Lies, damn lies and VSD efficiency!

04 July 2007

There is increasing concern that some large users of variable speed drives (VSDs) are (mis)using a ‘simplistic’ measure of efficiency to estimate total lifetime cost. Steve Barker gives the background to the issue and offers some advice about how real efficiency should be addressed

Inverter efficiency is notoriously difficult to measure, but many companies, particularly in the water industry, attempt to use crude and incorrect criteria as a means of comparing life cycle cost of different drives. What do I mean by crude attempts? In short, companies are considering power ratings, quantities of each rating and operating hours before multiplying this information with electricity cost to achieve a ‘total operating cost’. Naturally, the operating cost across the lifetime is > 99% of the total cost which leaves the purchase price at <1%.

However, this calculation is completely incorrect, ignoring the reality of manufacturer efficiency figures and missing the whole point about ‘system efficiency’ – namely that it is very easy to improve the apparent efficiency of the drive, but this is usually achieved at the expense of motor efficiency and may well reduce the overall efficiency. From my point of view, as an individual who makes his living advising companies on energy efficiency techniques, the really shocking point is that some very large users have managed to convince themselves that this factually incorrect calculation is a correct and rigorous technique, despite the fact this totally avoids any semblance of factual accuracy.

The root cause of the problem is that many large users are misusing, manufacturer catalogue efficiency data. A recent report by the University of Nottingham concluded that “the efficiency figures quoted by manufacturers of voltage source inverters are not accurate enough to allow for a direct prima facie comparison between products with any degree of confidence”.

There are numerous reasons why manufacturer stated values of efficiency cannot be directly compared to one another and therefore cannot be used to calculate relative lifetime cost.

Firstly, no standardised method of efficiency for inverter driven systems exist and therefore it is impossible to compare data from different sources. Efficiency measurements can also be artificially influenced - by switching patterns, for example.

Consequently, manufacturers usually model/estimate the losses in the separate components of a voltage source inverter and then sum them to get a figure for overall power loss. The complexity of the various loss mechanisms in an inverter inevitably lead to an estimate of efficiency that is not accurate enough to allow for a comparison between products.

Secondly, inverters are inherently efficient devices with circa 2 -3% losses being typical across most manufacturers. Fundamentally, the more efficient a system, the harder it is to measure its efficiency or loss accurately. This is because conventional techniques, such as direct measurement of input and output electrical quantities rely upon calculating the difference of two very similar numbers.

As previously mentioned measures to improve inverter efficiency may reduce motor efficiency and therefore total system efficiency. Also, instrumentation types, such as transducers and monitors, can greatly affect measured results – far in excess of expected tolerances.

It is also worth noting that input/output methods are only useful for comparative measurements (not absolute values), while attendant system disturbances, such as power frequency harmonics, greatly affect any measurement systems. Products with the lowest intrinsic efficiency (active front end, for example) may actually provide the highest system efficiency due to elimination of harmonic losses.

Finally, the cost of purchasing equipment being used in crude calculations can be misleading as the capital cost of an inverter is negligible compared to the cost of energy consumed over a ten year lifespan subject to a high duty cycle of use. Furthermore, in industries with high downtime costs, reliability is crucial and the effect of one fault could result in a financial cost significantly more than the purchase cost of the inverter.

The danger of such simplistic lifetime calculations is that the real factors which affect energy efficiency and lifetime costs are ignored and such calculations are open to exploitation by unscrupulous suppliers. Which begs the question, why do manufacturers state efficiency figures at all? In short, a measure of the heat loss of the VSD is necessary to allow panel manufacturers to provide the correct cooling and airflow arrangements to prevent excess temperatures inside cabinets. This is the only purpose of stated VSD efficiencies.

Such arguments offer an unwelcome distraction from the real benefits - unlike high efficiency motors, the benefit doesn’t derive from the product efficiency. The whole point of adding a VSD is to vary the speed of the load resulting in energy savings due to reduced mechanical power demand. Engineered correctly, the savings will be an order of magnitude greater than the intrinsic VSD losses!

Efforts to increase overall efficiency should concentrate on overall system issues. This involves the matching of all components to the required load characteristics (especially pumps), motors and the necessary torque/speed profile. Careful attention must be addressed towards correct system co-ordination and the attendant compatibility issues. Experience also shows that installation, commissioning and maintenance remains a major issue in achieving continuing levels of high efficiency.

Anecdotal evidence from one large UK utility company suggests that efficiency gains of up to 8% were obtained by simply re-checking the correct installation and commissioning of VSD systems. Also, many UK water companies continually suffer from the extremely expensive effects of incorrect system co-ordination.
It is clear, therefore, that whilst being arithmetically convenient, it is totally misleading and incorrect to compare inverter efficiency values. The only sensible approach is to apply technically rigorous engineering solutions to achieve ‘best practice’ system efficiency.

Steve Barker is Energy and Power Quality Manager at Siemens Automation & Drives.