Industries / Industrial equipment and services / Industrial services
Vibration analysis
Vibration analysis measures how rotating machines such as motors, pumps, fans and gearboxes shake, and reads the pattern to find a fault early enough to plan the repair.
1939
T. C. Rathbone, chief engineer of the turbine and machinery division of the Fidelity and Casualty Company of New York, an insurer, published "Vibration Tolerance", curves that tie a machine's vibration to its condition.
24
Milestones achieved across 6 domains.
6
To watch for: announced, not yet confirmed outside.
12
Open deliverables. Nobody has shipped them.
In one minute
Where it started
Machines were first judged against limits drawn from experience. In 1939 T. C. Rathbone, an engineer at an insurance company, published curves that tied a level of vibration to a machine's condition, and from 1943 piezoelectric accelerometers turned shaking into an electrical signal. The fast Fourier transform of 1965, and the analyzers built on it, let analysts split a reading into its frequencies and name the part that was failing.
Where it stands
Technicians walk routes with handheld collectors that store their readings, and the most important machines carry fixed or wireless sensors that report to analysts elsewhere. Analysts are certified to a shared body of knowledge, and balancing and laser alignment correct two of the commonest causes of vibration. Outside tests show the limits: in a 2012 blind test on a wind turbine gearbox damaged in service, the best of 16 teams found five of its seven damage sites.
What is open
Much of what the trade relies on has never been measured in public: how far readings scatter between technicians, how well sensors hold their calibration after years on a machine, what a wireless sensor that wakes once a day misses, and how often the textbook fault rules are right. Wide-band sensor chips, robots on plant rounds and AI models that run the analysis make these tests, and some of the work itself, possible with tools sold today.
Domains
AchievedWatch forOpen
Domain
Progress
Last achieved
Watch for
Open
Sensors
Sensors turn a machine's shaking into an electrical signal; the usual one is an accelerometer, which measures how fast a surface speeds up and slows down as it moves.
4 · 1 · 2
Last achievedChip sensors tried on bearings2024
Watch forChips to replace piezo sensors
2
openCollecting data
Collecting data is taking vibration readings from the same points on each machine at set intervals, and getting them into a form an analyst can compare over time.
4 · 1 · 2
Last achievedVideo checked against sensors2015
Watch forRobot rounds in cement mills
2
openReading faults
Reading faults is working out, from a vibration recording, which part of a machine is failing and how badly.
4 · 1 · 2
Last achievedBlind test on real damage2012
Watch forLanguage models over diagnostic tools
2
openWireless monitoring
Wireless monitoring leaves sensors on the machine that send their readings by radio or over a network, so that nobody has to walk to the machine to measure it.
4 · 1 · 2
Last achievedRadio measured inside plants2017
Watch forTen years on one battery
2
openBalancing and alignment
Balancing adds or removes weight on a rotor so its mass is spread evenly around its axis, and alignment puts two coupled shafts on one line.
4 · 1 · 2
Last achievedMisalignment costs bearings, not power2001
Watch forBalancing with no cables
2
openReporting
Reporting is judging whether a reading is acceptable, writing down what is wrong and how urgent it is, and showing the people who decide on repairs what the work is worth.
4 · 1 · 2
Last achievedWhat maintenance pays, surveyed2020
Watch forAI reports on computed evidence
2
open
Open in vibration analysis
3 of 12 open deliverables
Open · Sensors
Sensors re-checked after years
A public table of how much permanently mounted accelerometers drift in real plants. Collect sensors taken off running machines after five or more years in service, recalibrate each against a reference accelerometer, and record the change in sensitivity next to the sensor's model, age, mounting and the temperatures it saw.
Built with what exists
A back-to-back comparison calibrator, a small shaker that drives the sensor under test and a reference sensor together; a reference accelerometer with a current certificate from an accredited calibration laboratory; a data acquisition card; Python with pandas.
Open · Collecting data
Measure how much readings vary
A published measurement-system study for route vibration readings. Factories call this a gauge repeatability and reproducibility (R&R) study: it splits the scatter in repeated measurements into the part due to the instrument, the part due to who takes the reading and, here, the part due to how the sensor is fixed. Five or more technicians each measure ten marked points on a small, guarded, running test rig three times, with a magnet on bare paint, a magnet on a flat pad and a screwed-in stud.
Built with what exists
A bench rig of motor, shaft and bearings, such as a SpectraQuest Machinery Fault Simulator or a home-built one; an industrial accelerometer with magnet and stud mounts, such as PCB Piezotronics' 603C01; a data acquisition board for such sensors, such as Digilent's MCC 172 on a Raspberry Pi; Python with SciPy and statsmodels for the analysis of variance.
Open · Reading faults
Score the textbook fault rules
A public scorecard for the rules analysts are taught for reading a spectrum, each written as code and run on recordings whose fault is known. Fluke's guide gives examples: unbalance shows as a peak at once per turn (1X) growing in the radial directions but not the axial one; misalignment as raised 1X along the shaft and twice per turn (2X) across it; looseness as many harmonics of 1X. Report each rule's hit rate and false-alarm rate, and whether a threshold tuned on one data set holds on another.
Built with what exists
MaFaulDa, a free database of 1,951 recordings from a SpectraQuest fault simulator covering normal running, imbalance, horizontal and vertical misalignment and bearing faults; NREL's gearbox recordings; Python with NumPy and SciPy, with computed order tracking where speed varies.
Where it sits
Broader