Designing for electrical system flexibility

Current building codes and standards allow designers to build flexibility into the electrical system.

By Neal Boothe, PE, exp U.S. Services Inc., Maitland, Fla. September 30, 2013

A positive consequence of today’s building codes is typically a built-in degree of flexibility in the electrical system. In the U.S., these codes have stringent requirements for how electrical loads are calculated, and they require that certain demand factors are used to help delineate connected load (the sum of the building load: lighting, HVAC, and appliances that are physically connected to the building’s electrical system) vs. demand load (the load that a building will likely ever experience at any one time). As any electrical designer will tell you, using these allowed demand factors assures that the electrical system is sized to meet the NFPA 70: National Electrical Code criteria. 

For example, consider the NEC requirements for determining the electrical load of the receptacles in a building. Per NEC 220.14, most receptacles must be calculated at 180 VA (1.5 A at 120 V) throughout a building. This load is then summed for every receptacle throughout the building. Per NEC 220.44, the maximum demand factor that can be applied to nondwelling receptacle loads is 100% for loads of up to, and including, 10 kVA, and 50% for loads of more than 10 kVA. Next consider a 100,000-sq-ft office building with a total connected receptacle load of 200 kVA. This code dictates that the demand load must be at least equal to 105 kVA. Now, consider how many receptacles in a typical building do not even have appliances, such as computers, radios, or TVs, plugged into them, and for those that do, how often the equipment is used. Also, there may be a refrigerator plugged in that cycles on for only a few min every hr. That load is considered constant but runs only intermittently. As a result of these factors, the actual receptacle load should always be less than the value allowed by code. 

Another example is air conditioning loads. When designed, these systems are sized assuming a worst-case scenario of having a maximum number of people in a space, with all the equipment running simultaneously, on the hottest day of the year to determine the heat in a space that the air conditioning must handle. Realistically, most days won’t come close to this hottest day of the year. And even on the year’s hottest day, when will every room be at maximum occupancy with all the equipment running at the same time? These air conditioning loads will nearly always run at a reduced rate compared to their capacities. However, the electrical system is designed to accommodate their full capacity per code. Similar stories might exist for other building loads, such as lighting, kitchens, laundries, etc. 

As a result of these differences between code allowed diversity and the actual loads likely to be encountered when a building is in use, a degree of flexibility is inherently built into the design. 

Building in flexibility

There are other ways to design an electrical system to build in flexibility. These concepts may increase the overall cost of the system but can have huge returns if they prevent the need for new electrical costs down the road. Planning and forethought improves electrical design flexibility.

Plan now for future loads in panels: Is there space available for future loads? Filling the electrical panel today prevents the addition of loads in the future unless new panels are added. Having full panels, which leads to the need to add panels for any future work, will cause more disruption and cost to a facility during future renovations and additions. 

Understand possible future loads: If the panel does have expansion capacity, consider what the future load is likely to be. Will it be a small 20-A breaker for more receptacles or lights, or will it need to accommodate a large new motor? How much power would this equipment draw? Make sure the panel (and upstream equipment) has enough capacity—ampacity as well as physical space for a breaker—to accommodate this. 

Electrical room locations: Early in a project, work with the architect to include as many localized electrical rooms as possible. If successful, typically those rooms can be a little smaller so that the overall area needed doesn’t change significantly. It’s much easier to add new loads or otherwise modify the electrical system later when there is a nearby electrical room compared to running several hundred ft away when one isn’t close to an addition or renovation. 

Electrical room sizes: Determining the sizes of electrical rooms also requires coordination with the architect. Try to design electrical rooms so that there is future space for another panel or two to be added in the room. Later, if another panel becomes necessary due to new loads, it’s a lot easier to accommodate if it can fit into an existing electrical room vs. asking for more space and possibly disrupting important building functions. 

Considering these concepts during design can lead to more flexible building electrical systems for many years to come. For example, if the panel is already designed, ensuring it has space for future breakers is a minimal expense. On bid day, this is an item that may not have any additional cost. Also, by asking for a room that is just 2 ft longer during design (for example, 10 ft by 10 ft instead of 10 ft by 8 ft), you may be able to accommodate two more panels later and save the need for another electrical room altogether (more than offsetting the initial investment).

Generator sizing

By sizing electrical systems per national codes and standards, such as NEC, sufficient spare capacity is typically provided without the need to intentionally oversize the electrical system. This is because these codes must be broadly written so that they can apply to diverse building types without causing issues with electrical capacity. For example, an electrical engineer may design a building that requires a 3,000-A electrical service based on calculations, loads, and diversities required by codes. However, after it is in operation, that building may pull only 40% to 50% of this load. While this may sound like the system is oversized, in reality this size was necessary to meet the applicable codes, which are written around worst-case scenarios and inherently provide future flexibility. 

Whereas generators for emergency use are smaller and supply only the necessary power for providing safe egress from a building, using calculated loads and NEC-allowed diversities is still a prudent practice. For example, in a large office building, there may be a small generator that provides power only to loads such as exit signs, egress lighting, fire alarm systems, and other similar loads. When these emergency loads are needed and normal power is unavailable, it is very likely that all these loads will be needed at the same time, so the generator must be sized for this. This is covered by NEC 700.5, which reads that the capacity for emergency generators “shall have adequate capacity and rating for all loads to be operated simultaneously.”

However, consider an emergency generator that may be designed for other applications where maintaining a fully functional building is necessary for an indefinite period of time—not just for emergency evacuation. In this scenario, these generators are much larger and see a more diverse and less constant load because they see normal operating practices, not just a 15 min building emergency egress. 

Keep in mind that generators are basically large engines that convert mechanical power into electrical energy. As engines, these generators are much more efficient running close to their capacity than they are running at significantly reduced loads. Imagine owning a Ferrari and only driving it in stop-and-go, rush-hour traffic and never getting it above 15 mph. Not only would that defeat the purpose of owning the Ferrari, it also could affect the performance of its engine over time. 

Often, codes recognize that sizing emergency generators for these applications does not have to be done via the same calculations that are used for sizing other parts of the electrical system. Because hospitals represent buildings with large emergency loads that must stay operational indefinitely (even without utility power), they will serve as an example (see Figure 1). 

Article 517 of the NEC applies specifically to health care facilities. Section 517.30 (D) reads that generator sizing per NEC 700.5 shall not apply to hospitals. It allows several different options for sizing a hospital generator, including:

  • Prudent demand factors and historical data
  • Connected load
  • Feeder calculation procedures described in NEC Article 220
  • Any combination of the above. 

This allows the engineer either to apply other NEC diversity factors or to rely on historical data and prudent demand factors.

The NEC’s commentary elaborates on why it allows this variation: “The intent of 517.30(D) is to permit the sizing of generators based on actual demand likely to be produced by the connected load of the system at any one time. This method of calculation facilitates practical sizing of generators in health care facilities and helps eliminate prime mover operational problems associated with lightly loaded generators.” 

NFPA 99: Code for Health Care Facilities further explains sizing of hospital generators in its commentary: “…generators should be sized for the actual demand rather than the connected load. All too often, authorities having jurisdiction require generators to be sized based on a mathematical summation of the calculated loads modified by NFPA 70 [NEC] demand factors. Such designs often result in generators that are very large relative to their actual demands. Such designs will impair the reliability of the generators over time.” 

Hospital generators don’t have to be sized in the same manner as an electrical service or even generators for other types of buildings. 

Consider a scenario that requires sizing a larger generator for a nonhospital building with more of the building (nonemergency type loads) on the generator. One may be concerned that using NEC load calculations could cause an oversized generator. NEC Article 702 addresses optional standby systems: loads that aren’t required to be on the generator but may be selected to be for another reason. There is an allowance on how to size a generator that handles these optional loads. NEC 702.5 (B) reads: “the calculations of load on the standby source shall be made in accordance with Article 220 or by another approved means.” Using Article 220 would require following NEC demand factors. Using another approved method opens the door for discussion on how to handle these loads and sizing this generator. As this code is vague at best, consultation with the authority having jurisdiction on how to size this generator would be prudent to prevent problems during code review.

Divide and conquer generator loads

As generator oversizing is not recommended, how should engineers approach a large generator load that will experience significant diversity (will not be fully loaded all the time), but must accommodate heavily loaded conditions? 

In these situations, parallel generators often become the solution (see Figure 2). This practice allows you to install multiple smaller generators instead of one larger generator. This solves the problem of maintaining a larger emergency capacity on the system while lowering the concern of lightly loaded generators. For example, two 500 kW generators provide the same amount of emergency power as a single 1,000 kW generator (see Figure 3). However, if the demand load is less than 50% of the 1,000 kW total, only one generator is needed. Also, in the event of a maintenance issue with either generator, emergency power is still available to the building while repairs can be made.

For even more redundancy, an n+1 system of parallel generators could be employed. Again, assuming a total demand load of 1,000 kW at peak emergency demand, a system with three 500 kW generators could be used. Any time less than 500 kW of actual demand is needed, only one generator would need to run. Any time the load grew above 500 kW (but less than the maximum of 1,000 kW), a second generator would come online. This would always leave a third generator that provides redundancy. The loss of any single generator unit would not affect the building’s emergency capacity (see Figure 4).


Electrical systems should be designed to provide adequate flexibility for the inevitable changes in today’s buildings, such as changes in equipment, additions, or renovations. Fortunately, current electrical codes provide some level of built-in spare capacity as they provide conservative diversity and load guidelines in initial equipment sizing. 

However, other considerations, such as physical space for equipment and spare capacity inside electrical equipment for future breakers, should be considered when designing electrical systems. Furthermore, care must be taken in the sizing of generator systems. Although spare capacity may be needed, it must be carefully weighed against the possible adverse side effects of generators that are too lightly loaded. This is especially true for buildings that stay at full capacity in emergency conditions, such as hospitals, or for buildings that include standby loads (not just emergency egress type loads) on the generator system. In these cases, the designer must be savvy enough to balance proper generator sizing against the need for emergency and/or standby power requirements. In these cases, the use of paralleled generator systems can be employed to keep generators running closer to their capacities—which allows them to run more efficiently and safely—while keeping needed flexibility in the electrical system.

Boothe is a principal and electrical engineer at exp, where he specializes in the design of hospital electrical systems. He has more than 20 years of experience, including more than 200 projects ranging from new, greenfield hospitals to additions and renovations of existing facilities.