Line balancing aligns the effective capacity and response of connected modules. OEE—availability × performance × quality—is one framework for separating time loss, speed loss and rejected output, but the underlying stop and production data is what guides action.
1. Separate rated speed from effective capacity
Rated speed may describe a machine under a defined ideal condition. Effective capacity considers the actual product and pack, cycle pattern, replenishment, minor stops, quality checks and the margin needed to recover after an interruption.
For indexed machines, the number of packs per cycle and cycle time matter. For continuous machines, component feeding and stable product presentation may be the limiting factors. Record assumptions beside every proposed rate.
| Rate | Meaning | Use |
|---|---|---|
| Rated speed | Published or configured maximum under stated conditions | Initial machine comparison |
| Effective machine rate | Expected output for the actual format and normal interventions | Line capacity model |
| Sustained line rate | Good packs maintained over an agreed production period | Business and acceptance target |
| Shift output | Good packs after planned and unplanned losses | Production planning |
2. Map the complete flow and constraint
Draw each operation from component supply to finished-unit removal. Add cycle rates, buffer capacities, manual tasks and expected stops. The lowest effective rate may be the obvious constraint, but a feeder or case-packing task can limit the line even when its machine rating appears sufficient.
Repeat the exercise for different formats. A filler may constrain a large dose, while capping or labelling may constrain a smaller decorative pack.
- Product supply and batch change
- Container and cap feeding
- Filling cycle by dose
- Closure and label format
- Inspection and reject handling
- Collation and case packing
- Operator replenishment and removal
3. Use buffers for defined disturbances
Accumulation allows upstream and downstream modules to continue for a limited time when an adjacent machine stops. It is useful for frequent short disturbances, but it consumes space and can create back pressure, scuffing or product-age issues.
Calculate buffer need from the disturbance duration, pack rate and acceptable pressure. Decide whether the buffer should protect the constraint, allow replenishment or separate a batch process from continuous equipment.
4. Coordinate machine response and controls
A balanced mechanical layout can still perform poorly if machines stop and restart independently without a clear line philosophy. Define blocked and starved states, sensor positions, restart conditions, low-level warnings and the response to quality rejects.
Avoid excessive stop propagation. A downstream machine may be able to complete the pack it contains safely before stopping, while an upstream filler may need to finish a dose or protect open containers. These details are application-specific.
- Starved and blocked status
- Run permissives and controlled stop
- Buffer high and low levels
- Feeder low-level warning
- Fault and reset ownership
- Reject full or inspection fault response
5. Understand the OEE categories
OEE multiplies availability, performance and quality. Availability compares run time with planned production time. Performance compares actual speed with the defined ideal rate while running. Quality compares accepted output with total output.
The result is useful only when the underlying definitions are consistent. An inflated ideal rate can make performance look poor, while excluding changeovers from planned time can hide an important operational loss.
- Define planned production time
- Use a credible ideal rate by format
- Count rework and rejects consistently
- Keep loss categories stable over time
- Review the component measures, not only the multiplied percentage
6. Capture a small set of reliable stop reasons
Start with categories operators can identify consistently: no bottles, no caps, product supply, filler fault, capper fault, labeller fault, downstream blocked, quality hold, changeover and cleaning. Add detail only when it will lead to action.
Automatic stop capture may identify the first device to signal, not the root cause. Combine controls data with operator context and observation of the line.
| Level | Example | Decision supported |
|---|---|---|
| Area | Upstream, filling, closing, downstream | Where to investigate |
| Cause | No caps, label break, product low | What condition stopped flow |
| Detail | Cap elevator empty, roll end | Specific corrective action |
| Duration/frequency | Long rare or short frequent | Priority and solution type |
7. Treat changeovers as part of capacity
A flexible line can lose more output through format and product changes than through running speed. Observe the full sequence from the last good pack of one format to the first approved pack of the next.
Separate internal work that requires the line stopped from external preparation that can happen while the line runs. Organised tools, labelled parts, recorded settings and pre-staged materials can improve repeatability without changing core machinery.
- Last-to-first-good measurement
- Cleaning and product clearance
- Tool and change-part collection
- Mechanical adjustment and recipes
- Quality approval and first-off samples
- Return and storage of previous format parts
8. Improve the constraint, then rebalance
When the current constraint improves, another stage will become limiting. This is expected. Recalculate the line and observe the new stop pattern rather than assuming the same improvement will continue indefinitely.
Possible actions include maintenance, improved component quality, feeder capacity, conveyor changes, control tuning, change-part redesign, added nozzles, replacement machinery or a different production schedule. Select the least complex action that addresses the measured loss.
- Measure the current baseline
- Identify the active constraint by format
- Confirm the dominant loss mechanism
- Implement one controlled change
- Verify good output and unintended effects
- Rebalance and repeat
9. Use acceptance tests that reflect line behaviour
A very short run may demonstrate individual cycles but not component replenishment, buffer behaviour, normal minor stops or end-of-line flow. Select an acceptance period and material quantity that represent the agreed duty while remaining practical.
Record interventions, rejects, stops and the line state at the beginning and end of the run. Agree how planned pauses, sample checks and customer-supplied material defects are treated.
Key takeaways
- Use effective machine rates and sustained good output in the line model.
- Identify the constraint for each priority product and format.
- Add accumulation for specific short disturbances, not as a substitute for capacity.
- Define line stop, restart, starvation and blockage behaviour in the controls scope.
- Use OEE components and reliable loss data to direct improvements.