Can we detect false brinelling using vibration survey?

Can the vibration from a running machine be transmitted to its standby unit and cause false brinelling to occur?

How long does it take for false brinelling to occur?

Yes.

Yes.

Depends on severity of vibration, static load, etc. Probably anywhere between a week and a number of months.

If yes, what can we do to prevent it from happening in the standby machine?

How much vibration can the standby machine tolerate before false brinelling occurs?

Old merchant navy method was to make stop blocks out of teak wood with a spliced rope and prop up the shaft and when machines are swapped the block will fly as long as the rope let it, if it´s put in the correct way, that is and there should be no one standing beside. Municipal heatplant method is to go and turn the machine once a week. Last method is to let the valve of the spare pump/fan bleed/leak so the standby machine is always turning. I think I remember to have seen figures about 0.3-0.5 mm/s as where problems may start but not for sure. This has been up before somewhere so do a search and you may find more. Olov

Thank Olov!

So the problem that I raised up is not quite new, ah?

Basically, you mentions 3 methods:

1) Make stop block to prop up the shaft. - Any diagram to show this?

2) Turn the machine once a week. - This is probably the origin of some recommedations in the internet. Once a week appears too frequent. How about once a month?

3) Let the valve of the spared pump/fan to bleed/leak. Is it the inlet valve?

Thanks for the vibration figures of 0.3 to 0.5 mm/s. Walt has mentioned 1G peak in the motor forum and Electricpete said it should be lower. I wonder whether these are within the same magnitude after conversion.

This message has been edited. Last edited by: Josh,

Josh,

1. Pending device design, it should be pressing the shaft so it has no freedom to move and thus make no marks from the hi local pressure that the contacting balls in the bearing make. Not so hard so the motor burn when trying to start when forgetting to release the thingy, think of a wedge shaped thing made in wood. Watch the safety.
2. Pending the severity of the problem, spare rollers for papermachines that was stacked near the machine suffers from the same so now some have a small motor giving continous slow rotation and that works. Municipal heatplant people may sometimes have time to spare so they may overdo it. Goal is to keep some lube film existing where it should be.
3. Any way system permitting so you get normally a small backflow rotating the device slowly.
4. As always, Walt is correct, 1g must be one limit as that will make the thingy lift and bang down making the marks, what that would be in mm/s depend on the frequency, possibly the speed of the other machine, there are conversion calculators on the net for that. Olov

Olov,

1) I cannot visualize the ship method. So I draw something as attached. Is it correct? It has been said shafts should be locked in position during transportation as well.

2) The motorised slow rotation will certainly benefit big machines like the paper rollers. How did the heatplant personnel rotate the shafts? Using a chain wrap?

3) What is the small backflow rotating the device slowly?

4) Thanks

1. More like half of what you have like the wedge you have stopping things rolling on the floor. I will try to draw something later. Rope was to catch the thing when it commes flying after a start.
2. "By hand" so chain wrap would work.
3. If you open the shutof valve so it leaks or as I have heard of but not seen, drill a small hole so you get a small flow "the wrong way" so the pump or fan will slowly rotate backwards, not all processes or plants accept this behaviour.
4. You are welcome.
4b You could also make the companion machine vibrate less or better isolated :-)
If you add mass like my favourite concrete in a girder frame it will get lower levels.

Olov

1) If I prop up the shaft using my set up above, the contact point in the bearings will be on the top. So I'm waiting to see your sketch.
2) Ok
3) Drill a small hole where? In the shutoff valve?
4b. How to reduce vibration of or isolate the standby machine? Using an an absorber?

Have a customer with a couple of installations where the 'spare' fan has no isolation damper; a 'blank' has to be installed to isolate the fan which is time consuming, and even worse, requires a shutdown if the operating fan has to be shutdown. So they just let the spare fan freewheel... backwards. That will be normal for fans sharing a discharge duct or plenum.

It's not the best solution, but it's better than the false brinelling which has been a serious problem in the past.

On another pair of fans (overhung, roller bearings, continuous oil mist lubrication) we've seen false brinelling after only 6 months. Acceleration levels went from 1 g (new bearings, fan balanced) to 60 g's... the fan just sat there for 6 months. So, yes, it's a very real problem.

I think the "best" solution is to put these units on VFD's and use all of them, with no 'spare' unit. Unless it's a high-head machine and you need the speed, it lets you operate at a lower average speed, and you have "spinning reserve"... if one unit fails, the other automatically picks up the load, with little or no loss of production. Of course, in the short term, this is an expensive solution.

Rusty,

I don't understand why the spare fan freewheel backwards. Pls explain.

About the VFD idea, you may do so for fans but I think not for pumps which are of my concern. We may have spinning reserve but you are running both machines which defeats the idea of having redundancy in the design (ie. 2x100%).

This message has been edited. Last edited by: Josh,

If you have two pumps in parallel, normally there is a closed valve or check valve in the standby pump discharge to prevent backflow (backflow would reduce system efficiency and can reverse rotate the standby pump.)

If you have two fans in parallel, normally there is a closed damper or backdraft damper in the standby fandischarge to prevent backflow (backflow would reduce system efficiency and can reverse rotate the standby fan.). We have a few fans where the standby fan rotates backwards due to leakage in the backdraft damper.

Even though backflow is often considered undesirable, Rusty points out it provides an advantage if the standby machine rotates backward slowly which prevents false brinneling. Also perhaps in the case of check valves or backdraft dampers you can size a hole in disc of just enough size to rotate the standby machine without losing any more power than necessary. I have seen check valves with holes deliberately provided to provide small backflow in case where it is desired to keep the fluid in the standby pump warm to prevent unacceptable system thermal transietn upon start.

It can be a little extra strain on a motor to start when the machine is rotating backwards which may or mahy not be a problem. If rotating backwards quickly, there is a possibility of tripping overloads if start is too prolonged.

Ok thanks Pete. I understand now.

How to estimate the size of the hole on the check valve disc? Its location in the middle of the disc?

I just can't see where allowing the sister unit to turn backwards slowly would be a viable option. Too many bad things can happen, like the possiblity of spinning the impeller off the shaft, the extra start-up torque as mentioned above, possible bearing damage from not having enough load to make the bearing spin and maybe not enough speed to get good oil circulation. I think that the best solution would be swapping the pumps at a resonable time interval.

Good points that you want to try to anticipate all possible adverse consequences before you perform a modification to provide backflow, and weigh those risks against the benefit you are expecting.

I agree it would probably be a bad choice for sleeve bearings since hydrodyamic lubrication tends to go away at low speeds. I was assuming the discussion was about rolling beairngs, but you know what they say about assumptions.

I also agree periodic swapping is a great thing, not just for bearings, but to make sure your standby is really functional. Unfortunately at our plant we have a few machines in our plant that Operations Department refuses to shift because they consider there is unacceptable risk inherent in the swapping process. One example is stator cooling water pumps. We have had the check valves stick open on the secured pump. You can't close the discharge valve until the pump is secured. Once the pump is secured, you have already put the plant in jeapardy if that check valve fails open... low stator cooling water dp after a short time delay initiates a turbine trip. 1300+ MWe going down.

Sorry Pete, we were talking about rolling element bearings.
I remembered from a time in the past when an inlet steam valve began leaking through on a Clark Air Compressor that I performed maintenance on. It was only turning 40 or 50 rpm when I found it, but I did not know how long that had been happening and of course the operators did not notice it. It wiped out all the bearings on the air compressor and the turbine because the oil pump was not running. Just a nice little story but really does not apply here.

quote:

Too many bad things can happen, like the possiblity of spinning the impeller off the shaft

Vibeguy, my original post was before Josh specified this was a "pump" question. Of course, letting a pump freewheel backwards is not an option, for the reasons you stated. But this may be an option with fans.... not the best option, but an option that may be preferable to false brinelling when that is a known problem.

My experience is that an idle spare is often not reliable when called upon to replace a failed unit, usually because it has been sitting idle for too long (for the reasons others have mentioned above). And when units are routinely 'swapped' (weekly, monthly, etc), then chances are good they are both worn out at the same time since they have the same service hours. My opinon, is that perhaps the best option is to run the 'spare' one day a week, or one week a month. This will 'prove' the unit for service, and uncover any problems that may prevent it from being a reliable backup.

As for running both units at reduced speeds, this was usually not possible before the widespread use of VFDs. We know the forces on equipment due to imbalance vary with the square of the speed. By extension, we might reasonably conclude that service life will vary inversely as speed squared. (I once had a Toshiba motor engineer tell me that his opinon was that motor life varied inversely with the cube of speed.) With this in mind, reducing the speed of a unit with rolling element bearings is one of the best ways to extend the life of the unit. Of course, both units will 'age' at the same rate, so that is a consideration. But it is a practical way to deal with the false brinelling problem, where it is know that it will occur.

'Blocking' up the idle shaft to take the load off the bearings doesn't seem practical to me, and is certainly not safe unless the spare is "locked out" (and thus unavailable for immediate service).

Rusty

I agree with paragraphs 1 and 2 except that I prefer to run the duty pump with mech seal for 2 months and then test run the standby unit for 1 day. After 1 month on standby, turn the shaft by 270 degrees to lube oil the bearings and move the false brinneling location.

Does your paragragh 3 apply to both fans and pumps? Selecting a motor with lower speeds eg 1800rpm rather than 3600rpm would be better, won't it?

How to block a shaft in a fully assembled units during storage and transport?