Process parameter setting of paddy post harvesting process to maximize whole rice recovery and minimize energy consumption

Abstract

Reduction of post-harvesting losses and energy consumption emphasized mainly on the present research work. The essential requirement in India to fulfill the food for the growing population is paddy. The postharvesting losses and energy consumption in paddy processing is very high at present. Post-harvest in Paddy process includes Cutting, Field drying, Transporting, stacking, Pre-Threshing, Threshing, Drying, Par Boiling, storing and Milling. The post-harvest loss varies between 20- 30% in which 10-15% of losses considered as milling in the Paddy Processing Industry. It shows that it is vital to improve the milling methods in Paddy Processing. The detailed literature has identified significant post-harvesting losses and energy consumption in Paddy processing. In this regard Drying and Milling are considered as the most crucial process which influences the post-harvesting losses and energy consumption in the paddy processing Industry. In this research work, as a part of Paddy post harvesting model, the detailed study has been conducted to identify the different milling and critical parameters. The critical parameters are Soaking Temperature, Soaking Duration, Parboiling Temperature, Parboiling Duration, Drying Temperature, and Drying duration. The regression model has been developed to find the broken rice percentage, using critical parameters. The data collected from the Paddy processing Industry were pre-processed using the regression model. The pre-processed data were validated using a normality test. The preprocessed data were processed in the WEKA software to develop an Artificial Neural Network(ANN), model. Four different algorithms such as Gaussian process algorithm, Multilayer perception algorithm, Radial Basis Function Network (RBFN) algorithm, and Sequential Minimal Optimization (SMO) iv algorithm were discussed and applied in the ANN model. Multilayer perception algorithm was found to be suitable to predict broken rice percentage based on the accuracy of the results. newline