Analytical convolution model for shipping water evolution on a fixed structure

The shipping of water is a problem that affects naval and offshore structures. Estimating its propagation on the decks of these structures by using analytical methods has been a main concern of projects. However, classical approaches disregard the decay tendency of water elevation time series and tend to overestimate the resultant water on deck. This paper is concerned with estimating the evolution of water along the deck of a fixed structure due to shipping water events. An analytical convolution model is proposed to estimate water elevations. The model considers the freeboard exceedance time series and the mean shipping flow velocity as inputs and the frictional effects of the bottom by resistance coefficients, which enables an approximated representation of the water elevation time series over the deck. It was validated with experiments of isolated shipping water events that were generated with the wet dam-break approach. The results obtained with the proposed model captured the experimental results, approximating the peak values and the decay trend of time series. Improvement of the proposed approach over classical models to represent shipping water elevations was demonstrated by comparing the results obtained with those of the dam-break model of Stoker.     

降雨对地电阻率干扰的分析

在一些地电阻率观测台站受本身台址条件的限制，观测深度较浅，因此，观测值受到降雨等气象因素的影响。本文选取昌黎台ＥＷ向１９８３－１９９４年观测资料月均值，用褶积滤波和多元回归方法对ρｓ进行了降雨校正，将将之作为一个灰箱系统，将资料外推至１９９５４上，然后与原始观测值作残差分析，取得了一定效果。     

用褶积积分和低通滤波方法处理沂水泉氡观测资料

多年来的观测分析表明,影响沂水泉氧值变化的最大干扰是降雨。由于这种干扰存在着滞后效应,所以用一般相关分析难以将这种影响排除干净。本文采用褶积积分方法较好地排除了降雨对沂水泉氧观测的影响。并用低通滤波方法,探讨了沂水泉氧的长趋势变化。     

一维边缘检测算子的递归算法

本文对边缘检测理论和算法作了理论和实验方面的深入研究，选用Ｄｅｒｉｃｈｅ依据Ｃａｎｎｙ设计思想导出的边缘检测算子，证明了递归算法代数卷积算法实现边缘检测的基本原理，导出了几种算子逆归算法的实用公式，实验证明递归算法比卷积算法实现边缘提取在速度方面有着明显的提高，为一维影匹配中数字影像特征提取提供了快速而实用的边缘检测方法。     

地震油气储层的小样本卷积神经网络学习与预测

地震储层预测是油气勘探的重要组成部分,但完成该项工作往往需要经历多个环节,而多工序或长周期的研究分析降低了勘探效率.基于油气藏分布规律及其在地震响应上所具有的特点,本文引入卷积神经网络深度学习方法,用于智能提取、分类并识别地震油气特征.卷积神经网络所具有的强适用性、强泛化能力,使之可以在小样本条件下,对未解释地震数据体进行全局优化提取特征并加以分类,即利用有限的已知含油气井段信息构建卷积核,以地震数据为驱动,借助卷积神经网络提取、识别蕴藏其中的地震油气特征.将本方案应用于模型数据及实际数据的验算,取得了预期效果.通过与实际钻井信息及基于多波地震数据机器学习所预测结果对比,本方案利用实际数据所演算结果与实际情况有较高的吻合度.表明本方案具有一定的可行性,为缩短相关环节的周期提供了一种新的途径.     

基于偏微分方程的模型光滑处理在偏移成像中的应用

本文针对射线类偏移成像当中的速度模型光滑处理问题,借鉴数字图像处理当中的偏微分方程法,基于能量泛函,应用变分方法导出基于速度模型的偏微分方程实现射线类偏移成像当中的速度模型的光滑处理.由于偏微分方程法具有线性叠加特性、模型解的唯一性和局部特征保持性,因此,应用该算法可以实现基于原始速度模型空间结构的模型光滑处理.通过在原始速度模型以及光滑处理后的速度模型上计算速度的空间分布以及地震波走时、射线路径可以得出,偏微分方程法对速度模型的光滑处理能够很好地保持原始模型的空间结构,偏移成像结果也证明了该方法的实用性.     

Measurement and modeling of bed shear stress under solitary waves

Direct measurements of bed shear stresses (using a shear cell apparatus) generated by non-breaking solitary waves are presented. The measurements were carried out over a smooth bed in laminar and transitional flow regimes (~ 10⁴ < R_e < ~ 10⁵). Measurements were carried out where the wave height to water depth (h/d) ratio varied between 0.12 and 0.68; maximum near bed velocity varied between 0.16 m/s and 0.51 m/s and the maximum total shear stress (sum of skin shear stress and Froude–Krylov force) varied between 0.386 Pa and 2.06 Pa. The total stress is important in determining the stability of submarine sediment and in sheet flow regimes. Analytical modeling was carried out to predict total and skin shear stresses using convolution integration methods forced with the free stream velocity and incorporating a range of eddy viscosity models. Wave friction factors were estimated from skin shear stress at different instances over the wave (viz., time of maximum positive total shear stress, maximum skin shear stress and at the time of maximum velocity) using both the maximum velocity and the instantaneous velocity at that phase of the wave cycle. Similarly, force coefficients obtained from total stress were estimated at time of maximum positive and negative total stress and at maximum velocity. Maximum positive total shear stress was approximately 1.5 times larger than minimum negative total stress. Modeled and measured positive bed shear stresses are well correlated using the best convolution model, but the model underestimates the data by about 4%. Friction factors are dependent on the choice of normalizing using the maximum velocity, as is conventional, or the instantaneous velocity. These differ because the stress is not in phase with the velocity in general. Friction factors are consistent with previous data for monochromatic waves, and vary inversely with the square-root of the Reynolds number. The total shear stress leads the free stream fluid velocity by approximately 50°, whereas the skin friction shear stress leads by about 30°, which is similar to that reported by earlier researchers.     

高分辨率非线性储层物性参数反演方法和应用

对于陆相沉积环境下的复杂隐蔽岩性储层,由于观测信息不准确,如信息重叠、信息缺失和噪音污染,以及岩石物理关系模糊等原因,储层横向预测存在不惟一性、不稳定性和不确定性.基于线性假定的常规储层横向预测技术已不适用于复杂隐蔽岩性储层的勘探.本文采用一种非线性储层岩性物性褶积模型,建立波阻抗与孔隙度/泥质含量的函数关系;通过多级结构分解和双向边沿子波检测来刻画复杂岩石物理关系;通过Caianiello褶积神经网络实现确定性反演、统计反演和非线性理论三者有机结合;最后联合应用基于逆算子的反演方法和基于正算子的重建算法实现了综合地质、测井和地震波阻抗信息进行高分辨率储层物性参数反演.非线性储层物性参数反演采用多井约束机制和分频反演方式,在陆相和近海油气勘探资料的实际应用中,取得了明显应用效果.     

PZ-600型全身CT机反投影原理及其框图

目前大部分CT机常用的成像方法仍是卷积反投影法。对于在医院工作的某些工程技术人员来说是很想了解该方法在计算上是如何实现的。笔者在对我院 PZ-600型 CT 机的有关资料进行了分析和整理后,试图较直观地说明 PZ-600型 CT 机反投影成像的具体实现过程。PZ-600型机也有专用硬件,本文介绍其硬件框图。     

用CT技术反演面波Q值空间分布及其研究前景(英文)

本文运用CT技术反演地壳面波品质因数Q值空间分布,采用在医学上X—CT技术中的卷积反投影方法,引用 Radon 变换反演和图像重建方法来拟合地壳面波 Q 值的空间分布。由600次地震实际的2200条地震波射线的面波振幅、周期资料,进行实际地壳 Q 值反演,首先获得四川盆地附近:北纬26°—34°,东经103°—112°范围的面波 Q 值分布,作出该区 Q=210-255的等值线图,其平均 Q 值为225,在四川盆地中部面波 Q 值偏高,这可能与该盆地中部沉积物介质比较均匀,非弹性成分较少有关。然后以上述经纬度范围的四川盆地为基地,将面波扫描的方框分别向北、向南、向西、向东移动各2度,再做四次 Q 值成像,取其中间重迭部位,计算 Q 值的平均值及其残差,得每经纬度为半度的像素中观测误差的空间分布图。求得该范围的平均误差为8%。进一步将平面范围地震面波 Q 值成象扩大成球面上的投影成像:系将上述单框 Q 值图像沿经纬度方向各移动2度进行叠加,使若干个平面小单框复盖在预想的球面上的大区域上,除去边缘部分之外,在该范围的经纬度网上都叠加有一个以上的 Q 值数据,通过计算每个网点上的平均 Q 值及其标准误差的等值线,可获得球面上大范围的 Q 值成象图。得到了7×7框范围内的面波 Q 值分布图。本文探讨了 Q值成像在改进面波震级标度和解释地震震级的路径校正值以及为地震监测预报服务的研究前景。