位移预测是边坡地质灾害监测预警的关键,本文以温州绕城高速公路边坡为例,提出了一种新的基于图深度学习与北斗监测的边坡多因子位移预测方法。首先基于北斗高精度监测点位的空间位置对整体监测体系的图结构进行建模,构建图节点之间的...位移预测是边坡地质灾害监测预警的关键,本文以温州绕城高速公路边坡为例,提出了一种新的基于图深度学习与北斗监测的边坡多因子位移预测方法。首先基于北斗高精度监测点位的空间位置对整体监测体系的图结构进行建模,构建图节点之间的邻接矩阵。再对北斗高精度位移、降雨量、地下水位与土壤含水率多因子监测数据进行去粗差、插值与归一化等时序数据处理,并进行时空相关性分析,结果表明位移主要受连续两个月的降雨量、三级边坡的地下水位与土壤含水率的影响。将最先进的基于图深度学习的GTS(Graph for Time Series)预测模型引入边坡位移预测中,提出适用于北斗高精度边坡变形监测的GTS-BDS位移预测模型。当预测时长为1 h时,其均方根误差(RMSE)、平均绝对误差(MAE)与平均绝对百分比误差(MAPE)指标评价分别达到0.301、0.154与3.5%,均优于LSTM与T-GCN等模型。本文所提出的位移预测方法充分利用了北斗高精度及其他传感器监测点位之间的空间拓扑与监测数据的时序特征,从整体监测体系的角度提升边坡位移预测的准确率与可靠性,在边坡安全预警中具有良好的应用前景。展开更多
An accurate measurement of the refractive index is necessary for the optical design of both deep ultraviolet laser diodes and light-emitting diodes. Generally, the refractive indices along different crystallographic a...An accurate measurement of the refractive index is necessary for the optical design of both deep ultraviolet laser diodes and light-emitting diodes. Generally, the refractive indices along different crystallographic axes of anisotropic thin films are measured using variable angle spectroscopic ellipsometry.However, there are still some limitations concerning this method. Here we proposed a potential method to measure the band edge refractive index of wide bandgap semiconductor. An aperiodic oscillation emission phenomenon due to the Fabry-Perot effect was observed in the fluorescence spectrum of an AlN film with a thickness of 1500 nm. Based on the characteristics of the fluorescence spectrum and the definition of Fabry-Perot effect, we obtained the ordinary refractive index of the AlN thin film near the band edge directly. This refractive index measurement method is a supplement to the variable angle ellipsometry,and it is a more direct and effective method for transferred film and thinner samples to measure the fluorescence spectrum.展开更多
文摘位移预测是边坡地质灾害监测预警的关键,本文以温州绕城高速公路边坡为例,提出了一种新的基于图深度学习与北斗监测的边坡多因子位移预测方法。首先基于北斗高精度监测点位的空间位置对整体监测体系的图结构进行建模,构建图节点之间的邻接矩阵。再对北斗高精度位移、降雨量、地下水位与土壤含水率多因子监测数据进行去粗差、插值与归一化等时序数据处理,并进行时空相关性分析,结果表明位移主要受连续两个月的降雨量、三级边坡的地下水位与土壤含水率的影响。将最先进的基于图深度学习的GTS(Graph for Time Series)预测模型引入边坡位移预测中,提出适用于北斗高精度边坡变形监测的GTS-BDS位移预测模型。当预测时长为1 h时,其均方根误差(RMSE)、平均绝对误差(MAE)与平均绝对百分比误差(MAPE)指标评价分别达到0.301、0.154与3.5%,均优于LSTM与T-GCN等模型。本文所提出的位移预测方法充分利用了北斗高精度及其他传感器监测点位之间的空间拓扑与监测数据的时序特征,从整体监测体系的角度提升边坡位移预测的准确率与可靠性,在边坡安全预警中具有良好的应用前景。
文摘为解决湖南某纸业公司260 t/h循环流化床锅炉燃用运行燃料时飞灰和炉渣含碳量高、排烟温度偏高以及热效率偏低的问题,针对锅炉特点,采用单因素变化对比试验方法,考虑流化风量、氧量、一二次风配比、床温和污泥掺烧等方面,选取19种工况进行燃烧调整试验,以优化锅炉燃烧,提高热效率。试验结果表明:流化风量、床温、过量空气系数和一二次风配比对锅炉燃烧效率影响明显;当流化风量控制在100000~110000 m 3/h,氧量为3%,一二次风配比为54∶46,床层温度为940℃时,固体未完全燃烧损失可降低约1.2%,排烟损失降低约0.9%,锅炉整体热效率提升约2%;锅炉掺烧造纸污泥量以3.5 t/h为宜,可保证锅炉蒸汽参数和效率,每月可节省约60万元燃料成本,具有较好的经济性。
基金This work was financially supported by the National Natural Science Foundation of China(91333207,61427901,61604178,91833301 and U1505252).
文摘An accurate measurement of the refractive index is necessary for the optical design of both deep ultraviolet laser diodes and light-emitting diodes. Generally, the refractive indices along different crystallographic axes of anisotropic thin films are measured using variable angle spectroscopic ellipsometry.However, there are still some limitations concerning this method. Here we proposed a potential method to measure the band edge refractive index of wide bandgap semiconductor. An aperiodic oscillation emission phenomenon due to the Fabry-Perot effect was observed in the fluorescence spectrum of an AlN film with a thickness of 1500 nm. Based on the characteristics of the fluorescence spectrum and the definition of Fabry-Perot effect, we obtained the ordinary refractive index of the AlN thin film near the band edge directly. This refractive index measurement method is a supplement to the variable angle ellipsometry,and it is a more direct and effective method for transferred film and thinner samples to measure the fluorescence spectrum.
