In open-pit mining, production rarely slows because one activity is completely incapable of operating. More often, performance deteriorates at the interfaces: trucks wait for shovels, ore waits at the ROM stockpile, crushers run below plan, and information arrives after the operating decision has already been made.
Value Stream Mapping (VSM) makes this system visible. It shows how material and information move from drill and blast through loading, haulage, crushing, concentration and dispatch. More importantly, it connects operational data to customer demand, allowing improvement teams to distinguish a genuine bottleneck from a local inconvenience.
This worked example uses an illustrative copper mining operation. The numbers are designed to demonstrate the method and must be replaced with verified site data before investment or operational decisions are made.
1. Define the Value Stream and Customer Requirement
The fundamental purpose of VSM is to study a complete value stream, not an isolated department. In this case, the selected stream is:
Drill and blast → Load → Haul → ROM stockpile → Crush → Mill and flotation → Concentrate stockpile → Rail dispatch
The customer is the concentrator and downstream smelter. Its critical requirement is stable feed and concentrate availability, not simply the maximum number of truck movements.
The project team defines the following CTQs:
- 60,000 tonnes per day of crushed ore feed
- Ore delivered within the required grade and size range
- Reliable concentrate shipment windows
- Safe, compliant and predictable production
- Reduced lead time from blasted ore to dispatched concentrate
With 12 available production hours per day:
[
\text{Required takt throughput}=\frac{60,000\text{ tonnes}}{12\text{ hours}}=5,000\text{ tonnes/hour}
]
Here, takt is expressed as the required production rhythm. Throughput is the actual tonnes produced per hour or per shift.
The scope should cover one ore family, one production route and one representative operating period. Mapping an entire mine, all ore types and every logistics route in a single workshop usually produces a visually impressive but operationally weak map.
2. Build the Current-State Map at the Gemba
A current-state map should be built from observation, dispatch records, maintenance data, plant historians and operator knowledge. Do not map the process solely from standard operating procedures.
Capture:
- Equipment count and availability
- Actual cycle times and queue times
- Payload and tonnes per movement
- ROM and coarse-ore stockpile levels
- Crusher and concentrator uptime
- Grade, size and contamination defects
- Approval and communication delays
- Planned versus unplanned stoppages
The information flow begins with the daily mine plan and plant feed requirement. Dispatch then assigns trucks, while crusher and concentrator teams communicate feed constraints. If these signals are delayed, haulage may continue pushing ore toward a stockpile that cannot release material at the required rate.
A useful Time Observation Sheet separates value-added time from waiting, movement, queueing and rehandling. A simple process map should also distinguish work in process: such as blasted ore, loaded trucks and ROM inventory: from finished concentrate.
3. Worked Example: Confirming the Haulage Bottleneck
The illustrative current-state data is:
| Process step | Effective capacity | Key observation |
|---|---|---|
| Drill and blast | 6,200 t/h | Sufficient capacity, but batch release creates variation |
| Loading | 5,800 t/h | Shovel occasionally waits for available trucks |
| Haulage | 4,800 t/h | 40 trucks; average cycle time is 120 minutes |
| Primary crushing | 5,400 t/h | Capacity is adequate at demand level |
| Milling and flotation | 5,200 t/h | Narrow capacity margin |
| Concentrate dispatch | 5,600 t/h | Shipment windows are not currently limiting |
The haulage calculation is:
[
40\text{ trucks}\times240\text{ tonnes per load}\div2\text{ hours}=4,800\text{ tonnes/hour}
]
Required demand is 5,000 tonnes per hour, so the haulage system has a shortfall of:
[
5,000-4,800=200\text{ tonnes/hour}
]
Across a 12-hour production period, that equals 2,400 tonnes of unmet movement capacity. The problem is not necessarily a shortage of trucks. It may be a combination of road condition, queuing, loading delays, speed restrictions, dispatch logic, dump congestion and avoidable empty travel.
The map shows:
- 18,000 tonnes of ore waiting at the ROM stockpile
- Average ROM waiting time of 24 hours
- Trucks waiting an average of 14 minutes at the shovel
- Trucks waiting an average of 19 minutes at the crusher dump point
- Total lead time from blast release to concentrate dispatch of 52 hours
- Approximately 12 hours of processing and movement time
- Approximately 40 hours of waiting, storage or queue time
The bottleneck is confirmed when capacity, utilization, queue accumulation and customer impact point to the same constraint.
4. Identify the Eight DOWNTIME Wastes
Use the eight wastes: often remembered as DOWNTIME: to examine every interface:
- Defects: Incorrect ore routing, excessive fragmentation or off-specification concentrate.
- Overproduction: Hauling more ore to the ROM stockpile than the crusher can consume.
- Waiting: Trucks waiting for shovels, dump points, instructions, maintenance release or approval.
- Non-utilized talent: Dispatchers and operators observing recurring delays without a formal improvement channel.
- Transportation: Excessive empty travel, inefficient routes and unnecessary rehandling.
- Inventory: Large ROM, coarse-ore and concentrate stocks that hide flow problems.
- Motion: Avoidable travel to obtain permits, radios, samples or operating instructions.
- Extra-processing: Duplicate sampling, repeated grade approvals or unnecessary ore movements.
The distinction between value and activity is essential. Customers are willing to pay for correctly sized, correctly graded concentrate delivered reliably. They are not paying for avoidable queues, duplicate approvals or excess stockpile movement.
5. Analyse the Data, Not the Assumptions
During the Analyse Phase of DMAIC, the team identifies root causes using statistical and visual tools.
Use an Average (Mean) cycle time as a baseline, but pair it with a Box Plot to reveal spread, skewness and outliers. A Z-Score can identify unusually long haul cycles across different routes or shifts.
For repeated observations:
- An X-bar Chart, used with an R Chart, can detect shifts and trends in average cycle time and within-sample range.
- Attribute Data, such as Pass/Fail for ore routing or correct/incorrect dispatch instructions, supports defect analysis.
- ANOVA compares mean cycle times across routes, shifts, truck classes or loading areas.
- Bartlett’s Test checks whether group variances are equal before applying ANOVA.
- Bias in payload, dispatch timestamps or GPS measurements can make the constraint appear larger or smaller than it is.
The relationship can be expressed as Y = f(x): concentrate availability and shipment reliability are outputs influenced by inputs such as truck cycle time, payload, road condition, shovel availability, crusher demand and dispatch rules.
A Voice of the Customer requirement: stable feed and on-time concentrate: must be balanced with the Voice of the Business, including cost, safety, asset utilisation and capital return. The Voice of the Process, revealed by actual data, determines whether current performance meets those requirements.
6. Design the Future-State Map
The future state should not simply add trucks. It should create a controlled flow around the constraint.
Potential future-state changes include:
- Introduce a crusher-led pull signal based on ROM level and hourly feed requirement.
- Use dispatch rules that limit queue accumulation at the crusher.
- Reduce haul-cycle variation through road maintenance, traffic separation and route standardisation.
- Review loading and dump-point sequencing to reduce avoidable waiting.
- Improve blast fragmentation consistency to reduce crusher interruptions.
- Create an Andon or digital visual signal when the crusher, shovel or haul route deviates from standard.
- Apply Autonomation (Jidoka) where equipment or software can detect abnormal conditions and trigger a controlled response.
- Standardise escalation so an operator can stop, signal and correct a problem rather than passing the defect downstream.
In the future-state scenario, average haul cycle time falls from 120 to 106.7 minutes. With the same 40-truck fleet:
[
40\times240\div1.778=5,400\text{ tonnes/hour}
]
This creates an 8% capacity margin above the 5,000-tonne-per-hour requirement.
Current Versus Future State
| Measure | Current state | Future state |
|---|---|---|
| Haulage capacity | 4,800 t/h | 5,400 t/h |
| Average haul cycle | 120 min | 106.7 min |
| ROM inventory | 18,000 t | 8,000 t |
| ROM waiting time | 24 h | 8 h |
| Truck wait at shovel | 14 min | 7 min |
| Truck wait at crusher | 19 min | 8 min |
| Total lead time | 52 h | 31 h |
| Waiting and storage time | 40 h | 19 h |
| Concentrate dispatch reliability | 91% | 98% target |
The future state must be validated with a pilot, capacity checks and safe operating reviews. Once haulage is improved, the concentrator may become the next constraint. This is the logic of the Theory of Constraints: improve the limiting factor, then reassess the system.
7. Sequence Kaizen Actions and Governance
A practical kaizen sequence is:
- Week 1: Verify the map, measurement system, data definitions and baseline.
- Weeks 2–3: Remove dispatch, queue and route causes that require minimal capital.
- Weeks 4–6: Pilot pull-based haulage control and standard work at one mining area.
- Weeks 7–8: Compare throughput, lead time, queue time, safety and quality results.
- After approval: Scale the changes, update control plans and monitor performance.
Formal Approval checkpoints support governance, safety and capital control, but poorly designed approvals can themselves become bottlenecks. Define decision rights, escalation limits and response times.
Break-even Analysis can compare the cost of road improvements, dispatch technology or additional trucks with the value of recovered tonnes. The Business Case should include throughput, recovery, operating cost, safety, maintenance and shipment reliability: not only truck productivity.
Agile methods complement Lean Six Sigma here. Use short improvement sprints, visual backlogs and frequent operator feedback while retaining DMAIC discipline, verified data and controlled implementation.
8. Sustain the Gains
A control plan should monitor:
- Hourly haulage throughput
- Average and 95th-percentile cycle time
- Truck queue minutes
- ROM stockpile tonnes
- Crusher feed stability
- First Pass Yield for correctly routed, correctly sized ore
- Rolled Throughput Yield across drill-and-blast, haulage, crushing and processing
- Concentrate shipment adherence
The Zero Defects philosophy associated with Philip Crosby reinforces the principle of doing work correctly the first time. In mining, that means preventing incorrect routing, avoidable rehandling and specification errors rather than relying on downstream correction.
A White Belt can support basic awareness and DMAIC participation. A Yellow Belt can collect data and support local kaizen. A Green Belt can lead the project analysis and pilot. A Black Belt can lead complex cross-functional work, mentor Green Belts and connect the project to broader operational strategy.
For professionals ready to build these capabilities, Lean 6 Sigma Hub’s CSSC-accredited Green Belt Online Training covers process mapping, data collection, measurement systems, capability analysis, root-cause identification, hypothesis testing, piloting, SPC and control plans. Explore the full Lean Six Sigma training and certification pathway to match your role and project responsibility.
Map the value stream, measure the constraint, and earn the certification that enables you to lead measurable improvement.
Kaizen. Kai-Care. Kai-Done. ( Lean Six Sigma)
