How to Use a Digital Twin to Test Shift and People Decisions
A digital twin helps with shift and people decisions by running the production plan against your shift calendar. It shows which station or transport limits output, whether the plan fits the shifts you have, and how on-time delivery and work in progress change when you test a different shift pattern, one extra person or carrier, or a task moved to a cart. It does not optimize crew size or recommend a headcount. The model shows the effect of each option on fixed nominal times, and your team decides.
What the twin does with people, and what it does not
The twin treats shifts and people who move material as part of the flow. The shift calendar sets the working days, start and end times and breaks. People who carry material between stations are transport resources in the model, with their trips and how busy they are. The production plan then runs through that week.
What you get is the effect of a decision you describe: this shift pattern, this extra person, this task moved to a cart. What you do not get is an answer to "how many operators are enough". The twin does not search for the smallest crew, does not size teams and does not issue staffing recommendations. Headcount involves skills, contracts, absence, safety and people's careers. Skills, contracts and absence are not modelled in detail, and people decide them. The model gives the team facts about the line. People make the decision.
Why a staffing table is not enough
Most plants plan people with a table: stations in rows, shifts in columns, a number of operators in each cell. The table shows coverage. It does not show flow.
An illustrative line has six stations, and on paper every station is covered in every shift. On the floor, the third station waits for material because the person who carries parts is busy with another trip. Meanwhile parts pile up at the station before it. By the table the line is fully staffed, yet output is lower than planned.
People interact with machines, buffers and transport. When material is not where the work is, the effect shows up somewhere else: a queue at the next station, a machine waiting for material, a cart with nothing to pick up. That is a flow problem, and it needs a model that runs through time.
What to put into the model
- shift calendar: working days, start and end times of each shift, and breaks
- people who carry material between stations, as transport resources working in the shifts you set
- stations and routings: the operations each product goes through, with nominal cycle times
- transport: forklifts, hand trolleys and carts, and the trips they make between stations
- the production plan: the orders for a typical week, with due dates, plus a peak week if you expect one
Nominal times are enough to start. Enter routings, stations and times by hand, or bring them in by import (import scope to be confirmed). The simulation is deterministic: times are fixed, so the same input always gives the same result, and a difference between two variants comes from the change you made. See Averages vs simulation over time.
What the model shows about people and shifts
For each run you see throughput, work in progress, lead time, on-time delivery and station utilization, plus the bottleneck with its cause and the periods when it limited the line. For shift and people questions, read the results this way:
| Result | What it tells you about shifts and people |
|---|---|
| Bottleneck and its cause | Whether the line is limited by a station, by transport or by a station waiting for material |
| On-time delivery against the plan | Whether the plan fits the shifts you have, or which orders finish late |
| Work in progress and lead time | Whether parts pile up in front of a station during some shifts |
| Station utilization across the shifts | Which stations are busy, waiting or idle, and in which periods |
| Periods when the bottleneck limited the line | Whether the problem sits in one shift, at shift change or across the whole week |
The most useful signal is a station that waits for material or transport. It means the constraint is how material reaches the station, not the machine. An extra person or carrier may help there. An extra person at a station that is not waiting for anything will not. For more on finding the real constraint, see How to Find Production Bottlenecks Before They Happen.
Real case: more waiting than work at the bottleneck
A plastic injection molding company in central Poland, supplying the household appliances industry, has over 200 people on three shifts and a 20,000 m2 site that includes an 8,000 m2 warehouse. The company wanted to know whether to automate and robotize several stations on one line: is it technically feasible, what happens to productivity, which automation options exist and what each would cost.
DBR77 built a digital twin of part of the plant, with material flows and the cycle time of each task. The model identified which tasks had the greatest procedural, economic and ergonomic value for automation. A later iteration simulated automated transport between stations. For each variant, the team got reports on efficiency, productivity and working time. What the analysis found:
- The bottleneck was a manual step done by the first operator.
- In that step, the operator had 90 seconds of activity and 98 seconds of waiting.
- Replacing that operator with an autonomous cart increased efficiency by over 100%, which created the conditions for the investment to be economically viable.
- Automating intralogistics, including non-ergonomic tasks, and automatic quality control were proposed as next steps.
The telling number is that the operator at the bottleneck waited longer than they worked. A staffing table would have shown one person, fully assigned, with no trace of those 98 seconds. The model showed where the time went. The decision to automate that task was the company's.
See shift and automation questions in a model
Line changes, shift patterns and automation decisions from other plants are on the use case pages, written up like the injection molding case above.
Shift and people questions worth testing
Each variant is a copy of the base model with one change. All of them are compared side by side on the same measures.
| Variant | Question it answers | What to compare |
|---|---|---|
| Peak week with the current shifts | Does the plan fit the shifts we have, and where does it break first? | On-time delivery, late orders, bottleneck |
| Extra shift or extra working day | Does more working time fix the late orders, or is the constraint elsewhere? | Throughput per week, on-time delivery, WIP |
| Shorter shift or fewer working days | How much of the plan still ships on time? | On-time delivery, lead time, bottleneck |
| One extra person at the bottleneck station | Does output rise, and where does the bottleneck move? | Throughput, bottleneck, station utilization |
| One extra person carrying material | Do stations still wait for material, and does transport load drop? | Station waiting, transport load, on-time delivery |
| Transport task moved to a cart | Does freeing the person raise output, or does transport become the limit? | Throughput, cart utilization, bottleneck |
Read the results as answers to the question you asked, not as a verdict on the right crew. If one extra person at the bottleneck moves the constraint to transport, the model has told you that. Whether to add the person, change the shift pattern or buy a cart is a decision for the team.
Shift questions often come with a change in demand. A new contract or a seasonal peak raises the question of an extra shift or a working Saturday. How to Test Capacity Before the Next Demand Shift explains how to run several demand scenarios against each capacity option.
Showing the line to new people in 3D and VR
The same model helps before anyone stands on the floor. New employees can walk through the line in 3D and see where stations, buffers and transport paths are. With a headset, the session can switch to VR, so people see the layout at real scale. Operators and maintenance staff can also check access and space around a planned station before it is built.
VR shows the space. The numbers on throughput and delivery come from the simulation.
Using the results with the team
Shift and people changes affect people, so how you present them matters. A model helps because the discussion moves from opinions to what the line does. It shows where the line waits for material or transport and where the limit lies elsewhere, and a limit that is not about people is often the more important finding.
Share the model with shift leaders before the decision. They know where the real breaks and handovers are. If they recognize their line in the animation, they will trust the numbers. If they do not, ask what differs. That gap usually points to an assumption the model is missing, and sometimes to the real constraint on the line.
FAQ
Does the digital twin tell us how many operators we need?
No. The twin does not optimize crew size or recommend a headcount. You describe a variant, for example one extra person at a station or a different shift pattern, and the model shows its effect on throughput, on-time delivery, WIP and the bottleneck. Your team makes the decision.
Is this the same as shift scheduling software?
No. Scheduling software assigns named people to shifts. The twin tests whether a shift pattern can carry the production plan, and where the line is limited.
Do we need detailed time studies to start?
No. Nominal times are enough to start. Compare the first results with what the line really produces, and measure the tasks where the two differ most, such as manual steps at a suspected bottleneck.
What does the model know about people?
The model knows shifts and working days, breaks in the calendar, and people who carry material between stations as transport resources, with their trips and how busy they are. You see stations waiting for material or transport, station load, throughput, WIP, lead time, on-time delivery and the bottleneck with its cause. Skills, contracts and absence are not modelled in detail, and people decide them.
How often should shift variants be rerun?
Whenever the product mix, demand, shift pattern or line setup changes noticeably. The base model stays, so each new question is a new variant on a copy of it.
Conclusion
Pick the line where shift leaders say they are always short of people. Before you change the roster, run next month's plan through the current shifts in a model and check one thing: does the bottleneck station wait for material or transport, or is it busy the whole shift? The answer tells you whether a change in shifts or people can help at all.
Test your next shift change first
Weighing an extra shift, a working Saturday or one more person at a station? Talk to an expert to see that kind of change tested on a model with a shift calendar and transport.
Sources
- DBR77, Case study: Development of an automation concept (summary on this site)
- DBR77 Digital Twin, Averages vs simulation over time
- DBR77 Digital Twin, VR review
Want to test a decision from your plant?
Describe the decision to a DBR77 expert and we will show how DBR77 Digital Twin compares the options on output, waiting time and transport load.
Questions
Have a question about this topic?
Tell us what it is about. We will reply within one business day.
Digital Twin newsletter
Short notes on testing layout and flow decisions on a model, and news from DBR77 Digital Twin. You can unsubscribe at any time.