Powered by FOSS

The In-line NIR learning curve

It is easy to talk about the benefits of In-line NIR, but what does it really take to make it work? Our question and answer session with a major aqua feed producer sheds light on some of the practical considerations when going in-line.

Since 2012, EWOS Chile Alimentos Ltda, who manufacture balanced feed for salmonids, have been using in-line NIR for measuring protein, fat and moisture in raw ingredients during the mixing phase. Nirperformance talked to Head of the laboratory, Mario Vega Marín about the challenges of installing in-line NIR and why it’s worth the effort.

Where in the production line is the In-line NIR installed?

We installed the NIR in one of the manufacturing lines after the mixer. It is important for us to carry out the monitoring at this location mainly to check the level of protein in the diet, as it is a parameter that cannot be corrected later in the process. So any improvements that can be made at this point are vital for ensuring the nutritional levels required in each diet.

Using in-line NIR analysis in a balanced food plant seems very complicated. Why do you think it’s worth trying?

The effort required to maintain adequate calibration and monitor protein levels is compensated by the costs associated with protein levels above or below those expected. Losses are generated without adequate monitoring, because the final off-specification product has to be reprocessed at a high cost that is very difficult to reverse.

EWOS Chile Alimentos LtdaThe Chilean subsidiary of the Norwegian company EWOS. EWOS is a leading supplier of feed and nutrition for the international aquaculture industry.EWOS have produced fish feed since 1935, and currently operate in all four of the world’s major salmon farming regions: Norway, Chile, Canada, and Scotland.EWOS also operate in Vietnam, producing feed mainly for tropical species.


What were the major challenges and limitations you experienced in practice?

The equipment and calibration curves that we are currently operating with are optimised and tailored to our products. The main issues and challenges we faced were related to the installation and adjustment of our facilities and control room software, in order for the final reading of the equipment to be in line with the data of the control room operator, who is the end user of data delivered by in-line NIR.

Prior to that we had a tough job in terms of quality control, to be able to adjust calibrations to readings increasingly aligned with our laboratory testing equipment.

The work is linked with various areas of our business, which is why production, maintenance and quality control staff, as well as contractor personnel, were involved in different tasks in order to implement the project.

How did you achieve the correct calibrations for the new in-line NIR application?

Firstly, we started with the calibrations provided by the supplier. However, our product is clearly not 100% comparable with these global calibrations, which is why our GH and NH values were initially quite high. Although we were able to make an initial Bias adjustment, it was upsetting to see those high GH and NH outlier values in the statistics. The software tool, helped us to transform our database from our old bench top NIR system into a database equivalent to the new in-line solution, which improved our predictions. A new database was then created with exclusive in-line equipment readings as a result of the on-going work of quality control, and within a few weeks we were able to implement calibrations that were more robust, with predictions more closely targeted at laboratory values. GH and NH values have also returned to normal (less than 1), which puts us at ease now that we trust the tool’s predictions.

Currently, the maintenance and updating of calibrations are part of quality control and are carried out through random sampling by our process control staff.

In-line NIR directly applied to the manufacture of balanced food is a bit like moving the laboratory to the factory. How was this implemented by your team? What kind of training was required?

Actually, it was not so difficult, because production staff is always willing to acquire new tools that help improve process control. The preparation work that had to be undertaken to be able to read the equipment on the operator’s screen was performed in collaboration with maintenance staff, and the idea was to take readings to the control room using the easiest method. Ultimately, training is carried out simply to tell the operator, “just press the button on your screen and you’ll see the final reading for your product.”

In the case of quality staff, we had to devise a plan for taking random samples with a view to implementing new adjustments or calibrations, if required. However, the training is generally quite basic and anyone can learn to use the equipment, so the task was not difficult.

As for the operators, they seem pleased with this new tool. According to control room operators Marcelino Perez and Jose Cisternas: “the tool was very useful when we had to make adjustments to the process, because we could see the answer in real-time on our screens.”

Based on your own experience, what are the main points that other balanced food companies should consider if they want to use in-line NIR in their manufacturing processes?

Firstly, they should be aware that a good in-line reading from the equipment is made possible by having a plan for maintaining and updating calibrations. They should also be fully aware of the critical control points, so it can be a measurement that really helps improve their production processes. No less important is all the after sales support that the company provides as part of its service and this an important issue when deciding which brand will give you the most confidence in that respect.

Leave a Reply

  • (will not be published)