
Position
Place the sensor on the appropriate fixed housing / bearing area.

Solutions
Concrete outcomes, not abstractions: continuous condition monitoring, early anomaly detection, and explainable alerts that tell a maintenance team what changed, where, and what to do about it — before an unplanned stop.
What MachBeat does
Detect the change, explain it, and turn it into planned work. Each capability builds on the last.
Continuous vibration monitoring on rotating assets, with operating-state context. Temperature can be included when a compatible measurement source is connected.
Recent readings are compared with configured thresholds and available statistical baselines. Deviations appear as investigative cues, not guaranteed predictions.
Alerts include the available threshold and signal evidence with suggested checks, giving reliability teams a clear starting point for investigation.
Link alerts to work orders, assign follow-up, track parts, and record completed work in one maintenance workflow.
Record meter readings and review energy use alongside machine condition to support operational investigations.
The whole loop exists to convert a catastrophic-failure event into a planned inspection window — detect, investigate, plan, maintain.
Maintenance action
The goal is planned maintenance — not failure theater.
Condition departs from baseline.
Review the evidence on the asset.
Schedule an inspection window.
Correct on schedule.
Deployment
Deployment is physical and practical. Timing depends on the asset and site — we don’t invent “5-minute” claims.

Place the sensor on the appropriate fixed housing / bearing area.

Secure the device so vibration transfers into the sensor.

Configure connectivity to the MachBeat platform.

Confirm live data, then commission a baseline before trusting “normal.”
Healthy-data commissioning
Mounting the sensor is not the same as learning the machine. On the machine page the team declares health, captures a window, reviews quality, and approves the baseline. The steps below are that flow.
Type, criticality, rated power, rated speed, sensor location, and foundation. These fields decide which engineering limits are even eligible.
Vibration, temperature, RPM, energy, and load are each connected or not. Operating-condition detection stays partial if speed exists and load does not. Frequency diagnosis stays off if sampling quality is too low. Energy detection stays off if power is not connected.
Inspected, no known fault, no abnormal noise or heat, maintenance done, valid operating state. “Not sure” does not create a trusted baseline. It stays in reference mode.
A useful window is often 5–14 days, with a minimum of stable running and more than one speed or load. Five days is a possible minimum under good data, not a promise. Startup shocks, stops, and bad samples are left out.
Stability, how many operating states were seen, sensor quality, maintenance during the capture, and how many outliers were removed. If high load was never seen, the recommendation is to keep capturing.
Approve, reject and recapture, or use a temporary reference baseline. After later maintenance the old baseline is kept. It is not silently replaced.
The four methods and the alert policy are on Platform.
After the alert
A vibration cue only pays if the part, the person, and the claim are ready. These four lines sit on the same machine as the sensor. Parts used on a work order are recorded today. Forecasting, dealer replenishment, price, and warranty claims are scoped on a pilot — they are not a live global catalog engine.
Demand and stock for the last mile: the bearing, seal, or belt the next job will need. The work order already records parts used. A forecast of what to hold is agreed per plant, from that history, not from a generic optimizer.
Parts on the job · forecast scopedReplenishment across a dealer or branch network when sales and issues say a part is running out. Alignment uses the stock and usage the customer already keeps. MachBeat does not ingest a live point-of-sale feed unless that feed is in the pilot scope.
Scoped with the networkA price that follows the part, the machine, and how soon the stop will cost more than the spare. Lists and overrides live with the customer’s catalog. MachBeat does not publish an autonomous price for every SKU.
Catalog rules on requestClaim, cost recovery, depot repair, and the technician’s job on one asset. The work order, assignee, and result are in the platform. Automated claim adjudication and depot routing are a later workflow, not a shipped claims desk.
Work order live · claims scopedWhere it sits
A wide cloud program covers a whole catalog. MachBeat starts on one rotating machine and keeps the evidence next to the job.
| Dimension | A wide cloud program | MachBeat |
|---|---|---|
| First asset | Broad plant, ERP, and catalog programs | Motors, pumps, fans, compressors, gearboxes |
| First day | Often a long baseline and integration project | A mounted sensor, a reading, and a work order on that asset |
| Hardware | Often a lease or a large installed base | The plant buys the node. The monthly fee is the platform |
| What the alert says | A score, then a separate system for the job | The signal evidence and the work order on the same machine |
| Parts and warranty | Full planning, pricing, and claims suites | Parts on the job today. Planning, price, and claims scoped per pilot |
The left column describes a category of program. It does not name a vendor.
Engineering extension
For selected hydraulic and circulating-oil applications, MachBeat is assessing how validated fluid sensors could surface water, particle cleanliness, viscosity or lubricant-delivery changes alongside machine vibration.
The current V1 node does not measure oil properties. Sensor selection, safe sampling point, interface, oil-specific reference limits and lab correlation must be agreed before a pilot can make condition claims.

The best pilot asset is the one whose failure hurts most. Install MachBeat there and see the loop run end to end.
Start with one machine · no obligation