Behaviour of ferrocement box girder elements

Abstract

The experimental investigation has been carried out by testing near . prototype size ferrocement box girders. Two box girders were tested under uniformly distributed load (ud!) over the entire top flange and two other under udl over half flange width and full span. Two box girders were joined at the. level of top flange. The combined box girder was loaded and unloaded under various load combinations in the uncracked stage. The girder was later subjected to monotonically increasing sustained loads of short durations. One composi te box girder made of bottom flange and side webs of ferrocement and top flange of reinforced concrete was cast and tested under udl over the entire top flange. After unloading it was again subjected to sustained loads of short durations upto the penultimate load and for ten and half. months at maximum applied load. For the elastic analysis of ttreedimensional ferrocement structures such as folded plates and shells of various shapes, beam theory, membrane theory, membrane-bending theory and finite element method rove been used by variOUS researchers. The analysis in the cracked range has been reported only in the case of folded plates using beam theory to predict deflections, first crack Ferrocement has been used in a variety of applications such as boats, tanks, silos and roofs. For roofing purposes, its behaviour has .been investigated by many researchers in the form of channel sections, ribbed slabs, folded plates and shells of various shapes. Only channel sections and ribbed slabs provide a flat top surface. Due to their small flexural rigidities, these elements undergo large deflections and cracking at service loads. To reduce del!ection and cracking and also to have a flat top surface, a new type of roofing/flooring element in the form of box girder shape has been investigated experimentally and analyticaUy in tre present study. load and ultimate load. Even the elastic analysis of folded plates using beam theory is an approximate one because it does not take into account the distortion and warping of the cross-section. The analysis in the cracked range is even more approximate as it does not take into account the changing rigidity of the material at different sections, material anisotropy and yielding of the reinforcement leading to local redistribution of stresses. To overcome above deficiencies, finite element method has been used to predict the behaviour of ferrocement box girders through the elastic, cracked and ultimate stages. Further to economize the finite element solution, the conventional layered approach has been suitably modified for thin ferrocement plated structures. Instead of considering the element to be consisting of. suitable number of layers of mortar and reinforcement, the element is assumed to be consisting of single mortar layer in the uncracked stage, uncracked and cracked mortar layers in the cracked stage and smeared layers of wire mesh and skeletal steel. In the cracked stage, the depth of cracked/yielded/crushed mortar is determined. The stiffness of the element in the cracked stage is obtained by adding the contributions due touncracked mortar layer, cracked/ . yielded mortar layer and unyielded layers of wire mesh and skeletal steel. The finite element analysis has been carried out under dead loads and monotonically increasing live loads. A rectangular flat shell element capable of representing membrane action, bending action and the interaction between membrane and bending. action is adopted. Only. material nonlinearity due to cracking of mortar, tension stiffening effect of mortar between the cracks and the nonlinear stress-strain relationships for the mortar, wire mesh and skeletal steel is considered. Since the box section provides large flexural and torsional rigidity, the deflections in the cracked range are found to be small and hence, geometrical nonlinearity is not considered. Also not considered in the analysis are bond slip between the reinforcement and mortar, time dependent and thermal effects. An incremen tal itera ti ve procedure capable of taking advantage of both thc tangent and constant stiffncss approach has bcen used for the nonlinear analysis. A general computer program has been developed to facilitate corilputer aided analysis. The validation of the proposed analytical formulation has been checked by comparing the predicted results with the reported experimental/analytical results of typical test problems taken from the literature as weli as with the experimental results of the present investigation. The predicted values from the proposcd analytical method ilrc gcnerilily in good agreement with the experimental values except neilr the ultimate . failure load where predicted values are on the flexible side. The experimental .investigation shows that the limit state of serviceability for ferrocement box girders is governed by the maximum crack width. At the recommended crack width of 0.1 mm, the span/deflection ratio is much above the value of 250 as permitted by 1.5. Code. At a span/deflection ratio of 250, the load taken by the girders is .close to the yielding of the reinforcemen t. The double cel! box girder under various combinations of symmetric and unsymmetric loads in the uncracked stage has behaved as one single unit by undergoing downward deflections along the entire length and width. This demonstrates the large load distribution capability of the box section. Replacing the top ferrocement flange by a reinforced concrete one results into the lowering of the first crack load and ultimate load. The failure of the girders is characterized by well distributed flexural .. 5.0 to 6.0 ti mes' the The ulti rna te load was found The instantaneous deflection of' the girder is reduced due The sustained loading also leads to an increase in the width of cracks and the region of cracks formation. However, the ultimate load of the girder IS The increase In deflections or strains due to monotonically increasing method is the prediction of cracks on .the bottom surface of the top flange with the experimental crack-patterns. The added advantage of the analytical , The predicted crack-patterns of the bottom flange and the side webs to be 2.0 to 3.0 times the first crack load. sustained loads of short duration is maximum in the ini tial portion of the to the sustained loading at lower load levels as compared to the instantaneous deflection that would have occurred under monotonically increasing loads. (being inside the box) at ultimate or near ultimate loads. cracked range. not affected by monotonically increasing sustained loads of short duration.