Construction of water content prediction model for tobacco roaster based on back propagation neural network

China is the world's largest tobacco producer, with tobacco production accounting for one-third of the world's total production. Tobacco drying is one of the most core processes in the tobacco production process, and its drying control strategy directly determines the quality of tobacco drying, the quality of finished cigarettes, and the level of the whole line process. Therefore, it is highly valued and concerned by tobacco production enterprises. The moisture content of exported tobacco is one of the core indexes used to test the quality and taste of tobacco. Predicting the moisture content of exported tobacco is essential for controlling tobacco quality and conducting experimental research. In this study, an electromagnetic eddy current type of tobacco dryer is used as the object, and a Back Propagation Neural Network-based tobacco water content prediction model is constructed to provide more accurate prediction and decision-making basis for tobacco production and management.