CABLE BENDING ANALYSIS THOUGH FEA

Single strand cable (spiral) bending and OEC (Overhead Electrical Conductor) bending are somewhat similar problems. This is the reason why the following new paper is noteworthy within the context of this blog. It emanates from a Slovak team: S. Kmet, E. Stanova, G. Fedorko, M. Fabian, J. Brodniansky. Title : “Experimental investigation and finite element analysis of a four-layered spiral strand bent over a curved support”. Published in “Engineering Structures”, Vol. 57, December 2013, pp. 475-483. Abstract can be read online at www.elsevier.com/locate/engstruct

Its objective is to compare an FEA numerical model with some experimental data. The selected cable is a four-layer steel cable. Apparently (although not said explicitly), it is a parallel lay cable, thus interlayer contact consists of line contacts, instead of point contacts when dealing with cross lay systems (as in most OEC’s). FE model uses more than 1.5 million elements and close to 2 million nodes. Software: CATIA (for geometry input) and ABAQUS. Test specimen length is around 5.3 meters and passes over a curved saddle at midpoint. Strain gages are positionned at 3 sections: (I) at center of saddle, (II) in saddle vicinity, and (III) away from saddle (where cable is practically straight). No details are given on the number of gages, and on which wire they are bonded. Apparently, there are just three of them, at positions (I) (II) and (III), glued on that wire corresponding to the highest point from cable section neutral axis, on the convex side. A tensile force is applied symmetrically on specimen and increased from zero to 600 kN. Strains are given at five levels: 200, 300, 400, 500 and 600 kN, at gages (I) (II) and (III).

A numerical investigation is also presented of the effect of the saddle radius of curvature. In the numerical model, contact conditions between cable and saddle are assumed to be known (except that cable may slip on the saddle). Cable is assumed to take the saddle radius of curvature, over the saddle.

There is apparently a very good correlation between test data and FEA results. However, some adjustments are made to the test data which would require more explanation. Besides, the “theoretical stress” at position (III) is calculated as N/A where A is the strand cross-section area, taken as the sum of the 61 wire cross-sections. The helical effect is not taken into consideration. If one considers such effect, wire stress at position (III) should be quite higher.  Unfortunately, no comparison is made with results which could be obtained (more simply) with some of the analytical models mentioned in the Introduction.