VFD, motor strategies for energy efficiency
- Understand when to specify VFDs.
- Learn which types of motors require VFDs.
- Know how to achieve efficiency when specifying a VFD to meet load conditions.
The optimist says the glass is half full, the pessimist says the glass is half empty; the engineer says the glass is twice as large as it needs to be.
This joke’s underlying truth may seem straightforward, but as engineers, we often lose sight of these types of basic guiding principles when selecting equipment for a particular application. So while we will often specify variable frequency drives (VFDs) for motors as a “cover-all” solution to all our energy efficiency and control considerations, those same generic standard practices often have a rate of return less than expected or are just plain ineffective in doing what we thought they would do. To effectively specify a VFD, we need to go back to basics and methodically work through a few key steps:
- Understand the load (operating power, torque, and speed characteristics)
- Understand duty cycle (what percentage of operation at 100% load, 50% load, etc.)
- Once you have a solid grasp of the two items above, take a step back and determine what you’re trying to accomplish by using a VFD (energy savings, soft start, controllability, etc.)
- The value of the reliability of that VFD
- Specifying and controlling the VFD to produce desired results
Understand your load
Understanding your load is the first step to determining what you can gain by applying VFDs. First, a quick review of motor basics. In building hydronic and air systems, the most common type of motor providing power to a load is a 3-phase ac induction motor. That motor has a characteristic synchronous speed based on the quantity of “poles” in its design and the frequency of the electrical supply. But in an ac induction motor, the rotor speed never reaches synchronous speed—the rotor always slightly lags the frequency of the rotating field. This concept is known as “slip.” As such, the base speed of a motor is defined by the following equation:
For 60 Hz North American electrical supplies, base synchronous speeds without slip are: 2-pole, 3600 rpm; 4-pole, 1800 rpm; 6–pole, 1200 rpm; and 8–pole, 900 rpm. As load on a motor increases, the amount of slip will also increase. Slip will typically be 1% to 3% of the speed. It should also be noted that based on this equation, the speed of the motor will change proportionately to any electrical supply frequency change. The ability to change speed based on changes in the electrical supply frequency is the root concept of VFDs.
Once we understand what the base speed of a motor is, we can then address the power and torque that can be delivered by that motor. Power delivered by the motor is defined by the following equation:
So, for a given power rating, the motor base speed is inversely proportional the effective torque rating for that motor. For example, selection of an 1800 rpm versus a 1200 rpm motor reduces torque by a third.
Depending on the National Electrical Manufacturers Association (NEMA) design type, motors have different torque-speed relationship characteristics. The NEMA motor standard MG-1 defines five primary standard induction motor types, designs A through E. The letter designation for these design types should not be confused with motor winding insulation temperature rating classes. The characteristics of each design type are shown in Table 1.
If the load’s starting torque requirements are known, this basic understanding of the operational characteristics for each motor design type can provide basic guidance in selecting the proper motor NEMA design type. Once a load’s starting requirements are determined, the next step is looking at the load’s running requirements. In building systems, excluding constant horsepower and constant speed/torque loads, typical loads that can take advantage of VFDs can generally be divided into two primary categories:
- Variable speed, variable torque (fans, blowers, and centrifugal pumps)
- Variable speed, constant torque (positive displacement loads such as screw compressors, reciprocating compressors, or elevators).
So to support a characteristic load, we select a motor to meet a specific starting requirement and running output power, torque, and speed. However, through the affinity laws, we recognize that there are significant potential energy savings associated with reducing a motor’s speed and, by association, horsepower. So if we can define the required change in motor speed to meet the change in flow for a centrifugal load, the change in required power is proportional to the cube of the change in speed from one system point to another. The change in required torque is proportional to the square of the change in speed from one system point to another. These relationships can be expressed through the following equations:
This nonlinear relationship between power and speed can be exploited for significant energy savings if the speed of the motor can be changed. Figure 2 more clearly illustrates this relationship.
For positive displacement loads that require constant torque throughout an operation speed range, the potential savings resulting from reduced speed is not quite as attractive. With these types of loads, the change in required power is directly proportional to the change in speed. While the potential savings is not necessarily as great as that for a centrifugal load, there is still potential to significantly increase energy efficiency by reducing motor speed.
The most efficient motor is the one that never turns on. And conversely, the least efficient motor is the one that doesn’t turn off when it should be off. As engineers, we recognize that except for constant speed/constant torque load, the worst-case design scenario usually represents a relatively small percentage of the overall operating hours for any given building system.
In general, the characteristic load profile for the motors in any given system can and will vary dramatically depending on building usage and weather conditions. VFDs have given us the capability to change to the output of our motors to more closely match the load at any given point. However, this capability has also given us a crutch in allowing systems based on oversized equipment to remain functional. Numerous computer programs are available to size HVAC system components and model hourly energy requirements for unusually complex systems. However, beyond a cursory review of a system sizing summary and recommended equipment selections, how much time is really spent trying to understand the exact load profile and how to optimize the equipment selections to provide the best efficiency at the load where the system spends most of its time?
While it’s recognized that there’s value to the standard practice of including safety factors and redundancy in a system design, there’s another side to that story. To illustrate this, here’s an extreme example: Is operating a 50 hp motor at 80% load going to be more or less cost-effective than operating a 100 hp motor at 40% load? While the difference in energy costs may be marginal, the difference in initial capital costs for installation is not. How many projects have been sent back for value engineering because the design could not meet the construction budget?
There are certain situations where using a VFD may not result any energy savings and may only serve to increase the cost and complexity of an installation. The primary benefit of using a VFD is being able to operate a motor at a reduced speed. Per the affinity laws, operating at a reduced speed results in a dramatically lower power requirement, and by association, a reduced energy usage and operational cost. However, if a motor is serving a load like a toilet exhaust fan that operates all of the time at a fixed speed to meet a minimum code exhaust requirement, what energy savings benefit does a VFD offer if it only operates at 100%?
Motor efficiency is a simple ratio of total input energy to useful output power. Motors are not 100% efficient, but recent federal efficiency standards have resulted in noticeably increased efficiencies for motors across the board. The Energy Policy Act of 1992 (EPAct 1992) and the subsequent Energy Independence and Security Act of 2007 (EISA) mandate minimum full load efficiency requirements for all general purpose 3-phase motors from 1 to 200 hp rated up to 600 V that are manufactured or imported into the United States. Compared to pre-EPAct motors, these standards represent incrementally increased efficiency across the board, the magnitude of which is typically in the mid- to high–single-digit percent range. The greatest efficiency gains are typically realized in smaller-sized motors below 50 hp.
As demonstrated in our overview of the affinity laws, tremendous energy savings potential exists in reducing motor speed and power. But was this the true value of that reduced power usage? Table 2 shows some simplified examples of electrical cost compared to equipment cost.
While these examples use generalized costs for electricity and equipment, they serve to demonstrate as an order of magnitude estimate, that the yearly cost for electricity can easily approach double the cost of the motor itself. While the exact equipment cost and the utility rates can affect the results in either direction, it should be remembered that the typical service life for a general duty totally enclosed fan-cooled (TEFC) motor is approximately 20 years. Given the load profile, does adding a VFD make sense? Does the VFD have an acceptably short length of time for return on investment?
VFDs aside, the internal efficiency of a motor changes with load. Peak efficiency for most motors at a fixed speed is actually at about 75% load rather than at rated horsepower. However, the difference in efficiency between 75% and 100% load is negligible, typically below 1%. Most electric motors are designed to operate at 50% to 100% of rated load. As the motor size gets smaller, however, this efficiency drop off with lower load ratings may be something to consider. With motors below 25 hp, it’s not unusual to see up to a 5% to 10% difference in efficiency between 25% and 75% load.
VFDs aren’t black boxes
Compared to a VFD, motors are big, dumb hunks of iron, copper wire, and insulation. It should never be forgotten that VFDs are fairly sophisticated power electronics consisting of numerous discrete components. While the service life of a motor is dependent primarily on the condition of the rotor shaft bearings and insulation system for the windings, the service life of a VFD is directly related to the reliability of the individual VFD components—any of which can cause that VFD to fail. Depending on the size and age of a VFD, it is often more cost-effective to replace the entire VFD rather than attempt to replace individual failed components.
Historically, the focus has been on minimizing motor failures when designing and specifying equipment with VFDs. This is not unusual, given the numerous horror stories of motor failure after short periods of time due to a lack of consideration to possible harmful interaction between the motor and VFD. These issues of reduced motor life caused by VFD operation can be partially mitigated through several methods: by use of true inverter duty motors that conform to NEMA MG1-2006 Part 31, shaft grounding rings, conductive bearing lubricants, and keeping motor feeder lengths short. However, a fact that is often overlooked in this emphasis on ensuring the reliability of the motor is that the cost of the VFD can approach the cost of the motor. What do we do to ensure the reliability of that VFD and protect that investment?
Just like motors, VFDs are not 100% efficient. The rectifier-dc bus-inverter design of a typical VFD is not totally loss-less. Most VFD efficiency ratings are in the mid-90% range. The losses in a VFD are primarily attributed to conduction (electrical current flowing through the device) and switching losses (the power transistors on the rectifier input and inverter output). While higher carrier frequencies in a pulsewidth modulation scheme generally result in better VFD output waveform shaping, there is a downside. VFD conductive losses are fairly consistent, but switching losses are directly proportional to the carrier frequency at which the transistors operate.
A significant portion of the input power to the VFD is wasted as heat. That same heat and the ability to properly manage it has a direct effect on the aging of the VFD’s electronic components. But even with proper thermal management, the useful service life for a VFD can still be significantly less than that for the motor it serves. Many critical VFD components may require replacement in a time span as short as 5 to 10 years. These components include cooling fans, control boards, power capacitors, and more. Replacement of some components may represent a disproportionally large percentage of the cost of the VFD.
The key to reliable VFD operation is to maximize the life expectancy of the individual component by controlling temperature, humidity, and dirt/dust in the location where the VFD is installed. VFD failure modes caused by poor operating conditions are numerous. Fans and filters can clog with dust and dirt. Components can corrode due to high humidity. Poor power quality issues can fry control boards. The drive can overheat with no air circulation in high ambient temperatures. The list goes on, but the primary environmental consideration is always heat (see Figure 3).
While VFDs are designed to operate in up to a 104 F ambient temperature, the optimum ambient temperature for power electronics is 59 to 86 F. The typical rule of thumb is that for every 17 F reduction in operating temperature, the life of the device doubles. This serves to emphasize that the performance and overall reliability of the system can be dramatically impacted by something as simple as ensuring that there is adequate airflow around the VFD (see Figure 4).
We often put motors and VFDs in locations that are poorly suited to ensuring their reliable operation. Unfortunately, due to the nature of the loads that we apply VFDs to, there are often limited options as to where the VFD can be located. This may mean putting a VFD in an unconditioned pump house, on the roof in a packaged air hander. While these types of situations may represent a compromised design, it should be understood what effect this has on the reliability of the design and what additional levels of redundancy in the design are appropriate to account for this reduced level of reliability. Is a wrap-around bypass across-the-line contactor, or a separate standby VFD for emergency starting appropriate for the application?
Scalar and vector control
Assuming that we properly characterized our load type, found the right motor for the load, and determined that the load profile of the load makes a VFD useful, how do we control that VFD? There are two primary types of VFD control schemes: "Volts per Hertz" (also known as V/Hz, or scalar control) and vector control.
Scalar is the simpler (and cheaper) control methodology. The pulse width modulation used by modern VFDs to synthesize a sinusoidal output can be used to change not only frequency but also voltage. If the ratio of input voltage to frequency is known (e.g., 460 V/60 Hz = 7.67), maintaining the VFD output at a fixed voltage to frequency ratio theoretically can allow consistent torque over the entire operation speed range of a motor. At very low motor speeds/frequencies, if the voltage approached zero, the torque output of the motor would also approach zero. To address this, a fixed voltage (voltage boost) is added to the prescribed V/Hz ratio to maintain torque. As the speed is increased, the voltage boost is removed. In general, the primary benefit of this control methodology is that it is very simple and does not require direct motor feedback. If multiple motors are connected to a single VFD, this would be the likely control method. However, this lack of feedback to the drive is also one of its primary disadvantages. Using a set V/Hz ratio and no direct motor feedback, speed and torque regulation for a load becomes “best guess.” Scalar drives are generally recommended for turndown ratios of no more than 6:1. Other disadvantages include reduced ability to overspeed a motor, poor breakaway torque characteristics, and poor low speed torque—even with voltage boost features. With variable speed/constant torque load types, these downsides can often be deal breakers.
There are two forms of vector control, direct and indirect. There are two subsets of indirect, closed loop (feedback) and open loop (sensorless). All use the same basic concept. By shaping the VFD’s output voltage and frequency, we can separately address the magnetizing current (referred to as the “d” vector component) and torque producing current (referred to as the “q” vector component) in the motor’s stator. Whereas scalar controls use a fixed V/Hz ratio, the ability to decouple these components and independently address them opens a greater possible range of torque and speed control. However, as the motor interacts with a dynamic load, the VFD’s microprocessor needs some feedback reference signal to ensure that the motor is providing stable speed and torque to the load. When feedback is provided to the microprocessor, it can calculate changes on the fly to the VFD output to better regulate the motor speed and torque.
In direct vector control, you would want to directly measure the motor air gap flux within the motor. However, this is not very common due to the level of accuracy and cost associated with the additional sensors needed. In closed loop indirect vector control, a shaft encoder is added to the motor to tell the VFD exactly what the position and speed of the motor shaft is. Shaft encoders are typically mounted on the non-drive end of the motor. In open loop indirect vector control, there is no sensor on the motor. Rather, the VFD compares its output current to the motor and compares this to a mathematical model of the motor to determine if adjustments to the output are required. Since this is only an approximation of speed based on a model and not a direct measurement, speed regulation, and overall operational speed range, is not as good as with closed loop control.
Vector control methodologies are typically used to accommodate high turndown ratios (very wide operational speed range) or where precise torque control is required. Where a motor is required to produce rated torque at zero speed and “hold” a load for a period of time, there is typically no choice but to use a vector drive.
John Yoon is the senior staff electrical engineer for McGuire Engineers. He has nearly 20 years of experience in the design of electrical distribution systems. His project experience covers a broad spectrum, including high-rise building infrastructure renewal programs, tenant build-outs, mission critical data centers, good manufacturing practice (GMP) cleanroom facilities, and industrial facilities.