Ramp-up, reduced speed and learning curves
A machine does not always work at its rated speed: a new line ramps up after a launch, a worn tool or a sensor fault slows it without stopping it, or a new crew needs many units before it reaches the standard time.
Built in A field or setting does it directly.
How
All three are on the machine (Hardware → Timing & flow rules; Factory data → Stations), and stretch the time of a task started under them:
- Speed over the run (
speed_periods): from an hour of the run, a % of the normal speed,0 h: 60, 24 h: 85, 48 h: 100. Before the first hour 100%. A ramp-up after a launch, or a slow drift. - Reduced speed (
slowdown: { mtbf_s, duration_s, speed_pct }; Slows every, Slow for, Slow speed): spells on the calendar, at random, during which the machine works at that % of its speed. They have a random stream of their own.summary.slowdownsgives the spells and their time per station. - Slower when fuller (
load_slowdown_pct; Slower per extra job): on a machine with several positions, each other job running at once makes a job that much longer (an oven loaded fuller heats each part slower). - Learning (
learning: { rate, units }): each doubling of the units the station has made takesrateof the time (0.85 = an 85% curve), untilunits, from which on the task's time holds.
A task started in a spell or a period runs at that speed throughout. Auto Generate sizes for the normal speed: give the ramp-up machines yourself if the first days must meet the takt.
Check
A press at 50 s a stroke and a 60 s takt, 60% of its speed on the first day and 85% on the second, slowed to 70% for an hour about every 12 h. Over 72 h it makes 4,170 brackets against 4,320 at full speed, in four spells of 3.5 h together; from the second day on it keeps the takt.