Processing of Composite Rod with Embedded Optical Fiber

Abstract

Polymer composite rods have promising uses in telecommunications as cable strength members, in maritime structures because of their resistance to various environmental in uences, and as railway insulating rods. The use of optical ®bers as sensors for this structure enables the chance of real-time damage and strain assessment in structural applications. The investigations performed by past researchers differ in the variety of the laminate composite structure, optical ®ber coating, and type of testing (three-point bending, fatigue and, impact). One of the ®rst applications of optical ®ber sensors was that of the impact damage detection via ®ber breakage sensors [1, 2]. Waite and Sage [3] described the failure of silane solution-treated and acrylate buffered optical ®bers, which were either embedded in polymer composites or loaded individually. In another study by the same author [4], conclusive results were obtained by using hard clad silica and strippable co-polymer coated optical ®bers for the detection of fatigue damage. To the authors' best knowledge, no results have thus far been presented for embedding an optical ®ber in composite rod structures. It is dif®cult to integrate optical ®bers precisely into composites that do not have a ply structure. In our previous study [5], we detailed the development of an experimental facility for manufacturing composite rods with embedded optical ®bers by vacuum molding. Vacuum in®ltration involves placing the reinforcing ®bers and the optical ®ber in a die, applying a vacuum at one end of the die and a source of epoxy matrix at the other, and then using the vacuum to draw the resin through the die to impregnate the ®bers. The ®berous reinforcement used in the composite was E-glass ®ber Etex R2105 roving. An acrylatecoated multimode optical ®ber (with core=cladding= coating equal to 62.5=125=250 im) was chosen due to its relatively easy handling, simplicity of mounting, low cost of fabrication, and simple electronic detection. The matrix material was based on a diglycidyl ether or bisphenol A, together with an aromatic amine hardener (Araldite CY 223=HT972, Ciba-Geigy). The epoxy-amine mixtures were prepared by heating the resin in an oil bath to 70 8C, and adding the curing agent with continuous stirring until a clear homogeneous solution was obtained. The exural, impact, and fatigue resistance of a composite rod structure is related to the strength of the material at the surface. The specimen initially has a tensile force acting on the top surface, while the bottom surface is compressed. The fatigueor impact-induced rod bending causes tension on the rear surface, inducing strain in the matrix and the optical ®ber embedded at that location. When the strain in the exed rod exceeds the ultimate strain of the optical ®ber, it will fracture. Therefore, embedding optical ®bers near the die=composite interface is considered to be essential from the standpoint of strain sensing. The degree of cure of a composite is important in many aspects because the amount of cure can in uence the mechanical properties of the product and probably melt the buffer on the optical ®ber. In order to achieve a uniform degree of cure in the cross-section of a composite rod, the die temperature is the most important processing variable. Control of the die temperature requires information about the curing reactions taking place inside the die. A mathematical model has been developed to simulate the vacuum molding process, namely, the pro®les of temperature and the degree of cure in the radial direction in a die of cylindrical shape. A similar approach was taken by several authors [6±8], who were interested in modeling the pultrusion process. The kinetic parameters of the epoxy resin were obtained from the differential scanning calorimeter (DSC) scans. From the data shown in Fig. 1 and the peak obtained by scanning a standard Indium sample (Perkin Elmer DSC-2), the calculated reaction heats (yAH) at 423.15 K and 443.15 K were 579.3 J=g resin and 535.7 J/g resin, respectively. Other kinetic parameters were calculated by a nonlinear curve®tting method. The reaction rate equation used in this study was: