Plain bearings vs rotation speed = oil wedge

More dimensional details are needed:

- diameter of the bearings

- contact length shaft/bearing

- load on each bearing (presume you have 2)

- if possible clearances.

Roughly the loading capability is proportional to dynamic viscosity. Very un-precise you can expect that the admissible rpm to obtain same behaviour will be 3600*η(32)/η(68) at the running temperature. If speed is lower there is a risk but it cannot be estimated since we do not know the similitude criteria for the bearing under normal conditions. If it is overdimensioned one can use part of reserves.

Hello mayt2u

You don't advise whether this motor is shaft horizontal or shaft vertical, or in-between.

You don't advise loading either: vbelt or gear giving side loading, or generator giving symmetrical loading all around the shaft.

When you reply to nick name above, please advise motor shaft orientation, and load orientation, in addition to advising the other factors.

Thank you for

Kind Regards....

Here's something else to consider. When I went to VFDs in my plant, I was told that, unless the motor was designed for variable speed service, it wouldn't be a good idea to run it under 50% speed, else I'd be running a very real risk of the damages you're describing. Could that be your problem?

I would need to know the make, model, & specs on recommended lubricant, etc.

Two water pumps were mounted on a common skid frame, with the pumps arranged to be switched between 'duty' and 'standby' at long intervals. After a prolonged period of operation (several months), pump A was switched off and pump B was swithced on. Pump B was seriously louder than pump A.

The cause was eventually traced to the effect of pump A's vibrations producing microscopic flat spots in the bearing components of pump B, through inadequate mechanical isolation between the pumps.

The solution was to rotate the duty/standby function of the pumps at rather more frequent intervals than once every several months, so that flats could not easily develop in the bearings.

Hello PWSlack

This is a common problem with bearings of most types.

Ball and Roller type bearings can get false Brinnell Hardening between the stationary points due to both pressure and molecular movement, if left for long periods.

Likewise Babbitt type (soft Metal) bearings can also seize if unmoved for long periods, again due to the pressure and molecular movement.

When plant or equipment is left unused for long periods, it is necessary to regularly turn shafts so that bearings do not have this problem.

As you point out, the problem is aggravated by vibration.

Years ago it was found that stacks of gold bars in storage actually welded together, due to molecular movement, and the lower ones in a stack of ingots which had not been disturbed for several years, needed a pry bar to separate the bars, leaving quite ragged surfaces where the ingots had "molecularly welded" together.

Kind Regards....

It is quite interesting that this "Brinelling effect" was first detected during WW2 : at arrival in Europe many gear boxes were a lot noisier than they should have been. Investigation showed it was due to the high intensity of vibrations on bord the Liberty ships.

With respect to the problem of running at low speed there are 2 effects one is the bearing coming out of range and the second can be a too high temperature of the motor due to less cooling air. This is the reason the ramp time is limited according to the motor capacity to absorb heat without reaching a too high temperature. As it is known the life expectancy of wire insulation decreases with a power of temperature rise over the admissible limit and very rapidly.

I worked at a pumped storage hydro power station, we had six 300 MW sets, operating vertically, the rotor weighed 520 tons, and spun at 500 RPM, the mode of operation was, the machine would start in gen, mode by opening the MIV's, with approx, 850 psi behind them, the machine would spin up to 85% speed with the oil pump running, the pump would then switch off, and the angle of the thrust bearing would tilt slightly so that a wedge of oil would be drawn in to the bearing surfaces, the machine had ten thrust bearing pads, each one wedge shaped, supported on 52 springs, each one approx, 2 ins high, and approx, 3/8 to 1/2 dia,coils, after the first few weeks we started to have pad problems, on strip down we would find the white metal surface of the pads would be like broken paving, with the oil contaminated with white metal filings, for want of a better word! the pads weighed approx, 112 lbs, and were mounted under a highly polished collar, which was approx, 2 mtrs, in dia, the oil pressure on start up was 120 bar, which was sufficient to lift the rotor clear of the pads by 0.0015ins. from memory. The same oil pressure was used when we wanted to jack up the rotor for pad change, this would give us about 3/4 inch. clearance between pad and collar using special jacks located under a bottom collar.

When in pump mode the direction of rotation was changed, the machine now became a motor, requiring 260 MW to drive in this mode to lift the water back to the top lake, via the high pressure shaft.

After several attempts to overcome this problem incl, an experimental set of phos. bronze bearings were fitted, which "ran" in the first few revs.! the engineers then decided to have the oil pumps running continually,with it's own oil reservoir, in addition to the 3,000 ltrs, it had in it's own thrust bearing housing, also the cooling water system was changed to run as a closed circuit system, equipped with coolers, and filters, and pumped round the system, this method seemed to work quite well, reducing the down time for bearing renewal several months, the oil fiters were also changed on a regular routine basis of four weeks,

When I left, the powers that be decided to have a magnetic bearing set up, made in Japan, as far as I know, this system is still running successfully, this was at a cost of £3m I believe for just one machine.

Sometime later they employed an Iranian engineer, the story is, he looked at the problem that was, and designed a simple bracket to hold all the springs in place, so that the centre of the collar was more adequately supported,this meant that the springs were better contained,and stopped them working out to the free space on the periphery, the cost of this bracket just a few pounds I am told!

In operation if you observed the next machine to the one operating say in gen mode, it was possible to see and feel the rotor bouncing on it's springs, this also contributed to extra weight being imposed on the thrust bearings.

The gens. were made by GEC, the turbines made by Bovings, all this power station was housed 3/4 of a mile inside a mountain, at the time it was the biggest man made cavern in Europe, in gen spin mode the machine could generate 150 MW in 15 secs.with a max, 300 MW capacity, with a full top lake we could sustain this for 5 hrs, with all six machines generating.

I know this does not answer you problem, but thought it might be interesting. to know what problem we had with our bearings!

The way you describe the failure it could be possible that the bearings worked under cavitation conditions. I such a situation the implosion of air bubbles generates shock waves which destroy the bearings especially soft surfaces are subject to such destructions. The oscillations of the suspended mass could have generated the conditions for cavitation.

Your presentation presumes, that there is a definable minimum speed for an oil wedge to float the shaft. There is no such thing. If you slow down, it works, until it breaks down, then it is metal-to-metal contact. You have to go back under such conditions to much higher speeds to reestablish it. And it is load and vibration dependent.

There are bearings in the plain bearing family, that support a floating shaft down to zero RPM. They provide constant, pressurized oil flow to many small orifices around the plain bearing to provide a static floating of the shaft. This requires tight enough tolerances between shaft and bearing to maintain the floating oil pressure as the oil is spreading away from the orifice, providing adequate bearing surface. That is the way your car engine's shaft is lubricated.

Even if you are using needle bearings, you cannot go down to 0 RPM (and keep vibrating there) without bearing damage, unless employing pressure oiling.

There is a naughty way to get around you dilemma too. You might consider this, as I have difficulty believing you can modify how your motor is built. Get a torque converter, the type is used in every car transmission, just rated for you motor. It has an inherent slip in it. That way, when the load is at 0 RPM, the motor is still at somewhere 5-10% of its full RPM. The hydraulics convert the difference to heat.

From my perspective, since I am not a motor expert or an engineer, I always recommend to my customers that they use a lubricant with excellent boundary lubrication. From my experience, the heat is reduced approximately 20%. This heat reduction proves that it helps their situation dramatically.

I just happen to represent Schaeffer Oil and I have such a lubricant available. If interested, please get in touch with the equipment manufacturers recommendation. I can then come up with some suggetsions.

Before you blame the oil, what's the damage look like? In some VFD cases, electric arcs can develop between the inner and outer races of the bearings. These cause damage called "pitting" where the arc eats into the bearings races as well as the rollers or balls.

One cure is to put the VFD closer to the motor since long cable lengths can induce high voltages at the motor and/or drive. Another is to ground the shaft so that the current gets diverted to ground directly instead of through the bearings.

If for any reason you are obliged to run at very low speed it is better to use hydrostatic bearings either alone or in parallel to the hydrodynamic.

I can't help, but this discussion is running out of the questions szenario.

The thickness and load capacity of hydrodynamic fluid films for plain bearings are strongly related to:

- shaft speed

- shaft surface quality

- bearing diameter x width; and bearing clearance

- oil actual viscosity (viscosity grade and temperature)

- bearing load

Further more the method of oil cooling is important (start/rundown with cool oil is different from start/rundown with hot oil).

The minimum speed (where the fluid film fully is saparating the bearing from shaft surface is called transition speed and can be calculated f´rom above mentioned data (already asked with the firts reply to the initial question. Without these information we can't help mayt2u sufficiently.

By the way: the bearings heat generation (power loss) strongly is related with the oil viscosity. Using a VG68 instead of a VG32 is pushing the bearing temperature. I can't believe the VG 68 would be a good choice.

A. Caspers

Hi, sorry I've been away. The Great Outdoors of the North has kept me away from communication. Don't ask!

The answer to some of your questions are;

A) The Motor is a horizontal type.

B) The bearings are feed by a flinger ring.

C) The temperature of the bearings has never risen more than 56 C.

D) The manufacture is US Motors. This has no bearing (pardon pun) on the problem the motors have run successfully for 25,000 hrs without problems. It seems recently that with different operators this problem has come up.

E) The oil was changed to help the operators do what they had to do. I see now that it was a mistake to allow them to control the situation. There was never a problem with temp. just the contamination of the oil from the babbitt flaking.

F) The motors are run up on a weekly basis. The one leads for a awhile then the the other takes over for a month or so. Both have 60,000 plus hours each from start up, so they run pretty close together on hours.

Cheers beer time.

Need to know what speed, we can catch the wave at, in order to maintain! :)