Electric motors: cooling concepts

Thursday, March 29, 2012

Tags :  alternative-energy

kersThe rise of hybrid drive technology, both in the production automotive sphere and latterly in racing, is something I hope will breathe new life into both realms of engineering. Years of racing powertrain regulation have led to some fantastic pieces of machinery powering cars and motorcycles around the circuits of the world, but they have become increasingly irrelevant to the direction of mainstream production vehicles.

If we take the energy storage aspect of hybrid powertrains (electrical energy only - flywheels will be dealt with in another article), we are still some way from matching gasoline, diesels or alcohols. However, electric motors are more than competitive with internal combustion engines. While not at anything like the same stage of development as a race engine, an electric motor can deliver huge power per unit mass - 5 kW per kg was not at all unreasonable even a decade ago, and we can assume that the current batch of electric motors used in Le Mans and Formula One competition are at or above the same level.

Just as an engine has losses that detract from its output, so does an electric motor. There are losses in the rotor and the stator which are apparent as heat; if the motor is not to overheat rapidly, its output must be limited or the heat must be rejected. The power is limited often by the ability of the conductors in the windings to carry the current without overheating and melting.

The limits of passive air-cooled motors - those using finned stator cases - in terms of current capacity per square millimetre is in single figures, with 5-8 A/sq mm being typical*. This may be improved by forced convection, for example using fans or air ducts to channel air from the outside of the vehicle to the motor, thus keeping the cooling fins at a lower temperature. Finned motor cases lead to a large increase in diameter, although for low-cost static machinery, this is not a concern.

Jacketed motor bodies offer an effective way of taking heat away from the stator. As with a water jacket around a single-cylinder engine, the same principles apply with an electric machine but, if the windings are to be kept dry, this method relies on the conduction of heat through the steel stator 'stack'. An accepted level of current in the windings of such machines is in the region of 10-15 A/sq mm. Complexity is added here because we now need to have a cooler in the system, and a pump.


kers-cooling

A method of cooling the windings more directly without having them actually in contact with liquid coolant is to have enclosed coolant channels running along the stator slots and between the windings. This method most readily lends itself to conductors that are produced in rectangular section. By conducting heat through thin-walled coolant vessels rather than via the steel/iron stator, we can increase the allowable current capacity of the conductors to around 20 A/sq mm.

If we wish to go further than this limit, then we have to consider directly cooling the conductors with an electrically non-conductive coolant. There are many suitable fluids for this, ranging from pure water (deionised water is a poor electrical conductor) to special-purpose electrical oils, which are designed for cooling electric motors and transformers, and are hence often known as 'transformer oils'.

There are two common methods of cooling the winding conductors directly. The first is immersive cooling, where the winding coils are immersed in coolant which is then circulated via a cooler. The second method is to spray oil directly onto the end-turns of the coiled conductors. These are the areas which are most easily accessed, and in Fig. 1 here they are the parts of the coils at the end of this wound stator. Such cooling schemes where coolant removes heat directly from the coils allow a current capacity of about 30 A/sq mm.

* Gieras, J.F., "Advancements in Electric Machines", Springer, 2010, ISBN 9-0481-8051-1

Fig. 1 - Direct cooling of the stator windings allows high current densities to be used, resulting in a smaller motor for a given output

Written by Wayne Ward

Leave your comment

Comments

The picture shown is a wound rotor, not a stator.  Most high speed, high performance motors are brushless, and use wound stators with PM rotors.  Most of the heat load in these motors is produced in the stator's copper windings.  Getting that heat out of the copper windings can be difficult because the copper is electrically (and thermally) insulated from the iron stator.  The most common approach is to liquid cool the OD of the stator, and forced-air cool the copper windings and rotor.

Liquid cooling a rotating rotor is usually problematic due to windage losses.
Terry,

This is a stator, but one where the stator is on the inside of the machine rather than the outside as is conventional. There is a common assumption that the stator is always on the outside. To make this mistake is perfectly understandable; it does look like a conventional wound rotor from an automotive alternator or similar machine.

Liquid cooling of motor stators by the methods described in the article is a common way to achieve the very highest current densities, and for machines such as that pictured, liquid cooling of the stator OD is not possible. The data presented by Gieras and referenced in the text shows that stators with direct liquid cooling of the conductors can achieve double the current density of a stator cooled by an OD jacket.

Liquid cooled rotors are also used, although this increases risk, cost and complexity. I worked on such a motor almost 15 years ago and this ran to speeds in excess of 15000rpm. However, the technique did not have high 'windage losses' as you predict. This would only be the case if the rotor was to rotate within the cooling medium.
Wayne-

Thanks for the correction.  I'm a mechanical systems guy, and my knowledge of electrical machines is pretty limited.  

Most of the high performance motor/generators I've seen are high-speed, brushless PM designs.  They mostly seem to be slotted, wound outside stators with surface mounted PM rotors.  The OD of the stator is liquid cooled, and the copper and rotor surface are forced-air cooled.

As you noted, the motor output is limited by stator current density.  Which in turn is limited by the heat transfer rate in the stator.  Getting heat out of the stator iron, via an external liquid cooling jacket, is usually much easier than getting heat out of the copper.  The heat generation in the copper also tends to be a bit higher.

Even with a surface mounted PM rotor, there is also a mechanical speed limitation.  The highest speed PM rotor designs use carbon sleeves on the rotor OD, but they are typically limited to tip speeds below 250 m/sec.

Regards,
Terry
Hi,
do you have any idea how much heat will be generated in the rotor ? For example as a % of power input. I am designing a BLDC motor. Now that the magnets became very expensive, this matters even more. Magnets are available in temperature ranges from 80, 100, 120 etc degrees. I want to liquid cool the motor by direct submerging the stator coils in the cooling liquid. The rotor will have to remain air cooled. In my simulation software, there is no way to test the rotor temperature.
[...] motor to fit in the smallest space, and liquid cooling is quite likely to be required. In a previous article we noted that some liquid cooling strategies are more successful than others. Surely with the only [...]