数据清洗、特征选择和预测模型建立是基于数据采集与监视控制系统(supervisory control and data acquisition,SCADA)数据,实现风电机组异常状态预警不可缺少的重要环节。先结合孤立森林(isolation forest,iForest)和基于密度的空间聚类...数据清洗、特征选择和预测模型建立是基于数据采集与监视控制系统(supervisory control and data acquisition,SCADA)数据,实现风电机组异常状态预警不可缺少的重要环节。先结合孤立森林(isolation forest,iForest)和基于密度的空间聚类(density-based spatial clustering of applications with noise,DBSCAN)算法对SCADA数据异常点进行有效清洗,并采用随机森林算法(random forests,RF)与Person相关系数法优选模型输入参数;再进而基于Optuna优化的类别提升树(categorical boosting,CATBoost)算法,建立风电机组正常工况齿轮箱油池温度的预测模型;然后采用滑动窗方法,构建状态评价指标,并使用区间估计理论确定油温异常状态判别的临界阈值;实现油温异常预警;最后,采用某风电机组SCADA系统油温异常的真实历史故障数据进行检验,验证了该方法的有效性。展开更多
In this work,we investigate the impact of the whole small recess offset on DC and RF characteristics of InP high electron mobility transistors(HEMTs).L_(g)=80 nm HEMTs are fabricated with a double-recessed gate proces...In this work,we investigate the impact of the whole small recess offset on DC and RF characteristics of InP high electron mobility transistors(HEMTs).L_(g)=80 nm HEMTs are fabricated with a double-recessed gate process.We focus on their DC and RF responses,including the maximum transconductance(g_(m_max)),ON-resistance(R_(ON)),current-gain cutoff frequency(f_(T)),and maximum oscillation frequency(f_(max)).The devices have almost same RON.The g_(m_max) improves as the whole small recess moves toward the source.However,a small gate to source capacitance(C_(gs))and a small drain output conductance(g_(ds))lead to the largest f_(T),although the whole small gate recess moves toward the drain leads to the smaller g_(m_max).According to the small-signal modeling,the device with the whole small recess toward drain exhibits an excellent RF characteristics,such as f_(T)=372 GHz and f_(max)=394 GHz.This result is achieved by paying attention to adjust resistive and capacitive parasitics,which play a key role in high-frequency response.展开更多
文摘数据清洗、特征选择和预测模型建立是基于数据采集与监视控制系统(supervisory control and data acquisition,SCADA)数据,实现风电机组异常状态预警不可缺少的重要环节。先结合孤立森林(isolation forest,iForest)和基于密度的空间聚类(density-based spatial clustering of applications with noise,DBSCAN)算法对SCADA数据异常点进行有效清洗,并采用随机森林算法(random forests,RF)与Person相关系数法优选模型输入参数;再进而基于Optuna优化的类别提升树(categorical boosting,CATBoost)算法,建立风电机组正常工况齿轮箱油池温度的预测模型;然后采用滑动窗方法,构建状态评价指标,并使用区间估计理论确定油温异常状态判别的临界阈值;实现油温异常预警;最后,采用某风电机组SCADA系统油温异常的真实历史故障数据进行检验,验证了该方法的有效性。
基金Supported by the Terahertz Multi User RF Transceiver System Development Project(Z211100004421012).
文摘In this work,we investigate the impact of the whole small recess offset on DC and RF characteristics of InP high electron mobility transistors(HEMTs).L_(g)=80 nm HEMTs are fabricated with a double-recessed gate process.We focus on their DC and RF responses,including the maximum transconductance(g_(m_max)),ON-resistance(R_(ON)),current-gain cutoff frequency(f_(T)),and maximum oscillation frequency(f_(max)).The devices have almost same RON.The g_(m_max) improves as the whole small recess moves toward the source.However,a small gate to source capacitance(C_(gs))and a small drain output conductance(g_(ds))lead to the largest f_(T),although the whole small gate recess moves toward the drain leads to the smaller g_(m_max).According to the small-signal modeling,the device with the whole small recess toward drain exhibits an excellent RF characteristics,such as f_(T)=372 GHz and f_(max)=394 GHz.This result is achieved by paying attention to adjust resistive and capacitive parasitics,which play a key role in high-frequency response.
