Generator Sizing: Avoid the Pitfalls
Common generator application traps can cause significant project pain. The limitations of traditional sizing programs, load uncertainty, power factor, motor starting and generator transients, soft starting surprises, and harmonics are just some of the issues to contend with.
A November 8, 2008 webcast, hosted by Consulting-Specifying Engineer and sponsored by Generac Power Systems, focused on how to achieve a successful design by avoiding common sizing pitfalls. After the speakers concluded their presentation, a lively question-and-answer session ensued. This article captures many of the questions and answers that will be of interest to a wide audience of engineers who design and specify standby power systems. The questions are grouped by equipment types.
Isn’t a generator an inductive device in which the power actually lags voltage?
The power factor of the generator is determined by the load characteristics. Typical loads usually have a mix of motors and resistive elements. This means that most generators operate in a lagging power factor mode providing kW and kVAR to meet the power needs of the connected load.
If power factor correction capacitors are connected to an electric service, do they need to be dropped offline when switching to generator power?
Typically, yes. Generators are designed to work between 0.8 lagging to 1.0 (unity) power factor. Lagging power factor is when the current lags the voltage. Unity power factor is when current and voltage are in phase with each other. Leading power factor is when the current leads the voltage. Leading power factor load conditions will cause the automatic voltage regulator to reduce its output. This may lead to voltage instability at modest load levels and over-voltage shutdown at high load levels. To avoid these results, power factor correction may need to be switched off when operating on the generator.
If you have permant magnet generator (PMG) excitation, is there still the concern for leading power factor?
Yes. The power factor is determined by the load, not by the source of the power (generator or utility). The PMG has no effect, so the cause of leading power factor still will need to be corrected.
What happens when extremely low power factor loads (e.g., 0.5 constant operating power factor) are operated on a diesel generator that is tested and sized for loads operating at 0.8 power factor or higher?
When lagging power factor loads are applied to a generator set, the automatic voltage regulator increases power to the alternator field to maintain the desired output voltage. If the load is near the generator’s rated kW and at a low power factor, the increase in excitation needs and stator current could cause excessive heating. This is very uncommon and can be corrected by using an upsized alternator.
Does shutdown due to steady state overload typically damage the genset?
No. Overloading a generator will cause the engine speed to pull down. The genset then will typically shut down on low frequency or engine over-temperature. Gensets are generally engine limited. The alternator is largely unaffected due to excessive thermal capacity built into the system. This capacity is a result of a power factor rating of 0.8 and a UL2200 temperature rise limit of 120 F. Note that overload is not the same thing as short circuit. Short circuit events are not engine limited and require the generator’s internal protection to trip the unit.
Where can you find information on a generator’s harmonic impedance?
Alternator impedance can be found on an alternator’s specification sheet. It is normally referred to as alternator reactance. The reactance element that is used in harmonic analysis is the subtransient reactance (X”d). This is also the reactance that is used to determine instantaneous fault current levels. It is the reactance displayed by the alternator in the first few cycles of a transient event.
Is it true that both branches of a 220 V generator should be loaded equally (or at least as close as possible) on the 110 V circuits?
Yes. Balancing the load pulls power from all the stator coil groups equally. An imbalance loads one particular coil more than the rest. This is most critical in 240 delta connections that provide power from six coil groups.
AUTOMATIC TRANSFER SWITCHES (ATs)
Is the neutral always grounded at the alternator?
The neutral or grounded conduction is not always bonded at the alternator. In a separately derived system (a system with a 4-pole transfer switch), the neutral is bonded at the alternator or at the first disconnect. In a nonseparately derived system (a system with a 3-pole transfer switch), the neutral is pulled back through the ATS and bonded with the utility at the main service. Even though the bonding may be remote from the generator, the generator always requires a low-impedance equipment grounding conductor. A grounding electrode does not meet this requirement.
In a 480/277 V application supplying a data center with 480 V in and 480 V out UPS units, the manufacturer recommends a neutral input for the UPS primarily for return path for out-of-phase current during a static bypass event to prevent service GFI tripping. Do you recommend a 3- or 4-pole ATS for this configuration?
When evaluating whether to use a 3- or 4-pole transfer switch configuration, the current return path should be traced back to the source for both normal system neutral current and ground fault current. Once the current path for each is identified, the correct operation of ground fault indication and protection can be identified. In this particular example, the one line would need to be evaluated before a final recommendation could be established, but a correctly bonded and grounded separately derived system always minimizes the risks of unexpected current flow paths.
What type of time delay do you recommend between steps, for stepping loads (such as from sequenced transfer switches)?
To minimize the size of your genset, sequence the loads onto the generator. In most applications this happens naturally via natural load sequencing, but some applications require forced sequencing. The load steps should be long enough for the alternator to recover before adding the next step. This is typically a couple of seconds.
Where can you shed load? At the ATS?
The ability to shed load is typical in paralleled generator applications and used on some single generator projects. It is implemented to protect the generator system from overload transients and shutdowns. It can be implemented via shunt trip distribution breakers, building automation, load control circuits, and through controlling transfer switch loading. When load shedding with a transfer switch, verify that the switch can support a tripping to a disconnected position. The transfer switch would contain a load shed input contact that would be controlled from the generator or some type of system load controller.
When specifying the generator, do you always have to specify the battery charger, or will it automatically come with the package?
A battery charger should always be included in the generator specification. Typical sizes are 2 amp and 10 amp. NFPA 110 (Table 126.96.36.199) requires 24 hours recharge time for a Level 1 system.
Is the generator typically equipped with a main circuit breaker? If yes, is the breaker 100% rated?
The standard circuit breaker for a genset is an 80% rated, thermal magnetic breaker. This breaker is usually sized between 100% to 125% of the generator’s rated amps. Some engineers specify 100% rated breakers thinking that they are needed to ensure that the generator breakers don’t experience a false trip. This is not the case for multiple reasons: the generator is not operating at 100% load and running at 0.8 power factor, and breakers are located in a high-velocity airstream inside the alternator’s connection box.
How can you achieve selective coordination on an emergency system?
For systems requiring selective coordination, two options can be considered: setting the thermal magnetic breaker’s instantaneous trip above the alternator’s available fault current, or use an electronic trip breaker. The electronic trip option seems to be more common, given the added control over the breaker’s short- and long-time trip curves. Note that a PMG excitation system for the alternator also should be implemented to ensure fault current capabilities.
Does the National Electrical Code (NEC) require the generator to be sized for all loads connected at one time?
Sizing of the generator per the NEC depends on the application. Emergency systems (NEC 700) are required to be sized for connected load, unless the application is an NEC 517 healthcare application; in that case, the load is sized based on expected load levels and prudent design. If the application is NEC 701 legally required, the generator is sized for loads that are expected to operate at one time. For optional standby sizing, the code requires compliance with NEC 220 calculations or an authority having jurisdiction-accepted load calculation/sizing process.
To what maximum percentage of generator capacity should elevator loads be limited?
When sizing a generator for operation on an elevator, the type of elevator determines some of the sizing issues. Cable/traction elevators typically use four-quadrant VFD technology. This means that the alternator will need to be sized for the drive’s harmonic content and the system will need to dissipate the elevator’s regenerative power. When an elevator is lowered, the elevator motor operates as a break, resulting in the elevator drive pushing power back into the building. When the elevator is isolated on a generator, plan for other connected loads to burn away this power.
Do hydraulic elevators experience the same regeneration problems that traction elevators experience?
Generator sizing with elevators is mainly a problem with cable/traction elevators, not with hydraulic elevators. For applications with hydraulic elevators, regeneration is not a concern, but the generator needs to be sized to accept the elevator motor’s starting transients while maintaining voltage dip tolerances.
Larger, natural gas-powered generators seem to be growing in popularity. Are manufacturers improving the kVA levels and minimum start times?
It is true that we are seeing greater demand for larger natural gas systems. Most natural gas generators from manufacturers of standby generators meet 10-sec start requirements. As to the transient capabilities, natural gas generators historically have been less responsive than diesels, but a few suppliers have started optimizing the operating rpm of spark-ignited generators, resulting in diesel-like transient performance.
What precautions should be taken with exterior diesel gensets using belly-type fuel tanks in cold climates to prevent the diesel fuel from gelling? What temperature level can cause this to happen?
When the temperature drops, wax crystals can form in the fuel, coating the fuel filter and restricting the flow of fuel to the engine. This will prevent the engine from maintaining power output, resulting in an under-frequency shutdown. Diesel fuel No. 2 should resist gelling down to about 20 F, while winter blends of diesel fuel No. 2 and fuel No. 1 (kerosene) can typically handle temperatures down to -4 to -20 F. These characteristics can be lowered by additives, and by using heaters in the fuel filter.
What problems are associated with total harmonic voltage distortion (THvD) on the generator compared with the same distortion on the utility?
A standby generator has much higher internal impedance compared to utility transformers. Standby generators also tend to be sized much smaller than the transformer. The combined effects increase the level of THvD significantly.
Why is nominal voltage relevant to generator size for harmonics?
The nominal voltage of the generator is important because it affects its reactance—208 V configurations have 33% higher subtransient reactance. The level of voltage distortion is impacted by the level of total harmonic current distortion (THiD).
What THiD would be expected with 200% current limit on a soft start?
When sizing soft starters, THiD levels are impacted by the current limit setting. The THiD using a 200% current limit on a soft start would be about 20% (about 33% less than using a 300% current limit). Adjusting the current limit setting down on a soft starter helps reduce voltage distortion on transient voltage dip issues while operating on a generator.
What parameters must be matched to parallel different size gensets?
Paralleling of generators can be done over a large kW range. Paralleled generators are offered as small as 100 kW and as large as 3,000 kW. Whether to choose a paralleled system over a single large genset depends on the customer’s and engineer’s preference. Paralleling offers many advantages, including redundancy, scalability, and flexibility, that a single genset may not offer. Generators do not need to be the same kW size to parallel with each other. However, you do need to match the voltage, pitch (typically 2/3), frequency, and phase angles of the generators. After the first generator closes into the generator bus, the additional generators speed up their frequency until they are at a safe phase angle to close into the generator bus.
For large industrial facilities, should minimum generator sizing be based on largest starting load?
The size of the system should be based on the overall load totals, not just the largest starting load.
How do you size for a high-inertia motor load?
For existing buildings, a great way to accurately size the building load is to use a power analyzer. Make sure that you capture the peak loads of the building, so it isn’t a matter of how long the analyzer is connected. If the building is a restaurant, for example, peak loads may be on a busy weekend. If the building is a manufacturing plant, peak loads may occur at the end of the month when production is running at its highest capacity.
High inertia motor loads pull the same amount of starting kVA as low inertia loads, but what changes is how long the generator must supply starting current. High inertia loads take a long time to accelerate. This places a greater burden on the engine to maintain starting kW. Low inertia loads can be started with less engine capacity.
What is the minimum percentage of load the generator should run at?
Diesel generators need to run with at least 30% load to avoid wet stacking. This is less important for natural gas gensets. Also, you don’t want to size the generator to be 100% loaded; you want to allow for some load growth in the building and for loads cycling on and off.
If a user has a 1,000-amp service and wants to pick up the whole service, should the generator be sized for 1,000 amps, or for 80% of the 1,000 amps?
It is typically recommended that the generator be sized at 80% load. So if a customer has a 1,000-amp service, you can safely size the generator system to carry the full 1,000 amps, knowing that the load will never be over 80% of this amount (as limited by the distribution panels).
How much of an issue is altitude?
Altitude plays a role in generator sizing. Verify derating percentages with the manufacturer’s spec sheets. For example, a 100 kW diesel will derate -2.5% for every 1,000 ft above 3,500 ft. For a 100 kW natural gas (non-turbo), it will derate -3.5% for every 1,000 ft above 600 ft.
How do you break emergency loads into steps that will best reflect actual loading on the generator?
It is helpful to have load steps on a genset to minimize the size of the generator. Load steps can be accomplished with timers on motor starters or through multiple transfer switches.
Do you have experience and expertise with the topics mentioned in this content? You should consider contributing to our CFE Media editorial team and getting the recognition you and your company deserve. Click here to start this process.