enable uniform transfer of seismic forces from floor to walls and the resistance capacity of the vii building is fully utilized. The tests also shows that the model constructed according to the recommendations of the codal provisions have been effective in preventing the collapse and in reducing the damage mainly above lintel band. The damage in piers of wall still occur. The in-plane seismic performance of wall piers may be considerably improved by the provision of an additional band at sill level. The sill band increases ultimate load capacity of the model by about 20% and the deformation capacity by 1.20 times than the model without sill band. On the basis of quasi-static tests, three criteria emerge for retrofitting of wall (i) increase the stiffness of wall, so that the story drift is limited to 0.35% in order to avoid cracking, (ii) increase the ultimate strength of wall, (iii) increase the ductility of wall. Experimental tests suggested that retrofitting with welded wire mesh has increased the stiffness as well as ultimate load capacity. It has been observed from tests that grouting of cracks with epoxy-sand mortar has helped to improve ductility. The ductility of brick masonry models may be obtained by working out the ratios of ultimate displacement divided by the displacement of first cracking of piers which is about 1.8 to 2.0. An attempt is also made to make a comparative study of behaviour of two brick masonry models constructed with identical strengthening features by (i) quasi-static testing (ii) shock table testing. The shock table model responds with a significant higher initial strength and stiffness as compared to the quasi-static model subjected to equivalent lateral displacements. It has been observed that severity of damage is greater for static structure due to increased crack propagation.