Numerical prediction of cables fire behaviour using non-metallic components in cone calorimeter

Abstract

Electrical cables must fulfil the Construction Product Regulation (CPR) by testing them according to EN 50,399. Nevertheless, an unfordable trial and error procedure could stem from the definition of new cables. To achieve a better understanding of fire behavior of cables, researchers have been using techniques such as bench and reduced scale tests and computational fire models result in a way to minimize trial and error process. The present work proposes the combination of bench scale tests, using cone calorimeters and fire simulation modeling. In a first step, the thermal characterization of the cable parts is carried out, and then, in a second step, use these data to model complete cable samples in cone calorimeter tests. The simulations are compared with experimental data of two already rated cables. This process is intended to discard erroneous configurations, which display in bench scale signs of misbehavior compared with rated cables. This would avoid the manufacture of the complete cable if the results do not fulfill the requirements, and eventually, proceed to its production and test in full-scale when they do. This work has been carried out with two multi-core cables and the materials they were made of, and the results showed: a) the inverse modeling process to characterize materials parts obtained a fairly accurate approach, with small inaccuracies in the peaks of the curves; b) two simulation models (simple and detailed) were able to reproduce in general terms the heat release rate curve; however, they released more energy than experimental tests and some discrepancies in the peaks were observed. Despite its simplicity, simple model obtained results fairly close to the experimental curves and took less time to simulate.