Near-Infrared Spectroscopy and Chemometrics for Effective Online Quality Monitoring and Process Control during...

Area of Science:

• Analytical Chemistry
• Food Science and Technology
• Process Engineering

Background:

• Near-infrared (NIR) spectroscopy is established for fishmeal quality assessment.
• Its application for active online quality monitoring during processing is limited.

Purpose of the Study:

• To evaluate NIR spectroscopy combined with multivariate chemometrics for real-time prediction of fishmeal and oil quality parameters during processing.
• To focus on predicting changes in lipid quality, including fatty acid saturation and specific omega-3 fatty acids.

Main Methods:

• Utilized NIR spectroscopy to collect spectral data from fishmeal and oil during processing.
• Applied multivariate chemometric techniques, specifically Partial Least Square Regression (PLSR), to build predictive models.
• Validated models using independent test sets to assess prediction accuracy for various quality parameters.

Main Results:

• PLSR models accurately predicted proximate composition: water (R²=0.9938), lipids (R²=0.9773), and fat-free dry matter (FFDM, R²=0.9356).
• Successfully distinguished fatty acid saturation levels: SFA (R²=0.9928), MUFA (R²=0.8291), and PUFA (R²=0.8588).
• Predicted phospholipids (R²=0.8617) and omega-3 fatty acids like DHA (R²=0.8785) and EPA (R²=0.8689) effectively.

Conclusions:

• NIR spectroscopy, coupled with chemometrics, is a powerful tool for active online quality assessment in fishmeal and oil processing.
• This approach enables real-time monitoring and control of critical quality parameters throughout the production process.
• Facilitates improved quality management and product consistency in the fishmeal industry.