Downtime tracking: the time model, the codes and the log that make the hours add up

The time model

Downtime adds up only if calendar time is divided the same way every time, and the model in the figure is the usual one. Calendar time splits into scheduled downtime, when the plant was not meant to run, and available time, when it was. Available time splits into unscheduled downtime, when it was meant to run and did not, and operating time, when it ran. Unscheduled downtime is coded by cause, and operating time carries the rate and quality losses that OEE reads.

The one argument worth having before any event is recorded is what counts as scheduled. A weekly PM window is scheduled. A shutdown is scheduled. No feed because the mine did not deliver is a process cause, not scheduled, unless the site decides otherwise and writes it down. The decision is recorded in the code list, and the measure is only comparable with itself from that date.

The codes

Codes are the difference between a log and a dataset. A three-level hierarchy, category, cause group and cause, with every level chosen from a fixed list, lets the hours be added up by any level across shifts and months. The lists are short, because a list of two hundred causes is a free-text field with extra steps. The hierarchy below is a starting shape, and the third level is where a site puts its own equipment and its own recurring causes.

Two conventions keep the codes honest. A breakdown is coded to the equipment that failed, by functional location, so the hours join the work order history and the failure codes on the asset. And a stop with more than one cause is split into events, one per cause, rather than coded to whichever came first. The template's Codes sheet holds the hierarchy for a site to adapt.

Capturing the event

The times come from the machine wherever they can. A running signal in the control system, or the equipment status in the historian, gives an exact start and end for every stop, and many systems can create the event automatically from it. The cause cannot come from the machine. It comes from the operator or supervisor who was there, chosen from the list, with a sentence of description beside it, during the same shift, on the HMI, a tablet or the log.

Maintenance adds the work order number to every breakdown event, which joins the downtime to the failure code, the hours and the cost on the asset, and lets the bad actor list rank equipment by hours lost rather than by work orders raised. An event with no work order against a breakdown is a breakdown nobody fixed, or one that was fixed off the books, and either is worth a question.

Reading the hours

The summary is a table of hours by category and cause for the period, which the template works out, ranked. The top causes are the agenda. A breakdown cause at the top goes to the root cause analysis trigger and the maintenance strategy. A process cause at the top goes to operations and often to the design of the feed system or the stockpile. An external cause at the top is a case for a standby supply or a contract clause.

The same hours feed three measures. Availability, as operating time over available time, into OEE. The cost of downtime, which turns the hours into money and is the number that gets the conveyor's drift switches fixed. And the maintenance KPIs, where breakdown hours by equipment are the reliability measure that the MTBF and the bad actor list are read beside.

The system around it

At one line, a shared log is enough. At a site, the capture moves into the control system or an operations system: automatic stop detection from the running signal, a reason screen on the HMI that the operator completes before the restart or at the end of the shift, the code list held in one place, and the summary built as a report rather than a spreadsheet. That is the production reporting work Enco does, and the real-time production analytics case study is one shape of it. Whatever the system, the code list, the time model and the same-shift discipline are the parts that make the numbers mean anything.

Where tracking fails

• Free text instead of codes, so the hours cannot be added up.
• Scheduled never defined, so availability moves with whoever filled the log.
• Times from memory the next day, so the hours are wrong and the short stops are missing.
• Other and no cause carrying a third of the hours.
• Breakdowns coded to the area instead of the equipment, so the history cannot be read by asset.
• A stop with two causes coded to the first, so the second never appears.
• No work order against breakdown events, so downtime and maintenance history never meet.
• A monthly report nobody reads at a meeting where decisions are made.

Common questions

Key terms

Plain-language definitions from our glossary for the concepts this article leans on.

Standards and further reading

• Best practices, metrics and guidelines for maintenance and reliability (SMRP)
• ISO 22400-2:2014 Key performance indicators for manufacturing operations management (ISO)
• ISO 14224:2016 Collection and exchange of reliability and maintenance data for equipment (ISO)
• ISO 55001:2024 Asset management system requirements (ISO)

Related case studies and tools

• OEE Calculator (tool)
• Cost of Downtime Calculator (tool)
• Real-time production analytics (case study)
• Automated operations reporting (case study)

Related reading