In the second part of this blog series, Control Techniques’ chief engineer looks deeper at achieving greater energy efficiency using variable speed drives.
The previous blog gave an overview of the efficiency of drives and their applications, and how they are affected by the pattern of speed and torque combinations in a given application. We now look at how the proposed EU regulation and its related standards attempt to improve the energy efficiency of end applications by setting standards for performance and the provision of data. This includes the proposed new IE classes for drives.
There is a set of European (CENELEC) standards which were created as a starting point for a possible future regulation, which is EN 50598 parts 1 to 3 . It is likely that these will in due course be used as the basis for international standards (IEC), as are available for motors. The EU regulation is likely to use the IE class definitions from EN 50598-2.
Of this standard family, part 3 deals with environmentally conscious design generally, with the emphasis on the materials used in the product and the environmental impact of their eventual disposal. That is outside the scope of this blog. Part 1 deals with the design of complete systems which incorporate motors and drives. It aims to address the challenge for product standards which inevitably apply to the energy efficiency of specific products taken alone, when the actual purpose is to try to ensure the best use of energy in the end application, not the individual parts taken separately. It explains the issues which we covered in the previous blog, but in more detail, and sets out a methodology for assessing the energy efficiency of the complete system using the data for the drive as defined in more detail in part 2. Part 1 is intended to be used by technical committees working on energy efficiency of specific end applications. This is referred to as the “Extended Product Approach” (EPA). Part 1 contains useful tutorial material for system designers.
Part 2 is referred to as EN 50598-2 and it gives energy efficiency indicators for drives, the IE classes, which are likely to be the basis of a future regulation. The regulation  will apply to a drive as a product placed on the market in the EU. We have seen previously that the energy impact of a drive far exceeds its own energy consumption (loss) because it can allow major energy savings in the other components of a final application. The regulation cannot anticipate the wide range of different final applications, so it will aim to define the data which the drive manufacturer must provide to the purchaser. This data forms the “Semi Analytic Model” which is used for the EPA.
The groups responsible for creating the regulation and the standard are well aware that the energy benefit of using a drive where suitable far outweighs the losses, and they have aimed to keep the standard simple and practical. It is quite a long document, but much of the material is a detailed explanation of the sources of losses and the mathematical models to be used. It is a recommended read if you are interested in learning more on the subject.
In summary, according to the standard the drive manufacturer will have to provide the following:
It is likely that the regulation will prohibit the sale of class IE0 drives in the EU, and possibly set a timescale for prohibition of class IE1 drives. The standard does contain provision for possible future classes beyond IE2, but there is little benefit in trying to go further.
The purpose of the required matrix of data is to allow a user to predict the energy loss of their application, taking into account its specific torque/speed characteristic and pattern of loading, as we discussed in the previous blog and also explained in EN 50598-1.
To optimise energy efficiency, by far the most important aspect of a controlled speed application is to design the control function correctly so that the process is optimised and the output, whatever it is, is made available as required but without excess. The main skill of the drive system designer is to understand the overall process well enough to ensure that the motor speed and/or torque are set appropriately for the process. You can see from Table 1 in the previous blog that in this example the losses in the motor and drive are only 20.7% of the output, compared with 56% in the transmission and actuator. The 10% motor loss is typical of a modern IE3 class motor rated around 7.5 kW, and it is difficult to improve much on this. The drive loss is rather trivial. However a drive engineer might be able to find an opportunity to improve the whole system. Let us look at a new design where the motor and actuator are matched so that no speed-changing transmission is required. This might be made possible by using the drive capability to change the base speed of the motor. In that case Table 1 would become:
The efficiency is now improved from 56.5% to 67.9% and the loss reduced from 76.7% to 47.3% of output.
In this case we used the capability of the drive to move away from a restricted number of base speeds determined by the mains supply frequency and the number of poles in the motor. The drive also has programmable control capability so that the inputs from various process sensors can be used to help optimise the speed for the real operating conditions of the system. Finally, the drive can also act to optimise the operating condition of the motor, depending on the actual load.
The full-load loss in a 4-pole motor of IE3 class varies in the range of 14.5% for a rating of 0.75 kW to 3.8% for a rating over 185 kW. In the widely used and energy–intensive range around 5.5 kW to 55kW it is about 6 %. There does not seem to be much scope for further improvement here. The majority of the loss is copper (conductor) loss relating to the working current, which cannot be improved by any drive function. The best scope for improvement at high load is to use a permanent-magnet motor so that the motor power factor (cos f) can be close to 1 and the current therefore reduced.
It is however worth looking again at the fixed loss in the motor, because of the large class of applications where the operating torque is often well below its rated value. This might be in a fan or pump application where the normal delivery is less than the maximum possible, or in a constant torque application where the torque is usually less than the maximum possible. In that case, the magnetic flux density in the motor at its working voltage is higher than is necessary to achieve the required torque, and the fixed loss in the magnetic steel could be reduced by reducing the supply voltage and therefore the flux density.
To get a rough idea of the possibilities, take for example a fan application which usually runs at 50% rated speed and 25% rated torque. The power is therefore just 12.5%. The motor magnetic fixed loss is 2% of the rating, which seems to be trivial. However this is actually 16% of the normal power consumption. It would probably be possible to reduce the voltage by as much as 50% without increasing the current significantly, resulting in a fixed loss reduced to about 4% of consumption. The reduction of loss is small compared with the rated power, but it becomes significant compared with the actual average running power, which is what determines the owner’s power bill.
The traditional method for controlling the motor flux density in a variable torque application is the quadratic V/F mode, where the ratio V/F determines the motor flux density. Provided that the load is truly quadratic, i.e. torque proportional to the square of speed, and there are no load torque transients, this works well.
For constant-torque applications the Control Techniques Dynamic V/F function is very effective. This works by actively adapting the voltage to the motor current. It has the advantage that the flux is weakened effectively and automatically when the load torque is reduced, without any assumptions about the torque/speed characteristic of the load. However a sudden increase in load torque still results in a quick reaction, the flux is increased quickly so that the motor is unlikely to stall.
The inverter drive PWM switching results in an increased loss in the motor which is largely independent of load, i.e. it is an additional fixed loss. The higher the switching frequency the lower the added loss in the motor, but the higher the current-dependent switching loss in the inverter. At full load, research carried out in the development of EN 50598-2 showed that below 90 kW the best overall efficiency of the IE3 motor and the drive together at rated load is given when the switching frequency is about 4 kHz, the curve being quite shallow. This is why the loss measurement for the standard is made at these switching frequencies.
Figure 1 shows an example of the losses in a small motor and its drive as the switching frequency varies, both at full load (FL) and half load (HL).
The best switching frequency in this example at full load is about 5 kHz, while at half load it is about 7 kHz, because the drive losses at a given frequency are lower at part load. A drive which adapts its switching frequency to the motor current can give improved part-load efficiency, which again can be worthwhile in an application which spends much of its time at part load.
Control Techniques Unidrive M drives have an automatic switching frequency adaptation feature. The drive operates where possible at the highest switching frequency specified by the user, but reduces it if the drive losses become too high. This means that the switching loss in the motor is minimised unless this is resulting in excessive drive loss.