Humans can live and work virtually anywhere on the globe, regardless of climate. Despite inhospitable environments, inhabitants of frigid areas expect cars, buses, and trains to run even in the depths of winter. However, operating machinery in low temperatures poses some unusual problems. People in colder climates should rightfully expect machinery lifetimes similar to those of people living further south, barring unforeseen breakdowns.

In low-temperature conditions, lubricants must be carefully selected, taking into account the whole range of temperatures over which machinery has to function. Understanding what happens to the bearing system, including lubrication, is at the heart of keeping machines moving in chilly environments. Choosing the right type of lubrication can overcome most of the difficulties of working in the cold, but many designers are unaware of the problems that rise as temperatures fall.

Gauging the damage

Rolling bearings happen to be some of the more difficult mechanical components to design for reliable operation over a wide temperature range. Most bearings will start to rotate eventually, even at -25°C, if the engine or actuator driving them is powerful enough. For example, if train wheel bearings were originally filled with grease designated for operation over a medium temperature range (-20°C to 100°C), then as thelocomotive starts to pull the cars the wheels may rotate intermittently or with extremely high friction. Heat generated in the bearings warms the grease. Then the bearings will rotate. This heating process takes 60 seconds or less, but that’s enough time to cause severe and irreversible damage to the bearings.

Problems with lubricants in cold environments stem from low viscosity. Olive oil kept in the refrigerator at 8°C becomes solid and must be warmed to room temperature before it can be poured out of the bottle. The same thing happens to any engine or gear oil, but usually at a lower temperature than 8°C. Oils are classified by pour point, defined by a standard test in which oil is poured out of a thin pipe. The lowest temperature at which it can still flow out of the pipe within a specified time is the pour point. This number relates to the oil viscosity, or flowability. When the temperature falls, viscosity increases and flow becomes more difficult. Unfortunately, flow in a thin pipe is not very useful in determining lubricant behavior in a bearing.

Another measure commonly used to describe oil behavior at low temperature is the viscosity index (VI). One must be careful when looking at the VI of an oil, because it should not be used as the single measure of low temperature performance. VI describes how viscosity varies with temperature. High VI implies low variation, while low VI means viscosity varies greatly with temperature. High VI oils include esters, polyalphaolefines (PAO), and silicone oils. Mineral oils are often low VI oils. It should be noted that VI says nothing about the actual base oil viscosity at low temperature.

For greases, a way to assess low temperature behavior is to measure the friction torque of a small ball bearing packed with grease and rotated at 1 rpm at a set (low) temperature. This method is standardized and used worldwide. However, the degree to which the bearing is filled strongly influences friction torque, and in reality not many bearings are 100 percent filled. Bearing friction is related to the shear stresses that occur in the lubricant during rotation. These stresses will not be the same in a small ball bearing as in a large spherical roller bearing. Consequently, this test is also not reliable on its own for the assessment of low-temperature lubrication.

Truth is, there is no universal method to predict low temperature performance of lubricants in a bearing. Such situations may require a field test, which generates information on which laboratory test methods are the most useful. A “normal” engineering team will not have access to a wealth of grease test data but should keep in mind some simple guidelines. Generally, bearing lubricants should not have excessively high base oil viscosity. The consequences of excessive viscosity in a rolling element bearing are:

• An increase in friction due to the additional effort required to transfer thick oil away from moving rollers. These losses are called churning losses and contribute strongly to overall bearing friction. If the churning losses are very high, rollers can slide instead of roll. This is called roller skidding and occurs in the unloaded zone or at the entry to the loaded zone in a radially loaded bearing. Skidding can cause flat spots on the rolling elements, which lead to premature bearing failure.
• Bearing friction will increase due to increasing cage-rolling element friction, which is strongly related to oil viscosity. Cage-roller friction contributes to roller skid.
• Once the oil on a raceway has been pushed aside by a passing roller, it does not easily flow back again. Hence, the next roller will contact less oil, and the third might not even contact enough oil to provide surface separation. This situation is referred to as starved lubrication. If the rollers are moving at high speed, starvation occurs earlier, because time for reflow between each passing roller is shorter. Starvation can lead to surface damage.

Going for the grease

About 80 percent of all bearings are grease-lubricated. Grease is a two-phase lubricant, containing base oil (about 80 to 90 percent) and a thickener. There are also additives such as anti-corrosion agents, antiwear, and anti-weld (so called extreme pressure, or EP) additives. To provide adequate lubrication, the base oil and thickener must be allowed to flow to the bearing raceways. When base oil viscosity is high, due, for example, to low temperature, then the base oil cannot separate out of the thickener matrix. This is what is meant by the grease having a low base oil separation rate. Because of this low separation rate, the availability of lubricant in the raceway-roller area is small and the risk for starvation becomes high.

If the additives cannot be re-fed to the surface, roller skid becomes even more dangerous, because there are no additives to protect the skidding surfaces. Therefore, a highly viscous oil, either pure or as base oil in a grease, will give rise to high friction in a rolling bearing, and increase the risk of bearing damage. This is true at low temperatures but also under other conditions where high viscosity has been selected. Obviously, the statement of “lower viscosity, better surface separation” is contradictory to the common belief that higher viscosity is better, but it is nevertheless true. In recent years, SKF has studied lubricant starvation and conditions when it occurs and has concluded that there is an optimum base oil viscosity for every bearing. Too high and the rollerraceway contacts starve and frictional losses go up; too low and the separating film becomes so thin that surface irregularities touch, increasing friction. In both cases, wear and premature failure result.

Information for this article was provided by Minett Media, Cambridge, England for SKF.