In this post we offer an introductory overview of regenerative drive operation (also referred to as “Active Front End” or “AFE”), covering the basic principles and requirements.
Basic AC drives are not regenerative, meaning they cannot return energy to the power source. The AC motor and the inverter itself are inherently capable of regeneration, but the input rectifier of the drive is not. If the application causes mechanical energy to be returned to the motor, with power which exceeds the motor loss, then normally there has to be a brake resistor fitted to absorb and dissipate the unwanted energy. Without it, the returned energy causes the DC voltage to rise until a protective trip stops operation to prevent damage.
Sometimes it is desirable to avoid the use of a brake resistor, and in some applications it is a specific requirement to recover the returned energy. One method is to use a common or shared DC bus between a number of inverter drives – this is a particularly good solution in applications such as a winder/unwinder, where there is circulating power . If the regenerated energy needs to be returned to the AC supply then this is the role of a regenerative drive. The simple rectifier is replaced by an inverter, sometimes referred to as an “Active Front End” (AFE), as shown in the simplified schematic Figure 1. The Control Techniques regenerative arrangement (“Regen mode”) uses the same kind of inverter as for the motor drive to implement the AFE.
Another benefit of the Regen mode is that unlike a simple rectifier it does not inherently generate harmonic current in the supply. In some applications this is the most valuable feature.
Figure 1: Simplified schematic of regen inverter drive system, showing the input stage as an active rectifier
The inverter uses PWM to generate a three-phase sinusoidal voltage set, which is synchronised with the AC supply. The series a.c. choke shown in Figure 1 has two important roles:
The a.c. choke has to sustain magnetic flux at both the line frequency and the much higher PWM switching frequency. This is the reason for its specialised design. A conventional choke designed only for 50/60 Hz operation would suffer excessive power loss in this position, leading to overheating as well as poor efficiency.
A simple diode bridge rectifier works automatically to draw the required input current to maintain the DC voltage at an average level determined by the AC voltage and the effect of any AC inductances. The regen inverter requires continuous active control of its AC terminal voltage in order to maintain the necessary values of active and reactive current.
The active current determines the power flow. This then controls the DC terminal voltage, using a control loop with a set reference DC voltage.
The DC supply in a regen system must have a voltage which exceeds the peak mains supply voltage by at least small margin, and is less than the OV trip level. If it ever falls below the peak supply voltage then the inverter trips quickly, usually with instantaneous overcurrent, because it cannot control the current in the freewheel diodes.
The power grid was not designed to accept returned energy from loads, and the power companies have strict rules to ensure that local generation does not put the safety and stability of the grid at risk. Therefore any intentional generator has to comply with regulations and technical standards, and may require some form of independent approval. Regeneration might also be prohibited on a small power system such as on a ship, because of the risk that the regenerated power might exceed the load and cause an over-speed of the generator. On the other hand, if a particular load occasionally returns some of its own stored energy to the power network, and its power output is much less than the normal power consumption of the site where it is connected, the effect of regeneration is just a temporary reduction of power consumption and has minimal impact.
In the majority of regen applications the power rating of the drive is much less than the other loads on the site power system, and the stored energy is also small, so there is no risk of disturbance or malfunction. However it is worth looking at the aspects which are usually covered by regulations, because they give an insight into some of the risks of large-scale or widespread regeneration. We will look more closely at the regulations in the follow-up blog.
In an AC power transmission network the voltage drops primarily in the series inductances of the lines, cables and transformers and the lagging (inductive) component of load current has the greatest effect on the voltage magnitude at the load. Large generators may be required to regulate their power factor to help in voltage regulation. More commonly they are required to operate with unity power factor.
The Unidrive M inverter module has optional control of reactive current as measured at its terminals, which by default is set to zero. The external filter has series inductance and parallel capacitance which contribute small levels of reactive power. At rated load these roughly cancel to give nearly unity power factor. At reduced load the capacitance dominates, resulting in a lagging power factor (considered as a generator, i.e. VAr is positive). In many applications the reactive current reference is left at the default value of zero. If reactive power or voltage must be controlled, a control loop can be created in the inverter which adjusts the reactive current to meet the requirement.
Unlike a diode or thyristor rectifier the Regen inverter generates negligible levels of true supply harmonics, i.e. of current at integer multiples of the supply frequency caused by non-linearity. This is because it works by generating an accurate sinusoidal voltage using PWM, so the only unwanted frequencies are products of intermodulation between its own PWM switching frequency and the mains frequency. These frequencies are much higher than the mains frequency so they can be readily filtered. There will be a small residual level of these frequencies, which are still sometimes loosely referred to as “harmonics” although this name is not strictly correct.
Excessive emission of switching frequency current into a power system can have serious effects, and needs to be avoided. These frequencies are too low to cause electromagnetic disturbance, and are not currently covered by EMC regulations, but they can cause malfunctions of waveform-sensitive equipment such as UPS, and they may also disturb sound systems because they lie within the audible frequency range.
The recommended arrangements in the Regen user guide include switching frequency filter capacitor and choke components to reduce the emission to a level which is low enough for most purposes. It is important to include this filter in any installation unless there is a clear reason for knowing that it is not required. The only situation we have encountered where this can be known confidently is where the inverter will be supplied by a dedicated local generator which does not supply any other equipment.
Occasionally there is very sensitive equipment connected to the same LV supply circuit so that the standard filtering arrangement is not sufficient. CT has experienced this with a church sound system. In this case the filter attenuation can be increased by adding additional filter capacitors to those existing. Alternatively the switching frequency can be increased, which means that the attenuation of the standard filter is increased and also the sensitivity of human hearing is reduced.
The mains supply will have existing harmonics, which will result in some harmonic current in the inverter and its filter. Although these are quite small when compared with those caused by a simple rectifier, they can still cause contractual disputes where there is a strict harmonic specification to meet. It can be difficult to tell whether harmonic current is generated by or absorbed by the inverter, and where the current is not in phase with the voltage the concept of direction is in any case wrong.
It is also important to appreciate that harmonic current caused by existing harmonics is independent of load power. This means that when the load power is small compared with the rating, the harmonic current looks high if it is expressed as a percentage of the fundamental current. Most harmonic analysers give a simple reading of THD and individual harmonics as a percentage of the fundamental, as their default function. It is obvious that when the fundamental is small, the harmonics and THD will appear to be very high in percentage terms. The waveform observed on an oscilloscope will also appear distorted at light load, simply because the harmonics are fixed but the fundamental is a function of the power.
It is common in our experience that during commissioning when aspects such as harmonics and THD are being checked, the full load is not available for operational reasons. This can cause misunderstandings and disputes. It should be clearly understood that the harmonic specification relates to operation at rated current (i.e. at rated power).
(Part 2 of this series is available here)Browse all technical blog posts