Most data center managers are tasked with striking a balance between maintaining high levels of uptime and controlling both CAPEX and OPEX. When confronted with the dilemma of aging UPS infrastructure (a key component for keeping data centers up and running) data centers managers have a choice of either pursuing a “run to fail” approach, upgrading the existing system, or buying new UPS infrastructure.
Making the right decision will depend on a number of unique factors. These factors include the criticality of the data center, future load requirements, whether or not executives are planning to outsource some of the data center operations, and whether cost savings surrounding energy efficiency are a factor.
Let’s look at the three decision scenarios to determine under which circumstances each decision makes the most sense:
- Run to fail – Under this scenario, the older UPS in kept in service until “end of life” is reached. Assuming the UPS does not unexpectedly fail or drop the load, it is the lowest cost option with the least amount of disruption to day-to-day operations. However, the risk of sudden failure impacting the load is higher. As a UPS ages, maintenance costs will rise. Experience has shown that “time & material” maintenance expenses more than double once a UPS reaches 10 years of age.
Although risk of downtime is increased under such a scenario (especially if the overall degree of redundancy in the data center is weak), “run to fail” serves as an appropriate option when no budget exists for either an upgrade or purchase of new equipment. In that case, what funds are available should be spent investing in a service contract that will handle sudden failures. If the facility will soon be consolidated or if a short term move to the cloud is anticipated, the “run to fail” scenario may also present a sensible option.
- Upgrade – Whether a UPS upgrade can be cost justified depends on the manufacturer and model of the particular UPS. Some UPSs are more modular and scalable than others. Typical parts and subsystems that are replaceable include battery strings, DC capacitors, fan assemblies, AC capacitors, Power supply units (PSU), intelligence modules and controller boards, inverter assemblies, static switches, rectifier SCR, and IGBTs. Compared to the “run to fail” scenario, downtime risk is reduced when the existing UPS is upgraded.
Upgrades are less time consuming than a “buy new” scenario. An upgrade also presents less of a risk to the load. Upgrade tasks can be performed in a few hours with the UPS in maintenance bypass while the critical load is powered by the utility or a redundant power source.
- Buy new – Replacing an older legacy UPS with a new UPS offers high short duration risk during the changeover, but the lowest long-term risk profile. Initial capital expenditures will be higher than performing an upgrade, but this needs to be balanced against the lower operating expenses (such as lower service costs and higher efficiency performance). Purchase of a new UPS becomes a more attractive option when the vendor of the existing UPS no longer offers support and spare parts are unavailable. Oftentimes with older UPS models, the capacity and efficiency ratings do not address current or future needs. In some regions, government incentives or tax discounts may exist for companies that install newer, high-efficiency UPSs. In some situations, the fear of system failure is so acute that only a new UPS can provide the peace of mind required to minimize high levels of stress.
When dealing with an aging UPS, there is no one right answer. Understanding the current situation and future requirements will largely determine whether it makes more sense to “run to fail”, upgrade, or buy new. For more details on managing aging data center UPS systems, download the Schneider Electric White Paper 214, “Guidance on What to do with an Older UPS”.