About this example
The plant, equipment and numbers below are invented for illustration. They are not drawn from a real facility, and no real results are claimed. The aim is to show the sequence of events that sensor-based maintenance typically follows, so the steps are easy to picture.
The setting
Imagine a mid-sized food ingredient plant, which we will call the Riverside plant. A transfer pump moves liquid from a mixing tank to a holding tank, and the line cannot run without it. The pump is driven by an electric motor through a coupling. The maintenance team has fitted it with an accelerometer on each bearing housing, a temperature probe on the motor-side bearing, and a current reading from the motor starter. Readings are logged automatically and reviewed weekly.
When the sensors were installed, the pump was in good condition. The team recorded a baseline: the vibration level, bearing temperature and motor current at normal load. All values below are hypothetical and described in relative terms only.
Week by week
Weeks 1 to 8: steady. Readings stay close to the baseline. Small daily swings follow the load and the ambient temperature. No action is needed, and that is the expected outcome for most of a machine’s life.
Week 9: a faint change. The weekly review shows that vibration at one bearing has crept up slightly, mainly in a high-frequency band. The overall level is still well below the alarm limit. Because the change is small, the reviewer simply flags it for a closer look next week. Looking at one reading alone would not have been convincing, but a trend line makes the shift visible.
Weeks 10 to 12: the trend continues. The high-frequency component keeps rising. The bearing temperature, at the same load, is now a little higher than the baseline. Motor current has not changed noticeably. Two independent signals, vibration and temperature, now point toward the same bearing. This pattern is consistent with early bearing wear, although it could also come from other causes such as lubricant loss or a mounting problem.
The decision
The reviewer raises the issue with the maintenance planner. They consider:
- How critical is the pump? The line cannot run without it, so an unplanned failure would stop production.
- Is there a spare? A spare bearing set is in stock, which is the sort of item a team usually identifies as a critical spare for this reason.
- When is the next convenient window? A planned production gap is coming in two weeks.
- What could be inspected without stopping? A technician takes a closer look at the lubrication condition and listens with an acoustic tool during a short check.
The check suggests the bearing is running dry of grease. The team cannot be certain whether damage has already started. They decide to add the repair to the planned gap rather than waiting for the alarm.
The repair
During the gap, a trained technician follows the site’s lockout and tagout style procedure and the pump manufacturer’s manual. They isolate and de-energize the equipment and check that it is safe before opening it. They then inspect the bearing, replace it, check shaft alignment and relubricate to the manufacturer’s guidance. The old bearing shows early surface wear, in this story, and the cause is recorded as insufficient lubrication.
After restart, the readings return close to the baseline. The team updates the work order with what was found, which makes it easier to see later whether similar pumps share the problem and whether the grease schedule needs review.
What the story illustrates
- Trends beat snapshots. The fault was visible because readings were compared with a baseline over time.
- Two signals are better than one. Vibration and temperature agreeing made the warning more credible.
- A warning is not a diagnosis. The inspection was needed to confirm the likely cause.
- Timing is the main benefit. The repair happened in a planned gap, with parts on hand, rather than in the middle of production.
- Records close the loop. Writing down the cause helps prevent a repeat.
What this story cannot show
Real faults are messier. Some give days of warning rather than weeks. Some sensors are installed badly or fail themselves, and some alarms turn out to be false. A hypothetical example also hides the cost of setting up monitoring, training people to read the data and keeping the records organized. Whether the effort pays off depends on the plant, the equipment and how costly a failure would be, and no outcome can be promised.
If you wish to apply these ideas, start with the equipment whose failure would hurt most, gather a baseline while it is healthy, and agree in advance who reviews the readings and what happens when a warning appears. Always rely on the manufacturer’s guidance, qualified personnel and local regulations for any actual maintenance work. This article is general information, not engineering advice.