Comparison between predictions and measurements of the superconducting performance of Nb3Sn cable in conduit conductors with transverse load degradation

Nb 3 Sn cable in conduit type conductors (CICC) are a popular choice for high current conductors operating in steady or slowly changing current/field conditions. They provide a low cost way of bundling strands together to achieve large conductor currents at high field (up to 13 T), allowing a reduction in coil voltage requirements. The extensive strand contact, although giving relatively high AC losses, gives resilience against current non-uniformity and local strand damage, and there is good stability to electromagnetic disturbances due to the extensive helium contact area with the strands. In the last 5–6 years there has been clear evidence of Nb 3 Sn CICC performance degradation compared to expectations based on strand measurements. Investigations have linked this to filament bending and fracture under the mechanical loads, and particularly their accumulation from top to bottom of a cable. These effects were not widely expected to occur in small cables, and even in large ones their extent was expected to be limited. However, in 2006 testing of a small 15 kA conductor intended for a dipole showed severe degradation that continuously increased with cycling and some large 70 kA CICC have also shown a higher than expected performance drop. Experimental evidence strongly suggests that the degradation is dependent not only on the cable pattern but also on the sensitivity of different types of Nb 3 Sn strands to bending and pinching. The variability of results, with some strand–cable combinations giving a satisfactory performance, has provided an incentive to improve the modelling of the degradation to provide better predictions at the strand level of the characteristics needed for successful cable performance. Modelling was being developed as early as 2005 with an analytic cable mechanical model but a simplified and non-physical procedure to then derive the superconducting performance. The mechanical model was successfully benchmarked against overall mechanical property measurements on cables. This paper continues the development towards a complete predictive cable degradation model, this time covering the link between filament bending and damage (available from the earlier mechanical model and also increasingly from more sophisticated numerical simulations) and the overall superconducting performance, benchmarked using data from strand tests.