Vibration analysis of rotating machinery is very challenging; there is a lot to know and there is a lot of pressure to get it right.  That knowledge can be gained through experience and formal training.  The best analysts result from a combination of the best training (and lots of it) and the richest experience.

To a reasonable extent, the analyst is in control of the experience they gain.  Their experience is based on where they work, what types of jobs they perform, how much effort they put into the accuracy of the diagnosis, who they work with, and whose advice they seek and accept.

But the analyst is in total control of the training they receive.  The analyst controls where they receive their training, and how often they are trained.  Training can involve formal classroom training and e-learning or distance learning, but it also involves reading articles and books, participating in forums, and attending conferences.

The focus of this article is on the formal training: classroom, e-learning and distance learning.  In the author’s experience, analysts generally fall into three categories:

• Type One: Analysts that rely on their experience and recognize fault conditions because they have seen them many times before on the same type of equipment.
• Type Two: Analysts that rely on “cookie-cutter” patterns that you see on wall charts. To this group I will add the BFD analysts  – Bearing Fault Detector analysts.  And I will add the OMG analysts – “OMG, the vibration is high; we better stop the machine and find out what is wrong!”
• Type Three: Analysts who understand the machine and the analyzer and who can determine the condition of the machine using a combination of knowledge and experience.

Of course, you may consider that many analysts fall into more than one category, but I will now narrow the focus a little further.

Type One analysts

There are a lot of analysts with a tremendous amount of experience who can quickly diagnose faults.  But in the author’s opinion they fall into two categories: those who rely on experience alone and those who also understand the machine and tools and know why the vibration changes.  It is this first group that we will call “Type One” analysts.  The second group falls into the “Type Three” analyst category.

Type One analysts often work on the same type of machines for many years; often the same machines for much of their career.  Some 10-year analysts have one year of experience repeated 10 times.  They have seen failures in the past and they can remember how the vibration changes as the faults develop.  They are valuable people to have on staff, but they are hard to find – you either have one on staff or you don’t.  It takes a lot of trial and error to become a Type One analyst.  It takes the support of a patient manager/supervisor to accept the trial and error – not that any analyst is perfect, but when you rely on experience alone you have to experience success and failure to learn.

In the author’s experience, Type One analysts usually dismiss the need for training, and totally dismiss the need for certification.  You will hear them say “What I know can’t be taught in the classroom!” and “You can test what I know in an examination room!” – and they are right; you can’t teach 10+ years of application-specific knowledge, and you can’t test people on the specifics of their application.  Training and certification have to be broad-based.

Type Two Analysts

There is no doubt that vibration analysis is difficult to master. There is a lot to learn and there are a lot of difficult concepts to understand.

It is also true that some managers do not believe in the principles of predictive maintenance. Unless the bearing is about to fail they won’t take any corrective action. Fault conditions such as unbalance, misalignment, and other conditions (that will ultimately damage the machine) will be ignored unless the vibration levels are very high.

As a result of these challenges, vibration analysts can take a very simple approach to vibration analysis. Rather than focusing on detecting bearing faults at the earliest stage, and rather than making the effort to determine whether the machine has a fault condition that will result in premature failure of the bearing (or other component), instead they just try to stay one step ahead of the guy who listens to the bearing with a screwdriver. It may seem a little bit harsh describing vibration analysts this way, but there are a lot of vibration programs that take this approach.

It is true to say that if you can monitor a machine and see gross changes in the vibration, whether it is a result of the bearing failure or not, and the machine can be stopped before it catastrophically fails, then that vibration program is providing a valuable service – avoiding the downtime, and the resulting secondary damage, and the failure that may lead to injury or environmental harm, is a worthy goal. But you cannot argue that such a program is gaining the greatest benefits from the vibration monitoring technology.

These vibration analysts may evolve into Type One vibration analysts given enough time. They will begin to recognize some of the tell-tale signs that a problem is developing, and if they had taken the time to find out what was actually wrong with the machine during earlier failures, then that experience may be put to good use.  But sadly, most type two vibration analysts will not look at the bearing that is removed from the machine to determine how severe the damage was. They will not ask the craftspeople what was actually wrong with the motor when it was removed from service. Type Two vibration analysts will just continue monitoring the vibration, looking for the next problem.

In defense of the Type Two vibration analysts, they are often not given the opportunity to do a better job.  They need support from management so that they know that their contribution is valued. They should be working in an environment where condition based maintenance and reliability improvement is given a high priority.

Likewise, they should be given adequate training. A two or three-day course learning how to operate their vibration analyzer and software does not count. While that training is useful, it does not help a person become a better vibration analyst. The vibration analyst needs to fully understand why the settings of the vibration analyzer are important, what the failure modes are of the machine, how the vibration changes with the various failure modes, and more. Of course, this form of training can only be taken if approved by management.

Type Three Analysts

It is the author’s opinion that the Type Three analysts are the best analysts of all.  They have sought to gain as much experience as possible, and they have sought the best training possible.

Type Three analysts will seek out the tough jobs and will utilize all the tools (spectra in acceleration and velocity, linear and logarithmic spectra, high frequency data (enveloping, PeakVue, Shock Pulse, etc.), phase and time waveform data) and they will leave no stone unturned when trying to diagnose a fault.

Type Three analysts will also give the “customer” the clearest and simplest diagnosis and recommendation, whether the customer is a supervisor, plant or production manager, or a literal consulting client.

In order to be a Type Three analyst you need the right type of training.  Focusing on detecting bearing faults alone, or focusing on “cookie-cutter” patterns that you see on wall charts is not enough.   Type Three analysts:

1. Understand the machine failure modes.
2. Understand how the vibration of the machine will change as the fault develops.
3. Understand how and where to measure the vibration in order to get the best possible data.
4. Understand the analyzer and the analysis options well enough to be sure that they are capturing and seeing all the details revealed by the machine.
5. Understand why the vibration changes the way it does so they can “reverse-engineer” what is going on inside the machine.

Yes, the wall charts can act as great reminders of what the patterns may mean, but the moment you rely on those wall-chart patterns you will be guessing what is wrong with the machine.

Let’s explore that a little further.

Machine vibration and failure modes

Machines vibrate in three dimensions.  The nature of the fault and the nature of the machine dictate how much it vibrates in the vertical, horizontal and axial directions.  But the nature of the fault (i.e., the failure mode) also dictates whether a spectrum can provide all of the information that will allow the analyst to diagnose the fault.  When impacts, rubs, cavitation, turbulence and certain other types of events occur, the time waveform will reveal key information.  And with a large number of fault conditions where there are directional driving forces, phase measurements provide key information.  And with yet a different group of fault conditions, very high frequency data will reveal the earliest and most comprehensive details of the fault condition.

A Type Three analyst knows all of this and structures his or her work accordingly.

Measurement location

At each point on the machine, the vibration depends on a great many things: what type of bearing you are testing (for example sleeve bearing, deep groove bearing, thrust bearing, or floating bearing), how the vibration is transmitted from the rotating elements to the measurement location, how the sensor is mounted, and the type of sensor.  The data is only as good as the data collected.

Type Three analysts ensure their data is collected properly so that they have confidence in their data, and they know they have all of the data (with the exception of phase data and other special test data) they need from their routine measurements, whether they collect it or not.  (They also understand that a great deal can be gained just by standing next to the machine and using their sense of hearing, smell, sight and, when safe, touch to more fully understand the condition of the machine.)

Analyzer settings

One way to look at vibration analysis is that the machine is telling a story.  Sometimes the story is simple: not much changes from moment to moment.  Sometimes the story is boring: it won’t have changed much from last time.  But sometimes it is neither boring nor simple: it can take the machine longer to tell the story (because the vibration changes from one rotation to the next), the story can involve noise and/or short duration events (e.g., impacts) and/or high frequency (or low frequency) sounds, and more.  Some machines always tell an interesting story, and most tell an interesting story when they have a problem.

When you take a measurement, you have to be prepared for the interesting story.  You have to collect the entire story so that you can perform analysis on data that is both repeatable and comprehensive; the data must correctly reveal all the machine has to say.

Too many analysts use analyzer settings that are incapable of listening to the entire story. In an attempt to make the measurement as short as possible, and to make the analysis as easy as possible, they collect data that does not capture the entire story from the machine. Understanding the machine failure modes are to and understanding the analyzer settings ensures that the measurements taken capture all of the data necessary.