LONG TERM BEHAVIOUR OF HIGH PERFORMANCE CONCRETE BRIDGES

Abstract

In last two decades use of cement with cementitious materials in HPC has been increasing world over to improve the workability of fresh concrete and durability of hardened concrete to achieve superior dimensional stability and to reduced porosity for long-term sustainability. 1-lence, the use of varieties of locally available auxiliary cementitious materials for partial replacement of cement in concrete mix is common to produce low cost HPC. The problems associated with use of auxiliary cementitious materials lies in variation of its physical, chemical and its inherent mineralogical composition properties. In fact, the properties of cernentitious material employed greatly influence the time-dependent properties of HPC i.e. shrinkage and creep. Also, where the environmental impacts are quite severe i.e. temperature and relative humidity varies in wide range, it can be very difficult to prevent the micro-cracking on shrinkage in normal/high strength concrete. In such condition, HPC is one of the best options but this has been not fully exploited due to lack of understanding of its shrinkage and creep behaviour, especially indigenous HPC material resource. Also designers are not sure whether the provisions made in design on the basis of existing shrinkage and creep prediction models is adequate to ensure satisfactory long-term performance of modern HPC construction using mineral/chemical admixtures. In this regard existing material models namely Ad, 133,fib and GL have been chosen and statistically evaluated using indigenous shrinkage and creep database to find out their order of performance in local condition. The study finds that thefib model performance is the best for shrinkage and creep prediction for I IPC though it is far from satisfactory as errors are relatively higher. Further, second best option for shrinkage is 133 model and for creep is ACI model for HPC database. Use of HPC with locally available auxiliary cementitious materials has increased. The difficulties associated with indigenous cementitious materials lies in variation in its physical/chemical/mineralogical properties which may greatly influence the time dependent properties like shrinkage and creep in HPC. Towards this, seven different HPC mixes of M40 to M60 grade using different cementitious materials namely fly ash (FA), silica fume (SF) and ground granulated blast-furnace slag (GGBS) along with ordinary Portland cement (OPC) have been prepared to investigate the microstructural performance of concrete properties i.e. durability, shrinkage, creep, total deformation as well as mechanical property i.e. compressive strength. Also, the test data of shrinkage and creep has been used to evaluate the relative performance of the different prediction models for HPC. This study may be helpful to increase the material understanding, proper material selection and guideline for HPC mix design. lIT Roorkee ii l Abstraci A reasonably accurate estimation of actual shrinkage and creep is an important design parameter to ensure that structures built in such concrete perform satisfactorily specially in long-term i.e. it does not exhibit unduly large deformation or any distress like cracking etc. during its anticipated service life. Further, a very limited % of NU-ITI/RILEM data is for - concrete with auxiliary cementitious material and hence all the existing prediction models based on regression analysis of entire/partial data may not ensure satisfactory prediction of shrinkage and creep of modern HPC in general and especially for local geographic conditions. In the present research work appropriate concrete material models have been proposed to predict shrinkage and creep of HPC using Artificial Neural Network (ANN). The ANN models are trained tested and validated using 106 different experimental measured set of data collected from different literature and own experimental database. The developed models consist of 12 basic input parameters which include quantities of ingredients namely OPC, FA, SF, GGBS, water, and other aggregate to cement ratio, volume to surface area ratio, compressive strength at age of loading, relative humidity, age of drying commencement and age of concrete. The Feed-forward backpropagation networks with Levenberg-Marquardt training function are chosen for proposed ANN models and same implemented on MATLAB platform. The results shows that the proposed ANN models are rational as well as computationally more efficient to predict time-dependent properties of shrinkage and creep of HPC with high level accuracy. The concept outlined in this research work is not restricted to experimental. The experimentally measured results have also been used as a powerful tool in analyzing the long-term behaviour of HPC bridges i.e. camber deformations, fiber stress distribution and prestress losses. In present study, the incremental time-step analysis method has been used to obtain long-term behaviour of simply supported post-tensioned I-girder at different age of maturity of HPC concrete. The long-term behaviours have also been predicted on the basis of existing material prediction models i.e. ACT, B3,Jlb and GL and the developed ANNs. To study the significance of research work, four different HPC mix proportions of M50 grade shrinkage and creep experimental measured database in present work have also been considered for a critical appraisal. It is found that the proposed ANNs using in developed procedure of incremental time-step analysis method in Excel programme can provide a reliable long-term behaviour of HPC post-tensioned I-girder bridges without any substantial loss of accuracy while second best option is observed by fib model but errors are relatively higher. Also, results obtained by incremental time-step analysis indicated that the HPC ITT Roorkee iii I I> a e Abstract material properties and its influence should be considered carefully in analysis to obtain long- term behaviour of HPC bridges. Long-term behaviour prediction of HPC bridges involves major time-dependent concrete material properties i.e. shrinkage and creep, geometric nonlinearities and different changes in configuration of the structural component with partially/fully during construction. In fact, during actual construction the behaviour of bridges always continues to change with time period. Therefore, the time-dependent concrete material properties and sequence of stage construction in simulation of HPC bridges is very important to predict long-term behaviours. In this regard, the present work a post-tensioned I-girder bridge constructed in Alabama, USA and a typical cable-stayed bridge being constructed over river Ravi at Basoli near Jammu, India has been chosen as a case study for long-term deformations prediction of bridge deck in stage-by-stage construction using time-dependent concrete material property of shrinkage and creep. The three-dimensional finite element modelling of bridge has been constituted using computer programme CSiBridge considered from project drawings. All the geometric nonlinearities have been taken into account in analysis with P-delta (large displacement) criteria. In simulation short-term and long-term deformations of bridge deck has been predicted. The result shows that the actual stage-by-stage construction simulation is very important to determine long-term behaviour of concrete bridges. Also, it is demonstrated that the time-dependent analysis carried out in stage construction and after completion to be useful to identify the problem and take remedial solution on them.