ECOnti
Digital Twin for Continuous Bioproduction

A continuous E. coli process only pays off if every step stays stable for weeks. In ECOnti, a digital twin built from CFD, hybrid process models and plant automation kept the whole chain in balance.

Read the Paper Contact Our Team

Challenge

Continuous bioproduction depends on long-term stability.

A continuous process makes more product with smaller equipment, and it costs less to run. That only works while growth, production, primary recovery and purification stay in balance for weeks. So operators need models that tell them how each step will react when conditions change.

In this FFG-funded project, enGenes Biotech covered E. coli expression and BOKU and Tosoh the purification. Qubicon brought the automation, Novasign the hybrid process models and SimVantage the mechanistic CFD.

Process chain from raw material through growth, production, primary recovery and purification to product, orchestrated via Qubicon
Continuous process chain, with every step connected to the automation system

Approach

CFD Knowledge Inside the Process Twin

  • One process twin per unit operation, plus one for the whole chain
  • CFD results for mixing time, gas holdup (gas volume fraction) and shear rate built into the twin, so it also helps with scale-up
  • Model predictive control, which uses the twin’s forecasts to plan the next control step
  • CFD model outputs streamed live into the Qubicon automation system while the plant runs
  • A surrogate model trained on CFD results that finds production-scale operating conditions with the same mixing and oxygen transfer as in the lab

What CFD says about a reactor usually ends up in a report. Here it feeds the model while the process runs. When an operator moves a setpoint, the twin shows how mixing and shear will change before the plant gets there. That helps with real decisions, such as adapting the feed strategy or changing the stirrer-speed protocol.

Process parameters (left) and CFD model outputs (right) in the Qubicon system, recorded during operation and played at 32× speed. 45 seconds, no sound. © Qubicon

Results

30 Days of Stable Continuous Production

With the twin in the loop, the continuous E. coli process ran stably for 30 days. Operators saw mixing time, holdup and average shear rate from the CFD model right next to the process parameters in the control system.

ECOnti project film about the upstream process that ran stably for more than 30 days. 1:42 min, unmute for sound.

The consortium also compared the continuous chain with an optimized fed-batch process at 1,200 kg Protein A per year. Continuous operation needs 876 L of fermenter volume instead of 4,449 L, so the plant gets by with 25–39% less capital investment. In a larger reactor the cells often run short of oxygen or see gradients they never saw in the lab. The CFD-based scale-up model from SimVantage carries the lab operating conditions over to production scale, which lowers that risk before the plant runs.

  • 30Days of stable continuous E. coli production
  • 26–37%Lower operating cost than fed-batch
  • 45%Lower facility energy demand than fed-batch

Publication

End-to-End Continuous E. coli Biomanufacturing of Recombinant Proteins via Integrated and Automated Digital Control

Simon et al. · Microbial Cell Factories 2026 · BMC, open access
ECOnti consortium with BOKU Vienna, enGenes Biotech, Novasign, Qubicon and Tosoh Bioscience

Read the Paper All Publications

Facing a similar question? Let's talk.

Tell us about your reactor and your bottleneck. We will show you what simulation can quantify before you change anything in steel.

Book a Demo Contact Our Team All Case Studies