A spectrum tells you how a machine is moving. It does not tell you why. The distance between those two statements is where most machinery investigations succeed or fail.
01
What a spectrum actually establishes
Synchronous energy is consistent with unbalance, but it is equally consistent with a bowed rotor, a shifted coupling, thermal growth, or a soft foundation. Each of those produces the same line on a plot for a different physical reason.
The investigation only advances when the analyst asks which mechanisms can be excluded by data that already exists: run up and coast down behavior, phase relationships, load dependence, and temperature history.
A dominant frequency narrows the field of candidate mechanisms. It does not select one.
02
Operating context carries as much weight as the signal
Machines are diagnosed in the condition they were operated, not the condition assumed by a textbook example. Process upsets, control action, lube oil condition, and alignment history all shape the response the sensor records.
- Was the change gradual or step wise
- Did the change track load, speed, or temperature
- What maintenance preceded the change
- Do process records corroborate the machine record
03
Reaching a conclusion that survives challenge
A supportable machinery conclusion identifies the mechanism, explains why the alternatives are inconsistent with the evidence, and states the limits of the data. That structure is what allows an operator to act and a reviewer to verify.
Continue Reading
- Pressure Vessel Failures: What the Evidence Actually Shows
Pressure Systems
- Hidden CAPEX in Industrial Asset Acquisitions
Asset Advisory
- Arc Fault in Switchgear: Reading the Electrical Record
Electrical & Power