Faster and more accurate analysis of physicochemical properties of dairy products

SUMMARY

This user case explains how a dairy company implemented Chemometric Brain’s quantitative analysis to predict the physicochemical properties of a batch of mozzarella cheese in order to identify whether new batches conformed to previously produced batches and had the same performance when used, for example, as melting cheese. Chemometric Brain’s calibration models are more flexible than those of other NIR software and provide this identification more quickly and accurately.

CONTEXT

A dairy company was interested in determining the composition of a batch of mozzarella cheeses produced at their facility to ensure that the product properties conformed to the company’s previously established quality parameters, had the same performance (e.g. on melting) and were therefore replicable for more efficient and safer manufacturing.

To solve this, Chemometric Brain quantitative analysis was implemented, which allows instant information on a wide variety of physicochemical properties such as moisture, protein, fat, pH, dry matter or salt (sodium chloride), among others.

Chemometric Brain offers a wide range of calibration models that food companies can use for their NIR spectra, but, compared to other software, these calibrations do not need to be periodically updated by the manufacturer, but are updated automatically and the user simply has to incorporate the reference values into his NIR spectrum. This means that Chemometric Brain offers more flexible, faster and more accurate models than other NIR software.

TESTED PRODUCT

Mozzarella cheese manufactured by a dairy company.

ISSUE

Prediction of different physicochemical properties in a few seconds by using the quantitative calibration models developed by Chemometric Brain software as a service (Saas).

APPROACH

Our customer wanted to ensure that new batches of mozzarella produced at their facilities conformed to those produced in previous processes, ensuring that, for example, they would melt similarly.  To address this problem, Chemometric Brain provided this company with a large number of calibration models – based on multiple varieties of ingredients and foods – for its NIR spectra. This enabled to quickly predict the value of each of the physicochemical properties.

Figure 1 shows an example of the calibration model used to predict the fat content of a new batch of mozzarella with an unknown value. On the right side of the image, you can see the results obtained using the quantitative calibration models available in the Chemometric Brain libraries for other nutritional components such as protein, dry matter, sodium chloride and pH. Data provided by Chemometric Brain SaaS prove that the new batch is within the reference values and can therefore be considered as conforming to those previously produced.

CONCLUSION

Chemometric Brain’s multiple quantitative calibration models for common food ingredients and products in the food industry make it easy for companies to determine the composition of food products (fat, protein, ash, viscosity, dry matter, moisture, etc.) handled in their facilities. By implementing Chemometric Brain SaaS, customers will be able to predict a set of quantitative values for different physicochemical properties without need to use other not so flexible calibrations typically employed by other NIR devices.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>