期刊文献+
共找到2篇文章
< 1 >
每页显示 20 50 100
基于SCADA数据和单分类简化核极限学习机的风电机组发电机状态监测 被引量:11
1
作者 金晓航 泮恒拓 +2 位作者 许壮伟 孙毅 刘伟江 《计算机集成制造系统》 EI CSCD 北大核心 2022年第8期2408-2418,共11页
为了检测风电机组发电机异常,减少由故障引起的机组停机事件的发生,提出一种基于数据采集与监控系统(SCADA)数据分析和单分类简化核极限学习机(OC-RKELM)建模的发电机状态监测方法。首先,在风电机组SCADA数据中选取与发电机健康状态相... 为了检测风电机组发电机异常,减少由故障引起的机组停机事件的发生,提出一种基于数据采集与监控系统(SCADA)数据分析和单分类简化核极限学习机(OC-RKELM)建模的发电机状态监测方法。首先,在风电机组SCADA数据中选取与发电机健康状态相关的特征,参照风电机组的性能特性同时结合局部异常因子算法对数据进行清洗;其次,利用机组历史正常的行为数据训练OC-RKELM模型,探明机组正常工作的行为规律,进而基于该模型在线监测发电机运行状态,当其工作异常时及时实现预警;最后,与其他单分类方法进行了对比分析,并利用t-SNE可视化技术对所提方法的分析结果进行了验证与评估。结果表明:OC-RKELM具有较好的健康状态监测效果,比其他单分类方法能更早发现风电机组发电机工作异常。 展开更多
关键词 风电机组 状态监测 SCADA数据 单分类模型 极限学习机
下载PDF
C-Nanostructures Cluster Models in Organic Solvents: Fullerenes, Tubes, Buds and Graphenes
2
作者 Francisco Torrens Gloria Castellano 《Journal of Chemistry and Chemical Engineering》 2013年第11期1026-1035,共10页
The existence of nanographene in cluster form is discussed in organic solvents. Theories are developed based on the columnlet, bundlet and droplet models describing the size-distribution functions. Phenomena present a... The existence of nanographene in cluster form is discussed in organic solvents. Theories are developed based on the columnlet, bundlet and droplet models describing the size-distribution functions. Phenomena present a unified explanation in the columnlet model in which free energy of Cgraphene involved in cluster is combined from a volume part proportional to the number of molecules n in cluster and a constant. The columnlet model enables describing distribution function of Cgraphene clusters by size. From purely geometrical considerations the columnlet (Cgraphene), bundlet (single-wall carbon nanotube), CNT (carbon nanotube), SWNT (single-wall C-nanotube), and carbon nanobud, CNB (carbon nanobud)) and droplet (fullerene) models predict dissimilar behaviours. The interaction-energy parameters of Cgraphene are taken from C60. An CNB behaviour or further is expected. The decay of solubility with rising temperature is smaller for Cgraphene than for SWNT and CNB and, furthermore, than for C60, in agreement with lesser numbers of units in Cgraphene clusters. The discrepancy between the experimental data of the heat of solution of fullerenes, CNTs, CNBs and graphenes is ascribed to the sharp concentration dependence of the heat of solution. The diffusion coefficient drops with temperature result greater for Cgraphene than CNB and SWNT than C60 corresponding to lesser number of units in clusters. The aggregates near (C60)13, SWNT/CNB7 and (Cgraphene)3 could be representative of the droplet, bundlet and columnlet models. 展开更多
关键词 NANOSTRUCTURE graphene solubility columnlet cluster model bundlet cluster model droplet cluster model nanobud.
下载PDF
上一页 1 下一页 到第
使用帮助 返回顶部