Software: Time to make better use of NIR data

The widespread use of NIR can generate treasure troves of measurement data, but only if there is time to use the information to improve the quality of feed and the efficiency of the feed mill. In this interview for Nirperformance, software experts Bob Brill and Paul Smolders explain how modern software systems can transform NIR data results from a static list of figures to timely, value-adding data integrated into feed formulation and laboratory operations. The potential for improving feed quality and saving costs is significant, both for individual producers and the feed industry as a whole.

Imagine doing a feed formulation based on a fresh set of the latest raw material sample data combined with reference analysis results, supplier history and much more. Modern software can make this happen, but in around 90% of the feed industry, feed formulations that will decide the success or failure of production runs worth thousands of dollars are still based on disparate pieces of data that could be several weeks old.

On the one side you have the NIR instruments and on the other, the formulation software systems. If only the two worlds could come closer together, it would create a lot of value for feed millers.

Some are doing it, but as software consultant Bob Brill and Adifo product manager Paul Smolders explain, only about 10% of the industry is really exploiting NIR data today by getting it into the feed formulation process in time. “In many cases, people lack the knowledge about the opportunities to transform the data into useful information and how to go to formulation,” says Smolders.

Brill adds how NIR is being used largely as a defence mechanism to keep unwanted raw material out of the supply chain especially in light of some of the horror stories connected with adulteration. “We are trying to get them to go to the next step and put all that information to good use in the feed formulation process,” he says.

A way to do this is with the LIMS software system offered by Adifo which Smolders describes as “A platform to collect quality NIR data and a tool to transform that data into value for the company.”

All in one place: an example of how various sources of data such as NIR test data, reference analysis and supplier identity can be presented in the LIMS software. Example shows NIR-fat result (Y-axis) per sample (X-axis) taken over several days. Blue dots are Fat results for 3 samples measured with wet chemistry reference method.

One place, one picture

The LIMS software is used primarily by quality controllers to collect and manage data from various sources including a lot of NIR test data along with wet chemistry results, supplier identity, origin and such like. Once approved, this data is used by nutritionists to update the values for formulation. The software system gives a full picture and most importantly captures the latest information allowing the nutritionist to act. “Not only do you get the full picture. Without it you would not be able to use the most current information generated by the samples,” says Brill.

In line with this theme, Smolders describes the various NIR data inputs and how important it is to have an effective way of handling the variables involved.

One factor is that a single sample cannot be looked at in isolation in the feed matrix. “One sample represents maybe one truck, but in your silo you have maybe material from 20 trucks,” he says. “You have to find a way to take a number of samples representing what you have in stock.”

Another consideration is the source of material, for example, if a feed company has five different plants in different regions of a country sourcing local material. Values for samples from the different sources need to be averaged out with a standard safety margin, generating the new plant-specific matrix values. Add another material like Soya which is sourced globally and used across all plants and you have another setup. Then add several different parameters for the various samples like moisture, protein and fats and it is clear that quality controllers and nutritionists need the help of software tools if they are going to keep track of what is going on in the supply chain.

“Combining all that data into one database is quite a challenge if you have to rely on excel sheets, one for NIR data, one for wet chemistry and so on,” says Smolders. “Without the software they can’t use the data because it is spread across different storage systems. The LIMS system is relevant for any company having to manage around 1000 or more samples a year.”

Closing the loop by comparing values of finished products against expected ones

The LIMS system can be integrated with feed formulation software so that once the feed formulation software has generated the expected values for the finished products, they can be made available as target values in the LIMS system and compared with the real ones from samples. The source of any significant deviation can then be investigated and taken into consideration regarding the nutritional values used for the formulation. For example, for poultry products it is quite important to have the right energy value for broilers.

Example of sample data used for updating the formulation matrix.

A key goal with the system is to help feed producers master the variety of raw materials that can be seen from NIR measurements. “The NIR instrument helps to identify the quality, but then you need to find out what to do with that variability,” says Smolders.

Putting the NIR data to work

The ability to combine data and, most importantly, to compare actual data against targets, gives interesting possibilities for feed producers to adopt a more aggressive approach to different aspects of their business. Even so, barriers to adoption remain. One mental hurdle appears to be that many are afraid of the ramifications of changing the data directly.

However, progress is being made in moving towards a more reactive feed formulation approach. The impact of potential changes can be projected side-by-side with current production values to simulate the effect of making adjustments in the formulation in what Brill calls a ‘what if’ test. He says: “We are getting to the point now where people see that it is possible to move the data and some say I am not going to use it, but I am going to simulate using it to check the outcome in the formulation, what would have happened if I did use the NIR data. The formulation systems we have allow them to do that, but ultimately it is all about getting rid of the fear of change.”


Join our community of experts and professionals now to learn more about the many uses of NIR technologies in feed production while networking with your peers.