Answers to the most common questions.

No. Most projects start small. It often begins with a single area, a layout, a material flow or a specific question. The digital twin grows step by step – from a digital inventory to a data-driven representation of production.

It makes production systems not only visible but testable. By connecting the production model, real-time data and simulation, alternative scenarios can be run through digitally before they cost time, budget or output in reality. Starting pays off earlier than many expect: those who begin early have the basis to assess later decisions realistically.

A simulation answers one question and is then finished. A digital twin stays connected to the equipment and keeps answering new questions. The transition is gradual – almost every twin begins as a simulation for one specific question. The decision to continue it usually comes once the first benefit has become visible.

Yes. Virtual commissioning is fundamentally worthwhile for every machine. The effort occurs once, the benefit remains across the entire machine lifecycle: development, testing, service, retrofits and optimizations can be validated digitally – often before real parts are ordered or assembled. The ROI shows after 3-5 projects at the latest.

That is the decisive question, and it determines the effort. The simplest level checks signals: your controller switches outputs, the model returns the matching inputs, the step sequence runs through. The next level emulates drives with acceleration, ramps and real travel times – that makes cycle times and collisions reliable. The deepest level represents sensors, gripping processes and material behaviour. We recommend the level that answers your question, not the deepest one.

The effort depends on four things: how good and how complete the data situation is, how much CAD preparation is needed, what the control architecture looks like – one PLC or several – and how deep the simulation should go. For a single machine or cell the range is therefore between EUR 2,000 and EUR 20,000. The upper end is open: large plants where every signal is represented and higher-level systems are connected go beyond that. After the first conversation we tell you which depth answers your question – and what it costs.

We work regularly with Siemens and Beckhoff. Beyond that, our tools connect to Rockwell and Allen-Bradley, B&R, Bosch Rexroth, CODESYS, Fanuc, Mitsubishi, Omron, Schneider Electric, VIPA and WAGO. The connection runs via native drivers, via simulation units such as PLCSIM Advanced, SIMIT or Logix Echo, or via OPC UA, Modbus TCP and MQTT. If a coupling is missing, we build it.

Yes, using the real controller rather than an imitation: KUKA.OfficeLite, ABB RobotStudio, FANUC RoboGuide and Epson RC+, plus controllers from Kawasaki, Yaskawa, Stäubli, Denso, Mitsubishi and Universal Robots. We run the emulations in containers, which lets several controllers operate in parallel within one model. If one is missing, we build the connection.

Yes, in two places. In engineering: fe.screen-sim has an MCP interface to AI systems that lets models be generated, checked and analysed – this shortens model building considerably. And in the result: AI can propose parameters, sequences, batch sizes, vehicle counts. The difference to a pure AI solution is that we compute every proposal against the digital twin before it reaches the plant. A prediction without a system model remains an estimate. Together with partners we continue to develop such solutions.

No. Many projects initially start with a digital model or individual data sources. Real-time data, MES, ERP, PLC, warehouse or fleet management systems can be integrated step by step later – depending on the desired use case.

Preferably STEP or JT – the neutral formats that work everywhere. We also accept hall floor plans as DWG and point clouds from laser scans. Native design data we convert beforehand. And importantly: we do not need manufacturing-accurate models. For simulation the envelope geometry matters, not every screw – we deliberately reduce over-detailed data so the model runs in real time.

That is the normal case. Existing halls we capture by laser scan: this produces a point cloud and from it a dimensionally accurate floor plan – usually more accurate than the existing as-built drawing, because that one does not know about the latest modifications. If only 2D drawings exist, we model. We have an extensive library of machines, conveyor and transport systems, and whatever is missing we rebuild from drawings and dimensions. What we do not offer is classic design engineering: we build simulation-capable models, not production-ready parts.

Three things: CAD data of the equipment or hall, the control program or a functional description, and a contact person authorised to make decisions. If something is missing, we find a way – a hall we can scan, a controller we can also reproduce from a functional description and terminal diagram. It extends the start-up phase but does not prevent the project.

Mainly fe.screen-sim, Visual Components and realvirtual.io. The task decides which tool is used – not the other way round. If you already have a simulation environment in house, we work our way into yours, so the model stays with you rather than with us.

Not for the pilot project. We provide the environment and you carry no licence costs. If you then want to operate the twin yourself, you need your own licence – either you already have one, or we recommend the one that fits your use case. If you prefer to leave operation with us, you need none.

You receive the model and can keep operating it – the handover is part of the assignment. What stays with us are our methods, libraries and reusable components. Models we build for you usually remain available at our end because we keep working on them; at your request we delete our copies entirely. Non-disclosure agreements apply as a matter of course.

Both, depending on the task. Laser scans we do on site, model building remotely. Workshops, virtual commissioning and training of your staff take place on site again, because knowledge is handed over there. Travel costs are invoiced separately.

Yes, this is one of the most frequent reasons for a project. Quotations state vehicle counts based on assumptions about routes, intersections, load changes and charging cycles. In the simulation we compute the same transport task with your layout and your volumes. The result regularly differs – upwards as well as downwards.