WP 2 Optimal process control and raw material differentiation

Work Package Leader: Dr. Jens Petter Wold, Nofima

Participating partners: Nofima, SFA, DTU, Prediktor, Industry partners


Development of optimal process control concepts involving novel rapid and on-line methods based on spectroscopy/machine vision and interpretation of large process data based on Big Data to ensure
a) the right and timely differentiation of raw material;
b) desired quality of thefinal product so that discard and food loss in processing stage is minimized. This impliesmeasurement of the quality of the raw material continuously in several points of the line.


T2.1 Optimal sorting of raw materials:

Develop dynamic and flexible sorting concepts that consider all key factors, such as e.g. raw material quality measured by intelligent sensors, production capacity, product selling price, market demand, end quality classes etc. The concepts will be developed and tested based on relevant case studies defined in WP1. Extensive simulations based on the same case studies will give insight into the performance and robustness of the systems.

T2.2 Optimal interpretation of process data based on Big Data:

One or more relevant food industry processes will be chosen and analysis will be based on existing data or data obtained from large on-line measurement surveys of relevant quality features. Data will be integrated and analyzed to gain knowledge about the process and raw material variability, model relations between process settings and quality attributes, and identify critical parameters and process steps. The use of these data to learn about – and improve – production and handling of raw materials prior to processing will be studied.

T2.3 Novel rapid and on-line methods for differentiation:
Based on case definitions from WP1, methods that measure relevant surface, sub-surface (or deeply embedded) quality properties will be developed. The main goal is to obtain robust quality measurements from complex samples in motion at high speed. New tools based on in particular on-line near-infrared spectroscopy and hyperspectral imaging will be developed for grading and differentiation of raw materials from chicken/meat/fish/dairy so that one can acquire the desired quality of the final product.

T2.4 Process control systems:

Process systems based on Raman spectroscopy are promising for foods, and novel applications for quality determination will be developed. A traditional limitation with Raman is tiny sampling volume and relatively time consuming sampling.

Acquisition of high quality spectra at short integration time, speed of measurement and representative sampling will be developed. New Raman applications will be developed for intact meat and fish measurement of fatty acid composition, pigments and protein composition.

Interrelation with other WPs:
Close interrelation with WP1 and WP3. WP2 will generate data to WP4 so that the raw material is differentiated according to final product requirements.

Milestones: M2.1:
Measurement data collected and dynamic/flexible sorting concepts developed(Q8)
M2.2: Concepts for intelligent sensing and process control for chosen cases developed (Q12).
M2.3: Concepts for efficient handling and interpretation of large process data based on Big Data developed (Q10).
M2.4: Concepts for rapid and on-line differentiation developed (Q14).

D2.1:Optimal concepts for optimal process control.
D2.2: Intelligent concepts for multispectral/machine vision measurements for selected cases.

D2.3: Concepts for efficient handling of large process data based on Big Data strategies.
D2.4: Novel rapid/on-line methods for differentiation.
D2.5: 5 manuscripts submitted to peer-reviewed journals;
D2.6: 1 MSc Thesis.

About the Author: