Gelymar II: detection of minimal deviations in raw materials through qualitative NIR analysis

As we explained last week, Gelymar, a leading company in the supply of thickeners and stabilizers for the food industry, has adopted NIR technology to gain efficiency in quality control.

Beyond its use to predict and classify the different types of carrageenans supplied to its customers, we described how Gelymar uses qualitative models to control quality.

With qualitative analysis, each sample or batch is compared to reference samples previously validated and classified as compliant, which have been included in a qualitative model. This model recognizes product samples and detects deviations.

In the case of Gelymar, to detect possible non-conformities in the carrageenans, 5 different qualitative models we recreated. To create the models, a set of samples previously identified as compliant from different batches had to be collected, with the aim of capturing the variability of the product.

From there, Chemometric Brain’s technical team began to build the qualitative model with the compliant samples provided by the client, using principal component analysis (PCA) as the statistical method. The models were configured with "good" product samples, whose chemical footprint meets the quality standards previously approved, based on laboratory tests such as pH, gel strength, or particle size, among others.

Once the natural variability of the analyzed carrageenans was covered, the use of qualitative models allows the client to reduce the number of laboratory analyses, implementing an easier control that provides a conclusion on whether the sample passes or not, just in seconds.

In this way, Gelymar ensures that its customers always receive the same type of product with the ideal properties to be incorporated into the production process of their final products, thus avoiding possible claims or problems.

‍