The effect of vertical internal baffles on the particle mixing and graindrying characteristics in a batch fluidized bed column is investigated. Experimental work wascarried out in a 3m high rectangular fluidized bed d...The effect of vertical internal baffles on the particle mixing and graindrying characteristics in a batch fluidized bed column is investigated. Experimental work wascarried out in a 3m high rectangular fluidized bed dryer of cross sectional area of 0.15 m x 0.61 mat different operating conditions using paddy, a group D particle, as the fluidizing material. Theresults of the study showed that the fluidized bed dryer system with vertical internal baffles gavebetter particle mixing effect in the bed of particles than that without vertical internal baffles.This is due to the fact that the vertical internal baffle act as gas bubble breakers by breaking upthe large gas bubbles into smaller ones. The smaller bubbles cause a more vigorous mixing in the bedof particles before finally erupting at the bed surface. This improves the contacting efficiencyand enhanced the heat and mass transfer of the fluidized bed system. Thus a higher drying rate wasobtained in the falling rate period because the higher contacting efficiency increases theevaporation rate at the particle surface. However, the drying rate in the diffusion region showslittle improvement because the moisture diffusivity does not depend on the contacting efficiency.The fluidized bed dryer with vertical internal baffles could therefore be used in the initial rapiddrying stage in a two stage drying strategy for paddy. The insertion of vertical internal bafflesinto a fluidized bed system improves the processing of Group D particles in a fluidized bed systemespecially if the system is large in scale.展开更多
复杂产品生产数据具有高维度、不平衡的特点,为在复杂产品的生产阶段有效识别关键质量特性,及时进行质量控制,论文提出了一种基于聚类欠采样的改进随机森林算法(Random forest algorithm base on K-Means clustering under sampling,KMU...复杂产品生产数据具有高维度、不平衡的特点,为在复杂产品的生产阶段有效识别关键质量特性,及时进行质量控制,论文提出了一种基于聚类欠采样的改进随机森林算法(Random forest algorithm base on K-Means clustering under sampling,KMUS-RF),利用K-Means算法对多数样本进行聚类,并根据聚类结果进行多次欠采样形成多个平衡数据集,以随机森林为基分类器进行识别,最终根据分类过程中的特征重要性输出关键质量特性集。算例表明,KMUS-RF算法相比现有的多种分类器有良好的整体分类性能,并能显著降低复杂产品分类的第二类错误率,满足产品实际生产需求。展开更多
文摘The effect of vertical internal baffles on the particle mixing and graindrying characteristics in a batch fluidized bed column is investigated. Experimental work wascarried out in a 3m high rectangular fluidized bed dryer of cross sectional area of 0.15 m x 0.61 mat different operating conditions using paddy, a group D particle, as the fluidizing material. Theresults of the study showed that the fluidized bed dryer system with vertical internal baffles gavebetter particle mixing effect in the bed of particles than that without vertical internal baffles.This is due to the fact that the vertical internal baffle act as gas bubble breakers by breaking upthe large gas bubbles into smaller ones. The smaller bubbles cause a more vigorous mixing in the bedof particles before finally erupting at the bed surface. This improves the contacting efficiencyand enhanced the heat and mass transfer of the fluidized bed system. Thus a higher drying rate wasobtained in the falling rate period because the higher contacting efficiency increases theevaporation rate at the particle surface. However, the drying rate in the diffusion region showslittle improvement because the moisture diffusivity does not depend on the contacting efficiency.The fluidized bed dryer with vertical internal baffles could therefore be used in the initial rapiddrying stage in a two stage drying strategy for paddy. The insertion of vertical internal bafflesinto a fluidized bed system improves the processing of Group D particles in a fluidized bed systemespecially if the system is large in scale.
文摘为识别绿色能源产品太阳能光伏组件的复杂工序关键质量特性(critical quality characteristics,CTQ),提出基于状态空间算法(state space algorithm,SSA)构建质量关系模型.引入偏最小二乘算法(partial least-squares regression,PLSR),将其导入质量关系模型,进而建立复杂工序产品CTQ识别模型,最后应用变量投影重要性指标(variable importance in projection,VIP)验证提取的CTQ有效性和正确性.以太阳能光伏组件生产过程为例进行了实证分析.结果表明:新方法在克服各工序质量特性多重相关的情况下,不仅能有效地识别出光伏组件复杂工序输出的质量特性对终端产品质量的影响,还能提取出光伏组件复杂工序中关键质量特性;该方法的可行性得到了验证.
文摘复杂产品生产数据具有高维度、不平衡的特点,为在复杂产品的生产阶段有效识别关键质量特性,及时进行质量控制,论文提出了一种基于聚类欠采样的改进随机森林算法(Random forest algorithm base on K-Means clustering under sampling,KMUS-RF),利用K-Means算法对多数样本进行聚类,并根据聚类结果进行多次欠采样形成多个平衡数据集,以随机森林为基分类器进行识别,最终根据分类过程中的特征重要性输出关键质量特性集。算例表明,KMUS-RF算法相比现有的多种分类器有良好的整体分类性能,并能显著降低复杂产品分类的第二类错误率,满足产品实际生产需求。
