EDGEBIC Downtime Explained: The Hours a Working Day Loses

EDGEBIC by User Solutions models an interruption inside a working day as a subtraction: the lost hours come out of the shift's capacity while the rest of the shift stays schedulable. A two-hour maintenance slot on a nine-hour shift leaves seven hours the scheduler can book. That fills the gap between a calendar saying the plant is open and a plant where one machine is genuinely not available from eight until ten.

Downtime sits alongside holidays and capacity overrides in the capacity model, and each of the three answers a different question.

Downtime Is Not a Holiday

The difference is whether the working day survives.

Using a holiday for a two-hour maintenance slot costs you the other seven hours, and the scheduler will move work it could have run. Shortening the shift instead moves the start or end time and misrepresents when the machine is actually free. What you want is a day that keeps its shape and loses some of its hours.

How It Enters the Capacity Number

A day's capacity for a work center starts from the shift's gross hours. Breaks come out. Any partial-holiday overlap comes out. Downtime comes out. What remains is then scaled by the work center's instance count and its utilization setting to produce the hours the scheduler is allowed to book.

The order matters more than it looks. Because the reduction is applied before the instance and utilization multipliers, it scales with the size of the work center: two hours lost on a work center with three machines is six machine-hours of capacity, not two. That is the correct behavior and it surprises people who expect a flat subtraction.

One consequence is worth stating. A capacity override does not stack on top of a subtraction: if you have stated an explicit number of hours for a day, that number is the answer and the day's arithmetic is not re-derived from the shift. That precedence is covered in daily capacity overrides.

Which Screens You Actually Use

This is where expectations need to be set precisely, because the engine models more than the current release exposes.

A partial plant holiday is the plant-wide route. You define the window once and every shift is charged only its own overlap with it, so a 10:00 to 10:45 window costs the day shift 45 minutes and costs an evening shift that never touches those hours nothing at all. This is the right tool when the whole site is affected: a utility outage, an all-hands meeting, a site-wide safety stand-down.

A per-day capacity override is the single-machine route. You state the hours a work center can genuinely offer on a given shift and date, and that number replaces the formula for the days it covers. A four-hour service on an eight-hour shift becomes 4; a machine on the bench all day becomes 0. This is also the documented way to model a closure that affects one machine rather than the plant, since per-work-center holiday rows have no screen of their own in the current release.

Recurring windows and shift-scoped closures are supported by the scheduling engine but have no configuration screen today. A standing two-hour Wednesday slot is therefore entered as per-day overrides on the affected Wednesdays rather than as one repeating rule. That is more entries, and it is better to know it than to hunt for a menu that is not there. A date-range override that spreads a total across a period exists in the same category: modeled by the engine, without a visible entry point in the current release.

Planned Against Unplanned

The distinction between downtime you chose and downtime that happened to you is about understanding the plant rather than about arithmetic. For the plan, both are simply hours the machine does not have.

For everything else the split is the difference between "we lost 60 hours on the press last quarter" and "we lost 40 hours to maintenance we chose and 20 to failures we did not." The first is a complaint; the second is a maintenance case. On the execution side that record comes from the shop floor, where every pause an operator records carries a reason drawn from a machine, material, quality, or waiting category, which is exactly what makes the resulting analysis meaningful. The vocabulary is covered in planned downtime and unplanned downtime.

Watch the Oversized Entry

An entry larger than the day it applies to is the trap worth naming. A capacity override of zero closes the machine's day completely, and it holds until someone resets it, so a temporary entry left in place quietly keeps a healthy machine out of the plan for as long as nobody notices.

When a work center shows no capacity on a day you expected it to run, an override left behind from a past service is one of the first things to check, alongside a holiday covering the date and a shift that was never assigned to that machine.

It Does Not Touch What Already Happened

A capacity change entered today does not rewrite yesterday. Completed and in-progress operations keep their actual dates and logged hours, and a reschedule replans only what has not started. This is the same rule that governs why actuals are immutable everywhere else in the system.

Downtime also does not remind anyone that maintenance is due. EDGEBIC holds no service interval and raises no maintenance warning at the terminal, so the cadence lives in your maintenance system and the capacity entry is how the agreed window reaches the plan. That division is covered in preventive maintenance and EDGEBIC.

Where It Fits

Shifts declare when the plant is open, holidays remove the days it is not, partial holidays and per-day overrides account for the hours an otherwise normal day loses, and the scheduler plans against what is left.

Get those right and the capacity number the plan is built on matches what the plant can actually deliver, which is the entire premise of finite capacity planning. Two hours a week does not sound like a scheduling problem until you multiply it: across a year on a constraint machine, a standing slot nobody entered is roughly a hundred hours the plan promised and the plant never had, every one of them quietly borrowed from a job that then ran late. The complete EDGEBIC guide maps how capacity feeds the rest of the system, and /edgebic covers the platform as a whole.