【摘要】 基于Mie光散射理论,开发了石油炼化装置烟气催化剂颗粒浓度在线检测系统,采用了基于最小二乘支持向量机(least square support vector machine,LS-SVM)的软测量技术。利用LS-SVM优良的非线性映射和强大的泛化能力,建立了颗粒浓度的软测量模型,通过LS-SVM的学习和支持向量的自适应更新,实现催化剂颗粒浓度的最佳估计。仿真和实际运行结果表明所开发的监测系统可对烟气催化剂颗粒浓度进行有效测量,其测量精度比经验公式提高4.2倍。更多还原

【Abstract】 A on-line measurement system based on Mie light scattering theory is developed for measuring the catalyzer particle concen-tration of the flue gas flow in oil refining equipment.Compare with the common catalyzer particle concentration measurement system with empirical formula used in other systems,a soft measuring technology based on least square support vector machine(LS-SVM),which has better nonlinear approximation and generalization capability,is adopted to establish the soft measuring model between the par-ticle concentrations and the change of laser light energy such as scattering,incidence and transmission,and an optimal estimatation is obtained after the model trained and updated adaptively.The simulation results and the performance of practical application in industrial field show that the proposed system can measure the catalyzer particle concentration effectively,and the measurement accuracy enhance 4.2 times than that with empirical formula.更多还原