Machine Learning Based Grain Moisture Estimation for Real-time Monitoring of High-Temperature Paddy Drying Silo

Hot air drying in a storage silo is an alternative approach to paddy drying, which can help in keeping moisture content in safer ranges with periodic reconditioning stored grains over a longer storage period by blowing hot air through grains under controlled environmental conditions. In this study, a methodology has been developed to predict under-treatment paddy moisture content based on historical patterns, internal and external environmental parameters. The problem has been simplified from the multivariate regression problem to multiple univariate regression problems per forecasting timesteps. Machine learning models have been trained using supervised learning techniques on the data obtained from developed internet of things (IoT) enabled smart silo-bin and compared based on Root Mean Square Error (RMSE) with a shallow neural network providing the highest accuracy and lower computational time.