论文部分内容阅读
针对铅锌烧结过程透气性的预测具有模型不确定性和输入变量不确定性等特点,建立了综合透气性智能集成预测模型。首先建立了基于满意聚类的T-S综合透气性预测模型,针对聚类后各子模型结论参数的辨识工作计算复杂、容易陷入局部极值的问题,将混合粒子群优化算法用于这些结论参数的辨识;然后利用灰色理论建立了时间序列综合透气性预测模型;最后利用信息熵技术将2个预测模型进行集成,以获得集成预测模型。选取实际生产过程中100组合格的数据,分别用以上3种预测模型来预测相应的综合透气性,其相对误差的平均值分别为2.1%,3.2%,1.8%。实验结果表明,本文提出的集成预测方法能够有效地克服不确定性带来的影响、提高综合透气性的预测精度。
In order to predict the permeability of lead-zinc sintering process, the model has the characteristics of uncertainty of model and uncertainty of input variables, and an integrated prediction model of gas permeability is established. Firstly, the TS comprehensive gas permeability prediction model based on satisfactory clustering is established. According to the problem that the identification of the concluding parameters of each sub-model after clustering is complicated and easy to fall into local extremum, the hybrid particle swarm optimization algorithm is applied to these conclusion parameters Then, the gray model is used to establish the time series integrated air permeability forecasting model. Finally, the two forecast models are integrated by using the information entropy technology to obtain the integrated forecasting model. Select 100 sets of qualified data in the actual production process, respectively, using the above three prediction models to predict the corresponding integrated breathability, the average relative error of 2.1%, 3.2%, 1.8%. The experimental results show that the integrated prediction method proposed in this paper can effectively overcome the impact of uncertainty and improve the prediction accuracy of comprehensive permeability.