The art of over-dimensioning

A common question with electrical motor applications is how much safety margin one should reserve? In other words, how large a motor should be selected so that all the requirements are met and everyone is happy? Normally, we just select a motor from catalogue, and we don’t really think about how much margin there actually is.

Normally, the first margins are already put on the load. Many pump, blower, or compressor application requires some design margin, which guarantees that the system delivers the required air flow (or whatever) at all possible conditions. Hence, the motor has to deliver higher torque (typical margin is 10%) than the load actually requires. Further, in VFD applications there may be several different operating points, and the highest of which is not necessarily the most frequently driven.

When built according to IEC 60034-1, electrical motor has plenty of margin. First, the ambient temperature is set to 40 °C and the altitude to 1000 m above sea level. If these values are exceeded, de-rating is necessary, but if they are not, no actions are even considered. However, the NEMA MG-1 standard defines that motors at sea level can have 8 °C higher temperature rise. Second, the motors are stamped for temperature class B (130 °C) with nominal temperature rise of 80 °C when measured with resistance method, which includes a 10 °C safety margin. When measured with Pt100 sensors, the temperature rise can be 10 °C higher (90 °C for class B). In reality the motor temperature rise can be whatever as long as it does not exceed 80 °C, so in surprisingly many frames, there is plenty of margin in the designed temperature rise. Hence, we may assume that there is a 10 °C margin in the motor temperature rise. Third, nowadays a common practice is to build the motors with class F (155 °C) insulating materials, even though the maximum operating temperature is given by class B. This yields to a margin of 15 °C. This is actually not needed by the IEC standard, but by many customers. It is also required by API 541 standard for large motors in oil and gas industry.

Now, let us consider a typical motor application in southern Finland. The nominal load is 100 kW, but under the worst case conditions, the load can increase to 110 kW. The motor is specified to have additional 10% torque margin at nominal speed. A common interpretation is that the required motor power is 110 kW + 10% = 121 kW, and the correct motor is the first frame size larger than that, so 132 kW (four-pole IE3 motor is considered). The frequency converter and cables are chosen accordingly.

Now, the site is located close to the shore line, only few tens of meters above the sea level. The motor is assembled indoors with the ambient temperature of room temperature, 20 °C. Actually, these conditions are not only valid for southern Finland, but to a surprisingly many locations in Europe. The ambient temperature provides a temperature margin of 20 °C, the altitude 8 °C, the temperature rise measurement 10 °C, and the motor materials 15 °C. These put together yields 53 °C. So, theoretically and considering the stator winding alone, the motor could run for 25 years continuously in a room heated up to 73 °C, which corresponds to a sauna with comfort setting.

Another way to illustrate the margin is to convert the temperature to power. According to the tables in Siemens and Hoyer motor catalogues, an ambient temperature difference of 10 C corresponds roughly to a power difference of 5%. So, a 10 °C increase in ambient temperature requires a power de-rating of 5%. With this relation, the above 53 °C yields 26.5% power margin. Considering that the motor could be this much overloaded, the required motor power for the 121 kW load is only 95.7 kW. The motor frame size would then be 110 kW. If we consider that the motor torque margin can be put on the nominal load instead of the maximum load, it leads to 87.0 kW, and to a 90 kW motor. We could further consider only the designed load point of 100 kW placing all the safety of the load into the safety of the motor. This leads to the required motor power of 79.1 kW, but the frame would again be 90 kW, since the next frame downwards (75 kW) is too small.

Obviously, if a motor is overloaded, it has to be able to deliver the power requested by the load. IEC 60034-1 states that the motor should have at least 60% margin between the load torque and the breakdown torque. For example, a 90 kW four-pole IE3 motor manufactured by ABB has a breakdown torque of 2.9 times the nominal torque. There is 190% of margin, which is more than enough for this particular case. If the motor is driven by a VFD, the torque margin could even be less, although this logic is usually not accepted since the VFD-fed motor is laid out is if it was a DOL-connected one.

In this example, a 100 kW load can – under these operating conditions – most likely be run by a 90 kW motor. But the normal practice is to select a 132 kW motor, which is now 47% over-dimensioned.

An obvious consequence from over-dimensioning is increased costs. The motor and the VFD frames are too large. The cables are chosen for 231 A instead of 158 A (again, ABB four-pole IE3 motor), so they are 46% more expensive, if the cable price goes together with the amount of copper. But the other and not so obvious consequence is the motor efficiency. It is well known that the induction motor efficiency drops at partial load. Now, a 132 kW motor operates at 100 kW, so at 75.7% of the load. Can the motor still pass the IE3 efficiency class requirements, or is it actually an IE2 motor? And what if there were more operating points on lower loads?

Over-dimensioning is generally accepted, because it brings safety. A rotating device usually takes a longer time to break, if it runs cooler. This is not only valid for failures in the stator winding, but for various other kinds of failures as well, including bearing failures. Sometimes, over-dimensioning is even justified because the bearings provided by the given motor manufacturer are known to fail prematurely. Of course, this is also a way to improve quality, but why is the customer willing to pay for it?

Thick margins can be well understood for critical applications, but not for common ones. If the weather conditions are very extreme on one day in a year on average, why is it stated as a fixed requirement? As if the weather is extreme all the time. Also, why are the rare conditions caused by failures or malfunction in other systems (such as sudden voltage reduction, blockage of filters etc.) considered as if continuous operation? As if no-one repairs the other systems. Why is it not accepted that under these rare conditions the load could be de-rated, or that the motor could run slightly hotter for a few hours or days? It hardly has any influence on the specified 25-year lifetime.

The above is merely one example. It shows how the motor selection can sometimes cause severe over-dimensioning. Yet, there is a message. Nowadays, one can assume that the design of an induction motor is more or less cemented, and no major breakthroughs can be expected. Further, the quality of the motors is constantly improving, and failures occur less and less frequently. Still, the general principles of motor dimensioning together with the IEC motor standard advise to set up the system as in 1970s with everything DOL-connected. The frequency converters are usually not taken into account. Perhaps, the next big motor innovation could be linked to the optimisation of the entire system including the VFD, the cables, the motor, and also the load.


By | 2017-09-13T09:19:09+03:00 September 6th, 2017|General|0 Comments

About the Author: