Edge Twins HPC · Project

Affordable Digital Twins That Run Directly on Factory Machines, Not in the Cloud

manufacturingPrototypeTRL 4

Imagine you have a virtual copy of your machine that lives right inside the machine itself — no expensive cloud servers needed. Right now, digital twins require massive computing power and teams of specialists, which prices out most small and mid-size companies. This project figured out how to shrink those heavy simulations down so they can run on cheap, low-power hardware sitting right next to the equipment. Think of it like compressing a huge video file so it plays smoothly on your phone instead of needing a cinema projector.

By the numbers

consortium partners

33%

industry participation ratio

SME in consortium

EUR 99,875

EU contribution

total deliverables produced

demonstrator deployed on real devices

The business problem

What needed solving

Most companies — especially SMEs — are locked out of digital twin technology because it requires expensive cloud infrastructure, specialized expertise, and ongoing subscription costs. Running real-time simulations of physical equipment demands massive computing power that only large enterprises can afford. This leaves smaller manufacturers and operators unable to benefit from predictive maintenance, real-time optimization, and virtual testing.

The solution

What was built

The project produced an alpha demonstrator of an open-source digital twin builder designed for edge deployment — meaning the twin runs on low-cost hardware attached directly to the physical asset. They delivered 5 deliverables total, including a report on demonstrator results with analysis of application downscaling and deployment on real devices.

Audience

Who needs this

SME manufacturers wanting predictive maintenance without cloud costsAutonomous vehicle developers needing on-board real-time simulationBuilding management companies monitoring HVAC and energy systemsIndustrial IoT integrators looking for open-source digital twin toolsEquipment OEMs wanting to embed digital twins in their products

Business applications

Who can put this to work

Automotive manufacturing

SME

Target: SME automotive parts suppliers running CNC machines or assembly lines

If you are a small auto parts manufacturer dealing with unplanned machine downtime and no budget for cloud-based monitoring — this project developed an open-source digital twin builder that runs directly on your shop floor equipment. Instead of paying for expensive cloud subscriptions, the twin lives on a low-cost edge device attached to the machine, giving you real-time performance data. The consortium tested this with a demonstrator deployed on real devices.

Building management and HVAC

mid-size

Target: Facility management companies operating commercial buildings

If you are a building manager dealing with energy waste across dozens of HVAC systems — this project developed software that creates digital twins of your equipment running on cheap local hardware. Instead of sending sensor data to the cloud for analysis, each unit monitors itself locally and flags problems instantly. The approach uses reduced order modelling to cut the computing power needed by orders of magnitude.

Logistics and fleet management

any

Target: Companies operating autonomous or semi-autonomous vehicle fleets

If you are a logistics operator exploring autonomous vehicles but struggling with the cost of real-time simulation infrastructure — this project built an open-source tool that puts a digital twin directly on the vehicle. The edge-based approach means each vehicle can run its own real-time simulation without depending on network connectivity. The project specifically identified autonomous vehicles as a target application area.

Frequently asked

Quick answers

What would this cost us to implement?

The digital twin builder is designed as open-source software, so there are no licensing fees for the core tool. Your main costs would be integration, edge hardware (low-cost devices), and adapting the twin to your specific equipment. The project was funded with EUR 99,875, indicating a lean feasibility-stage effort rather than a full commercial product.

Can this scale to hundreds of machines on a factory floor?

The core idea is specifically designed for mass deployment — putting a twin on every physical asset rather than running everything centrally. Reduced Order Modelling compresses heavy simulations so they run on constrained compute environments. However, the project delivered an alpha demonstrator, so scaling to hundreds of units would require further engineering.

Who owns the intellectual property and can we use it?

The project aimed to build an open-source software tool, which means the code is intended to be freely available. The underlying technology leverages the ExaQute FET Proactive project's high-performance computing software. Specific licensing terms would need to be confirmed with the coordinator, CIMNE in Barcelona.

How does this compare to existing digital twin platforms like Azure Digital Twins or AWS IoT TwinMaker?

The key difference is edge-first design. Major cloud platforms require constant connectivity and cloud compute costs that scale with usage. This approach puts the twin on the device itself, eliminating cloud dependency and recurring costs. Based on available project data, this targets users priced out of enterprise cloud solutions — specifically SMEs.

What stage of development is this at?

The project ran for 18 months (June 2020 to November 2021) as a feasibility study and produced an alpha demonstrator tested on real devices. It is a Coordination and Support Action, meaning the focus was on evaluating business viability rather than building a market-ready product. Further development would be needed before production deployment.

Is there regulatory or standards compliance built in?

Based on available project data, the project focused on technical feasibility and business viability rather than regulatory compliance. Any deployment in regulated industries like automotive or building systems would require additional certification work specific to your sector.

Consortium

Who built it

This is a compact, Spain-only consortium of 3 partners — 2 research organizations and 1 industry player (which is an SME). The coordinator, CIMNE in Barcelona, is a well-known computational engineering research center. The 33% industry ratio and presence of an SME partner suggests some commercial grounding, but the single-country composition and small team size (typical for a CSA feasibility study) mean this hasn't yet been stress-tested across diverse industrial environments or markets. A business partner would be working with a research-heavy team that has strong technical depth but would need commercial partners for market deployment.

CENTRE INTERNACIONAL DE METODES NUMERICS EN ENGINYERIACoordinator · ES
BARCELONA SUPERCOMPUTING CENTER CENTRO NACIONAL DE SUPERCOMPUTACIONparticipant · ES

How to reach the team

CIMNE (Centre Internacional de Mètodes Numèrics en Enginyeria), Barcelona, Spain — a leading computational engineering research center

Order a report on this consortium See CENTRE INTERNACIONAL DE METODES NUMERICS EN ENGINYERIA profile →

Next steps

Talk to the team behind this work.

Want to explore whether edge-based digital twins fit your operations? SciTransfer can connect you with the Edge Twins HPC team and help evaluate the fit for your specific use case.

Order a report on this topic More in Manufacturing & Industry 4.0