Quality Prediction Based on ANN in Tobacco Redrying Process

Abstract

To improve uniformity of quality in tobacco redrying process, an ANN-based method is proposed to predict the possible output of post-redrying. Firstly, this paper presents an intelligent process control model with ANN. And then, it describes the three-layers network and improves BP algorithm by using the method of variable learning rate, in which 10 process parameters were selected as inputs, the moisture and temperature of post-redrying as outputs, and one hidden with 25 neurons is designed. Finally, the ANN is trained to map the nonlinear relationship between inputs and outputs. With the analyses and comparisons, the result obtained shows that the prediction of possible output has higher accuracy by this improved ANN-based method. Such model designed can be suitable for quality improvement in the tobacco redrying process.