We work in a mission-critical world where reliability is not just a commodity, but an absolute necessity. However, a system’s reliability declines exponentially as equipment ages. The only way to combat this loss of reliability is through an organized system of maintenance. The quality of this maintenance can significantly impact the long-term system availability.
The importance of an electrical preventative maintenance (EPM) program is a widely accepted tenet of data center professionals. We know that equipment lasts longer and performs better when covered by an effective maintenance program. While we recognize that equipment maintenance is crucial to the long-term operation of a data center, what many overlook is how the consideration of maintenance in the initial design enhances reliability.
There is a key opportunity in the early stages of data center design to make both scheduling and implementing maintenance more realistic. This includes specifications for complete system start up and benchmark testing. Due to the high cost of downtime while a data center is out of commission, proper planning of all the components (inspection, testing, and repair) of the maintenance program is vital to the overall success. When it is an afterthought in the design, the necessary downtimes will increase and maintenance often will either be delayed or eliminated.
The first step to an effective EPM program is the architecture of the data center itself and certain design elements should be evaluated to augment the maintenance program’s success. Many equipment features permit easier maintenance during operation, such as dual feeds, draw-out circuit breakers, infrared windows, hinged doors, and barriers between sections. For example, draw-out equipment can be quickly removed from service for testing, while fixed-mounted equipment usually requires an extended shut-down for removal from the equipment.
Another consideration is designing the data center to allow shutdowns of a small portion of the system for maintenance by grouping the equipment. If all of the personnel and test equipment can be staged and located in one part of the facility, the process will become smoother and quicker.
Data center designers should take into account the maintainability of the system in addition to its reliability. In an ideal world, the initial design of a power system should go beyond value-engineering and account for both the loads being serviced in the facility and the ease of maintenance. This requires consideration of equipment isolation for access, working clearances, equipment handling, equipment spares, personnel training, maintenance schedules and inspection checklists.
Further, new and existing government regulations need to be addressed as they affect the system’s maintainability as well. Arc flash calculations and procedures must be up to date following the NFPA 70E and the OSHA requirements. Per these regulations, the necessary personal protective equipment worn by the maintenance personnel when they work on the equipment is another factor influencing the ease of maintenance.
Because the EPM program should be tailored to each individual system, details such as types of equipment, availability of downtimes, operating conditions, and the nature of the loads must be considered. Hostile environments, overloads, and severe duty cycles can increase the need for regular maintenance. Clean and dry conditions could require annual maintenance, while a hot, dirty environment will require more frequent maintenance. In addition, if equipment feeds critical loads, it should be maintained regularly to decrease the likelihood of failure. Two systems with identical designs could have different EPM programs if they are operated in a different manner.
The risk of failures greatly decreases not just with the frequency, but with the quality of an EPM program. The quality of the maintenance program is bolstered when it is incorporated into the initial design of the data center and incorporates key elements to make it successful. Maintenance concerns are best addressed if they are brought to the forefront during the concept and design phases of a project, not after the system has been commissioned.
If you would like detailed information on data center maintenance, please refer to the application theories, Maintaining the Long Term Reliability of Critical Power Systems and Power System Equipment Maintenance and Long-Term Servicing Requirements for Data Centers, at sedatacenters.com.