基于模糊c均值聚类的RBFN的混炼胶粘度在线估计

Online Estimation of Mixing Smelting Viscosity Based on RBFN of Fuzzy c-Means Clustering

介绍了将基于模糊c均值聚类 (FCM)算法的多模型建模方法 (简称FMM)与径向基函数神经网络 (RBFM)相结合 ,先用FCM算法将训练集聚类 ,再用隶属度将子模型的输出结合起来 ,从而完成软测量模型的建立。这种方法不仅增强了在对象的整个输入空间的预测精度 ,同时减少了隐层节点数目 ,加快了学习速度。算法仿真表明 ,所提出的算法是处理橡胶混炼粘度软测量建模的一种很有效的方法。

The model of software measurement is established by combining multiple model modeling method based on fuzzy c-means clustering algorithm (FCM) and radial basis function neural network, firstly, clustering the training set with FCM algorithm, then combining the outputs of sub-models with subordination? This method not only enhanced the prediction accuracy of object in the whole input space, but also reduced the number of implicit nodes and increased the learning speed. The simulation shows that the algorithm is a effective measure for modeling of soft measurement of viscosity of mixing and smelting rubber. Key words Soft measurement Viscosity Online estimation RBF network Fuzzy c-means clustering