STABILITY ANALYSIS OF LAMINATED COMPOSITE BOX BEAMS

Abstract

Laminated composite box beam is the basic building block. It can be used both as a compression member and flexural member. Over the past decade FRP laminated composites are gaining popularity in marine, aerospace and civil industry for numerous structural applications. Being lightweight it is easy to handle and transport leading to short installation times and speedy construction which is a major. criteria during disasters. Also the quality of work is improved in terms of corrosion resistance , endurance limit, damage tolerance, aesthetic appearance, longer life, low maintenance, fire retardancy etc. For an instance laminated composite FRP deck weighs approximately 80 percent less than a concrete deck. Reducing the dead load increases the allowable live load capacity of bridge without significant damage to the existing superstructure, thus lengthening its service life. Versatility of the material is such that it can be used even in earthquake prone areas for seismic retrofit or for remote installations in difficult terrains. Further the flexibility to tailor different properties of the structural elements by varying fibre orientation, number of layers etc. to achieve the strength and stiffness requirements in the required direction has enhanced their usage. Need of appropriate structural form with FRP is required to utilize its advantages. Thin-walled box beam is chosen for study as it offers excellent flexural and torsional stiffness with minimum weight. Use of FRP in box beam form combines the advantages of both the material and form and results in an efficient generic structural element. FRP box girders comprising of FRP stiffened panels are predominantly under membrane compression (top flange), in plane shear (web), and membrane tension (bottom flange). Hence the use of laminated composite materials for these structural elements is aptly suited as they provide flexibility to designer to specify different material properties for different elements of the beam cross-section. This enables the shape of the cross-section to be exploited to the fullest by judicious arrangement of plies within the laminated composite panels, which form the beam. The vast array of design variables for composites contrasts sharply with traditional materials where choices are limited and hence complicates the design. Due to large number of variables human intuitions will not work and use of computational procedures coupled database should be considered in improving design. Currently there are no proven analysis procedures, design standards or specifications for FRP. FRP being a high strength material, thickness required to satisfy the strength constraint is very less resulting in thin walled structure where the design is governed mainly by stability requirements. Stability of plate increases with increase in plate thickness but a more economical solution is obtained by keeping the thickness of plate as small as possible and increasing the stability by introducing stiffeners. In the present work studies are carried out on unstiffened, stiffened and multi-cellular box beams using finite element analysis and an effort is made to understand the stability characteristics of these thin walled structures. The FEA generated database is used to develop the computationally efficient analysis approaches in the form of neural networks. The ANN based analysis tool can be further used to predict the local buckling strength of FRP shapes (beams and columns). It is observed that ANN takes fraction of second to predict the output function while FE analysis takes ten to twenty minutes. This proves the computation efficiency of developed artificial neural networks. Also for stiffened box sections few important parameters affecting the buckling behavior are identified and guidelines for better stiffener proportioning are developed which will be useful for the designer. In nut shell present work is carried out to understand the buckling behavior of laminated composite box sections with the help of exhaustive numerical studies. Additionally, ANNs are trained which can be used in predicting the local buckling strength. Also some guidelines for better stiffener proportioning are suggested which can be used by the designer and for further analysis and research on the subject. W