Generators and transfer switches for mission critical facilities
Transfer switches and generators are specified into commercial buildings—specifically health care facilities—to provide optional standby, emergency, and legally required power.
- Understand the basic requirements of generators when used for standby or back-up power.
- Learn which code pertains to the design of generators and transfer switches.
- Obtain an overview of the types of systems supplied from transfer switches.
Most commercial building applications require some form of an alternate power source for life safety purposes and to comply with NEC 700 and various other building codes. For small facilities, this typically is achieved by using emergency battery packs in select light fixtures for egress, exit signage, and integral battery backup for life safety equipment such as the fire alarm system (see NFPA 70: National Electrical Code (NEC) 700.12(A)&(F)).
If the facility is larger or maintenance of a large quantity of individual battery packs is a concern, a central battery inverter system could be installed (NEC 700.12(C)). As the name implies, a central battery inverter replaces the individual batteries scattered throughout the building with one or more central locations for the batteries. This type of system must be listed for the purpose and requires dedicated distribution from the inverter system to serve the emergency egress lighting fixtures and exit signs. Alternatively, a standby generator system may be used in lieu of a battery-based inverter system (NEC 700.12(B)). Batteries are expensive and require replacement every 3 to 5 years. Although a generator system also requires scheduled maintenance and testing, it provides the building with additional uses beyond life safety. Before delving into source of power and transfer equipment, we will explore the three different types of code-defined systems.
NEC Chapter 7 has requirements for three distinctive systems that may be served from a standby generator.
- Legally required standby
- Optional standby.
The emergency system loads are addressed in NEC Article 700. Only those loads as defined in Article 700, Part IV are allowed to be connected to this system. Generally, loads specified for “emergency use” that may be connected to this system include emergency light fixtures and exit signs for the designated paths of egress and fire alarm systems. Depending on the codes adopted by the local authority having jurisdiction (AHJ), the loads classified as “emergency use” may differ slightly, but for the most part the International Building Code (IBC), International Fire Code (IFC), and NFPA are fairly consistent.
NEC Article 701 provides the criteria for any legally required standby loads. The NEC defines these loads differently than Article 700 does. They are not designated as “emergency use” and are generally mandated by the AHJ. Legally required standby loads could include other loads considered critical to building evacuation and firefighting, such as the HVAC equipment that provides smoke management during and after a fire, and communication systems.
The systems addressed in the first two articles of NEC Chapter 7 are code required where applicable, but NEC Article 702 allows a third system to be provided for any other loads a building owner may deem critical to its business or when disruption may cause significant financial loss. This system is called the optional standby system.
Health care facilities’ design differs as NEC Article 517 contains more prescriptive requirements for an essential electrical system. This system comprises two components: the emergency system and equipment system (NEC 517.30(B)(1)). These systems may be combined if the total maximum demand does not exceed 150 kVA (NEC 517.30(B)(4)).
The emergency system is composed of two separate subsystems, one for the life safety system and the second for critical branches (NEC 517.30(B)(2)). NEC 517.32 explicitly details the only functions that may be connected to the life safety branch. NEC 517.33 lists the functions related to patient care that may be connected to the critical branch. Although an exhaustive list is not provided, NEC 517.33(A)(9) gives performance criteria to allow the designer to make judgments for minimal additional loads it deems critical to patient care to also be connected to this branch of distribution. NEC Article 517 is comparable to Article 700, but with additional functions and more stringent wiring methods (NEC 517.26). NFPA 99 Appendix A, A.184.108.40.206.7.3(2) also is a good reference point.
The equipment system functions are addressed in NEC 517.34. For the most part, this system serves mechanical equipment deemed necessary for proper function of a hospital with inpatient care. NEC 517.34(B)(8) also allows the engineer to use his or her judgment to connect additional loads not specifically listed.
Continuous, prime, standby
There is often confusion regarding the differences between continuous, prime, and standby power systems—and when each type is appropriate. Most manufacturers publish data for the same generator set with both a prime and standby rating; it’s physically the same piece of equipment. The main differences are rated capacity and warranty implications.
A standby generator is rated with the higher capacity but is warrantied for use only for a limited number of annual hours. The average varying load is typically required to fall below 70% to 80% of the generator nameplate rating. This type of generator is the most common application in typical commercial applications where there is a fairly reliable utility. If the project is located in an area where utility historical data indicates a large number of average yearly outages and/or substantially long outages, the installation may exceed most manufacturer warranty terms. So, for this type of environment, a standby generator may not be the best long-term option. In addition to potential warranty issues, the generator life may be greatly reduced and incurred maintenance costs may be substantially higher.
A prime rated generator is provisioned for continual run time with a correlating reduction in the nameplate capacity as compared to a standby unit. Again, the average varying load is typically required to fall below 70% to 80% of the generator nameplate rating. The lower average loading placed on the generator puts less stress on the components while operating. The manufacturer de-rates the engine for prime applications due to more hours on the engine to ensure the operational life is not compromised. A short duration overload of up to 10% is generally acceptable.
A continuous rated generator is just that: These generators have upgraded components and a larger cooling system that are designed to withstand the stress of operating at 100% of nameplate capacity for an unlimited number of hours. The loads, however, need to be fairly steady or nonvarying, per manufacturer guidelines.
All of these systems typically are connected to the generator through the use of transfer switches. In most cases, automatic transfer switches (ATS) are code required, except in the case of optional standby (NEC 700.6) and some health care equipment system loads (NEC 517.34(B)) where manual transfer switches are permitted. ATS equipment is available in closed- and open-transition configurations. Because a standby generator system that serves emergency and legally required standby loads (or in the case of health care) requires monthly testing (NFPA 110-8.4.2), a closed-transition ATS provides a system with the least disruption to building occupants during testing. The operation of a closed-transition ATS momentarily parallels the utility with the generator before breaking connection from utility. This is often referred to as a make-before-break action. It maintains a reasonably uninterrupted service to downstream loads during routine testing and transfer back to a restored utility following a loss of power sequence.
An open-transition ATS typically is used when the serving utility company does not allow the service to connect close-transition with a generator. In this case, as the transfer occurs, the load momentarily breaks from the utility before making it to the generator source. Re-transfer back to the utility also has the momentary break. It is important to coordinate this design aspect with the local utility company early in the design and to keep the client aware of any limiting factors imposed by utility requirements.
Another option when considering the ATS is whether bypass isolation should be provided. This type of transfer switch includes a second “backup” transfer switching mechanism for use when the main transfer mechanism requires maintenance or replacement. The main transfer switch is manually bypassed to the second transfer switch and then isolated from the system. This feature enables the loads to continue to operate while the ATS is essentially taken out of service, while also maintaining manual transfer capability to generator source if utility power is lost. This type of ATS has a premium cost over a standard ATS and is often specified when the loads are considered extremely mission critical such as in hospitals, financial data centers, and casinos. Some jurisdictions mandate the use of bypass isolation switches; such is the case when applying them to health care facilities in California, which are governed by Office of Statewide Health Planning and Development (OSHPD).
Smaller systems where the generator is located in an outdoor enclosure are usually capable of including a few unit-mounted circuit breakers to serve an emergency, legally required standby, and optional standby ATS without additional distribution equipment. Space is limited, however, and in most cases where the generator is larger than a few hundred kW, a generator distribution switchboard will be required. A requirement that is often overlooked is that the circuit breaker and subsequent feeder that serves the emergency ATS must be located in a separate vertical switchboard section from the legally required and optional standby devices and feeders (NEC 700.9(B)(5)).
In some larger installations where a single generator is not sufficient or multiple gensets are required to provide system resilience, paralleling switchgear (PSG) provides a scalable and intelligent automation option. Typical PSG systems provide the programming and logic to synchronize the generators and optionally operate a transfer pair of circuit breakers. Some PSG designs include a utility feeder input on the load bus. The utility input circuit breaker must be opened and closed simultaneously with the main generator input circuit breaker such that only one input is closed onto the load bus. In this sense, this equipment could operate like an ATS. The PSG can be programmed to operate in manual open transition, manual closed transition, automatic open transition, and automatic closed transition.
Many configurations can be devised to meet the client’s expectations for system performance. Figure 3 represents a simple generator paralleling configuration with minimal additional distribution costs. Keep in mind that the functionality of this system will be limited to the synchronizing of generators only. A few generator manufacturers can provide synchronizing equipment without the need for sophisticated PSG programmable logic control (PLC) hardware. The ATSs perform the remaining sequencing and operate with little difference from a traditional one-generator configuration.
In Figure 4, the PSG has taken on a larger role in the sequence of operations. It includes a programmable logic controller (PLC) and automatic breaker control of sources and loads. There is a transfer pair that maintains the currently available source to all downstream feeder breakers, though some AHJs reject this technique, so engineers must work closely with the AHJ. During a loss of utility, all breakers open and the generator main and priority 1 emergency feeder breakers close. The individual generator breakers remain open until the generator is synchronized. Simultaneously, all generators are provided a start signal and the first one to stabilize closes to the generator bus. This must occur within 10 seconds (NEC 700.12). As additional generators synchronize with the first, they close to the generator bus. Additional feeder breakers may then close until all feeder breakers are closed or the maximum capacity of the generator system is reached.
Dynamic real-time load demand management functionality can be provided to add and subtract generators from the system as loading dictates. For example, if the peak demand in the system meets a specific programmed setting indicating all four generators are not required, the excess generator may go into a programmed shutdown sequence initiated by the demand management system. This configuration will normally provide a robust facility backup system because of aggregation of generator capacity. It also can accommodate large motor starting currents and nonlinear loads.
On-site fuel storage
The generator run time is largely dependent on the amount of fuel storage. The most common fuel source for a standby generator system is diesel. Natural gas and hybrid generators also are available; this article will focus on diesel.
The first priority for determining the amount of on-site fuel needed to be stored is to study the load applications. If loads are limited to backup for life safety to meet code minimum criteria, the fuel supply must provide a minimum of 2 hours of run time (NEC 700.12(B)(2)). However, if additional standby loads are served or the client indicates a need to provide additional backup time, the fuel storage must be sized appropriately to address the requirement. The amount of fuel stored on-site must be considered; diesel fuel does not remain very stable for long and will foul if untreated and/or stored for an extended period of time (usually more than 3 to 6 months). There are methods for conditioning the fuel, but these methods add complication to the system. To determine the amount of fuel required for the desired run time, obtain product data from a few manufacturers that meet the system specifications. A good rule of thumb is 7 gal/hr/100 kW of generator nameplate at full load. Again, fuel consumption should be reviewed with the specific manufacturer for the project’s generator.
The manufacturer data will most often include the amount of fuel consumption at 50%, 75%, and 100% loading. When determining the amount of fuel storage, the prudent design approach is to use the value for 100% loading. Even if the system is not fully loaded initially, this provides a safety factor to ensure the system run time does not fall below client expectations or code minimum requirements in the event additional loads are incorporated in the future.
There are multiple methods for on-site fuel storage. For exterior generators, a sub-base (or belly) tank is usually the least complicated. It should be sized to provide the amount of fuel required to operate the individual generator through the specified number of hours. In this sense, it is a self-contained fuel system and is a part of the packaged generator system. If a centralized fuel farm is to be provided, then additional components and design will be required.
The fuel storage and supply must be designed by a qualified professional. Each generator will require a day tank in the form of a small fuel storage tank located on or in close proximity to the generator to provide the initial fuel for startup. Day tanks are typically sized to store 15 min to 1 hr of fuel, based on the generator fuel consumption rate. The day tanks will have fuel piping from the bulk storage tank(s) with additional return piping from the generators back to the bulk storage tanks for fuel recirculation.
Recirculation back to the bulk storage tank is generally advised as opposed to being returned to the small day tank. The temperature of the unused fuel is elevated after passing through the engine and, if returned directly to the small day tank, may cause engine trip on fuel over temperature. Returning the unused fuel to the bulk storage reduces this risk because the fuel is cooled when mixed with the higher volume of fuel in the bulk storage tank. Day tanks add capacity to the system in addition to the bulk storage tank, which in most cases allows for a reduction of the bulk storage tank capacity equal to the aggregate total of day tank capacity.
Depending on the type of facility and operational expectations/requirements, the level of resilience built into the system varies. Cost typically is a concern, so if no resilience is expected, then a single generator is likely sufficient. If the client expects a higher level of redundancy, then a sophisticated PSG system can provide the flexibility and logic required for implementation of a highly resilient system.
There are a variety of approaches between a minimalistic single unit and a highly redundant PSG system. As shown in Figure 6, a 12.47 kV system where the estimated peak demand of priorities 1, 2, and 3 on the generator system is 16 MW. The breakdown of demand for each system and feeder is detailed in Table 1. Note the generator system includes some normal feeders (or low-priority optional standby), but the system capacity only accounts for the anticipated priority 1, 2, and 3 loads. The additional normal feeders are included to provide access to the generator for additional non-essential loads as system capacity permit. For lower priority feeders, this approach takes advantage of the diversity in the system loading rather than increasing capacity based on the peak requirements of all connected feeders. When the peak demand exceeds the available capacity of the generator system, the PSG real-time demand management functionality will shed the nonessential loads based on the established priorities.
There are a total of nine 2 MW/2.5 MVA generators connected between three generator buses. This provides a capacity that ensures the top three priorities always have access to generator backup, even upon the loss of one generator. However, it is not uncommon for a generator system to be lightly loaded as diversity of loads based on time of day or seasonal variations for demand is a factor that creates capacity for feeders to additional normal loads. Large PSG systems are, for the most part, not typical installations. However, in our experience with high-profile mega-resorts where the client has a budget that does not support complete property backup and has high expectations to maintain a seamless guest experience whenever possible, PSG systems have proven to be a viable, cost-effective approach.
The robustness of this system is twofold. It is technically an N+1 configuration for the three highest priorities through the use of load pickup and shedding. It also provides the client with the opportunity to maximize the system usage during off-peak times of the year and supply other loads. An additional benefit of the PSG system is flexibility for planned outages or for outages with extended duration as the owner can decide which loads, beyond code minimum, are desired based on the situation.
Determine which system—emergency, legally required standby, or optional standby—is required by code, and carefully select and specify the product to best suit the facility’s needs.
Robert R. Jones Jr. is an electrical project engineer at JBA Consulting Engineers with more than 10 years of design experience. Jones has experience in market sectors including hospitality, commercial, medical, and government projects. He specializes in medium- and low-voltage distribution systems, emergency/standby power systems, renewable energy design and implementation, circuit analysis calculations, and equipment space planning.