Background: Artificial neural networks (ANN) are a common mathematical tool widely used in many research fields. Since they are applicable to non-linear relationships and do not require preliminary assumptions, they are a particularly promising tool in relation to meat processing. Thermal denaturation contains a lot of information concerning the quality of meats. The aim was to create a methodology of kinetic analysis to obtain a quick and accurate tool for meat protein denaturation in non-isothermal conditions based on The Coats-Redfern equation with the use of ANN.
Methods: The analyses were carried out on samples of minced samples of Longissimus dorsi (pork). Thermal properties were determined using the differential scanning calorimetry (DSC) method with a Q100 TA Instruments apparatus. The data obtained was processed using the artificial neural network module in Statistica 13.0 software.
Results: The following models fit well with experimental data: F1 and F2 (r = 0.99, F Snedecor’s F statistics 836943.20 and 971947.41 respectively). Deviations from experimental conversion degrees were higher for model F2, while for F1, good conformity was obtained across the whole range of α(T).
Conclusions: This preliminary study confirmed that methods of traditional kinetics of processes in non-isothermal conditions based on the Coats-Redfern equation can be successfully applied to meat protein denaturation. The method of kinetic analysis allows a high level of accuracy to be achieved and meets the requirements of an efficient engineering tool.  .
Keywords: DSC; artificial neural networks; meat; pork; thermal denaturation.