MAS suite · Monitor

Four records carry a signal from the historian to a work order

IBM Maximo Monitor in production: a tag binding, a baseline band, an anomaly record and a work request in Manage. Each has an owner and one way of breaking. All four are taken apart below.

Two engineers on a walkway at dusk reviewing live plant condition signal, representing IBM Maximo Monitor

The operating envelope

What the ingest and alerting path actually does

Under 5 min
Signal outside its band to an open anomaly

A five-minute historian poll plus the detection window. Anything shorter is work at the source.

1 s to 15 min
Sample rate range, set per tag

The rate the source sustains under duty, fixed at binding rather than on a dashboard.

Tens of thousands
Historian points one asset class offers

We bind the subset that answers a maintenance question, not the whole point list.

How condition reaches the plan today

Operational tagging and asset numbering were built by different teams

Both conventions are internally consistent. Joining them was never anybody's job.

The signal exists and nobody owns it

The historian has held the trend for months and an operator noticed it. The planner who could have scheduled the work never heard, because no route runs between the two screens.

Tag names no asset record answers to

OT tag naming and the Manage numbering convention grew up years apart under different owners, and neither joins to the other without a person who knows both.

Condition travels by conversation

The evidence arrives as a screenshot in an email. By the time it is a work order the trend is attached to nothing, so the argument runs again at the next planning review.

What gets paid for is the second diagnosis, the deferred job that became a breakdown, and evidence nobody could find when the plan was challenged.

The teardown

Four records, in the order the data moves

Equipment is illustrative. The fields are the ones we populate, and every record names an owner.

  1. 01

    The tag binding record

    Source path: the historian identifier, with unit and sample rate. Bound to: one asset record in Manage, at asset level. Owner: a named reliability engineer, reviewed whenever tag or asset changes. Breaks when it binds to a parent location: triage becomes a walk round the station.

  2. 02

    The baseline band

    Normal range: the band for this class under its actual duty. Exceptions: seasonal behaviour and start-up transients that would alert every spring. Agreed by: the supervisor and engineer who act on it. Breaks when operations never signed it: the first false alert spends the credibility the second needed.

  3. 03

    The anomaly record

    Asset: the bound asset, with criticality, parent, PM schedule and last work orders beside it. Window: the departure range, trend retained. Rule: which rule fired against which band, so it can be questioned. State: open, reviewed or dismissed, with a person on each transition.

  4. 04

    The work request in Manage

    Raised against: the asset the signal was bound to, so findings return to the evidence. Justification: the anomaly and its trend window, attached rather than pasted. Requested by: whoever reviewed it. Monitor does not appear in that field. Breaks when evidence is typed into a description and the deferral is later argued from memory.

The write path

Monitor states the condition, a named person decides the work

One trend on a raw water pump, from band crossing to work request. Anything consuming crew time passes through a person.

Signal, person, record

Signal
Drive-end vibration on raw water pump 2 crosses the band agreed at baselining, nine days into a rising trend.
MAS Monitor
Anomaly opened against the bound asset, trend window retained, detection rule recorded.
Reliability engineer
Checks the operating log for a duty change, finds none, and confirms the departure is real.
Reliability engineer
Rejects the call-out route and requests a bearing inspection at the next planned window.
Maximo Manage
Work request written against pump 2, anomaly attached, requester and rejected option recorded.

The prerequisite, shown

The same trend read on two asset hierarchies

One asset class, corrected before the tags were bound. Nothing about the OT estate changed.

Anomaly A-4417, raw water pump 2

4 of 4 fields changed

Field Before After Set by
bound_to Pumping station location, covering four pumps Pump 2, at asset level Reliability lead, at the binding review
triage Walk the station to find which unit moved One trend and one operating log, reviewed at a desk Named triage owner per shift
work_raised Work order against the station Work request against pump 2, trend attached Reliability engineer who reviewed it
finding_lands On a record that cannot say which asset degraded On pump 2, beside the signal that raised it Close-out in Manage

Remediation is narrow: correct one class, bind at asset level, prove the loop, widen after. We quantify that in the readiness assessment rather than in month four.

Boundaries

Four limits on anomaly detection

It does not control anything

SCADA runs the plant and the OT team owns it. Monitor reads operating signal and issues no control action.

It cannot invent an asset record

An anomaly lands only where the asset exists in Manage under the right parent. Data work into Maximo closes that gap.

It states condition rather than probability

Monitor reports against the agreed band. Failure probability belongs to Predict, estate-wide ranking to Health.

It needs a triage owner per shift

With nobody accountable for reviewing an anomaly within a shift, the output is a queue. We design that role into the rollout.

IBM Maximo Monitor, frequently asked questions

What does IBM Maximo Monitor do?
Monitor ingests operating signal from SCADA, historians and edge gateways, scores it against the baseline band agreed for that asset class, and opens an anomaly against the bound asset record in MAS. A named reviewer decides whether work follows.
How is this different from a SCADA screen or a BI report?
SCADA runs real-time control in the OT estate. A BI report describes what already happened. Monitor scores operating signal against the asset, criticality and work history held in Manage, which is what lets a planner act on it.
Which asset class should be first?
The class where the signal is already reliable, the asset records are clean enough to bind a tag at asset level, and unplanned downtime costs more than seeing it sooner. A well-instrumented population of thirty beats the whole estate.
Does Monitor raise work orders automatically?
It can write to Manage, and the pattern we deliver keeps a named person in the path for anything that consumes crew time. That boundary is what makes the work request defensible at the next planning review.
Where does the OT cybersecurity boundary fit?
It is the binding constraint on every engagement. We design across the boundary with the OT team, normally through an edge or DMZ component. The zone by zone detail sits on IoT and OT connectivity, and sector constraints on Monitor for utility networks.

Bring one asset class and its tag list.

Name the equipment costing you unplanned downtime. We test whether those tags bind at asset level today, what the OT integration involves, and whether Monitor earns its licence on that class.

Bring this to the first call

  • The asset class, and roughly how many units you run
  • A tag list or historian point export for a handful of them
  • The Manage numbering convention, and where it duplicates
  • Who owns the OT boundary, and what they have declined
  • The last unplanned outage on that equipment, and what it cost