Optimization of process parameters for continuous kheer-making machine

If rice is cooked in milk, starch-milk reaction results into a thick product, which is very popular in India known as kheer. Conventionally, kheer is prepared by cooking rice in milk in an open pan over low fire followed by addition of sugar toward the end. The present investigation aims to optimize the process parameters (operating pressure and cooking time) for designing the pressurized cooking section of a continuous kheer-making machine. Sensory trials of kheer prepared conventionally and using pressurized methods were carried out and the data was analyzed using Fuzzy Logic. Sensory results of open-pan samples indicated that there is a small range of Whiteness Index (WI) and Hardness (H) values that is desirable in kheer. An Artificial Neural Network-Genetic Algorithm (ANN-GA) model was developed to further optimize the operating parameters to result in a product that would have the desired color and texture observed in kheer prepared conventionally. The developed ANN-GA model was successful in providing with input conditions leading to desired WI and H values. Finally, from the set of optimal input conditions, operating pressure of 0.27 MPa and cooking time of 7.5 min was chosen for designing the pressurized cooking section of the continuous kheer-making machine.

► Soft computing tools to optimize cooking conditions for Indian rice pudding-kheer. ► Use of 'Penalty Functions' to carry out the constrained optimization process. ► Optimization done without compromising the sensory acceptability of the product. ► Results would aid in developing continuous processes for manufacturing of kheer.