Manufacturing
IBM Maximo for manufacturing: OEE, reliability and the shutdown
Maximo for discrete and process manufacturing: OEE-driven configuration, reliability-centred maintenance in practice, planned shutdown discipline, and inventory tied to asset criticality.
Two numbers, one line
The availability figure maintenance owns, and the one production lives with
Both are calculated correctly. A maintenance system that produces only the first will not settle a Monday morning meeting.
What the maintenance function is accountable for
Was the equipment capable of running when it was asked to. Planned maintenance is excluded, because it was scheduled.
- Clock starts
- The asset is declared unavailable, usually when the work order is raised
- Clock stops
- The repair is complete and the fitter signs off
- Short stops
- Below the reporting threshold, so usually invisible
- Owned by
- The maintenance manager, defending a maintenance budget
What the production function lives with
Did the line make what it was scheduled to make, at rate, without intervention. Planned maintenance counts if it overran.
- Clock starts
- The line stopped producing, often several minutes earlier
- Clock stops
- The line is back at rate with good product, sometimes an hour after sign-off
- Short stops
- The largest single loss on many lines, and the hardest to attribute
- Owned by
- The plant manager, defending a production commitment
In conversation with Ivan
Ivan Milic
Chief Operating Officer · 20+ years in EAM and IBM Maximo
Ivan on why the shortest dropdown wins
- Where does a manufacturing configuration get decided?
-
At 02:14, by an operator with a supervisor waiting and a fitter on the way. Whatever the system asks for in the next thirty seconds is what the plant will believe about that stoppage forever. Sixty entries in the dropdown and the answer is whichever code is near the top and roughly plausible. Eight entries in shop-floor language and it is accurate.
- Why IBM Maximo on a plant that already has an MES and an ERP?
-
None of them wants to own the asset. The MES owns what was being made and when it stopped, the ERP owns money, the historian owns equipment behaviour. Not one will tell you the five-year maintenance cost of a specific pump, what was replaced in it, or which spare is on the shelf for it. Maximo is the one place a stop, the fix and the part sit on one record.
- What do you find wrong most often?
-
The PM library, and it is nobody’s fault. Routines are inherited from the original equipment manuals at installation and nothing is ever removed, because removing one feels like accepting risk. Twenty years later the crews know which to rush and which to sign off, and the completion data is fiction. Pruning it means telling people that work they have done for years was not justified by criticality.
- What will you not claim?
-
That Maximo raises OEE. It makes availability honest and shows where the losses sit, which is valuable and is a different thing. Recovering them means changing a maintenance regime, a changeover process or a piece of equipment, and those are plant decisions with their own owners and costs.
Ivan owns how environments are built, how upgrades are run, and how MaxIron managed operations work day to day.
The taxonomy
The reason code list decides what the plant can later prove
60
Reason codes in a list we typically inherit
8
Reason codes an operator distinguishes with the line down
A long list collapses into one generic code carrying most of the stops, so the reliability analysis that justified the platform cannot be run from the data the platform collected.
Illustrative of the range across discrete and process plants.
The mechanism
Five moves, starting on the shop floor
Stage one happens with operators present: everything downstream depends on data they have to enter accurately.
- 01
Design the downtime taxonomy with operators
A short unambiguous list of reason codes, tested on real stoppages before it is configured, with the sixty-first code resisted.
- 02
Rebuild the PM library against criticality
Every routine justifies itself against criticality and failure mode, and duration is matched to the access window that exists. A four-hour job raised into a forty-minute changeover teaches the plant that the schedule is advisory.
- 03
Draw the MES, historian and ERP boundaries
Production owns what was being made, the historian owns equipment behaviour, the ERP owns money, Maximo owns the asset, the work, the parts and the maintenance cost. Each direction is agreed once and built once, under Maximo integrations.
- 04
Re-cut the stores against downtime cost
Critical-spares levels set from what the line costs while it waits, with vendor-managed and consignment arrangements where the supplier carries the risk more cheaply.
- 05
Plan the shutdown in the system, then run it
Work packages, resource loading, contractor management and critical path in Maximo, with lessons learned recorded against the assets.
The test is the following January: a shutdown scope assembled from the record, and an availability figure production and maintenance both recognise.
Who keeps score
Six readers of a plant maintenance record
On most plants the strictest reader is the insurer. They want two things: the examination happened, and the defect was closed.
| Ref | Scorekeeper | Asks for | Will not accept |
|---|---|---|---|
| MF1 | Your insurer and their engineer surveyors | Evidence that statutory and policy-required examinations are on schedule and that identified defects have been closed. | An open defect with no target date. To an underwriter that is an accepted risk rather than a managed one. |
| MF2 | Health and Safety Executive with the environmental regulator, as the COMAH competent authority | Evidence that major accident prevention measures are maintained, tested and functioning, with findings acted upon. | A deferred safety-critical routine with no recorded risk assessment behind the deferral. |
| MF3 | Health and Safety Executive, PUWER 1998 and LOLER 1998 | Equipment maintained in efficient working order, lifting equipment thoroughly examined at interval, defects acted on. | An examination report filed centrally with no link to the item examined. If it is not findable from the asset it is not usable in an investigation. |
| MF4 | Health and Safety Executive, Pressure Systems Safety Regulations 2000 | A written scheme of examination, examinations carried out to it, and remedial action inside the required period. | An overdue examination sitting as an open work order with no escalation. The obligation does not queue. |
| MF5 | Your certification body, ISO 55001 | A line of sight from organisational objectives through asset management policy to the work executed, with performance evaluated. | A policy document with no trace into the maintenance regime. Certification looks for the connection. |
| MF6 | Not a regulator: ISO 14224, where a plant benchmarks reliability against comparable operations | Consistent equipment class and failure mode coding, captured at the point the work is done. | Free-text failure descriptions that cannot be aggregated across a fleet of similar equipment. |
Codes MF1 to MF6 are stable, so an auditor or bid writer can cite one by reference.
Boundaries
Three conditions we state at the start of a plant engagement
Cutting the PM library needs engineering authority
Retiring routines means telling crews that work they have done for years falls outside what criticality justifies. Without a mandate behind it the library grows back inside two years.
Stores accuracy is a habit built at the point of use
We design item masters, storerooms and cycle-count regimes correctly. Whether the balance is right at three in the morning depends on issues being recorded where the part physically moves.
Maximo in manufacturing, questions we are asked
- How many downtime reason codes should we have?
- Fewer than the list you have. The constraint is not analytical richness, it is what an operator selects accurately at two in the morning with the line down and a supervisor waiting. A short unambiguous list produces data a reliability engineer can trend; a long one produces one default code carrying most of the stops.
- Where does the OEE number come from?
- Availability comes from the downtime record, performance and quality normally from the MES or the production system. Maximo owns the maintenance side of that equation and the reasons behind the stops. We draw the boundary so the two systems reconcile automatically rather than being compared by hand at month end.
- How does Maximo handle a major turnaround or shutdown?
- Work packages, resource loading, contractor management and critical path are held in Maximo rather than in a parallel spreadsheet, and lessons learned are recorded against the assets. The next shutdown scope is then assembled from the record instead of from recollection.
Bring your reason code list and one PM schedule.
The reason code list shows whether the availability data can be trusted. The PM schedule shows whether the crews believe in it. We go through both and name the first change.
Bring this to the first call
- The downtime reason code list exactly as operators see it
- The PM schedule for one line, with completion rates if you have them
- Last month availability from maintenance and from production
- The critical spares list for one asset class, with the last stock check date