Predictive maintenance tools
Predictive maintenance tools divide into sensing hardware, analysis software and managed services. Buying the wrong layer for your constraint is the most common and most expensive mistake in this category.
| Topic | Key point |
|---|---|
| Three layers, sold as one category | Layer What you are buying Buy it when Sensing hardware Sensors, gateways, mounting, wireless infrastructure You have analysts but no data from the ass |
| Diagnose your own constraint first | Do you have condition data on your critical assets? |
| Requirements that apply whichever you buy | A healthy baseline. Everything is a comparison against a reference. |
| Questions that separate vendors | Which failure modes does this detect on our equipment class — and which does it miss? |
Three layers, sold as one category
| Layer | What you are buying | Buy it when |
|---|---|---|
| Sensing hardware | Sensors, gateways, mounting, wireless infrastructure | You have analysts but no data from the assets that matter |
| Analysis software | Trending, diagnostics, alerting, model management | You have data and people but no system to work in |
| Managed service | Hardware, analysis and the analysts | You have neither the data nor anyone qualified to interpret it |
Most disappointment traces to buying one layer while the constraint sat in another — most often buying software when the actual gap was competent interpretation.
Diagnose your own constraint first
- Do you have condition data on your critical assets? If no, you have a sensing problem.
- If yes, does anyone look at it? If no, you have a competence or capacity problem, and more software will not fix it.
- If someone looks at it, does anything change as a result? If no, you have an organisational problem — no route from alert to work order — and that is not for sale from any vendor.
Only if all three answers are yes is the constraint genuinely a tooling limitation.
Requirements that apply whichever you buy
- A healthy baseline. Everything is a comparison against a reference. Fitted to a degraded machine, any tool encodes the fault as normal.
- Operating context with every reading. Speed, load and process state, or variable-duty assets generate false alarms until nobody trusts the system.
- Measurement interval shorter than the P-F window. Sampling less often than the fault develops detects failures after they happen.
- A named owner for alerts, with authority to change the schedule.
- Qualified interpretation. ISO 18436 defines the categories; a no-code tool lowers the barrier to deployment, not the need for judgement.
Questions that separate vendors
- Which failure modes does this detect on our equipment class — and which does it miss?
- What lead time is realistic for those modes?
- What does the system need from us: sensors, context data, tag mapping, people?
- What happens on a variable-duty machine?
- Who interprets alerts, and what qualification do they hold?
- What does a false alarm cost us, and what is the observed rate?
The last question is the one most vendors are least prepared for, and the answer tells you the most.
Frequently asked questions
What tools are used for predictive maintenance?
They fall into three layers: sensing hardware such as wireless vibration and temperature sensors; analysis software for trending, diagnostics and alerting; and managed services that supply hardware, analysis and the analysts. Choosing the wrong layer for your actual constraint is the common expensive mistake.
How do I choose a predictive maintenance tool?
Diagnose your constraint before shortlisting. Do you have condition data on critical assets? If yes, does anyone look at it? If yes, does anything change as a result? A gap at the second or third question is a competence or organisational problem that no tool purchase resolves.
Do I need sensors on everything?
No, and instrumenting broadly is a common failure mode. Every monitored asset consumes analyst attention, so monitoring more than you can interpret produces an alarm queue rather than a programme. Criticality analysis comes first.
What is the most overlooked requirement?
A named owner for alerts with authority to change the maintenance schedule. Detection without a route to action produces reports rather than avoided failures, and it is the most common reason programmes are abandoned.
Related guides
Types of predictive maintenance
Seven predictive-maintenance techniques, what physical fault each one actually detects, the assets they suit, the warning time they realistically give, and where each one is blind. Technique is chosen by failure mode, not by asset type.
Setting up a condition monitoring programme
ISO 17359 sets out the general flow for condition monitoring: audit the assets, select measurements, establish a baseline, set alert criteria, then diagnose, prognose and act. Most failed programmes skip the criticality audit and start at sensor selection.
How to choose predictive maintenance software
Choose predictive maintenance software by starting from your critical assets and data, not the feature list: match the approach (sensor vs analytics) to those assets, check it integrates with your CMMS, insist on a clear pilot with a measurable target, and weigh total cost against failure cost.
Software that helps
Augury
Machine health monitoring for rotating equipment using vibration and AI.
Tractian
Sensor-based machine health monitoring with a fast retrofit setup.
Waites
Wireless condition monitoring and predictive maintenance.
Emerson AMS
Asset management and condition monitoring for process plants.