Optimization of Pneumatic Separating Process Parameters Based on BP Neural Network

In order to explore the relationship between the process parameters of the pneumatic separating process and the evaluation index of the pneumatic separating effect, and to find the optimal process parameters, using the primary flue-cured tobacco leaves in Yunnan tobacco area in 2021 as raw materials, a three-factor and three-level test plan was designed. The influence weights of feed flow, fan frequency and high-speed feed belt frequency on pneumatic separating process quality were studied by correlation analysis method and multiple linear regression method. Then the BP neural network model was built with the experimental data, and the optimal process parameters were predicted by the model, that is, the feed flow rate was 11500 kg/h, the fan frequency was 44 Hz, and the high-speed feed belt frequency was 44 Hz. The research result shows that, based on the pneumatic separating process parameters, the prediction model of various pneuma separating process quality indicators established by BP neural network has high accuracy.