The key question: how long does a fault take to develop?
Between the first detectable sign and functional failure there is an interval. A route is valid if the interval between measurements is clearly shorter than that time. If the fault can progress between one visit and the next, the route is not enough.
Route-based monitoring
- A technician walks the machines with a portable collector.
- It covers many machines with few instruments.
- It depends on the repeatability of the point, the mounting and the load conditions.
- It allows trending and diagnosis on machines of medium or low criticality.
Continuous monitoring
- Fixed sensors connected to a system that records, compares with alarms and protects.
- It captures transient events, startups, shutdowns and load changes.
- It can execute automatic protective trips.
- It requires greater investment, cabling and maintenance of the system itself.
How to decide
- Criticality of the machine for the process and for safety.
- Expected rate of degradation of the relevant failure modes.
- Conditions that make the route difficult: hazardous access, remote locations, variable operation.
- Cost of a failure versus the cost of the instrumentation.
A mixed scheme
It is common to monitor critical machines continuously and the rest by route, and to use semi-permanent or wireless systems at intermediate points. Continuous monitoring does not eliminate analysis: it generates data, but someone must interpret them and act on the alarms.
How to interpret the result
The sign that the route is no longer enough is that a machine shows repeated faults between visits, or that changes are always detected late. In that case, it is advisable to shorten the interval, add semi-permanent sensors or move to continuous monitoring at the critical points. Conversely, if a continuous system generates alarms that nobody attends to, its value is lost.