基于神经网络MIV算法的烟叶加料均匀性预测分析

为了预测分析不同工况下烟叶加料的均匀性,文中提出一种基于神经网络MIV算法的烟叶加料工艺参数优化方法。通过对烟叶加料的工艺参数进行神经网络训练,利用MIV计算出各参数与加料均匀性的相关性,确定影响加料均匀性的关键工艺指标。文中进行了相应的试验证明,结果显示相对误差被控制在5%以内。由此结果推出影响加料均匀性的关键工艺指标为:排潮开度、工艺流量、气体压力及料液流量。其中,排潮开度、工业流量及料液流量与加料均匀性负相关,气体压力与加料均匀性正相关;经过MIV进行变量剔除的神经网络的预测性较好。

Forecast and Analysis of Tobacco Leaf Feeding Uniformity Based on Neural Network MIV Algorithm

In order to predict and analyze the uniformity of tobacco leaf feeding under different working conditions, a method based on MIV algorithm was proposed in this paper. The neural network training was carried out on the process parameters of tobacco leaf feeding, and the correlation between each parameter and feeding uniformity was calculated by using MIV. The key index influencing the uniformity of feeding was found, and the error was less than 5%. The key indexes that affected the uniformity of feeding were the opening of discharge, process flow, gas pressure and feed liquid flow. The discharge opening, industrial flow and feed liquid flow were negatively correlated with the feeding uniformity. The gas pressure was positively related to the uniformity of feed. The neural network obtained by MIV method was more predictive.