Technical

Energy Efficiency With Variable Speed Drives (Part 1)

24 May , 2016  

In this article, Colin Hargis, chief engineer for Control Techniques, looks at the issue of energy efficiency with variable speed drives.

Some readers may be aware that a new EU regulation is in the pipeline which is intended to control the efficiency of drives and drive systems. The regulation is still under consultation so it does not yet officially have a number, but it can be identified through EC mandate M/476, and is referred to as “Lot 30” under the consultation process for the Energy related Products directive (ErP).  The consultation appears to be stalled at present, but technical standards to underpin the regulation already exist as EN 50598-1 and EN 50598-2. The intention is that drives will have to be allocated efficiency classes just like industrial motors, and at some stage it is possible that the lowest grade will be banned from the EU market. Manufacturers will also have to provide further data about part-load losses, to help users assess the overall energy efficiency of their application.

Before the regulation itself comes into force some drive users may be interested to learn more about the proposed regulation and the efficiency classes, to assess whether they are relevant to the energy efficiency  of their own end product or system. In this blog we look at some of the basics of drive system efficiency. In the next we will look in more detail at some issues arising from the new standards and the proposed regulation.

One of the main reasons for using a variable speed drive is to regulate the speed of a motor to match the demand on the end process which it is driving, so as to optimise energy usage. This is particularly valuable when moving fluids (gases and liquids) because viscous friction means that the power required to move the fluid around a circuit varies as a cube law of the flow rate, so that a small reduction in flow rate gives a big reduction in power used. Control methods such as dampers, valves and even variable guide vanes also result in unnecessary power loss.  This idea is so well known that there is no need to write much more about it, there are many useful guides available [e.g. references 1 & 2].  However with the forthcoming regulation in mind it is helpful to review a few principles, mainly to keep the effect of the regulation and the standards in perspective.

Energy loss in a drive system

The diagram, not to scale, illustrates broadly how power is consumed in a drive application. It is clearest to work in terms of losses rather than efficiencies. At every stage there is a power loss in the relevant device, which is generally expressed as a proportion of its rated throughput.

arrows pic

The lost power emerges as heat, usually of the surrounding air.  Sometimes the heat can be put to good use, but usually it must be considered as wasted, and it may even incur further costs if additional ventilation or cooling of the area is required. Actual losses vary greatly across various applications, but a typical breakdown for an air-moving application, operating at maximum output, is given in Table 1. Note that at each stage the loss incurred in a device is a function of both the useful system output and also the accumulated losses of all the other downstream devices.

first table

In this example, the overall efficiency is about 56.6 %. The largest loss is in the actuator, and the loss figure of 30% is typical for a fan for moving air. Air is a difficult fluid to move efficiently, a modern pump might have losses closer to 10%. All of the losses shown can be reduced by improved technology, and the attention paid to energy efficiency means that they all tend to be reduced over time as improved designs become cost-effective or are required by regulations.

Note that the drive loss is the smallest in the list, and this is realistic in the great majority of applications. The loss of 3% is rather trivial compared with the others. Modern drives have very low losses, the main driver for this being the desire for physically compact units, which means that he cooling devices (heatsinks and fans) have to be minimised in size, so the losses too have to be minimised.  The power throughput of the drive does include all of the other losses, so the headline drive loss of 3%, based on the drive data, becomes 5.1% when expressed as a proportion of the system output. The intelligent use of the drive can often allow useful reductions in the losses of the other devices, resulting in savings which far exceed the losses in the drive. However we do have to consider the actual operating conditions rather than only the maximum load condition.

Control and losses

The typical loss values discussed above are the “headline” values which are given at the rated load or throughput of each device. They are therefore relevant when the system operates at its maximum design output.  Many systems spend large parts of their lives operating below their rated load, because the demand varies but the system has to be designed for the maximum. Also, the performance is usually judged by the maximum throughput capability, so the supplier tends to over-size components to avoid the risk that the customer rejects the system if it fails to deliver the rated output during acceptance trials. A control system is therefore needed, with a method for adjusting the output. The control technique applied can affect the part-load efficiency a great deal. For example, it is well known that air dampers and regulating valves cause quite high part-load losses, because they result in a rise in pressure at the fan or pump which means that it has to develop more power than is needed at the delivery point. The variable speed drive avoids this additional loss.

The efficiencies of all components change with load. The details vary greatly, but generally the losses have the following elements:

  • A fixed loss which is independent of output. This tends to reduce the part-load efficiency.
  • A proportional loss which does not affect the part-load efficiency.
  • A rising loss (e.g. square-law) which tends to reduce the full-load efficiency.

The result is that there is usually an optimum efficiency output level, for example in a standard induction motor this is around 80% of rating. At a higher output the efficiency falls slightly. At a lower output the efficiency also falls, but the actual power loss does also fall.

Losses in a variable speed electrical drive system

Having summarised the general situation, we can now look in more detail at the electric drive system, i.e. the motor and drive. The output of the system is mechanical power at the motor shaft, comprising the product of torque and speed. Both motor and drive have loss elements which vary with torque and speed. Table 2 summarises these. For simplicity we assume that current is proportional to torque. This is a simplification because it ignores motor magnetising current.

2nd table

Note that we have to consider the effect of speed and torque both separately and combined. The resistive losses in the motor are almost entirely related to the torque, regardless of speed, and this is also true of the inverter stage of the drive. On the other hand the losses in the drive input stage rectifier are purely a function of the power throughput, i.e. the product of torque and speed.

This rather complex picture can be simplified when a specific type of load is considered where the torque and speed are related. For example a simple pump or fan which feeds into a process with little static head, so that the pressure is predominantly a square law function of flow rate, gives torque which is a square law function of speed. Conversely a process such as a conveyor system has a torque which is largely independent of speed, but depends on the demand on the conveyor. These two types of load are broadly referred to as “variable torque” and “fixed torque” drive applications respectively.

As well as the losses inherent in the drive and the motor taken separately, there are losses which are a function of the combination of the two. The key factors in the interdependence are:

  • Operating magnetic flux density, controlled by the drive and imposed on the motor, affecting magnetising losses. (Also affecting resistive losses for a given torque, there is an optimum flux density for any given torque and speed.)
  • PWM switching, controlled by the drive and resulting in additional losses in the motor
  • Power factor of the motor, reactive current imposed by the motor resulting in additional current losses in the drive

A simple standard for the efficiency of the drive would only deal with the losses in the drive alone, using a standardised motor load. A useful standard has to address the interdependence and manage trade-offs; for example that the chosen PWM switching frequency has to balance the desire to minimise losses in the drive, requiring a lower frequency, and the motor, requiring a higher frequency. It also has to allow for the designer of a complete system or machine to enable them to calculate the losses in the complete machine over its practical range of operating conditions.

In the next blog we will look more closely at the standards, particularly EN 50598-2, which specify energy efficiency classes for drives, and consider how they manage these requirements. We also look at functions available in the drive which can optimise efficiency, and in particular optimise the part-load losses which might be more important than they first appear.

References

[1] https://www.carbontrust.com/media/13063/ctg070_variable_speed_drives.pdf

[2]  http://www.gambica.org.uk/resourceLibrary/CEMEP_guide_to_energy_efficiency_with_electric_drive_systems.html

, , , ,


One Response

  1. […] previous blog gave an overview of the efficiency of drives and their applications, and how they are affected by […]

Leave a Reply

Your email address will not be published. Required fields are marked *

Colin Hargis

Colin Hargis