基于BP网络的复杂曲面钢板角变形量分析

通过计算外板展开所产生的变形量来确定水火弯板的加工工艺参数。以钢板厚度、角变形及角变形量作为输入参数,中性层位移系数作为输出参数,建立BP神经网络模型。通过实板实验数据对网络模型进行训练,得到中性层位移系数与这些输入参数的关系,提出了一种准确的计算复杂曲面钢板角变形量的方法。     

基于BP神经网络的SAR干扰效果评估

将BP神经网络引入SAR干扰效果评估过程，根据干扰效果评定诸因素构造合适的指标作为网络输入，网络输出为干扰效果所对应的等级划分，然后利用训练样本对网络进行学习和训练。仿真结果表明，这种方法是可行的，减少了评估过程中人为因素的干扰，使得评估结果更为准确、可靠。     

人工神经网络在计算波浪冲击力中的应用

本文采用人工神经网络方法,通过模型试验获得的试验数据,建立了规则波作用于开边界码头面板的波浪冲击力的预测模型.本文采用一个三层的BP网络,其输入层有三个分别表示:入射波的周期、波高、相对净宽;输出层有一个为波浪冲击力峰值;中间层神经元数通过逐次递增法选取.通过训练后,本文所得到的预测模型的预测结果与实验结果误差较小.     

基于人工神经网络的胶州湾浮游植物生物量模拟

文章首先介绍了BP网络数据标准化、隐层神经元选择、网络训练和有效性检验方法。以2003年胶州湾环境监测资料为基础,建立了多输入单输出的3层BP人工神经网络模型,采用8个水环境因子预测浮游植物生物量(Chla浓度)。检测集样本网络预测值与观测值的相关系数为0.8943,平均绝对误差为11.33%。为避免个别网络输入初值对输出的干扰,采取全局灵敏度的方法,分析了各水环境因子变化对浮游植物生物量的相对影响。结果表明,浮游植物生物量对各水环境因子变化响应的敏感系数顺序为DOCODPO4-PSSTpHOilDINSiO3-Si。     

基于物元分析法的公路桥梁损伤等级评定研究

在构建桥梁损伤等级评定指标体系的基础上，采用 AHP确定了各指标的权重，运用物元分析的方法建立了桥梁损伤等级评定的数学模型，并利用模型评定结果，得到了某桥梁的损伤等级。     

基于BP神经网络的维修性设计参数灵敏度分析

由于维修性设计参数存在模糊性和相关性,传统的评估方法往往很难明确表达出维修性设计参数对维修性的影响程度.通过采用综合评判和人工神经网络相结合的方法建立用于维修性设计参数灵敏度分析的网络.在网络中维修性设计参数作为输入变量,维修性评价值作为输出变量,通过参数值的变动得到不同的网络输出结果,利用网络的输出定量计算出参数的变动对维修性的影响程度,从而为优化维修性设计参数提供方向.     

AASHTO概率模型在船桥碰撞风险评估中的应用

近年来随着航道等级的提升,设计通航船舶尺度增大,要求的通航净空尺度增加,桥区通航水域条件发生显著变化。桥梁存在船撞风险,需对船撞桥梁风险实施评估、为实施防撞设施工程提供依据。国内外因船舶撞击而导致桥梁垮塌或严重破坏的事故逐渐增多,平均每年就有一座大型桥梁因为船舶撞击而遭受严重破坏甚至倒塌。北江航道乌石至三水河口航段经整治由Ⅳ级提升为Ⅲ级后,桥梁存在船撞风险。以船撞桥概率模型(AASHTO)为研究方法,分析了整治河段清远北江二桥参数对船撞桥概率的影响,计算了船舶撞击桥梁各涉水桥墩的年撞击概率,确定了存在较大船撞风险的桥梁与涉水桥墩,建立了船撞桥损伤概率模型,分析桥梁各部位抗撞能力、桥梁各部位船舶撞击力及各部位的年撞击频率,得出通航孔桥墩的年撞击倒塌频率。     

预制桩承载力时效的人工神经网络预测

分析预制桩单桩承载力时间效应的机理，说明其主要与桩长、桩截面积、土体摩擦角、渗透系数、弹性模量、休止期6个参数有关，建立了考虑时间效应的预测单桩承载力的误差反馈型神经网络模型。该模型充分考虑了桩周土固结对桩承载力的影响，输入层神经元为以上6个参数，物理意义明确且容易确定；训练模式采用共轭梯度法。对34根桩的计算表明该模型预测结果与实测值吻合，说明其是科学有效的。     

基于组合优化算法的船舶信息风险评估

为避免船舶信息风险引发重大船舶航行事故,研究基于组合优化算法的船舶信息风险评估方法。选取通信、环境、管理、人为这4个方面因素共17个指标,构建船舶信息风险评估指标体系,将其作为径向基函数（RBF）神经网络输入层输入数据,经隐含层映射操作后,通过输出层输出评估到的船舶信息风险等级,采用结合模糊C均值聚类算法和遗传算法的组合优化算法,合理选取RBF神经网络隐含层中心向量并寻优获取最佳隐含层基函数宽度和权值向量,提升船舶信息风险评估效果。实验结果表明：该方法可有效评估多艘船舶的信息风险,并可依据评估结果获取何种因素导致船舶信息风险,提出针对性指导建议。     

航海模拟器中船舶进港训练评估系统的研究

为了提高航海训练评估的效率,减少主观因素对评估结果的影响,实现自动评估成为当前热点。在调查大量的学员实际训练和考试评估数据基础上,利用概率论中的非参数假设检验方法,对学员的操纵误差予以分析,得到学员的误差分布函数,并结合教师在考试评估中评定成绩的概率,计算出评估的等级区间。最后利用加权平均法计算出学生的综合成绩,采用VC＋＋6．0实现的系统在航海模拟器上测试,得到了比较理想结果。     

平原河网地区内河航道设计最高通航水位取值

设计最高通航水位取值直接影响跨航道桥梁改建方案。针对平原河网地区跨航道桥梁改建实施难度大的问题,进行桥梁改建案例、典型水位曲线特点、船舶过桥现状、撞桥事故统计数据、内河航道导助航及管理手段的发展等方面的研究,探讨降低设计最高通航水位取值的可能性,并提出按通航保证率进行设计最高通航水位取值,并辅以完善的交通管理系统的方案。该方案可作为导助航体系完善、管理水平较高的平原河网地区设计最高通航水位取值参考。     

船舶监控网络中的冗余设计和实现

当前船舶自动化的发展方向是用计算机网络实现船舶机电设备与航海仪器的综合监视、控制和管理，即实现数字化船舶监控。作者结合三年多来在系统设计、软硬件开发和实际工程应用方面的经验，提出了下层使用现场总线、上层使用以大网实现船舶综合网络监控的系统设计方案，给出了其中上层计算机网络的系统组成、软件分布和确保系统高性能运行的一些技术手段，详细论述了系统中所采用的各种软硬件冗余功能和实现原理，给出了部分实现代码。     

Black-box modeling of ship manoeuvring motion based on feed-forward neural network with Chebyshev orthogonal basis function

Based on polynomial interpolation and approximation theory, a novel feed-forward neural network, the feed-forward neural network with Chebyshev orthogonal basis function, is proposed for black-box modeling of ship manoeuvring motion. The neural model adopts a three-layer structure, in which the hidden layer neurons are activated by a group of Chebyshev orthogonal polynomial activation functions and the other two layers’ neurons use identity mapping as activation functions. Weight update formulas are derived by employing the standard back-propagation (BP) training method. With the simulated 15º/15º zigzag test data as input and calculated values of the hydrodynamic forces and moment as output, the feed-forward neural network with Chebyshev orthogonal basis function and the BP neural network are applied to identify the nonlinear functions in the nonlinear hydrodynamic model of ship manoeuvring motion. With the simulated 20º/20º zigzag test data and 35º turning test data as input, the hydrodynamic forces and moment are predicted by using the identified nonlinear functions. Comparison between the calculated and predicted hydrodynamic forces and moment shows that the feed-forward neural network with Chebyshev orthogonal basis function is superior to the BP neural network in identifying the nonlinear functions of the nonlinear hydrodynamic model of ship manoeuvring motion and is an effective method to conduct the black-box modeling of ship manoeuvring motion.     

基于BP神经网络的船舶企业员工综合评价模型

为了保障船舶工业企业内部对其员工的使用管理和评价上的客观性和准确性,提出基于BP神经网络的综合评价方法,在建立了员工评价的指标体系基础上建立了基于BP神经网络的评价模型,将评价员工综合素质的指标值作为模型的输入量,评价结果作为模型输出量,用样本对神经网络进行训练,利用收集的中国船舶工业集团某造船企业管理执行层员工信息,使用MATLAB对模型进行模拟仿真训练,用验证样本对评价模型检测,检测结果是仿真输出与期望输入基本一致,说明网络合理,评价模型能够较好对船舶企业员工的综合素质进行评价.     

一种潜艇作战软件质量评价的新方法

提出一种基于专家知识与神经网络的潜艇战术应用软件质量评价方法.该方法运用专家知识对软件质量指标进行了修正,并作为模型的输入,利用模糊综合评判方法获取该模型的训练数据和测试数据训练神经网络,将训练好的神经网络用于对软件进行质量评价.经算例仿真证明该方法是可行和有效的.     

Distributed and Redundant Design of Ship Monitoring and Control Network

The world trend in ship automation is to integrate the monitoring, intelligent control and systematic management of the instruments and equipments both on bridge and in engine room. The paper presents a design scheme of the ship integrated monitoring and operating system based on two layers distributed and redundant computer network. The lower layer network is the field bus network connected mainly by CAN bus; the upper one is the PC local network in TCP/IP protocol, which consisted of a database server, monitoring and operating computers, industrial computers and a set of switches. Distributed schemes are fully applied to both software and hardware. This paper specifically describes the composition, software distribution and redundant technology of the upper local network and gives some important sample codes for the implement of the redundant and distributed design. The technologies here have been proved in the many applications and it may be applied to other industrial fields.     

基于BP神经网络的接触电阻测量与分等定级

提出了接触电阻的软测量方法，使用红外热像仪测量表面温升，采用有限元方法计算准备训练样本，使用BP神经网络建立起接触电阻与表面温升之间的非线性映射关系，以汇流条接触电阻测量为实例，其他电气设备接触电阻的测量也可类似进行，为提高工程应用的可操作性，将接触电阻分为表征接触程度好坏的I-V级，用模式分类方法对测得的接触电阻进行分类，实现了接触电阻的定量评价和分等定级。     

Effects of a classroom-based bridge resource management training on knowledge,attitudes, behaviour and performance of junior naval officers

This study asses the effectiveness of classroom-based bridge resource management (BRM) training for junior naval officers, in which general principles of human behaviour and performance in teams and under stress were conveyed. Although BRM training is recommended by the International Maritime Organization and is increasingly common in seafaring, very little is known about whether the adaptation of crew resource management (CRM) training from aviation to the maritime domain has been successful and what type of training is effective. A study with a quasi-experimental, two-factorial mixed design was conducted with BRM training as the between factor and time as the within factor. For 117 study participants, evaluation criteria were assessed on all levels as defined by Kirkpatrick (Train Dev J, 178–192 1979): subjective training evaluation, knowledge, attitudes and behaviour as well as performance while commanding a vessel during a real-world exercise. BRM participants showed better subjective training evaluations and more BRM-related knowledge than controls. Training did not produce differences between groups regarding BRM-related attitudes, the demonstration of non-technical skills or the overall success in the real-world exercise. Overall, BRM training effectiveness was rather low, which can most probably be attributed to the focus of training on generalizable knowledge, skills and attitudes at the expense of their specific application to the context of the real-world exercise. In the design of BRM and CRM training courses alike, the effective application of general principles to a given context must be defined, and the application must be emphasised during training delivery.