Battery Bank Earth Fault

Recently we had a really unusual fault with one of the UPSs that we maintain. The UPS is a 200kVA Delphys Green Power model from Socomec. It is about 4 years old and has been very reliable and easy to maintain to date.

We had performed a routine half yearly service the week before and found the UPS and batteries to be in good condition. So we were surprised to receive an after hours call to say that there was a rectifier alarm. A quick investigation didn’t identify a cause and after a restart the UPS was running correctly again.

24 hours later, another after hours call, the rectifier alarm was back. Again the UPS restarted fine but this time I organised to carry out a thorough examination the next day. With the help of the Socomec support staff we poured over the parameters and configuration of the UPS. Nothing unusual. We scoured through the log files and debug menus. The only pattern we could see was that the faults had occurred a minute after the boost charge mode had finished.

So we forced the UPS out of boost mode and sure enough the rectifier alarm sprang up. Again we combed through the log files and debug menu. But there was nothing to indicate what the root cause of the fault was.

I went out to the battery bank to see if there was any clues out there. Sure enough there was a plastic smell that hadn’t been present before. I checked the string voltage and the shelf voltages but they all appeared normal. I scanned the strings for temperature to see if one cell was overheating. As I relayed these details back to the Socomec engineer, he suggested that a battery short to earth would bring up a rectifier alarm. I measured the DC voltage from a battery terminal to earth and to my surprise there was 170V present (zero volts expected). Working back along the string I found that the top battery in the string had shorted to earth.

Battery fault between two batteries and earth

Peering along the side of the Vision HF12-890W battery I could see some moisture and debris about half way along the length of the battery.

Close up of battery short to earth

With the battery disconnected I was able to get a better look at the break down.

It is hard for me to say exactly what the cause of this fault is. Clearly the plastic insulating outer shell of the battery has failed. Any mechanical damage or cracking of the plastic could be the cause however it is difficult to identify now that the battery has flashed over to earth. Some possible causes or factors that would make this more likely to occur include:

  • A manufacturing defect in the battery that has taken 4 years to wear through
  • A drop or placement on a hard edge during installation (the battery hasn’t been moved since it was installed 4 years ago)
  • The battery may have been placed on a sliver of metal or some other contaminant piercing the plastic during installation
  • The lack of temperature control of the room housing the batteries could lead to earlier ageing of the plastic as well as increased expansion and contraction of the battery cell
  • The modern transformer-less UPSs may place higher insulation demands on the battery bank due to the non-isolated nature of the power transmission from mains to the batteries.

I haven’t witnessed a fault like this before, though I have heard a few stories from others. Fortunately this UPS has three strings of batteries so we were able to isolate this string and still leave the customer with several hours of runtime. A week later, equipped with new batteries I was back onsite, and after a clean up of the shelf and some extra insulation the UPS was back running at its full capacity.