BARRAGE-FOUNDATION INTERACTION UNDER STATIC AND DYNAMIC LOADING

Abstract

Linking of rivers across India to solve flood, drought, power and other waterrelated problems of the country is presently being taken up by the Government of India for implementation in near future. The Government is addressing this mammoth task in a scientific manner, which it deserves. The project, it is hoped, will transfer 'surplus' waters of the eastern rivers - especially the Brahmaputra and various tributaries of the Ganga - to the drought-affected western and southern parts of India through various links between rivers. This project will, therefore, entail the construction of several hydraulic structures including barrages across various river basins in India. Considering the importance ofthe barrages which are likely to be built in large numbers during the river linking project, a detailed study ofatypical barrage raft floor has been taken up to contribute towards a betterunderstanding of the structure and its behaviour. Barrages are amongst the important hydraulic structures constructed across a river to raise the river water level onthe upstream side of structure, in order to feed an off-taking canal from its upstream side at one or both of its flanks. The floor of the barrage is a critical component of the barrage structure. The barrage raft floor is exclusively supported by foundation soil/rock and hence the term interaction has been used in this context, for the mechanics of interaction between the foundation soil, the raft floor and superstructure. The problem ofanalysis ofbarrage including interaction ofbarrage raft floor with foundation soil/rock is different in many aspects from that of other hydraulic structures and buildings, as different hydraulic structures serve a different purpose and each problem is site specific. 11 The Larji barrage, which is a part of power development project in Beas Basin, has been selected for study. It is located 20 km upstream of Pandoh Dam at a distance of about 190 km from Shimla in Himachal Pradesh. It consists of 5 bays and each being 11m wide. There are two expansion joints first between bay 2 and bay 3 and second between bay 4 and bay 5. The plan dimensions ofbarrage raft floor are 49.526m in the direction offlow and 82m across the direction offlow. The barrage raft floors for bays 1to4 rest completely onalluvial soil but bay 5 is founded on hard rock mass. The alluvium zone consists of sand and gravel intermixed with silt and distributed cobbles and boulders. The hard rock mass consists of grey dolomites and pink limestone. This barrage has been adopted to investigate the effect of diverse foundation conditions in detail and their effect on barrage raft floor since the foundation media is representative of the most commonly encountered conditions in the field. The work presented herein, deals with the assessment ofthe static and dynamic behaviour of barrage raft by appropriate analysis and corroborated by experimental model testing. Three-dimensional finite element analysis has been carried out incorporating spatial distribution ofgeometry ofbarrage resting on varying foundation media. Three separate composite models consisting ofbays 1-2, bays 3-4 and bay 5 considered independently, and one hyper-composite model consisting of all the bays raft floors i.e. bays 1-2, bays 3-4 and bay 5 taken together have been developed for analysis. The soil media, cut-off, pier, abutment wall and beams have been discretized by using eight noded isoparametric brick elements (SOLID 45) and the barrage raft floor has been discretized using four noded isoparametric three-dimensional shell elements (SHELL 63) to simulate the behaviour ofbarrage raft floor by plate bending elements in ANSYS. in Finite element analysis was conducted to study the effect of the foundation media on the moments and deformations of the raft floors, in the various models, by varying the modulus of elasticity and Poisson's ratio of the foundation soil, the thickness of the raft floor, effect of rock intrusion in the foundation media and proximity effects orstructure-structure interaction. In the analyses the soil was assumed to be linear elastic. The foundation soil was subsequently also modeled as nonhomogeneous and elastic perfectly plastic to simulate field conditions, to the extent possible. Dynamic finite element analysis was also carried out for the determination of free vibration characteristics and dynamic stresses. The objective of the aforementioned study being the identification of factors affecting the static and dynamic behaviour of the raft floor. A comparison of moments and deformations as obtained from finite element analysis and Hetenyi's method, for few representative load cases, was necessitated by the fact that the Bureau of Indian Standards (IS: 11130 - 1984) criteria for structural design of barrages recommends Hetenyi's method for analysis of the barrage raft floors. Further, an independent model was experimentally tested under static loads and the same was analyzed by both the analytical techniques in order to establish the adequacy of the analytical outcomes. The static interaction analyses, considering the variation of parameters like elastic modulus of soil, thickness of raft floor and intrusion of rock indicated a significant influence on the moments and deformations of barrage raft floor. The Poisson's ratio, on the other hand, did not affect the moments but affected the deformations of raft floor. The intrusion of rock into the foundation soil has a significant impact on the moments and deformations of bays 3-4 raft floors. The IV moments and deformations have also beeninfluenced by the proximity effects and nonhomogeneity of foundation soil media. The elasto-plastic considerations for foundation media contributed to the creation of plastic zones in the vicinity of the cut-offs. The elasto-plastic analysis as compared to the elastic analysis marginally affected magnitude ofmoments and deformations. The dynamic studies indicated that the increase in modulus of elasticity of soil leads to higher natural frequencies. The additional mass under wet condition reduces the natural frequencies of each corresponding mode in contrast to dry condition. The natural frequencies of the system increase due to consideration of non-homogenous foundation media, which has greater stiffness. The dynamic analysis indicates that the variation of modulus of elasticity significantly affects the magnitude of dynamic stress of the barrage raft floor. The dynamic stresses evaluated from the hyper-composite models indicated a major deviation from stresses obtained from the composite models. The comparative studies between Hetenyi's method and finite element analysis highlighted the shortcoming of the Hetenyi's method. Hetenyi's method is unable to account for the geometrical disposition of the raft floor, spatial variation of stiffness as well as variations in foundation media. Further, the considerations of proximity or structure-structure interaction and the inability to simulate loads along the flow direction are additional shortcomings of the Hetenyi's method. Experimental investigations on an independent barrage raft model with cut-off exhibit rigid behaviour. The model placed near the test tank wall exhibits significant differential settlement due to the confinement effect. Further, the deformation data obtained from the finite element analysis and Hetenyi's method has been compared with that observed inexperimental investigations. The finite element analysis exhibited better approximation ofsettlement compared to the experimental values in contrast to Hetenyi's method. Hetenyi's method found to be lacking in lateral boundary considerations and exhibited its limitations. The present study critically examines and enhances the understanding of the behaviour ofa barrage raft floor, which is the most critical component ofa barrage. The comparison of the conventional Hetenyi's method with the various applicable finite element models indicates clearly the limitations of the practice adopted in design offices. In the light of the aforementioned findings it has been recommended that Hetenyi's method should be restricted to the preliminary investigations of design parameters ofbarrage raft floor and the finite element method should be adopted for analysis ofbarrage raft floor in details. The above contribution, itis expected, will draw the attention of the analyst and designers, in this field, to the governing issues in the analysis and design ofraft floors for barrages leading to an improved design criterion. vi