- Line balancing distributes work across stations so no station exceeds takt time and idle time is minimised.
- Theoretical minimum stations = total work content divided by takt. Reality is higher because of precedence and physical constraints.
- Balance efficiency = total work content divided by (number of stations x the longest station time).
The line's output is set by its slowest station. Every improvement anywhere else changes nothing until that station is under takt, which is why balancing starts by finding it.
The measurements you need
- Takt time: available production time divided by customer demand.
- Work content per element, measured by observation, several cycles, at each station.
- Precedence relationships: which elements must happen before which.
- Physical constraints: elements tied to specific equipment or locations.
- Walking and handling time between elements, which is real work content people forget to count.
The calculation
Divide total work content by takt time to get the theoretical minimum number of stations, and round up. If total work content is 300 seconds and takt is 60 seconds, the theoretical minimum is five stations. You will normally need more, because elements cannot be split arbitrarily and some are locked to equipment.
Balance efficiency
Total work content divided by the number of stations multiplied by the longest station's cycle time. Five stations with a 300-second total and a 70-second bottleneck gives 300 / 350, or about 86%. The 14% is idle time distributed across the faster stations, and it is the loss balancing exists to reduce.
Redistributing
- Start at the station exceeding takt. Move the smallest element that brings it under.
- Respect precedence, or you will create rework.
- Watch for elements that cannot move because they need a specific machine; these often dictate the achievable balance.
- Consider splitting an element rather than moving it whole, where the work genuinely divides.
- Recheck every station after each move, because relieving one bottleneck exposes the next.
What makes it hard in practice
Product mix. A line running several variants has a different work content per variant, so a balance optimised for one is wrong for another. The usual approach is to balance for the weighted average mix and accept that individual variants run above or below takt, provided the mix is scheduled so the line averages out over a shift.
When not to balance
If station times vary substantially cycle to cycle, balancing a set of averages produces a line that stalls constantly. Stabilise the work first, usually through standard work, then balance. Balancing an unstable process optimises noise.