During a routine TEGG survey of a Low Voltage electrical panel that supplied mechanical and chiller plant at a colocation data centre high temperatures were seen on the gland terminations above the incoming cable chamber. Access was limited by other services in the plant room and viewing angles were not ideal. However, from the deterioration in the rubber shrouds over the glands it was apparent that the issue had been present for some time and if left would rapidly lead to catastrophic failure of the cables at the gland plate leaving the building without cooling and resulting in the data halls overheating.
Due to the severity of the issue the customer was notified immediately and because no alternative supply was available and the panel could not be shut down a sequence of monitoring and further investigation visits were arranged while remedial works were scheduled.
Detailed thermal and digital images were recorded to allow us to trend the temperatures over the following weeks. While the weekly thermal monitoring continued investigation into the cause of the heating was carried out. The panel was supplied directly from a dedicated 11KV transformer. Our investigations revealed that the parallel armoured single core cabling between the transformer and the LV panel had been run on ladder rack but had NOT been run in a trefoil configuration. This resulted in overheating of the cables and circulating currents along the armouring of the cores. Thermography and current readings taken from the individual cores also indicated that the cores were not sharing the loads evenly which compounded the heating effects.
A proposal was developed that would allow new containment and cabling to be installed between the transformer and LV panel. The new cabling was installed prior to a planned generator run when we were able to isolate the transformer and connect the new cabling.
We were able to virtually complete the new installation before the changeover which was completed over a weekend when the system could be supported by the site generator. This enabled a seamless transition. During the changeover works the existing cables were removed and the full extent of the damage was evident. The insulation on the lateral runs of cable had become so hot that it had melted and joined with the other cores. The cable glands at the LV panel were loose which accounted for the temperature seen during our survey. In addition to not being run in trefoil the circulating currents were flowing in the armouring through the high resistance joint of the loose cable glands were causing irreparable damage to the cables which would have eventually failed resulting severe damage to the LV panel and a much more time-consuming and expensive repair.