For the highest confidence level possible while doing predictive vibration analysis work, location and placement of your sensor are crucially important to you. However, sensor placement is sometimes a trade-off between time, safety, and precision.
Most equipment is mounted with the shaft oriented horizontally to the ground and the analyst must decide whether to take a radial horizontal measurement, radial vertical measurement, a shaft axial measurement, or some combination of the three. If the analyst has the time, there is good reason to take all three positions. When it comes to analysis, an analyst can always choose not to view data he or she has, but can never choose to view data he or she did not collect.
Where should I place my sensor? The answer to this question should come from a focused consideration of what each possible position offers to the overall analysis effort. The fundamental consideration in placing your sensor is that the vibration signal from the component or components of interest should take the path of least resistance to the sensor. This should be coupled with thought given to how every likely defect frequency presents itself relative to direction.
Many fault diagnoses are determined, or the level of confidence enhanced, by relating one locations result to another. For example, a high 1× vibration in the horizontal direction on a direct drive center hung rotor could mean a lot of things, if taken by itself. But if this information is coupled with the additional hypothetical data below:
- The axial vibration is also high: then possible misalignment or bent shaft become likelier.
- The axial is low, but the vertical is 1-½ times as high as the horizontal: Now looseness is probable.
- The axial is low, and the vertical is ½ of the horizontal: probable unbalance.
- The axial and vertical are both very low relative to the horizontal: In fact, the horizontal is 75 to 100 times the axial and vertical! Almost unquestionably there is a resonance problem.
In all of the above scenarios a phase and visual inspection are required to deliver a truly confident diagnosis, but the directional aspect of data analysis is clearly shown.
Trending puts still further demands on the analyst by requiring each successive measurement to be taken in exactly the same place as before, to the extent possible. For parameters to be reliably trendable, operating states and data collection states need to be the same from measurement to measurement.
A real danger for the analyst using mobile data collectors lies in the monotonous repetition of placing a sensor hundreds of times in a day of data gathering, and becoming careless when making that next placement. If feasible, routes should be kept small enough to help avoid the human factor. It’s too easy to day dream while collecting data. A route under 200 points helps keep the monotony away. The analyst should focus on both safety and data collection precision throughout each route. Happy defect hunting!