论文部分内容阅读
RBF-CSR是在分析RBF-PLS的基础上提出的新方法。它保留了RBF-PLS的优点:采用神经网络的结构, 又用数学方法直接求解,免去了ANN冗长的训练过程和其它诸多欠缺。RBF-CSR方法可以在更宽广的空间内寻找最优的网络参数,它所建立的模型具有很高的预报精度和良好的稳定性,又有简洁的解析形式,便于优化等进一步的计算和处理。该方法已成功地应用于裂解装置的建模。
RBF-CSR is a new method based on the analysis of RBF-PLS. It retains the advantages of RBF-PLS: the use of neural network structure, but also directly solve the mathematical method, eliminating the ANN lengthy training process and many other deficiencies. The RBF-CSR method can find the optimal network parameters in a wider space. The model established by the RBF-CSR method has high prediction accuracy and good stability, and has a simple analytical form for further calculation and processing such as optimization . This method has been successfully applied to the modeling of pyrolysis plant.