Standby Power Generation - Applications and Transients

Standby Power Generation - Applications and Transients

In today’s environment with a high reliance on reliable electrical power, it has become a common practice to include engine-generator sets in the design of new facilities. Their purpose is to provide power in the event of a utility power failure and, where local utility rate structures and policies make it a viable option, to reduce overall electrical energy costs. One of the major reasons for this emphasis on emergency/standby power has been the increased use of computers and digital controls in all aspects of commercial and industrial processes. Personal safety, production schedules, and in-process inventory are some of the items adversely affected by the lack of power integrity. It is the intent of this program to help you apply generator sets in a manner that will provide reliable power of the quality and capacity required by the end user.

The design of any system utilizing an engine-generator set must take into account the generator size, fuel selection and location of the unit. For the purposes of this discussion, we are going to focus our discussions on the sizing aspects only, and more importantly, to the potential problem areas that have historically surfaced throughout the industry.

General Sizing Considerations

By definition, a standby power system is an independent power system that allows operation of a facility, or parts thereof, in the event of a normal power failure. In order to assure that the final product can adequately provide the required power, an analysis of the load requirements is needed.

If the project is for a new facility, existing empirical data will not be available and assumptions will have to be made based on accepted engineering practices. Currently, there are no software programs that one can use to definitively design a new facility.

If the project is for an existing facility, data can come from a number of sources. One source may be the utility bills, which will provide the actual load requirements and the average peak demand. Utility bills spanning at least 12 to 24 months should be utilized to show possible seasonal requirements and fluctuating business.

The utilization of a power monitor / analyzer at the service entrance is another excellent source of information. In addition to providing a current load profile, the ability to capture available harmonic information will prove valuable later in correctly applying the generator to the system. Examination of the service entrance equipment will be required. The service entrance rating defines the absolute maximum power level and is required to size the automatic transfer switches.

Walking the facility is a critical part of the sizing process. While walking the facility, look for loads that will help define the size of the generator: largest single step load, largest motor, voltage sensitive devices, any significant harmonic distorting loads, building power factor, etc.

Discussions with the customer regarding future load growth vs. load management, size and amount of critical loads (i.e. phones and office lighting, computers, critical processes, and safety equipment, etc.) will have to take place. The end user’s participation in the process is essential, because ultimately the discussion will have to center on cost and risk management.

Power Factor

One of the items often overlooked in sizing a generator is the load's power factor. As a review:

Power factor is the ratio of kilowatts (kW) to kilo-volt-amps (kVA). It is also the cosine of the angle between the voltage sine wave and the current sine wave. For a building with a .9 lagging power factor, the current will lag behind the voltage by 26 degrees (figure 1).
Power factor is not usually a problem because most loads when aggregated together operate within the generator's rated power factor of .8 lagging to 1.0. When a load has a power factor less than .8 lagging, the alternator will need to be larger than the running load to meet the load's
kilo-volt-amp-reactive (kVAR) requirements. A typical low power factor load are motors running lightly loaded.

Figure 1. Lagging Power Factor (.9 PF)

Low power factor loads are not very common because most facilities with poor power factor utilize power factor correction capacitors to meet utility requirements. However, power factor correction capacitors can result in a condition known as leading power factor (current leads voltage) (figure 2). A careful review of the building's power factor correction scheme is needed to ensure that under no conditions does the load connected to the generator become leading. Leading power factor loads can cause the alternator to self excite resulting in voltage run-away.

Figure 2. Leading Power Factor (.9 PF)
Since most digital control panels on generators include over voltage protection, leading power factor will typically be denoted by an over voltage shutdown. Sources of leading power factor include some types of high intensity discharge (HID) lighting. When the primary load is HID lighting, inductive load may need to be added to the system to reduce the power factor to an acceptable level (.8 lagging to 1.0).

Basic Generator Set Concepts

Capabilities of both the engine and the alternator are considered individually and collectively when sizing a generator set. Engines produce horsepower (Hp) while controlling speed or frequency.

Remember that 1Hp is equal to .746 kilowatts (kWm) of mechanical power. Actual electrical output (kWe) will be something less due to system losses. Typical rule of thumb is that it takes 1.6 Hp to produce 1.0 kWe.

The alternator takes the mechanical energy (Hp) of the engine and converts it into electrical energy (kVA).

Remember that kVA is apparent power and is composed of real and reactive components. The real component is kW created by the engine. The reactive component is kVAR created by the alternator's excitation system.

Figure 3. Power Triangle

In addition to converting mechanical power into electrical power, the alternator system controls output voltage and provides high magnetizing current draws (kVAR) necessary to start electrical motors.

Motor Starting

Motor loads have a significant effect on the size of the generator set due to the high currents required while starting. The most common form of motor starting is full voltage -- across the line, because it is simple and low cost. In this method, full line voltage is supplied to the motor resulting in maximum starting. Using this method, an instantaneous inrush of current occurs when the motor is energized. This current is known as inrush current, locked rotor current (ILR), or starting current. This current is also typically referenced by its corresponding skVA. The "s" denotes kVA white starting.

skVA = 1.73 x VLL * ILR

The duration that the motor requires skVA is a function of motor speed and is independent of time (figure 4). Note that the motor maintains a relatively high level of skVA until close to nominal operating speed.

Figure 4. Motor Speed vrs. skVA

The magnitude of the skVA can be calculated from motor nameplate data. Most motors are identified with a NEMA starting code letter that describes their individual starting characteristics (figure 5). The letters correspond to a skVA/Hp. Due to motor design considerations, most motors crowd the upper side of each NEMA starting code range.

Figure 5. NEMA Starting Codes

As an example in using the NEMA starting codes, a 100 Hp motor, with a starting code of G, would have an approximate skVA of 600 (100 Hp x 6.0 skVA/Hp = 600 skVA). IEC rated motors, as well as high efficiency motors, tend to have higher starting current requirements than do standard NEMA units.

The alternator must have sufficient motor starting kVA capacity to limit voltage dips to acceptable levels. The basic question now is, “How does the generator respond to the starting requirements and why?” Inrush current to the motor causes a rapid dip in generator output voltage. Figure 6 shows a family of curves with each curve representing a different size motor load. As motor size increases, the voltage dip seen at the generator increases. The majority of the industry has established a maximum voltage dip of 35% as being the acceptable alternator limit. However, some sensitive equipment may only be able to handle as little as a 15% voltage dip.

Figure 6. Voltage Dip Characteristics

Some manufactures will specify alternator motor starting capabilities based on a 90% recovery voltage also known as 90% sustained skVA (figure 6, line c). At this skVA load level, the alternator will have a voltage dip between 40 and 50% and will only recover to 90% of the generator's nominal voltage. This is a valid way to specify an alternator's capabilities; however, "90% sustained" is often misunderstood as a 10% voltage dip by many people. It is also specifies the generator at the extreme limit of its skVA capabilities (figure 7).

Figure7. Alternator skVA Capability

The alternator's instantaneous transient voltage dip is a function of the impedance of the generator and the skVA required by the motor. Ohm's law states that when a current flows through an impedance, voltage will drop across that impedance (V = IxR). Figure 8 shows that as locked rotor current flows through the generator's internal impedance (x"d), a voltage dip will occur. The voltage dip occurs within a few sine wave cycles upon connecting the motor to the generator output.

Figure 8. Why Voltage Dips

This voltage dip can be minimized by upsizing the alternator -- larger alternators have lower impedances. The industry publishes a predicted voltage dip on most alternator data sheets. Figure 9 illustrates the effects of upsizing an alternator on minimizing voltage dips. Assuming that frequency is constant, these curves give a snapshot of what might be expected.

Figure 9. Upsizing to minimize voltage dips

Upon sensing the instantaneous transient voltage dip, the alternator excitation system responds by increasing excitation to recover to rated voltage. The electric motor load will accelerate towards running speed provided it produces enough torque to overcome its shaft connected load. For induction motors, torque is directly proportional to the square of the applied voltage. Motor acceleration is a function of the difference between motor torque and the torque requirements of the load. The alternator needs to recover to rated voltage as quickly as possible to avoid excessive acceleration times or motor stall.

The time that it takes the generator voltage to recover is a function of the relative sizes of the generator and motor, engine power (kW) and generator excitation capability. A few milliseconds after sensing the initial transient voltage dip, the voltage regulator applies full forcing voltage to the exciter which results in generator field current build up. Generator set components are designed and matched to achieve the shortest possible response time while maintaining voltage stability and avoiding engine overload. Depending on the severity of the motor load, the generator should recover to rated voltage in a few seconds, if not within a few cycles.

Reduced Voltage Motor Starting

When starting a motor with a reduced voltage or ramp type starter, the locked rotor requirements can be reduced significantly. However, two significant problems occur when using a generator in place of the utility. We’ll look at these problems and put safeguards in place to prevent them from occurring.

A simple example of a 70% auto-transformer type motor starter is as follows. The auto-transformer will reduce the voltage at the motor terminals by 30% or 70% of the nominal line voltage for a specific amount of time. That means that when the motor is turned on, it will see 70% of the line voltage. The reduced voltage starter has a built in timer. When the timer times out, the starter will switch from reduced voltage start to full voltage run.

During the initial stage, the line voltage is at 70%. According to Ohm’s law, the current will also be reduced to 70% of its nominal value. If voltage (V) is reduced and resistance (R) doesn’t change, then I (amps) will be reduced by the same proportion). Now, the skVA delivered to the motor is equal to voltage times amps; thus, 70% of the voltage times 70% of the amperage equals 49% of the skVA that would be delivered at full voltage. Figure 10 shows the reduction in skVA values for various voltage reductions.

Figure 10. skVA vs. Voltage

Figure 10 also shows that the motor's torque is also dramatically reduced. This means that the motor is going to come up to operating speed more slowly, but the reduced starter doesn’t know this. The starter only knows that it has a specific time delay. When the delay is finished, it disconnects the transformer section and reconnects across the line. The trouble occurs when the motor is not close to operating or synchronous speed. Because of the reduced voltage, it comes up to speed slowly and by the time the starter times out, the motor may only be at 75% of its operating speed (figure 11).

Figure 11. Switching to Full Voltage Run

This means that the locked rotor current will increase significantly when the motor becomes connected to the line rather than the reduced voltage transformer. When this happens, the alternator can no longer supply the required skVA because it has been sized for the lower reduced voltage skVA requirement, not the full locked rotor kVA. The voltage then dips beyond 35% and the contactor drops out. As soon as the contactor drops out, the voltage snaps back and pulls the contactor in and thus pulls the voltage back down. Basically, the contactor is clapping in and out at full locked rotor current. This is not good for the contactor, the motor or anyone in close proximity to the equipment. To prevent this problem from occurring, confirm that the motor can reach full voltage on reduced voltage. If the motor can not reach full speed on reduced voltage, size the generator as if the motor is starting across the line.

Engine Loading

Engines also respond according to the load requirements of the generator. The normal response is a change in frequency. There is a corresponding drop in frequency as the engine seeks to respond to the requirement for starting kVA from the alternator. The reaction is a function of the start power factor of the motor. Most typical three phase motors have a starting power factor (sPF) of .25 to .45. Given that:

PF = kW/kVA, then
skW = skVA x sPF.

If we take an example of a 100 Hp motor and assume a skVA of 600, we can calculate the skW. (600 skVA x .3 sPF = 180 skW) A good rule of thumb for estimating engine size is that the generator's kW should be twice the motors Hp. The following are some typical rules of thumb for starting and running motors on a generator:

kVAgen  Hpmotor x 6.0
kWgen  Hpmotor x 2.0

kVAgen  Hpmotor
kWgen  Hpmotor x .85

Voltage Regulator Unloading

Upon close examination of the corresponding voltage dip and frequency dip graphs of a block loaded genset, one could see that the deviations correspond (figure 12).

Figure 12. Block Load Characteristics

This is due to an effect called regulator unloading. The generators voltage regulator intentionally lowers voltage in an attempt to help the engine recover. If the voltage regulator does not shed voltage, some engines would not recover and would stall. The voltage regulator will begin to shed voltage at the point when engine speed intersects the voltage regulators under frequency protection (figure 13). Under frequency protection is also know as a volts per hertz because the regulator reduces voltage proportionately with a reduction in frequency.

Figure 13. Voltage Regulator Unloading

Remembering from the discussion earlier on reduced voltage starters that a reduction in voltage reduces the required skVA by the square of the remaining voltage requirement. As a result, the skW requirement is also reduced, allowing the engine to recover faster. For example, a 30% reduction in voltage by the regulator will result in a skVA reduced by a multiplier factor of .49 [(.70) 2 = .49]

Non-Linear Loads

The second significant component for generator sizing is the magnitude of non-linear loads to be carried by the engine/generator. Common non-linear loads encountered are computers, battery chargers, uninterruptible power supplies (UPS), variable frequency drives (VFD), medical imaging equipment and some forms of fluorescent lighting. EGSA defines a linear load as an “AC electrical load where the voltage and current waveforms are sinusoidal, and the current at any time is proportional to voltage.” It further defines non-linear loads as “AC loads where the current is not proportional to voltage.” The nature of a non-linear load generates harmonics in the current waveform, which leads to distortion of the voltage waveform (figure 14). It is this harmonic distortion that must be accounted for in the sizing of the generator.

Figure 14. Harmonic Distortion

Variable frequency drives (VFD’s) are utilized to control the speed of induction motors. There are three commonly used types of variable speed drives: VSI (voltage source inverters), CSI (current source inverters) and PWM (pulse width modulated), with the most common being the PWM. To help deal with the harmonic effects of nonlinear loads, IEEE 519 provides guidelines as to acceptable voltage distortion levels. For general loads, the acceptable level of distortion is 5%. For dedicated loads, the acceptable level is 10%. Since the generator is only in service for limited hours, most generator systems can withstand more than 10% distortion; however, the common practice in the industry is to limit the THVD (total harmonic voltage distortion) level to less than 15%. Above that level, the generator system could experience problems.

In analyzing nonlinear load applications, it should be noted that the same constraints that caused the generator's voltage to dip during motor starting will contribute to an increase in harmonic voltage distortion. The generator provides 60 hertz current to the nonlinear load. The nonlinear load produces harmonic currents as a byproduct of its operation. These harmonic currents flow through the generator's impedance resulting in a corresponding voltage distortion (figure 15).

Figure 15. Nonlinear Load Circuit

To minimize this distortion, the generator's impedance needs to be held within acceptable limits. In a typical application of a 6-pulse PWM unit, the base kVA of the alternator selected should be adequate to limit the subtransient reactance (X”d) to 6-8%, when compared to the total non-linear load. Generally, if the generator is sized to minimize harmonic distortion, alternator overheating due to harmonic currents is no longer an issue.

Other solutions for controlling harmonic distortion include active or passive filtering technologies to remove the harmonic currents.
Also, additional loads that can be added to the system prior to starting the nonlinear load will help minimize distortion.

A general rule of thumb is that VFD loads on a generator should be less than 50% of generator capacity to limit total harmonic voltage distortion to 15%. An additional point to remember is that in VFD applications, size the generator for the full nameplate rating of the drive and not the motor. Harmonic distortion levels may be higher when the drive is being utilized partially loaded and a larger motor may be installed in the future to utilize the full drive capabilities