Undervoltage Ride Through
This post talks about VFD capability to withstand a fault in the supplying grid without tripping the drive system – undervoltage ride through function [1].
Availability of drive system is very important for many users and the importance of high availability is expected to grow. It implies certain requirements on the VFD, such as robust operation and certain degree of immunity against external disturbances.
At the same time grid disturbances are not unusual. There are many sources of such disturbances, for example:
– atmospheric events
– switching transients
– sudden load rejection
– single phase to ground, phase to phase and three phase faults
– starting of heavy load (e.g. DOL motor start, inrush current of large transformer)
Some of these disturbances lead to temporary increase of grid voltage, others cause transient voltage drop. In this article we focus on disturbances associated with transient undervoltage. These situations may be referred as dynamic voltage drop, voltage sag or a grid fault.
Importance of ride through
Disturbances affect the power quality. The grid sees increased voltage variation and elevated negative sequence components as a response of the power system to such events. Majority of grid faults has a duration of less than 100 milliseconds. Unfortunately, when the VFD trips during such disturbance, it may take minutes, hours or even few days to resume the process. Because of that the users wish the VFDs to have a resilience against short time disturbances of supply voltage.
According to [2], with further reference to another study, the voltage drop has a typical duration at least 0.5 cycles to 30 cycles. In the United States, a typical voltage sag lasts 6 to 10 cycles, i.e. 100 to 167 milliseconds. Situation in Europe is similar with majority of disturbances lasting 100 milliseconds or less and seldom exceeding 200 milliseconds.
So what is the undervoltage ride through function all about? Simply said, the VFD shall drive through a grid voltage disturbance without tripping. The drive shall be able to bridge a certain period of time when the grid voltage is low and resume the normal operation once the grid voltage is back in the standard range.
Note that not every undervoltage disturbance triggers the ride through mode of the VFD. Grid voltages down to 90…85% of nominal value shall be managed by the drive without significant loss of performance. Less severe undervoltages are handled “normally”, i.e. VFD keeps providing power and torque to the machine. Depending on depth of voltage drop, on VFD dimensioning and load point the VFD keeps providing the required torque or the torque is temporarily reduced. VFD can also utilize its overloadability (if available) to maintain the motor torque during undervoltage conditions.
Energy in dc link and time constant
The energy in dc link is rather small and generally cannot be used as energy source during grid voltage dips. The time constant of dc link is very short. For majority of medium voltage VFDs it is in the range of several milliseconds [ms]. Time constant can be expressed as E/P ratio that also has units of seconds or milliseconds (Joule = Watt · second). The parameter says how quickly would the dc link be discharged if a rated power is drawn out by the load.
The example stated in [1] mentions time constant of 6 milliseconds. Depending on the VFD design the time constant is little shorter or little longer, but generally much too short to provide adequate resilience to grid faults.
An example of dc link discharge after a grid voltage disturbance is shown in figure 2. It depicts a dc link voltage (1/2 of dc link in a typical voltage source inverter with neutral point clamped topology). The grid disturbance appears at time point of 2 seconds. Prior to the disturbance the dc link was almost constant, apart from its characteristic ripple.
Delay between grid undervoltage appearing on V side of the transformer till start of dropping voltage in dc link is less than 2 milliseconds. About 8 milliseconds later the half dc link voltage reaches approx. 1900 V (~ 76% of nominal dc bus). Without any control intervention fro VFD control and protection system the dc link would reach an undervoltage trip level. As we can see, such VFD without having any ride through control feature could withstand only some 12-15 miliseconds of severe grid disturbance.
Limitation of dc link
Based on above description one could think of increasing the size of dc link in order to increase the time constant and by that also the robustness against voltage dips. Well, this might not be a viable solution. The size of the dc link is limited not because of cost savings or footprint. The main reason is protection of the converter. Larger capacitance means larger fault current and higher energy released e.g. in case of electric arc [3]. Therefore, the size of dc link cannot be simply increased without modifying other components as well. In the end, for larger dc capacitance a bigger frame size of VFD would be required. This is not a path to solve the issue. Even heavily oversized VFD could bridge only extremely short voltage dips.
Instead of increasing the dc link a control strategy is used to enhance the robustness against voltage dips.
Ride through strategy
Control strategy to ride through a grid disturbance is based on kinetic buffering, i.e. using the kinetic energy stored in the rotating mass of the electric motor and the driven load.
When a grid fault happens, the grid voltage collapses to some extend. The VFD senses reduced input voltage. Now let’s assume a diode front end (DFE) drive first. The dc link voltage drops. Rate of change depends on the time constant (E/P ratio) and actual motor load. Below certain threshold the VFD typically starts limiting the output power. The next threshold means that VFD enters the ride through mode. As soon as ride through condition is detected, the VFD sharply reduces the torque to zero. By doing that the output power is set to zero which prevents the dc link from being further discharged. So the inverter stops delivering active power to the motor. However, it keeps the motor magnetized. There are several good reasons for that. Let’s mention at least two of them:
(i) Since the motor is magnetized, the VFD can immediately start making torque again as soon as the grid voltage recovers to a level where the VFD exits the ride through mode.
(ii) As the inverter keeps modulating, the motor speed is still estimated (most motor drives are encoderless) that also helps to quickly re-accelerate after the voltage recovery.
Looking a bit closer into the mechanism the torque is not ramped down to zero, but is actually slightly negative. The value is very small, in the range of 1-2% of nominal torque. It means that the machine is operated in generator mode. This small portion of power is used to compensate losses of the inverter. Obviously, this is a very little power as the VFD has high efficiency and in this case operates at very light load.
We just said that the VFD is basically braking the motor a bit to take a small portion of the kinetic energy. This “braking” has negligible impact on the motor speed. Sure, the motor decelerates during the ride through mode, but that is mainly caused by the load torque that is usually much more significant.
Role of system inertia
Inertia and angular speed define the kinetic energy of the system based on well-known formula:
Just remember that in case of system with gearbox the inertia of motor and load shall refer to the same speed before summing them up.
The larger the system inertia (sum of motor and load inertia) the smaller the rate of change of speed. Systems with high inertia, such as e.g. fans and blowers, can hypothetically ride through quite long undervoltage disturbances. As long as the machine is spinning, the VFD can take small portion of power from the rotating system to cover the internal losses.
In contrast to fans and blowers, most pumps have relatively small inertia and lose their speed quite fast. There are also applications that might have sufficient inertia and corresponding kinetic energy, but presence of other issues is limiting the performance (e.g. compressor surge described later).
One important assumption is that the power supply for control system is not affected by the disturbance, i.e. there is an independent supply (“safe line”), eventually using a suitable UPS.
Zero torque ride through and partial torque ride through
The classical approach of voltage source inverters (but also current source inverters) leads to a zero torque ride through. During the fault the motor torque is zero (precisely said slightly negative) and the machine speed is dropping. In most applications such behavior is accepted. The user is happy that the VFD could avoid the trip. However, in some cases it is desired that the motor provides some partial torque during the grid fault. An example is a pipeline compressor. During the conventional ride through the torque is zero. As the pressure behind the compressor is suddenly than the discharge pressure, the gas tends to reverse the direction of flow. As a consequence, the inlet pressure increases and the flow is reversed again. The process of such periodic reversals is called compressor surge. It is an unstable operation and compressor is exposed to large vibrations with high risk of damage. A bypass valve (part of anti-surge protection) needs to open and the system is likely to be tripped.
Now whether or how quickly the compressor reaches the surge line depends on the design, initial state and also on the torque provided during ride through.Therefore, partial torque ride through has a major benefit as it extends the “time to trip” and increases the chance to survive the grid fault. In such systems with surge line relatively close to the normal operation points the surge phenomena is the main limitation regardless how long voltage dip the VFD could overcome.
Voltage source inverters (VSI) typically use zero torque ride through. Power to the motor needs to be ramped down in order to preserve energy in the dc link. Main concern of too low dc link is uncontrolled rapid charging after the grid voltage is restored.
Diode front end VSI drives have a passive rectifier. When the grid voltage drops, the dc link voltage is reduced accordingly. Boosting is not possible. On the other hand, diode rectifier is very robust and does not suffer from asymmetries and negative sequence component accompanying typical grid faults.
Active front end VSI does not necessarily have a fix ratio between grid voltage and voltage in dc link. The active rectifier (AFE) is able to compensate part of missing grid voltage. Therefore, AFE may boost the immunity against voltage dips. However, we shall mention that immunity against voltage dips involves both hardware and software. It is crucial that AFE drive software can handle the negative sequence and does not drive the input isolation transformer into saturation.
Current source inverters (CSI) often just bock pulses which basically leads to zero torque ride through similar to VSI. However, CSI drives do not have dc link capacitors and re-charging of dc link is not a concern. CSI technology can easily realize partial torque ride through. That is a major advantage in applications such as centrifugal compressors due to extended time to surge.
A more detailed description is provided for our premium subscribers.
Fault ride through and grid codes
In case of generating plants a grid code may become applicable. VFDs supplying power into the grid for a short time during regenerative braking are usually exempted from the grid code requirements.
In case the grid code applies, there are stricter requirements to be followed. The drive shall not just stay “operational” during the fault, but shall actively support the grid. The grid code defines minimum amount of reactive power that the converter shall inject into the grid. It may require certain overcurrent/overload capability of the converter.
Examples of such typical V-Q diagrams are shown in figure 4. The grid code requirements depend on the country of installation, on the installed power of the plant and on the system voltage level (for example there is a set of requirements for voltage levels up to 35 kV, another set for voltages up to 110 kV and one more set for voltages up to 420 kV). As you probably surmise, the higher the voltage level the stricter the requirements.
Summary
Key take-away from this article about undervoltage ride through capability:
1. Grid voltage disturbances occur now and then. It is important that the VFD can ride through most if not all of them avoiding multiple trip events per year.
2. Although most undervoltage events are rather short, the FD generally cannot handle them passively using energy from dc link to compensate the deficit.
3. Energy stored in dc link is limited. The capacitance installed in dc link cannot be arbitrarily increased due to protection considerations.
4. Overdimensioning of the VFD is not an economical way and does not considerably help to extend the duration of voltage dips that VFD could ride through.
5. Ride through strategies are mostly based on using small portion of the kinetic energy stored in rotating masses (called “kinetic buffering” by some manufacturers).
6. Most common performance is zero torque ride through. In some cases a partial torque ride through can have significant benefits. Example was given for a centrifugal compressor drive with risk of surge.
7. Grid code requirements shift the fault ride through to next level of complexity. In such case the VFD shall not just withstand the grid fault without tripping, but shall in addition support the grid (injection of reactive power during grid undervoltage condition).
The last point usually does not apply to industrial drives. However, it is a fundamental requirement for other power electronics based converters, such as STATCOM, wind converters or variable speed hydro plants with pump/turbine operation.
References
[1] T. Wymann, P. Jörg, “Power loss ride-through in a variable speed drive system”, PCIC Europe 2014
[2] E. Koneva, R. Osman, “Input power quality issues and how to specify variable frequency drive for weak input line conditions”, PCIC 2017
[3] How to choose a medium voltage VFD: Protection concept, https://mb-drive-services.com/choosing-mv-vfd-protection/
[4] How to choose a medium voltage VFD: Line side connection and power quality, https://mb-drive-services.com/how-to-choose-mv-vfd-line-conn/
[5] LCI versus VSI drives (series), https://mb-drive-services.com/category/lci-versus-vsi/
[6] Current source and voltage source inverters, https://mb-drive-services.com/current-source-and-voltage-source-inverter/
[7] Medium voltage AC drives, https://new.abb.com/drives/medium-voltage-ac-drives
1 Comment
VFD control: Introduction - MB Drive Services · November 26, 2020 at 10:53 pm
[…] or external disturbances or by nature of application. An example of such dynamic operation is a fault ride through mode described recently […]
Comments are closed.