Multi-dimensional Seismic Response Analysis of Base-Isolated Frame Structure with 3D Isolator

INTRODUCTIONAs a new and effective method for withstandingthe horizontal component of earthquake motion on abuilding structure, the base isolation technique hascome to the engineering application phase ( Hu,1988). However, for some important buildings andfoundation facilities in an earthquake region, it is im portant to consider the effects of 3D aseismatic analy sis, including the vertical component, due to themulti dimensional characteristics of earthquakes (Suand …     

惠民凹陷沙三段岩性油藏勘探中地震技术的应用

针对惠民凹陷岩性油藏勘探存在的难题,开展了岩性油藏勘探技术研究。以地震、地质资料为基础,采用井震联合及小波变换技术,完善了层序地层等时格架建立方法,确定了各三角洲的时空展布规律。应用多元综合标定、地震相划分、正演模拟和测井约束反演一系列技术,实现了不同沉积类型有利砂体的描述。形成了适合惠民凹陷沙三段岩性油藏勘探的技术系列,可以指导惠民凹陷沙三段岩性油藏的勘探。     

基于A- CGAN的深反射地震数据随机噪声压制方法研究

基于深度学习的地震数据噪声压制方法是当前地震数据去噪处理的重要方向。深度学习方法突破了传统滤波处理的局限,在对常规地震数据的噪声压制中表现出效率高、信噪分离效果好的特点。但针对深部弱有效反射数据,当前的深度学习方法特征提取能力有限,难以取得较好的去噪效果。笔者等结合深反射地震数据特点,针对当前深度学习噪声压制方法在特征提取及对数据集依赖上的局限,提出了基于注意力循环生成对抗网络(Attention Cycle- Consistent Generative Adversarial Networks,A- CGAN)的深反射地震数据随机噪声压制方法。借助循环一致生成对抗网络(Cycle- Consistent Generative Adversarial Networks,Cycle- GAN)的域映射思想,降低对数据集的要求。为了构建适用于深反射地震数据的去噪网络,从3个方面对Cycle- GAN进行改进：在Cycle- GAN的生成器（去噪器）中加入残差结构和注意力机制,用于加深网络深度和提高其特征提取能力;在Cycle- GAN的鉴别器中使用块判决,提高鉴别精度和准确度;在损失函数部分加入感知一致性损失函数,提升网络模型恢复纹理细节信息的能力。通过合成地震数据和实际深反射地震数据测试,验证了优化算法的有效性,体现了良好的应用价值。     

夜间灯光遥感对城市发展类动能与相似性评估

评估城市发展水平与制定城市规划和城市发展政策息息相关.已有研究表明人口、国内生产总值(GDP)等统计数据和夜间灯光遥感数据能够用来衡量城市发展水平,但大多研究只关注城市发展总量,忽视了发展速度对城市发展的影响.因此,本文基于2012年—2019年Suomi NPP-VIIRS(Suomi National Polar-...     

改进的能量守恒方法及其在CHAMP重力场恢复中的应用

An efficient method for gravity field determination from CHAMP orbits and accelerometer data is referred to as the energy balance approach. A new CHAMP gravity field recovery strategy based on the improved energy balance approach IS developed in this paper. The method simultaneously solves the spherical harmonic coefficients, daily Integration constant, scale and bias parameters. Two 60 degree and order gravitational potential models, XISM-CHAMPO1S from the classical energy balance approach, and XISM-CHAMPO2S from the improved energy balance, are determined using about one year＇s worth of CHAMP kinematic orbits from TUM and accelerometer data from GFZ. Comparisons among XISM-CHAMPO1S, XISM-CHAMPO2S, EIGEN-CGO3C, EIGEN-CHAMPO3S, EIGEN2, ENIGNIS and EGM96 are made. The results show that the XISM-CHAMPO2S model is more accurate than EGM96, EIGENIS, EIGEN2 and XISM-CHAMPO1S at the same degree and order, and has almost the same accuracy as EIGEN-CHAMPO3S.