Predictive maintenance that schedules the repair before the failure.
Predictive maintenance watches for early signs that a machine is wearing: rising vibration, a new sound signature, energy draw creeping up at the same load. Aqueduct reads vibration, acoustic and power data and raises a maintenance alert while there is still time to plan.
Wear shows up in the data before the breakdown.
What is watched
- Vibration
Accelerometers on motors, pumps, fans and compressors, trended against each machine's own normal.
- Sound signature
Acoustic sensors pick up bearing and belt noise that changes before failure.
- Energy drift
Power drawn at the same load, rising as efficiency falls.
An alert with time to act
- A rule on any metric
A maintenance rule fires on a threshold or trend, from the machine itself or a linked sensor.
- A planned work order
Routed to your CMMS so the repair fits a planned window.
- Runtime and health
Per-component health and runtime share, so wear is compared like for like.
What planned maintenance is worth.
savings from predictive maintenance over preventive maintenance alone.
Source: DOE FEMP O&M Guidefewer breakdowns in working predictive programs, in industry surveys cited by FEMP.
Source: DOE FEMP O&M GuideAn example
Illustrative: vibration on a chilled water pump rises steadily over three weeks while flow and speed hold steady. A maintenance alert opens a work order for the bearing in the next planned shutdown, and the trend returns to normal after the change.
Limits
Predictions need sensors on the machine. Equipment with no vibration, sound or power data can only be judged on faults.
Vibration and acoustic sensing is newly added to Aqueduct, so coverage is confirmed site by site. An alert points to wear, not a guaranteed failure date.
Related
Common questions
Preventive work runs on a calendar whether a machine needs it or not. aQ PredictIQ schedules work when vibration, sound or energy data shows wear, so service goes where it is needed. See condition-based maintenance.
Fault detection and diagnostics picks it up, names the failed component and routes the fix to your CMMS. Sudden failures with no prior trend still happen, so the two work together. See fault detection and diagnostics.
Usually for vibration and sound, which most machines do not report today. Power data often comes from existing meters. Sensors go first on critical rotating equipment whose failure stops a line. See aQ PredictIQ.
aQ PredictIQ needs enough of each machine's own history to know its normal, usually a few weeks of running across its regular loads. Threshold rules work from day one. See equipment uptime.
See Aqueduct on your own facility.
Bring a site, a utility bill or a question. We will show you where the loss is.