Real-time beet extraction optimization

Model Predictive Control (MPC) combined with Near Infrared Spectroscopy (NIRS) and Visual Smart Sensors (VSS), integrated on a universal IoT platform to improve yield, energy efficiency, and operational transparency.

At a glance

- Runs above your DCS and reads real-time plant data
- Optimizes KPIs relevant for plant managers and technical directors
- Economic objective: minimize total cost per tonne of beet

Operational motivation

Extraction plants must maximize sugar yield while maintaining stable juice and pulp quality, despite fluctuating beet quality, variable slice rates, delayed lab feedback, and operator-dependent control.

Control concept

The Extraction MPC uses dynamic process models and constraints to calculate optimal control actions for recirculation flows, water addition, temperatures, wash water distribution, and tower heating.

Measurement power

NIRS delivers near-continuous quality information, while VSS characterizes cossette geometry. Together, they enable faster disturbance handling and improved extraction kinetics.

Universal IoT platform

A scalable platform aggregates DCS, NIRS, and VSS data, runs MPC and analytics, and visualizes KPIs and controller actions via web dashboards.

Proven performance

Industrial results show an extraction yield increase of +0.2 to +0.6 percentage points. For a 10,000 t/day factory, this equals approximately €200k–€400k per 100-day campaign.

Implementation roadmap

1. Data analysis and dynamic model development
2. MPC configuration and KPI definition
3. Commissioning, tuning, and operator training
4. Continuous performance monitoring and extension to further process areas