With data center energy and operations costs rising, you need to find areas where you can reduce costs to compensate. Energy efficiency is certainly a big one, but another, often overlooked area is adhering to a routine equipment maintenance program in order to extend the life cycle of your infrastructure, avoid downtime, and lower your total cost of ownership (TCO).
Equipment breaks down for a number of reasons. Based on Schneider Electric and Hartford Steam Boiler’s expert assessment, some of the more common – and avoidable – reasons include:
- Faulty parts, which account for 34% of failures. Part of a routine maintenance program involves exercising equipment to determine whether certain parts are showing signs of wear and impending failure. It’s like replacing the timing belt on your car before it fails, rather than have it give out when you’re on the highway. Too many companies fail to take this step, and then wonder why devices don’t operate when needed.
- Improper maintenance work. This accounts for 17% of failures. You need to know who is doing your maintenance work, what their qualifications are, what training programs they’ve completed, and what safety ratings and policies they have in place to ensure they are well-qualified to work on your critical infrastructure.
- Environmental issues, accounting for some 8% of failures. For example: do substation rooms have positive air pressure, and is there dust, dirt or other contaminants that can get into the equipment? Is the temperature appropriate for the equipment in question? Humidity alone is responsible for 9% of failures. Moisture and electrical equipment do not go well together. You wouldn’t use a hair dryer when taking a bath, right? The dust and dirt that can accumulate in neglected electrical systems is a magnet for moisture that can promote system failures. Keep it clean and dry!
- Improper operation, where a worker is not trained in the proper operation of the equipment, contributes to 9% of failures.
Failure to properly maintain equipment can cause system downtime, which can cause costs to a business in a number of ways. Let’s take a production facility as an example: an equipment failure that causes a disruption in business operation can result in lost future profits, because no products can be made and only existing inventory can be sold. Fixed costs such as salaries continue accumulate, but are now essentially wasted because no product is being produced. Some variable costs, such as power and cooling, may fall because of reduced consumption, but others, such as maintenance, will rise sharply in response to the incident. The list goes on and on.
The goal of any facility should be to maximize the useful service life of all installed assets, perhaps none more important than critical data center infrastructure. A well thought-out maintenance program consists of three types of maintenance: reactive, preventive and predictive.
Reactive, or corrective, maintenance is typically performed in response to unplanned events, usually an outage of some kind. Such maintenance is twice as expensive as preventive maintenance, and 10 times as expensive as predictive maintenance.
Preventive maintenance means following standard procedures on all gear, regardless of its condition. Preventive maintenance costs half as much as corrective maintenance, because it involves fewer parts and is typically performed on a schedule during normal business. However, preventive maintenance costs five times as much as predictive maintenance, and about 60% of preventive maintenance is unnecessary.
Predictive, or condition-based maintenance is the least expensive form of maintenance in a well thought-out maintenance program, because predictive maintenance is performed only at the time the equipment actually needs it. To be done properly, this involves continually monitoring the health of the equipment, which is now more and more feasible since we’re able to thoroughly instrument just about any piece of data center equipment. This reporting can be sent to a central monitoring tool, such as a data center infrastructure management system, enabling data center operators to accurately predict when a piece of equipment may have an issue – or may not perform as it should – and therefore will need maintenance.
A proper data center maintenance program improves reliability of your critical infrastructure while enhancing safety and, ultimately, saving money. The right blend of reactive, preventive, and predictive maintenance is essential for a well-thought out maintenance program. Putting the most focus on predictive maintenance will save more money long term, and will provide the greatest improvement in TCO.