摘要
在三维冷态试验台架上对喷动流化床射流穿透度进行了实验研究,得出了射流穿透度随喷口尺寸、载气密度、喷动气速的增大而增大,随静止床高、颗粒尺寸、颗粒密度、流化气率、载气黏度的增大而减小的结论.在实验研究的基础上利用最小二乘支持向量机对射流穿透度与喷动流化床主要设计参数之间的数值关系进行了智能拟合,并利用自适应遗传算法优化了最小二乘支持向量机的初始参数.通过15个预测样本的检验,最小二乘支持向量机模型的预测平均相对误差减小至4.0%,其性能大大优于常用的经验公式以及神经网络.
Experiments for jet penetration depth measurement was carded out in 3-dimensional cold-testing bench of spout-fluid bed. It is concluded that jet penetration depth increases with the in- crease of nozzle diameter, gas density and spouted flow velocity and decreases with the increase of static bed height, particle size, particle density and gas flow rate. On the basis of experiments, rela- tion between jet penetration depth and the main designing parameters is intelligently fitted by least square support vector machine. Adaptive genetic algorithm was applied into the optimization of ini- tial parameters of least square support vector machine. Through examination of 15 forecasting sampies, the average relative error is reduced to 4. 0% by the forecasting model of least square support vector machine, which is more superior to commonly used empirical formula and neural network.
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2010年第1期139-143,共5页
Journal of Southeast University：Natural Science Edition
基金
国家重点基础研究发展计划(973计划)资助项目(2007CB210208)
国家自然科学基金资助项目(50776019)
关键词
喷动流化床
射流穿透度
最小二乘支持向量机
自适应遗传算法
spout-fluid bed
jet penetration depth
least square support vector machine
adaptive genetic algorithm
