针对电力机车牵引变流器中故障率最高的单相脉宽调制(pulse width modulation,PWM)整流器,提出一种流形学习算法融合多域特征的故障诊断方法。根据整流器在不同工作状态下的时域、频域和时频域特征构建多域特征向量;采用Hessian局部线...针对电力机车牵引变流器中故障率最高的单相脉宽调制(pulse width modulation,PWM)整流器,提出一种流形学习算法融合多域特征的故障诊断方法。根据整流器在不同工作状态下的时域、频域和时频域特征构建多域特征向量;采用Hessian局部线性嵌入(Hessian local linear embedding,HLLE)算法融合多域特征,根据故障样本数和聚类结果,解决高维数据中固有维数和最近邻数选取困难的问题,得到用于描述故障特征的最优低维特征向量,减少特征之间的冲突和冗余;采用支持向量机进行模式识别,实现对整流器的故障诊断。结果表明:对不同的输出电压,不同的训练和测试比,15种故障模式均具有较高的诊断率。与其他方法相比,本文方法具有较好的融合效果和较强的鲁棒性。展开更多
Water is an important component in agricultural production for both yield quantity and quality. Although all weather conditions are driving factors in the agricultural sector, the precipitation in rainfed agriculture ...Water is an important component in agricultural production for both yield quantity and quality. Although all weather conditions are driving factors in the agricultural sector, the precipitation in rainfed agriculture is the most limiting weather parameter. Water deficit may occur continuously over the total growing period or during any particular growth stage of the crop. Optical remote sensing is very useful but, in cloudy days it becomes useless. Radar penetrates the cloud and collects information through the backscattering data. Normalized Difference Vegetation Index (NDVI) was extracted from Landsat 8 satellite data and used to calculate Crop Coefficient (Kc). The FAO-Penman-Monteith equation was used to calculate reference evapotranspiration (ETo). NDVI and Land Surface Temperature (LST) were calculated from satellite data and integrated with air temperature measurements to estimate Crop Water Stress Index (CWSI). Then, both CWSI and potential crop evapotranspiration (ETc) were used to calculate actual evapotranspiration (ETa). Sentinel-1 radar data were calibrated using SNAP software. The relation between backscattering (dB) and CWSI was an inverse relationship and R2 was as high as 0.82.展开更多
文摘针对电力机车牵引变流器中故障率最高的单相脉宽调制(pulse width modulation,PWM)整流器,提出一种流形学习算法融合多域特征的故障诊断方法。根据整流器在不同工作状态下的时域、频域和时频域特征构建多域特征向量;采用Hessian局部线性嵌入(Hessian local linear embedding,HLLE)算法融合多域特征,根据故障样本数和聚类结果,解决高维数据中固有维数和最近邻数选取困难的问题,得到用于描述故障特征的最优低维特征向量,减少特征之间的冲突和冗余;采用支持向量机进行模式识别,实现对整流器的故障诊断。结果表明:对不同的输出电压,不同的训练和测试比,15种故障模式均具有较高的诊断率。与其他方法相比,本文方法具有较好的融合效果和较强的鲁棒性。
文摘Water is an important component in agricultural production for both yield quantity and quality. Although all weather conditions are driving factors in the agricultural sector, the precipitation in rainfed agriculture is the most limiting weather parameter. Water deficit may occur continuously over the total growing period or during any particular growth stage of the crop. Optical remote sensing is very useful but, in cloudy days it becomes useless. Radar penetrates the cloud and collects information through the backscattering data. Normalized Difference Vegetation Index (NDVI) was extracted from Landsat 8 satellite data and used to calculate Crop Coefficient (Kc). The FAO-Penman-Monteith equation was used to calculate reference evapotranspiration (ETo). NDVI and Land Surface Temperature (LST) were calculated from satellite data and integrated with air temperature measurements to estimate Crop Water Stress Index (CWSI). Then, both CWSI and potential crop evapotranspiration (ETc) were used to calculate actual evapotranspiration (ETa). Sentinel-1 radar data were calibrated using SNAP software. The relation between backscattering (dB) and CWSI was an inverse relationship and R2 was as high as 0.82.
