The design of an emergency power design extends beyond just the electrical aspects of the electrical generation and load transfer equipment. In the event a generator fails to start or enter an overload condition, steps must be taken to selectively reduce the emergency load to a level allowing the generator plant to support higher-priority loads. A brief description is given for some general emergency system design considerations followed by a discussion of selective load shed through the facility’s building automation system (BAS).

The essential electrical system (EES) provides three separate branches for distribution of generator standby power. Each branch is provided with at least one automatic transfer switch (ATS) as required by National Fire Protection Association (NFPA) 70 National Electrical Code (NEC) Article 517 for an EES greater than 150 kVA.

NFPA 99 allows a single generator to power the EES. Providing two or more generators operating in parallel can provide partial or N+1 redundancy. Incorporating parallel generators can mitigate problems that could otherwise result from generator failure with only a single-generator emergency power system (EPS).

Neither NFPA 99 nor Facility Guidelines Institute (FGI) standards require the EES to be supplied by multiple generators. When the EES is supplied by a single generator, NFPA 70, Article 700.3(F), requires a docking station to allow connection of a temporary generator to supply the EES while the permanent generator is undergoing scheduled maintenance or if generator failure occurs.

Operating in Parallel

The resiliency of the EES can be improved by combining two or more generators operating in parallel. In a parallel configuration, the EES demand load is distributed equally among the generators. All parallel generators should have the same kilowatt (kW) rating, and it must be large enough so if any one generator fails, the combined kW ratings of the remaining generators can continue supporting the full EES demand load without the need for shedding load.

For example, if the demand load requires two generators, a third is added to provide redundancy for a single generator failure. This is typically referred to as N+1 redundancy. The individual generator kW raring should also be large enough to properly start all life safety loads, critical loads, and the fire pump.

If the design uses paralleled generators without N+1 redundancy, a load-shedding scheme should be designed in the event a generator fails. The load-shedding scheme can incorporate BAS control to load-shed selected mechanical equipment supported by an ATS rather than by dropping all of the ATS load when it may not be necessary.

Another design consideration, whether using N+1 redundancy or no generator redundancy, is to limit the size of each ATS where practicable so that load-shedding can be more selective. For example, a mechanical load supported by a 2,000-amp ATS could be supported by two 1,000-amp switches. This would avoid dropping all 2,000 amps if only 1,000 amps or less is necessary to avoid generator overload.

When a hospital has multiple chiller systems, the designer may consider placing each chiller system (including any associated pumps, cooling towers, and basin heaters) on its own ATS. This will allow selective load-shedding rather than dropping all chiller systems. The designer may also consider providing a dedicated ATS for mechanical equipment serving surgical suites. This allows an equipment branch for each ATS with a lesser priority to drop load without disrupting surgical procedures.

Imaging equipment can be provided with generator power through an optional standby branch (NFPA 70, Article 702) provided with either a manual or an ATS. An ATS with a long delay for the lowest priority may be preferred, as the Imaging load can be automatically dropped to prevent generator overload.

System Reliability

An on-site, permanent load bank can be provided to allow required exercising and testing of the EES without disruption to the hospital’s loads. (Refer to NFPA 110 and NFPA 99 for EES exercising and testing requirements. All references to NFPA 70 NEC are from the 2020 edition. All references to NFPA 99 Health Care Facilities Code are from the 2021 edition.)

Life safety branch and critical branch ATSs may utilize closed-transition switching. This switching function allows loads to be switched between power sources without disruption to the loads served. This switching function is also referred to as “make-before-break” in that both sets of source contacts are momentarily closed at the same time. This function allows for a momentary paralleling of the generator plant and the utility and, therefore, requires coordination with the utility company. Some utility companies may prohibit the use of closed-transition transfer switches.

The BAS is equally important when designing EESs and must not be overlooked. The electrical and mechanical engineers must work closely when designing such a critical and potentially lifesaving emergency response system. It is crucial these systems are controlled and monitored to ensure the building’s systems are functioning properly should an emergency arise.

For the BAS to appropriately react and control equipment in an emergency, the BAS must monitor all input positions for each ATS. The ATS’s change of position tells the BAS to start a specific sequence of operation (SOO). Knowing the position of the ATS can also provide the facility operator invaluable troubleshooting information in the event of equipment malfunction. These BAS inputs include normal power position, neutral position, emergency power position, and emergency to normal pre-transfer signal.

Two Types of Automatic Transfer Switches

There are primarily two types of ATSs: closed and open transition. Closed transition is known as make-before-break in that the transfer switch makes a connection to the new power source prior to breaking its connection from the old one. A closed-transition ATS provides a continuous power source as it switches between power sources, meaning equipment is not required to shut down and restart. However, not all utility providers allow closed-transition transfer switches. An open-transition ATS is known as break-before-make in that the transfer switch breaks the connection to the old power source prior to making its connection to the new one. All connected mechanical equipment must shut down and restart as the power is disrupted, and the power source switches from utility to generator and vice versa.

When dealing with an open-transition ATS, the BAS will also lose power for a short period of time and shutdown, leaving the building uncontrolled until the BAS is able to restart once the new power source is connected. For critical applications, it’s imperative to minimize downtime for the BAS to reduce the risk to the health and safety of building occupants. For this reason, it’s recommended to include an uninterrupted power supply (UPS) in all BAS control panels. The UPS allows the BAS to ride through the power disruption and immediately begin an orderly startup-specific sequence of operation for emergency equipment.

FIGURE 1. The automatic transfer switch (ATS) and generator control points needed to successfully monitor and control essential electrical systems at the building management system (BMS). Images courtesy of SmithGroup

Orderly startup and shutdown SOOs are often ignored during the design process by the engineer of record and put into place during the commissioning of the building. However, this crucial piece of the emergency response system must be considered and planned early in the design process to ensure it aligns with the owner’s project requirements (OPRs). The orderly startup SOOs should clearly define the order in which HVAC systems must start in an emergency operation. In addition, an orderly startup SOO limits all equipment from simultaneously starting and prevents an inrush current from tripping electrical equipment. An orderly shutdown SOO defines the order to shutdown equipment to prevent potential equipment failures. Furthermore, orderly shutdown and startup SOOs may include load-shedding scenarios.

Although the aforementioned emergency events occur few and far between, the generators are required to be tested monthly and annually per NFPA 110. During testing, it’s imperative to minimize disruption to the operation of the building and to design a testing protocol that allows for the seamless transition between normal and emergency power. The generator test SOO should be developed as part of the construction documents and address how to test generators while minimizing disruptions to the building. An example of an orderly shutdown in a generator test SOO is as follows:

1. General description

  • a. The building operates under normal power and emergency power due to scheduled testing of the generators; and
  • b. The emergency generators are on.

2. Upon receiving a signal from the generator's test switch to "stop," the generator test mode will end, and the ATS will switch to normal power within five minutes (ADJ.), as timed by the electrical system. The BMS shall send an inhibit signal to the electrical system to prevent the electrical system from switching to normal power.

3. The BMS shall maintain the inhibit signal active as long as required to implement the following actions:

  • a. Disable AHU-1, AHU-2, AHU-3, and AHU-4;
  • b. Disable the chilled water plant;
  • c. Disable the steam/heating hot water plant;
  • d. Disable all fan coil units;
  • e. Disable all remaining HVAC equipment; and
  • f. All HVAC equipment that is not on emergency power but still has normal power available shall operate under a “generator test mode.”

4. After all HVAC equipment has been disabled, as described above, the BMS shall wait five minutes (ADJ) and deactivate the inhibit signal to the electrical system such that the electrical system can switch to normal power.

The first bullet of the sequence describes the general scenario when to apply the following SOOs. The second bullet defines when to initiate the following sequence. Further, the second bullet explains to send an artificial inhibit signal to be programmed into the BAS. The inhibit signal is a signal from the BAS to the electrical system to delay the ATS from switching positions from emergency power to normal power or vice versa. The inhibit signal allows the appropriate time for the BAS to perform orderly startup or orderly shutdown functions (in this case, orderly shutdown). The third bullet point defines the order in which to disable each HVAC system. The fourth bullet point explains that once all equipment is shut down and the inhibit signal is disabled, the system will then switch back to normal power and begin the orderly startup sequence of operation. A version of the above SOO should be modified to include a generator test orderly startup sequence.

FIGURE 2A & 2B. A visual representation of the building automation system (BAS) sequence of operations required for each type of automatic transfer switch (ATS).


Engineers have a lot to think about when designing and monitoring EESs. There are many different ways to design code-compliant systems; however, it’s imperative the electrical and mechanical engineers of record work together with the owner early in the design process to think about how these systems will work together to provide the greatest system redundancy and resiliency for the entire life of the system. It could literally be a life-or-death situation.