欧洲港口集装箱运输前景光明

随着经济一体化的整体推进,欧洲港口集装箱运输进入了良性发展轨道。据英国远洋航运咨询有限公司(OSC)预测,在未来12年内,欧洲港口集装箱吞吐量将为现在的两倍以上,其中支线运输量增长最快。 据预测,欧洲港口集装箱吞吐量将从1996年的     

基于随机森林的复坡堤越浪量预测研究

针对复坡堤越浪量的计算问题,提出了采用随机森林算法预测越浪量的方法。首先,通过对欧洲CLASH数据集进行筛选,挑选出符合复坡堤越浪量预测的数据;其次,对数据做无量纲化处理,建立以随机森林为基础的复坡堤越浪量预测模型,并通过网格搜索(GridSearchCV)方法对模型进行调参以改善模型的性能;最后,利用决定系数R~2来评估模型的精度,并将随机森林模型与集成神经网络模型做了预测能力的对比,同时还给出了随机森林模型各个特征参数对预测精度的重要性。结果显示,随机森林模型的决定系数为92.7%,集成神经网络模型的决定系数为87.7%,表明随机森林模型对越浪量具有更强的学习和预测能力。通过对特征重要性的分析,墙顶高程对模型预测精度的影响最大,堤顶高程次之,堤脚宽度影响最小。     

基于BP神经网络的公交动态行程时间预测方法研究

公交行程时间的精确预测对于提升公交吸引力具有重要意义。本文基于公交车到离站的历史数据,综合考虑时间周期、站点、站间距离、天气等多个因素,建立了基于BP神经网络的公交车静态行程时间预测模型,以该模型为基础,采用动态迭代的方法,叠加多个站间行程时间预测结果,进一步构建了面向连续站点的公交车动态行程时间预测模型,实现对跨越多个站点的公交行程时间预测。以青岛市125路公交为例对算法进行测试。在模型的横向对比实验中,本模型预测结果的绝对误差均在50 s以内,平均绝对误差百分比(MAPE)为11.74%,均方根误差(RMSE)为23.15,R2的确定系数为0.905 1,SVM的MAPE、RMSE、R2 误差指标分别为:12.38%、38.33、0.743 6,LR对应的误差指标分别为:12.50%、25.59、0.884 1;在静态模型与动态模型的对比实验中,动态模型预测结果的MAPE为11.75%,RMSE为23.15,静态模型对应误差指标分别为:11.63%、26.74。研究结果表明,基于BP神经网络的公交动态行程时间预测模型比传统的静态预测方法具有更高的预测精度。     

利用输沙量预测现代黄河三角洲的面积增长

依据１９７６－１９８９年黄河亚三角洲的面积数据及其相应年份的利津站黄河来沙量，利用单输入，单输出的线性系模型对黄河三角洲的面积增长进行动态预测研究。结果表明，这一线性系统模型能较好地反映黄河三角洲的面积增长规律，且优于灰色系统的ＧＭ（１，１）模型。     

港口堆场规模与集装箱吞吐量关系研究——以北仑港为例

针对目前尚不完善的集装箱堆场规划问题,讨论了港口堆场规模与集装箱吞吐量的关系,按堆场类型,分别研究了前方堆场规模、后方堆场规模与集装箱吞吐量的关系,同时考虑了堆场的工作时间、集装箱堆箱层数以及各种不平衡系数对集装箱堆场规模的影响,给出了前后方堆场规模与吞吐量等因素的关系式,对目前港口堆场规划具有一定参考意义。以北仑港为例,首先建立数学模型得出了目前北仑港各港区的堆场需求规模,再将结果与现状情况对比,作出分析,得出目前北仑港的堆场规模满足各港区的集装箱疏散,论证了数学模型的合理性。最后对未来北仑港各港区堆场规模进行了预测,并对未来北仑港集装箱发展提出了建议。     

海南省海洋经济发展趋势分析

对快速增长的海洋经济做出可靠的预测,可以深化对海洋经济发展规律的认识,对确定海南省海洋经济增长前景和目标、制定海洋开发战略具有重要的实践和参考意义。文章在对比不同预测方法的基础上,基于现有的数据基础,遴选出灰色系统模型与时间序列模型作为预测工具,对海南省海洋生产总值进行预测。研究结果表明:①与时间序列模型相比,灰色系统模型在海南省海洋经济预测方面更为有效。②目前海南省海洋经济正处于成长期,未来10年仍将保持快速增长。③根据预测结果,到2025年海南省海洋生产总值将达到3 340亿元,年均增长率约为12.5%。     

三参数模型法地层压力预测及应用

针对目前地层压力不能够精确预测的问题,提出了利用测井声波资料配合相关测井曲线预测地层压力的三参数模型法,该方法考虑了与地层声速相关的3个主要因素,即泥质含量、孔隙度、有效应力。将该方法应用于济阳坳陷东营凹陷,通过实测压力系数与预测压力系数的对比分析,结果表明:利用三参数模型法预测地层压力在提高地层压力预测精度方面具有一定的有效性与合理性。     

浮游植物密度的人工神经网络预测研究

将人工神经网络模型应用于藻类密度数据的预测计算之中,并利用遗传算法对其网络结构进行优化计算以保证计算结果的准确性且自动确定网络结构,分别对神经网络的“当天模型”和“预测模型”进行了计算。结果显示：几种人工神经网络模型在计算精度以及预测数据的趋势上都有较好的效果,目前国内学者使用的人工神经网络“当天模型”无法对其后数据进行预测,不能起到实际预测的作用,而经过遗传算法优化后的人工神经网络“预测模型”不仅达到了很好的预测效果,而且网络结构简单,适用于浮游植物密度的预测计算。     

基于GM(1,1)模型的海洋经济发展预测研究

文章采用灰色预测法,选取2001—2016年我国海洋生产总值作为基础数据,建立海洋经济发展预测模型即GM(1,1)模型;运用残差检验、关联度检验和后验差检验对该模型精度进行检验,同时运用该模型对已知年份数据进行预测并将预测值与实际值相比较,结果表明模型预测精度较高,对于短期预测具有可信度;在此基础上,运用GM(1,1)模型计算和预测2017—2021年我国海洋生产总值,得出我国海洋经济将保持稳步增长的结论。     

海洋内波发生概率的模糊逻辑推理预测模型

海洋内波发生的诊断及预测,因其生成机制的复杂性和实际观测资料的缺乏而被海洋学界公认为一个难题。为此,提出了一个基于模糊逻辑思想的内波发生概率诊断与预测新方法,并利用该方法建立了内波发生概率预测模型。从实际海洋水文资料中提取出影响内波发生的8个海洋环境因子作为模型的输入,内波发生概率作为模型的输出。最后将预测模型在西北太平洋不同海域进行了应用。结果表明,该模型可为海洋工程和舰船航行中规避和防范内波提供参考。     

长周期动力形态模型中地形演变方法探讨

PRD-LTMM是一个模拟和描述三角洲地形中、长期演变过程的动力形态模型.在长时间尺度的模拟过程中,必须从模型输入、模型物理系统本身到模型输出等各个角度考虑模拟约减处理.代表性输入条件是实现模型约减的基础,而地形时间的步长延长是提高长周期模型模拟效率的关键.延长技术通常有直接推延、时间中心推延、潮汐延展及泥沙输运率公式线性化等四种方法.通过系列敏感性试验,对PRD-LTMM模型选择潮汐延展技术来进行地形时间步长的延长处理,推延的时间长度为10 a,基于此条件下的PRD-LTMM在长周期模拟中不仅稳定性较好,而且结果也较精确.     

An Improved Method of Predicting Drag Anchor Trajectory Based on the Finite Element Analyses of Holding Capacity

Drag anchor is one of the most commonly used anchorage foundation types. The prediction of embedded trajectory in the process of drag anchor installation is of great importance to the safety design of mooring system. In this paper, the ultimate anchor holding capacity in the seabed soil is calculated through the established finite element model, and then the embedded motion trajectory is predicted applying the incremental calculation method. Firstly, the drag anchor initial embedded depth and inclination angle are assumed, which are regarded as the start embedded point. Secondly, in each incremental step, the incremental displacement of drag anchor is added along the parallel direction of anchor plate, so the displacement increment of drag anchor in the horizontal and vertical directions can be calculated. Thirdly, the finite element model of anchor is established considering the seabed soil and anchor interaction, and the ultimate drag anchor holding capacity at new position can be obtained. Fourthly, the angle between inverse catenary mooring line and horizontal plane at the attachment point at this increment step can be calculated through the inverse catenary equation. Finally, the incremental step is ended until the angle of drag anchor and seabed soil is zero as the ultimate embedded state condition, thus, the whole embedded trajectory of drag anchor is obtained. Meanwhile, the influences of initial parameter changes on the embedded trajectory are considered. Based on the proposed method, the prediction of drag anchor trajectory and the holding capacity of mooring position system can be provided.     

基于最小二乘和卡尔曼滤波方法进行原子时预报的研究

介绍和分析了最小二乘和卡尔曼滤波方法在时间预报中的应用。通过IGS站提供的钟差数据，分别运用这两种方法对其中六个原子钟进行了时间预报的实验。通过对预报结果进行分析，结论证明为了取得较好的预报效果，不同的预报方法和钟参数的模型对于观测数据的要求有一些差别。从实验数据所成图形来看，当采用一天的观测数据进行模型预报时，最小二乘法的预报精度比卡尔曼滤波法稍高一些。     

基于多种神经网络的风暴潮增水预测方法的比较分析

简要介绍了利用BP神经网络、小波神经网络、递归神经网络进行风暴潮增水值预测的原理。选取广东省珠江口以南的阳江站2017年风暴潮增水数据进行测试。结果表明,三种神经网络方法针对阳江地区风暴潮增水的预测均具有可靠性和实用性。以当前增水值为输入量的单因子模型更能反映真实风暴潮增水趋势,而从增水极值预测的准确性来看,以台风风力、气压、风向等相关参数为输入量的多因子模型优于单因子模型。BP神经网络更适用于多因子长时间预测,小波神经网络在单因子短时间预测上准确性更高,递归神经网络预测值与实测值相关性更强。在工程运用中,需根据地域时空特点、数据资料的丰富度与预测值评估指标选择合适的方法。     

基于EMD-LSTM模型半潜平台运动极短期预报

半潜平台在复杂海洋环境作用下,会发生不规则的六自由度运动响应。这种平台运动的不规则性和随机性对平台作业、栈桥控制以及直升机起落等造成极大的不确定和未知风险。因此,在极短期内准确快速预报平台运动响应对深海浮式平台作业和设备安全具有重要的实际意义。然而目前针对浮式平台运动响应的计算主要是基于势流理论,对确定波浪激励下的平台一阶运动和二阶慢漂运动进行求解,计算的时效性不能满足实际需求。此外,还需要对入射波浪时历进行准确预报,导致平台运动响应准确计算更加困难。针对上述难题,提出了基于EMD-LSTM模型进行平台运动极短期预报的方法。该方法以半潜平台模型试验数据为研究对象,首先对平台运动响应的时间序列进行预处理,接着采用经验模态分解算法(EMD)将时间序列分解成相对平稳的分量,再利用长短期记忆(LSTM)神经网络可以处理复杂非线性长时间序列的优点,对时间序列进行训练预测,最后进行仿真,同时与传统LSTM模型与EMD-BP模型进行对比,仿真结果表明基于EMD-LSTM模型的平台极短期预报方法精度较高,该方法是可行的,具有工程应用的实际意义。     

Application of deep learning technique to the sea surface height prediction in the South China Sea

A deep-learning-based method, called ConvLSTMP3, is developed to predict the sea surface heights(SSHs).ConvLSTMP3 is data-driven by treating the SSH prediction problem as the one of extracting the spatial-temporal features of SSHs, in which the spatial features are "learned" by convolutional operations while the temporal features are tracked by long short term memory(LSTM). Trained by a reanalysis dataset of the South China Sea(SCS), ConvLSTMP3 is applied to the SSH prediction in a region of the SCS east off Vietnam coast featured with eddied and offshore currents in summer. Experimental results show that ConvLSTMP3 achieves a good prediction skill with a mean RMSE of 0.057 m and accuracy of 93.4% averaged over a 15-d prediction period. In particular,ConvLSTMP3 shows a better performance in predicting the temporal evolution of mesoscale eddies in the region than a full-dynamics ocean model. Given the much less computation in the prediction required by ConvLSTMP3,our study suggests that the deep learning technique is very useful and effective in the SSH prediction, and could be an alternative way in the operational prediction for ocean environments in the future.     

Obtaining the distribution of quiescent periods directly from the power spectral densities of Sea waves

There is a growing practical interest in the ability to increase the sea states at which marine operations can be safely undertaken by exploiting the quiescent periods that are well known to exist under a wide range of sea conditions. While the actual prediction of quiescent periods at sea for the control of operations is a deterministic process, the long term planning of future maritime tasks that rely on these quiescent periods is a statistical process involving the anticipated quiescence properties of the forecasted sea conditions in the geographical region of interest. It is in principle possible to obtain such data in tabular form either large scale simulation or from field data. However, such simulations are computationally intensive and libraries of appropriate field data are not common. Thus, it is clearly attractive to develop techniques that exploit standard wave spectral models for describing the quiescence statistics directly from such spectra. The present study focuses upon such techniques and is a first step towards the production of a computationally low-cost quiescence prediction tool and compares its efficacy against simulations. Two significant properties emerge for a large class of wave spectral models that encompasses the ubiquitous Neumann and Pierson Moskowitz or Bretschneider forms. Firstly, the auto-correlation function of the wave profile that are required to produce the quiescence property can be obtained analytically in terms of standard special functions. This considerably reduces the computational cost making desktop computer-based planning tools a reality. Secondly, for each class of these parametric spectra, the probability of a given number of consecutive wave heights (normalised to the significant wave heights) less than some critical value is in fact independent of absolute wave height. Thus, for a broad class of practically interesting wave spectra all that is required to obtain the statistical distribution of the quiescent periods is simple rescaling.     

Study of tide prediction method influenced by nonperiodic factors based on support vector machines

Harmonic analysis, the traditional tidal forecasting method, cannot take into account the impact of noncyclical factors, and is also based on the BP neural network tidal prediction model which is easily limited by the amount of data. According to the movement of celestial bodies, and considering the insufficient tidal characteristics of historical data which are impacted by the nonperiodic weather, a tidal prediction method is designed based on support vector machine (SVM) to carry out the simulation experiment by using tidal data from Xiamen Tide Gauge, Luchaogang Tide Gauge and Weifang Tide Gauge individually. And the results show that the model satisfactorily carries out the tide prediction which is influenced by noncyclical factors. At the same time, it also proves that the proposed prediction method, which when compared with harmonic analysis method and the BP neural network method, has faster modeling speed, higher prediction precision and stronger generalization ability.     

基于对数比转换方法的沉积物粒级组分空间预测与底质类型制图

以废黄河三角洲表层粒度分析数据为基础,探讨了对数比转换和kriging插值相结合方法在沉积物粒级组分空间预测和底质类型制图中的应用。结果表明,基于沉积物粒级组分原始数据的kriging预测方法难以保证各组分预测结果的非负和定和要求,因而预测结果的可信度低;而对数比转换kriging方法不但满足非负和定和要求,而且还有着更优的组分预测结果和较高的底质类型制图精度。新方法对于开展定量化的沉积物粒级组分预测和底质类型制图具有参考价值。