We have been discussing how to best do (proximity probe based) radial shaft vibration trips. The proposed languague in the new version of API 670 suggests 1oo2. I personally like 2oo4 (any two of the four probes on a machine case vote to trip).
I am interested to see what the industry is doing. Thanks for your participation.
among three voting choices, chances for spurious trip is highest if 1oo2 voting philosophy implemented. i used to handle a steam turbine that has issue on one of its radial probes. 1oo2 arrangement. in that particular year, we had 3 unnecessary PLANT trips initiated by same radial probe. after a discussion, the team finally decided to remove the trip considering the train has been running good for years. it was not there in the first place (1980) but due to recurring issues on coupling failure after few years running time, the radial trip was implemented. the issue on the coupling resolved and since then there were no radial vibration issue but the radial trip interlock was still there until the issue on the probe emerged..
yea, i like 2oo4 too. the place where i work now has 2oo2..
Representing vibration equipment supplier and OEM (sometimes). Do not think that we ever dictated customer how to trip their machines. Personally like 2oo2 and 2oo4 with some modifications. I would recommend tripping on any High and HiHi (danger) levels combination, no matter does that happens at the same bearing or not. That will protect from spurious trips in case probe loop fails, but will give enough protection. As machinery diagnostics engineer I can't remember story when only one out of four probes was in HiHi, while other three shown low vibrations. But it is possible in case of large misalignment, for example.
At our facility we presently only trip on axial 2oo2 voting. We are in the process of implementing radial bearing shutdowns which is way overdue,
We are pursuing the X and Y on a bearing vote to trip philosophy. Personally, I have reservations regarding this way of thinking as far as machine protection.
I am no means an expert in this field, but observing shaft orbits it seems to me that potential bearing damage could occur in one direction and not the other. In other words, we may iniate a X vote with high amplitudes but may not encounter high amplitudes in the other Y axis (at the same bearing), which would not satisfy a 2oo2 vote at that single location. I may not be thinking correctly on this issue but have often wondered if this situation could occur.
As you know the typical arrangement does not include redundant transducers. Thus, voting 2oo2 doesn't seem right. In fact we have had relatively new equipment sustain greater damage due to the delay in time for the second channel to go into trip.
Of course our downstream guys think otherwise. In fact I thought that ShELL (77345 is that?) used 2oo2 after a discussion at last years Turbo Show in a discussion session.
Redundant transducers are an option in some cases, perhaps more than thought.
Similar to what you like, I have been pushing a 1 in trip and 1 or the other signals in alarm. This should cut out nearly as many false trips as the others with the potential to react as quickly as a 1oo1, which would be either my favorite or second to the 1 in trip and 1 in alarm.
In our latest version of our standard we settled on 2oo4. I like it in that to me radial vibraiton trip is unneccessary for most malfuntions, and it is only when things get bad enough to move more than one probe into danger that we need to shut. In 2oo4 if there is two probe on one end, or very directional vibration on both ends then you trip.
Remember if somthing goes high, the operator can with some analysis support by RE trip the machine at any time. This is only for situations where we want to take the decision out of their hands.
False trips in a 1oo2 configuration have bit me as well. Both upstream and downstream.
i see all radial trip i have here having no delay i.e at default 0.1 sec..
is it normal not to have delay? if it is not, what is the normal delay i should have?
i have seen increasing a real vibration only in one direction. (it wasn't discharche) so i'll choose for one probe. the importancy is to install your hardware cornfirming the API standards through required & trained personal=> invest not only in equipment also in human beings...
|Powered by Social Strata|