基于偏最小二乘的Kriging代理模型在加氢裂化建模中的应用

Kriging Agent Model Based on Partial Least Squares in the Application of the Hydrocracking Modeling

摘要: 提出了一种改进的代理模型方法（Kriging with Partial Least Squares，KPLS）。该方法在Kriging模型的基础上引入偏最小二乘的思想，利用偏最小二乘方法构建新的Kriging模型的高斯核函数。将该模型应用于加氢裂化过程建模，有效地提高了航煤、柴油质量收率的预测精度。采用 GLAMP（Global and local search strategy）优化算法对建立的KPLS模型进行优化，仿真结果显示航煤、柴油质量收率得到了显著提升。

Abstract: This paper proposes a modified agent modeling method,Kriging with partial least squares (KPLS).By means of Kriging model,we use the partial least squares method to establish a new Gaussian kernel function.Compared with the traditional Kriging model,the proposed KPLS model can effectively improve the accuracy of the fuel and diesel yield prediction.Besides,the GLAMP (global and local search strategy) search algorithm is used to optimize the KPLS model.The simulation results show that the yield of diesel and fuel is significantly improved.