Predicting high‐protein bar processing ability from rheological and tribological analyses

With the growth of the high‐protein bar market, predictive models for good processing ability would assist bar manufactures in development of bar formulations. The objective of this study was to create predictive models for high‐protein bar model formulations based on empirical testing and instrumental data. The predictive models generated had relatively high accuracy rates (> 85%). However, three misclassifications were seen for oil‐ and shortening‐based formulations, leaving gray areas of predictive values and indicating that data from additional formulations is needed to improve model accuracy. Model validation testing showed that cold flow was best for predicting processing ability of oil‐based formulations. For shortening‐based formulations, wear rate and G 3 ′ / G 1 ′ at 4% strain and 10 rad/s best predicted processing ability. These models provide valuable information about ingredient ranges and instrumental tests that could be used to assist in the determination of processing ability. Practical applications The results of this study may be used by high‐protein bar manufacturers and manufacturers of similar products to determine the processing ability of novel formulations with a rapid test that requires only a small amount of sample. There are no similar methods currently used except for manipulating samples by hand. Judging processing ability from hand manipulation requires significant experience working with the product, whereas the tests used to create data for the models outlined in this article can be performed by inexperienced personnel to rapidly screen novel formulations.