How applicable are tier ratings to UPS systems?
Established by the Uptime Institute (TUI) in the 1990s, data centre tiers are a convenient system to organise types of data centre infrastructure based on a set of operating criteria. The classification ranks facilities from 1 (l) to 4 (llll) with 1 being the lowest rank and 4 being the best performing level.
The exact method of assigning tiers is kept a secret by the Uptime Institute however, the most important metrics are common knowledge: Service availability and uptime guarantees, redundancy levels, cooling capacity and power infrastructure, security, staff expertise and maintenance protocols to name a few.
Customers typically use the tiers to help them to decide which facility to trust for their critical data. However, the facility tier rating is not as applicable to the performance of a UPS (uninterruptible power supply) in good working order. A UPS installation can be made to be as robust as necessary irrespective of the facility tier rating. It comes down to what is actually required.
One of the metrics, service availability refers to the five 9’s of uptime. Whilst the 99.9999% availability is a useful metric, it is irrelevant for power availability. A facility could see an outage every day of the year and still fulfil this metric or KPI. For UPS, the metric has got to be for it to fulfil its purpose irrespective of the overall facility tier rating.
The tier rating of the facility will determine what happens when there is an external failure outside of the control of the UPS, such as an issue resulting in no power to the UPS (the batteries only have a finite charge in them and once it’s gone, the load will lose power.) It does not determine what is within the control of the UPS and how the system is made as resilient as possible within the different design configurations.
A more appropriate way to measure and configure UPS is to limit the effects of any faults as far as possible, this is where the tier ratings do have a small impact – fault tolerance and concurrent maintainability.
A UPS will likely have a fault at some point, they are in constant 24/7 operation and are made up of consumable electrical components. The internal makeup of the UPS such as PCB’s, component ratings, airflow designs, etc, are ultimately what can affect the reliability of the single UPS.
Director at Power Control, Rob mather commented “The majority of UPS on the market today are relatively reliable compared to a decade ago. At that time there were some UPS where the reliability of the power would have been improved if the UPS was taken out of a circuit.
There are still differences, but the main commercial forces prevent poor units from gaining any market traction. For this reason, we will focus on how the overall UPS configuration and installation can be done to improve the reliability of achieving a supply to the load 100% of the time.”
A major cause of UPS faults is the batteries. Batteries are often an overlooked aspect of the overall UPS design, yet if they are done wrong, it has a huge impact on the ability of the UPS to support the load and potentially causes a fire risk.
Rob Mather commented “For this reason, I am always surprised when there is an installation with a common battery that has no redundancy but N+1 of UPS. Whilst this sort of installation makes maintenance easier, it does not really improve the reliability of the system considering that general batteries are more likely to be the cause of a fault than the UPS.”
How to increase UPS reliability
Design the UPS to be operated at a maximum 80% capacity. Constant operation higher than this will result in more wear and tear on components which mean the efficiency of the UPS will degrade sooner, and potentially result in needing to replace consumable items sooner. There are numerous reasons for doing this, and the only reason for not is cost. Furthermore, nearly all modern UPS will be more efficient at 80% load than 100% anyway.
Consider multiple battery strings with redundancy if having a single UPS system.
If only installing a single UPS system, consider a modular UPS with a redundant module, and redundant battery string. Or have an N+1 UPS system each having its own batteries, this means that maintenance can be carried out and in the event of a fault the load should remain protected.
Is it feasible to have a 2N installation from the UPS down?
Even if there is only a single main supply and LV panel available, splitting the supplies from the UPS will add an increased level of resilience. This will not be a full tier III infrastructure but is going to make the supply of UPS backed power more reliable without the full expense. This will enable concurrent maintainability for the UPS systems and provides redundancy in the event of a fault on the UPS.
Power Control ltd has been delivering UPS solutions into data centre facilities of all tier ratings for almost three decades. We work closely with our UPS manufacturing partners to design the most reliable configurations for all sizes and budgets. Our nationwide team of engineers are well experienced in installing, commissioning, and maintaining these solutions across all facility tiers. For more information, please contact us 01246 431431 or email info@powercontrol.co.uk