The Synthesis of Cyclodextrin Derivatives with Ethilenic Bond and Their Application on Silk

The cyclodextrin (CD) derivatives with ethilenic bonds mono-(2-O-tosyl)-6-acryloyl-β-cyclodextrin (β-CD-2-OTs-6-A) and 6-acryloyl-β-cyclodextrin (β-CD-6-A) were respectively synthesized.β-CD-2-OTs-6-A was prepared by the reaction propylene chloride with β-CD-2-OTs.β-CD-6-A was synthesized by the reaction propylene chloride with β-CD.The most suitable pH value for β-CD-6-A grafted onto silk fabric was proved to be 5.The optimal condition for grafting β-CD-2-OTs-6-A onto silk was proved:monomer concentration of 20％ on weight of fabric ( o.w.f),potassium persulfate of 1.85％ o.w.f,liquor ratio of 1:50,pH=5,and treated 30 min at80 C.Compared with β-CD-6-A,the grafting reaction has a higher graft yield while β-CD-2-OTs-6-A acts as the monomer.     

考虑样本加权的迁移学习暂态稳定评估模型更新方法

在电力系统暂态稳定评估模型的更新过程中，针对与潜在故障相关性较小的故障样本影响迁移效果的问题，本研究从原始样本的特征量出发，发现其分布差异能反映故障之间的相关程度，由此提出考虑样本加权的迁移学习方法，进一步提高更新后评估模型的性能.首先，通过预先训练获得一个独立的域判别器，以此衡量训练模型的各故障样本相对于潜在故障的相似程度.其次，将量化后的分布差异通过密度比估计的方式进行转化，得到训练模型的各故障样本所赋予的权重大小.最后，将权重引入迁移学习更新评估模型的损失函数中，实现样本筛选.所提方法的有效性在IEEE-39节点系统和华东某区域的实际系统中均得到验证.