Bottleneck Identification & Simple Simulation Workbook

Purpose

This workbook helps teams quickly find the true production constraint, verify it with simple flow metrics, estimate expected throughput changes from common interventions, and design short experiments to protect and improve the constraint without heavy IT work.

When to use this

• You're unsure which machine or station is limiting daily output.
• WIP or queues keep growing in one place and throughput isn't rising.
• You want a quick, data-informed experiment to validate balancing options.

What you'll need

• Simple stopwatch or cycle-time data (seconds or minutes per piece).
• Available time per shift (minutes or seconds) and uptime% estimate.
• Setup/changeover time estimates (if relevant for batch processes).
• Short observation of flow (WIP pileups, starved machines).

Key concepts & quick formulas (shop-floor friendly)

• Cycle time: average processing time per unit at the station (seconds/unit).
• Available time per shift: shift length minus breaks (seconds/shift).
• Capacity (parts/shift): Available time / Cycle time.
  Example: Available time = 28,800 sec; cycle = 120 sec → capacity = 240 parts/shift
• Demand (parts/shift): What you must produce that shift (customer or downstream requirement).
• Utilization: Demand / Capacity (expressed as a fraction or %). Stations with utilization near or above 1.0 are candidates for the constraint.
  Example: demand 200 / capacity 240 = 0.83 (83%)
• Effective capacity with uptime: Capacity × Uptime% (e.g., 0.9 uptime reduces capacity by 10%).
• Throughput: the actual output rate the line delivers. The constraint largely determines max steady-state throughput.

Step-by-step worksheet

Use the following steps to collect the minimum data and identify candidate constraints.

1. Pick a short window: Observe one or two shifts or a representative hour. Walk the flow from raw input to finished goods.
2. Record per-station data: For each station (or machine), note: cycle time (avg), available time/shift, uptime% (or observed availability), and typical setup time.
3. Compute capacity and utilization: For each station compute Capacity = Available time ÷ Cycle time. Adjust Capacity × Uptime% to get effective capacity. Then Utilization = Demand ÷ Effective capacity.
4. Look for flow signs: Is WIP piling upstream? Are downstream machines starved? Which station has the highest steady utilization or constant queues?
5. Validate candidate constraint: The true constraint will typically show: highest utilization, persistent queue upstream, and more frequent stoppages or quality gates limiting flow.
6. Run simple what-if calculations: Estimate throughput change if you: reduce cycle time by X%, reduce setup time, add an extra operator, or protect uptime by better PM. Use the capacity math above to estimate new throughput.
7. Design a short experiment: Pick one protection or improvement (e.g., dedicate an operator for the constraint for two shifts, add a small buffer, introduce a quick quality check upstream). Run it and measure throughput and lead time for a day or shift.

Example (worked)

Line demand: 180 parts/shift. Shift available time (net) = 28,800 sec.

Station A cycle = 100 sec → capacity = 288 parts/shift. Uptime 90% → effective capacity ≈ 259 parts.

Station B cycle = 140 sec → capacity = 205 parts/shift. Uptime 95% → effective capacity ≈ 195 parts.

Station C cycle = 160 sec → capacity = 180 parts/shift. Uptime 92% → effective capacity ≈ 166 parts.

Utilizations vs demand (180): A = 180/259 = 0.69, B = 180/195 = 0.92, C = 180/166 = 1.08 → Station C is the constraint (utilization > 1 and upstream queue observed).

What-if: Reduce Station C cycle by 10% → new cycle 144 sec → capacity = 200 parts; uptime 92% → effective ≈ 184 parts → utilization = 180/184 = 0.98 (constraint eased; throughput could increase if other stations remain stable).

Simple balancing & protection tactics to try (low friction)

• Small buffer before the constraint (WIP of a few pieces) to prevent starvation from upstream fluctuations.
• Priority rules: always feed the constraint first; delay non-constraint work during experiments.
• Short operator cross-training to handle the constraint when needed.
• Reduce or batch setups at the constraint (quick-change, SMED experiments).
• Preventive maintenance or daily quick-checks to improve uptime at the constraint.
• Quality gate upstream so bad parts don't reach and block the constraint.
• Measure impact for one shift and compare throughput, lead time, and WIP before and after.

Common pitfalls (mal-hungers)

• Fixing a non-constraint: optimizing utilization on a non-constraint can hide spare capacity and create problems elsewhere.
• Chasing utilization numbers alone: noisy utilization or minor measurement errors can mislead—use flow signs and queues too.
• One-off balancing without protection: rebalancing only once without rules and protection often shifts the problem or creates starvation.
• Over-optimizing single-machine uptime at the expense of overall flow and predictability.

Quick experiment template

1. Hypothesis: (example) Short daily 10-minute preventive check at Station C will reduce unplanned stops and increase throughput by X%.
2. Intervention: Describe exactly what you'll change, duration, and who is responsible.
3. Metrics to record: throughput (parts/shift), WIP before constraint, number of stops, average stop duration.
4. Run length: 1–3 shifts for a quick check; longer if results are promising.
5. Decision rule: If throughput increases by at least Y% or lead time decreases and no new issues arise, adopt the change; otherwise revert and try another experiment.

Next steps & suggested tests

• Run the worksheet for one shift and capture the numbers in the worksheet table (use a clipboard or the interactive form if available).
• Pick a single low-cost protection or improvement and run a short experiment per the template.
• If the experiment helps, codify the protection (standard work, buffers, priority rules) and spread to other lines carefully—monitor results to avoid shifting the constraint without seeing net gain.

Links & references

• Little’s Law (useful for linking WIP, throughput, and lead time).
• SMED quick-change methods for reducing setups.
• Simple OEE checks to cross-validate uptime impacts.

Printable worksheet (copy/paste into your clipboard)

For each station write: Station name | Cycle time (sec) | Available time/shift (sec) | Uptime% | Capacity (parts/shift) | Effective capacity | Demand | Utilization

Notes/observations: queues upstream? frequent stops? quality escapes?