Work Sampling in Machinery Fleets and Quarries
Study construction machinery utilisation using clearly defined categories, production data and costs.
Updated:
In quarries, aggregates and earthmoving, dispersed equipment and variable cycles can obscure shift use. Sampling helps estimate activity states; it does not replace all telemetry or stopwatch study or guarantee savings.
1. Operational scope
Define equipment, unit, period and exposure. A loader waiting for a truck does not establish personal operator failure. Confirm restrictions through process evidence. Use authorised safe observation points; do not approach blasting or vehicle zones simply to complete a reading.
Size the study around the operating question, target population and required precision. National production or investment figures do not replace local utilisation data or establish the return from an intervention.
2. Categories
| State | Operational rule |
|---|---|
| Main activity | Defined excavation, loading, loaded travel or processing |
| Flow waiting | Ready equipment with confirmed upstream/downstream constraint |
| Preparation/support | Positioning, refuelling, cleaning and checks |
| Breakdown | Confirmed equipment failure |
| Planned maintenance | Scheduled intervention |
| Logistics waiting | Supply/coordination under non-overlapping rules |
| Break, training or absence | Separated according to purpose and exposure |
| Other/unobservable | Documented reason |
An engine off does not establish breakdown or unjustified absence; safety, demand and operating decisions may explain it. Categories require training and agreement, not automatic inference.
3. Sample
Independent approximate planning uses n₀=ceiling(z²p₀(1−p₀)/e²). z=2 approximates two-sided 95.45%, not exactly 95%.
| p₀ | Absolute e | n₀ with z=2 |
|---|---|---|
| 0.50 | 0.05 | 400 |
| 0.50 | 0.03 | 1,112 |
| 0.30 | 0.05 | 336 |
| 0.30 | 0.03 | 934 |
| 0.10 | 0.02 | 900 |
With z=1.96, p₀=0.5 and five points, use 385. Requirements concern the overall or subgroup estimate, not every machine automatically. Adding two hundred readings does not automatically enable rare-event comparisons; define effect, power and design.
4. Rounds and coverage
Select times with known probabilities and cover relevant conditions. Fixed fifteen-minute intervals may align with cycles; a random start does not eliminate every periodicity. Plan feasible routes and retain actual times, delays and omissions.
Simultaneous machine observations need not create bias but may be correlated. Separate lists do not guarantee independence when equipment shares stops. A recorded seed supports reproducibility, not future comparability. An app does not guarantee exact alerts or error-free recording.
5. Simulated case
Assume fourteen mobile machines plus two mobile processing plants, and eighteen operators in three shifts. Including both plants gives sixteen assets, not fourteen. A fictional campaign collects 412 states across five morning and afternoon shifts. It does not automatically represent nights, weekends or equally precise individual-machine estimates.
| State | Records | Approximate share |
|---|---|---|
| Main activity | 215 | 52.2% |
| Flow waiting | 35 | 8.5% |
| Preparation/support | 29 | 7.0% |
| Breakdown | 47 | 11.4% |
| Planned maintenance | 17 | 4.1% |
| Logistics | 41 | 10.0% |
| People/breaks as defined | 25 | 6.1% |
| Other | 3 | 0.7% |
Shares are rounded from assumed counts. They do not establish a universal 65–70% target or 8.5% waiting as acceptable.
6. Improvement hypotheses and cost
Investigate hydraulics, truck coordination and road maintenance. Scenario reductions of 60% in breakdowns, 50% in logistics and 25% in flow waiting are hypotheses, not results. They do not automatically imply 62–64% main activity or 2,000–2,050 tonnes/day.
To assess €18,000 spares, forty workshop hours and eight road-maintenance hours monthly, include all costs and genuinely recoverable contribution margin. Under-three-month payback needs demand, capacity, cash flows and timing. Reassignment that consumes work is not cost-free.
7. Validation and tools
Confirm causes with records, compare compatible campaigns and assess uncertainty. Preserve safety, breaks and necessary tasks. Sampling does not guarantee legal acceptance or no reactivity.
Worksamp supports study configuration and activity capture; confirm features and terms. Stopwatch study can address specific cycles and telemetry can complement states. OEE also needs ideal cycle and output counts; operator observations alone are insufficient.