噪声信号源超磁致伸缩换能器设计方法研究

针对以往噪声信号发射换能器在实际工作中,存在着诸多影响正常科研试验的问题,并根据噪声信号的技术要求,以及探测方式的技术特征,我们对噪声发射换能器进行了重新设计,从选用超磁致伸缩材料作为换能器的材料入手,结合国内外一些新的设计理念,从材料性能分析,到理论设计,最后到可靠性实施方法,都进行了周密细致的研究,并完成了设计全过程。测试结果和使用性能表明,达到了预期的目的,这也是该材料在国内首次应用于单只宽频水声换能器中,开创了稀土超磁致伸缩材料应用的又一个先列。     

黄河中下游沿线城市水资源利用效率的时空演变及影响因素

采用超效率DEA模型、GIS空间分析方法、动态面板的系统GMM估计方法对2010—2017年黄河中下游沿线城市水资源利用效率时空格局及影响因素进行分析,结果表明:①黄河中下游沿线城市水资源利用效率整体由2010年的0.70波动下降为2017年的0.67,且区域内70%以上的城市水资源利用效率变动趋势与其一致;②黄河中下游沿线城市水资源利用效率呈"U"型区域两侧相对较高,中间区域相对较低的空间分布态势,且空间分布重心始终位于地理重心的北方;③环境规制、经济发展水平、产业结构优化、技术水平对水资源利用效率呈显著正向影响,农田水利设施建设、城镇化对水资源利用效率呈显著负向影响。研究结论可为黄河流域生态保护与高质量发展的推进提供决策支持。     

结合序贯指示模拟的贝叶斯高斯混合线性反演

高斯混合模型(Gaussian Mixture Model, GMM)可以用来描述储层性质的多峰分布特性,多峰特性主要是由于它们在不同离散变量内的变化而引起的.在高斯混合模型中,高斯分量的权值代表离散变量的概率.然而,基于高斯混合模型的贝叶斯线性反演可能会对某些点的离散变量错误地分类,进而影响连续变量的反演结果,尤其存在强噪声的时候.在本文中,我们考虑了离散变量的空间变化性,并将高斯混合模型与序贯指示模拟(Sequential Indicator Simulation, SIS)相结合来确定离散变量的后验条件权值,形成了结合序贯指示模拟的贝叶斯高斯混合线性反演方法.该方法能够准确地对离散变量进行归类,且具有良好的抗噪性.通过模型试算,我们证明了这种方法的可行性,并在实际资料中取得了较好的结果.     

基于AIS的海上渔船捕捞活动聚集区提取方法研究

船舶自动识别系统(Automatic Identification System, AIS)不仅是海上交通监管的有效工具,也为研究海上交通运输及其相关产业活动特征提供了一种良好的数据源。基于海上渔船AIS数据,本研究利用高斯混合模型(Gaussian Mixed Model,GMM)识别渔船捕捞活动状态,提出一种将核密度估计(Kernel Density Estimation,KDE)与热点分析(Hot Spot Analysis, HSA)相融合用于渔船捕捞活动聚集区提取的方法。结果显示：与单一KDE或HSA方法相比,二者相融合的方法将KDE的距离衰减效应与HSA统计指数相结合,在渔船捕捞活动聚集区提取中的应用效果较好、效率较高;采用该融合方法,基于2018年9—12月AIS数据,实现对渤海海峡周边海域渔船捕捞活动聚集区的提取,发现不同月份,渔船捕捞活动聚集区的分布范围和空间形态特征具有一定差异性,烟威近岸海域和渤海海峡是主要的捕捞活动聚集区,其结果可为该海域捕捞活动管理和海洋生态保护提供技术方法和决策支持。     

基于GMM的与文本无关的变阈值说话人确认

详细介绍了说话人确认的基本原理和主要技术方法,并通过对传统高斯混合模型(GMM)匹配方法的分析,指出了其运算量大、自适应性差的缺点,并在此基础上提出了解决方法-变阈值判决准则(DTSV).     

利用马尔柯夫随机场图模型的变化像斑类别判定方法

提出了一种利用马尔柯夫随机场(MRF)图模型获取变化像斑变化后类别的方法。用吉布斯(Gibbs)分布描述MRF图模型,通过计算最大后验概率获取变化像斑类别,以此类别为初始值,通过建立像斑类别空间关系概率矩阵,在初始结果基础上进一步判定变化像斑的类别。实验结果证实了该方法的可行性。     

利用行人轨迹挖掘城市区域功能属性

城市土地利用功能区是城市规划中的一个重要概念,遥感技术手段在城市土地利用类型识别和动态监测中取得了很大进展。然而,由于城市实际功能的复杂,往往很难从遥感影像中获得城市各个区域的社会、经济或文化等功能属性。互联网技术的发展和移动定位设备的普及,极大地便利了行人移动轨迹数据的获取。本文从行人移动规律和模式与城市功能分区之间高度相关的角度出发,通过机器学习的方法,从大量行人轨迹数据中挖掘隐含的城市功能属性与强度。该方法首先利用矢量栅格化和数学形态学方法,将城市不同等级的路网分割为互不相同的空间单元;其次,根据行人轨迹数据的时空分布特点,定义9个变量并构建高斯混合模型(Gaussian mixture model,GMM),对上述空间单元进行非监督分类,得到7种城市用地类型;随后,结合选定的60个样本区以及人为标识的6种功能区(教育用地、绿地休闲区、一般商业区、政府设施、中心商业区、住宅区),依据样本功能区GPS轨迹时间分布特征,最终对7种城市用地类型进行功能配对;最后,利用核密度估计方法进行功能区强度的可视化。该框架结合机器学习的优势,结果具有较高的准确度。     

Determinants of CO2 emission for post-Soviet Union independent countries

This article analyses the determinants of CO₂ emission for 15 post-Soviet Union independent (PSI) countries given their recent transition to market-based economies and their relatively high levels of corruption. The direct and indirect effects of economic growth on CO₂ emission for the PSI countries are derived using a multiple-equation generalized method of moment (GMM) approach to account for simultaneity among corruption, growth and CO₂ emission. A linear relationship between gross domestic product (GDP) and CO₂ emission was observed from the analysis. Furthermore, GDP influences CO₂ emission directly, but also indirectly through its impact on corruption. Similarly, corruption affects CO₂ emission directly, as well as indirectly through its impact on GDP. Political democracy and economic freedom increase CO₂ emission indirectly through their impact on economic growth. Improved energy efficiency and the EU climate policy reduce CO₂ emission, while inflows of foreign direct investment tend to increase CO₂ emission.

Policy relevance

First, PSI countries need to invest more in efficient energy technologies to mitigate CO₂ emission levels significantly. Second, PSI policies aimed at reducing deforestation (thereby increasing population density) may help mitigate carbon emission. Third, PSI countries would be well served to recognize the detrimental effects of foreign direct investment before embarking on a misguided policy path that attracts such inflows at any cost.     

Distributive justice in global climate finance – Recipients’ climate vulnerability and the allocation of climate funds

The ‘climate justice’ lens is increasingly being used in framing discussions and debates on global climate finance. A variant of such justice – distributive justice – emphasises recipient countries’ vulnerability to be an important consideration in funding allocation. The extent to which this principle is pursued in practice has been of widespread and ongoing concerns. Empirical evidence in this regard however remains inadequate and methodologically weak. This research examined the effect of recipients’ climate vulnerability on the allocation of climate funds by controlling for other commonly-identified determinants. A dynamic panel regression method based on Generalised Method of Moments (GMM) was used on a longitudinal dataset, containing approved funds for more than 100,000 projects covering three areas of climate action (mitigation, adaptation, and overlap) in 133 countries over two decades (2000–2018). Findings indicated a non-significant effect of recipients’ vulnerability on mitigation funding, but significant positive effects on adaptation and overlap fundings. ‘Most vulnerable’ countries were likely to receive higher amounts of these two types of funding than the ‘least vulnerable’ countries. All these provided evidence of distributive justice. However, the relationship between vulnerability and funding was parabolic, suggesting ‘moderately vulnerable’ countries likely to receive more funding than the ‘most vulnerable’ countries. Whilst, for mitigation funding, this observation was not a reason for concern, for adaptation and overlap fundings this was not in complete harmony with distributive justice. Paradoxically, countries with better investment readiness were likely to receive more adaptation and overlap funds. In discordance with distributive justice, countries within the Sub-Saharan Africa and South Asia regions, despite their higher climatic vulnerabilities, were likely to receive significantly less adaptation and overlap fundings. Effects of vulnerability were persistent, and past funding had significant effects on current funding. These, coupled with the impact of readiness, suggested a probable Low Funding Trap for the world’s most vulnerable countries. The overarching conclusion is that, although positive changes have occurred since the 2015 Paris Agreement, considerable challenges to distributive justice remain. Significant data and methodological challenges encountered in the research and their implications are also discussed.     

无源自适应磁流变阻尼器设计与研究

针对防震、航空航天、野战军械等领域对能源的限制,以及车辆减振中对阻尼器拉压行程不同阻尼力大小的要求,设计制作了一个无源自适应磁流变阻尼器。该阻尼器采用磁致伸缩逆效应及磁流变液的特性,把负载力的变化转化为磁流变液间隙处磁场的变化,解决了常规的磁流变阻尼器需要外配能源装置及控制电路的问题。对该阻尼器进行了建模,并进行了阻尼特性试验,结果表明:试验数据与模型符合得较好,阻尼力大小和负载有关,拉力越大,阻尼力越大,压力越大,阻尼力越小,具有负载自适应的特点,验证了设计的可行性。