Mechanical faults of high voltage circuit breakers(CBs)seriously affect the reliability of their operation,which may cause severe damage to power systems.In order to monitor operational conditions and detect mechanica...Mechanical faults of high voltage circuit breakers(CBs)seriously affect the reliability of their operation,which may cause severe damage to power systems.In order to monitor operational conditions and detect mechanical faults of CBs,a multi-parameter monitoring system is designed and a fault diagnosis method based on multi-mapping is proposed.The paper focuses on the trip/close circuits,the spring-charging mechanism and the transmission mechanism,and obtains four current signals and a vibration signal that can reflect CB conditions.For the current signals,a morphological filter is used to remove noise and then characteristics of the waveforms’shape information are extracted.For vibration signals,the wavelet packet transform is used to decompose the signal into various frequency bands,and the sample entropy of the low frequency bands and the wavelet energy of the high frequency bands are calculated,respectively.Based on these feature parameters,a multi-mapping strategy is proposed for CB fault diagnosis.Laboratory experiments have been conducted to obtain on-site signals under various conditions,and experiment results have verified that monitoring the aforementioned signals and using the corresponding feature extraction and fault diagnosis methods,the mechanical faults of high voltage CBs can be effectively diagnosed.展开更多
基于高通量测序的RNA-Seq(RNA-sequencing)是用于转录组研究的一种新技术,针对该技术在转录组表达分析研究中存在的读段多源映射和读段非均匀分布等难点,提出一个改进的转录组表达研究方法 LDASeqII(Improvement of latent Dirichlet al...基于高通量测序的RNA-Seq(RNA-sequencing)是用于转录组研究的一种新技术,针对该技术在转录组表达分析研究中存在的读段多源映射和读段非均匀分布等难点,提出一个改进的转录组表达研究方法 LDASeqII(Improvement of latent Dirichlet allocation for sequencing data)。模型利用剪接异构体结构信息对参数进行约束并进行外显子读段数目归一化处理,解决了读段非均匀分布下的多源映射问题。通过引入"伪外显子"和"伪转录本"分别处理接合区读段和噪声读段。将模型应用到真实数据集上,并与原LDASeq(Latent Dirichlet allocation for sequencing data)模型和目前流行的Cufflinks与RSEM(RNA-Seq by expectation maximization)方法进行对比。结果显示,改进方法获得了更为准确的转录本及基因表达水平计算结果。展开更多
近年来,基于单分子测序技术的ISO-seq数据以其超长读段长度被越来越多地应用于转录组新型异构体预测研究,但目前大多数研究工作只用到全长读段数据,丢失了非全长读段数据中较多有用信息,因而数据没有得到充分利用。针对这一问题,本文在...近年来,基于单分子测序技术的ISO-seq数据以其超长读段长度被越来越多地应用于转录组新型异构体预测研究,但目前大多数研究工作只用到全长读段数据,丢失了非全长读段数据中较多有用信息,因而数据没有得到充分利用。针对这一问题,本文在保留非全长读段的基础上提出了两个能同时预测异构体结构和计算其表达比例的模型基于狄利克雷采样的异构体探测与预测(Dirichletsampling for isoform detection and prediction,DSIDP)和基于马尔科夫链的异构体探测与预测(Markovchain for isoform detection and predition,MCIDP)。两个模型均从全长读段中建立异构体预测集,并采用全长读段和非全长读段计算异构体表达比例。DSIDP将所有读段比对至异构体预测集,并使用Dirichlet采样解决多源映射问题,MCIDP使用马尔科夫链模拟基因外显子之间的选择性剪切,该模型还能预测出数据中没有全长读段的异构体。本文采用模拟数据和真实数据验证了两个模型的有效性。展开更多
基金This work is supported by the Guangdong Innovative Research Team Program(No.201001N0104744201)the Fundamental Research Funds for the Central Universities,SCUT(No.2015ZZ019),China.
文摘Mechanical faults of high voltage circuit breakers(CBs)seriously affect the reliability of their operation,which may cause severe damage to power systems.In order to monitor operational conditions and detect mechanical faults of CBs,a multi-parameter monitoring system is designed and a fault diagnosis method based on multi-mapping is proposed.The paper focuses on the trip/close circuits,the spring-charging mechanism and the transmission mechanism,and obtains four current signals and a vibration signal that can reflect CB conditions.For the current signals,a morphological filter is used to remove noise and then characteristics of the waveforms’shape information are extracted.For vibration signals,the wavelet packet transform is used to decompose the signal into various frequency bands,and the sample entropy of the low frequency bands and the wavelet energy of the high frequency bands are calculated,respectively.Based on these feature parameters,a multi-mapping strategy is proposed for CB fault diagnosis.Laboratory experiments have been conducted to obtain on-site signals under various conditions,and experiment results have verified that monitoring the aforementioned signals and using the corresponding feature extraction and fault diagnosis methods,the mechanical faults of high voltage CBs can be effectively diagnosed.
文摘基于高通量测序的RNA-Seq(RNA-sequencing)是用于转录组研究的一种新技术,针对该技术在转录组表达分析研究中存在的读段多源映射和读段非均匀分布等难点,提出一个改进的转录组表达研究方法 LDASeqII(Improvement of latent Dirichlet allocation for sequencing data)。模型利用剪接异构体结构信息对参数进行约束并进行外显子读段数目归一化处理,解决了读段非均匀分布下的多源映射问题。通过引入"伪外显子"和"伪转录本"分别处理接合区读段和噪声读段。将模型应用到真实数据集上,并与原LDASeq(Latent Dirichlet allocation for sequencing data)模型和目前流行的Cufflinks与RSEM(RNA-Seq by expectation maximization)方法进行对比。结果显示,改进方法获得了更为准确的转录本及基因表达水平计算结果。
文摘近年来,基于单分子测序技术的ISO-seq数据以其超长读段长度被越来越多地应用于转录组新型异构体预测研究,但目前大多数研究工作只用到全长读段数据,丢失了非全长读段数据中较多有用信息,因而数据没有得到充分利用。针对这一问题,本文在保留非全长读段的基础上提出了两个能同时预测异构体结构和计算其表达比例的模型基于狄利克雷采样的异构体探测与预测(Dirichletsampling for isoform detection and prediction,DSIDP)和基于马尔科夫链的异构体探测与预测(Markovchain for isoform detection and predition,MCIDP)。两个模型均从全长读段中建立异构体预测集,并采用全长读段和非全长读段计算异构体表达比例。DSIDP将所有读段比对至异构体预测集,并使用Dirichlet采样解决多源映射问题,MCIDP使用马尔科夫链模拟基因外显子之间的选择性剪切,该模型还能预测出数据中没有全长读段的异构体。本文采用模拟数据和真实数据验证了两个模型的有效性。
