模糊C-均值聚类算法及其在船舶故障诊断中的应用

船舶设备故障的早期诊断和预测,对船舶的安全运行具有非常重要的意义。由于船舶设备繁多,运行环境特殊,各种设备的故障症状与故障原因之间的关系十分复杂,致使传统诊断方法在实际应用中效果不理想。因此,研究采用模糊C-均值聚类算法来实现船舶故障的诊断乃是非常必要的。将被诊断对象间故障和症状的特征通过建立模糊关系矩阵进行故障分类,用当前所得的故障征兆群与过去该设备故障征兆结果相对照,找出最相似的结果,从而确定其故障。通过船舶主机轴系诊断的实例,证明了该方法的有效性。     

船舶浮态计算方法及船舶浮态图谱

本文用化非线性问题为线性问题的线性化方法,对非线性隐式的船舶浮态方程组进行了数值解,并在深入研究的基础上提出了两种形式浮态图谱,以便能迅速而准确地确定船舶各种不同装载情况下的浮态。本文给出了五万吨油轮“西湖”号算例。     

大型集装箱船在斜浪工况下的扭矩特征研究及结构强度分析

随着大型集装箱船的发展,针对斜浪条件下扭矩载荷特性及其结构强度的分析研究日益重要.本文采用选定的某大型集装箱船开展动态载荷(DLA)分析,建立了水动力湿表面计算模型和质量模型,研究基于扭矩传递函数和主要载荷控制参数的斜浪设计波参数确定方法,分析超越概率水平对扭矩载荷计算结果的影响.考虑典型斜浪参数和超越概率水平,分析对比DLA扭矩与船舶规范(ABS和HCSR)扭矩的差异及原因,提出集装箱船扭矩载荷计算与应用的建议.在此基础上,选定3个斜浪计算工况(45°,60°和75°)开展结构强度分析,通过分析应力计算云图,研究扭矩载荷下船体结构响应的关键位置及其应力趋势.该研究可为大型集装箱船结构设计过程中船体梁扭转强度计算、舱口角隅设计、抗扭箱强度评估等方面提供有益的参考.     

基于CFD的内河船深浅水中操纵性预报

由于岸壁效应和浅水效应,内河船舶在限制水域作操纵运动时通常受到比在开阔水域中更大的水动力.这些水动力对船舶操纵性具有不利影响,有可能导致船舶碰撞或触底等海上事故.因此,为了在船舶设计阶段预报其操纵性能,考虑浅水效应和岸壁效应以准确计算内河船舶操纵运动水动力非常重要.本文基于CFD方法,通过对粘性绕流进行数值模拟,对长江中营运的三艘内河船舶的操纵运动水动力进行计算.首先,为了验证数值方法的可靠性,对标模KVLCC2纯横荡和纯首摇试验的水动力进行计算,并将计算结果与现有的试验数据进行对比.然后,对三艘内河船舶在不同水深下的静舵试验、纯横荡和纯首摇试验进行数值模拟,计算得到水动力及相应的线性水动力导数.最后,基于计算得到的水动力导数,获得Nomoto模型中的操纵性参数,对比分析三艘内河船舶在深浅水中的操纵性能.结果表明,本文方法可以揭示不同水深下三艘内河船舶的操纵性变化趋势.该方法可为船舶设计阶段内河船舶深浅水中的操纵性预报提供一种实用的工具.     

聚脲涂层舷侧板架抗撞性能试验研究

针对船舶舷侧结构抗碰撞问题,开展有无聚脲涂层舷侧板架落锤试验研究。以某型舰船结构为依据建立舷侧板架有限元模型,利用瞬态动力学软件MSC/Dytran对模型进行数值仿真并确定落锤高度及试验工况。在此基础上,制作模型板架进行有无聚脲涂层舷侧板架落锤冲击试验,分别获得有涂层和无涂层舷侧板架在碰撞冲击载荷作用下的损伤变形、破口大小及碰撞力,对比研究聚脲材料的抗撞防护性能。结果表明,聚脲涂层的存在能够加强舷侧板架的耐撞防护性能。     

海事重点监管船舶定位系统黄浦江水域应用研究

海事重点监管船舶定位系统是针对上海黄浦江闵行段水域海事重点监管船舶逃逸问题而建设应用的系统。该系统前端设备独立性强,携带、装卸方便,应用功能尤其是报警功能全面及时可靠,已在黄浦江闵行段水域的海事监管应用中取得良好的效果,但仍存在硬件性能不足、定制化功能待完善、应用管理机制未建立等问题。文中针对该系统存在的问题提出了提高硬件性能、完善定制化应用功能及建立完善应用管理机制等建议,以期为海事重点船舶防逃逸监管提供一种高效、完善的解决方案。     

水力式升船机单井塔柱内对拉钢索优化布置

单井型塔柱结构极大改善了水力式升船机竖井水位同步性问题,但随着提升高度增大,竖井内水深越大,侧壁位移、弯矩过大等问题更加突出。以150米级单井塔柱为例,提出了竖井内布置对拉钢索的设想和简化布置方案。基于钢索拉力权重分配系数的优化方法,取得了钢索等应力的密度分布最优解,并离散转化为等效的等荷载布置方案,可在实际工程中应用。经有限元模型结果对比分析,等荷载布置与简化方案相比,最大位移可减小47.44%,纵竖向弯矩极值差别不大。与无钢索方案比较,最大位移、纵向弯矩、竖向弯矩可分别减小93.5%、74.36%、75.84%。高强钢索预应力方案值得进一步研究。     

引江济淮工程派河口船闸通航水流条件及改善措施

派河口船闸是引江济淮派河口枢纽重要组成部分,位于派河入巢湖口门段,建设条件复杂。综合已有工程、河道边界和相关规划,提出了复线船闸平面布置初拟方案。采用定床河工模型试验的技术手段对船闸下游引航道与口门区通航水流条件开展了详尽研究,探明了节制闸泄洪时下游通航水域纵向和横向表面流速、流态特征,分析了通航水流条件不满足规范要求的原因,指出可通过岸线平顺、新开泄洪分流通道和控制临界安全泄洪流量等措施进行改善。优化后的方案可满足设计要求,为工程设计和今后运行管理提供技术支撑。     

Surf-riding and oscillations of a ship in quartering waves

The behavior of a ship encountering large regular waves from astern at low frequency is the object of investigation, with a parallel study of surf-riding and periodic motion paterns. First, the theoretical analysis of surf-riding is extended from purely following to quartering seas. Steady-state continuation is used to identify all possible surf-riding states for one wavelength. Examination of stability indicates the existence of stable and unstable states and predicts a new type of oscillatory surf-riding. Global analysis is also applied to determine the areas of state space which lead to surf-riding for a given ship and wave conditions. In the case of overtaking waves, the large rudder-yaw-surge oscillations of the vessel are examined, showing the mechanism and conditions responsible for loss of controllability at certain vessel headings.List of symbols c wave celerity (m/s) - C(p) roll damping moment (Ntm) - g acceleration of gravity (m/s²) - GM metacentric height (m) - H wave height (m) - I _x ,I _z roll and yaw ship moments of inertia (kg m²) - k wave number (m^–1) - K _H ,K _W ,K _R hull reaction, wave, rudder, and propeller - K _p forces in the roll direction (Ntm) - m ship mass (kg) - n propeller rate of rotation (rpm) - N _H ,N _W ,N _R hull reaction, wave, rudder, and propeller - N _P moments in the yaw direction (Ntm) - p roll angular velocity (rad/s) - r rate-of-turn (rad/s) - R(,x) restoring moment (Ntm) - Res(u) ship resistance (Nt) - t time (s) - u surge velocity (m/s) - U vessel speed (m/s) - v sway velocity (m/s) - W ship weight (Nt) - x longitudinal position of the ship measured from the wave system (m) - x _G ,z _G longitudinal and vertical center of gravity (m) - x _S longitudinal position of a ship section (S), in the ship-fixed system (m) - X _H ,X _W ,X _R hull reaction, wave, rudder, and propeller - X _P forces in the surge direction (Nt) - y transverse position of the ship, measured from the wave system (m) - Y _H ,Y _W ,Y _R hull reaction, wave, rudder, and propeller - Y _p forces in the sway direction (Nt) - z _Y vertical position of the point of action of the lateral reaction force during turn (m) - z _W vertical position of the point of action of the lateral wave force (m) Greek symbols angle of drift (rad) - rudder angle (rad) - wavelength (m) - position of the ship in the earth-fixed system (m) - water density (kg/m³) - angle of heel (rad) - heading angle (rad) - _e frequency of encounter (rad/s) Hydrodynamic coefficients K roll added mass - N _v ,N _r yaw acceleration coefficients - N _v N _r N _rr N _rrv ,N _vvr yaw velocity coefficients K. Spyrou: Ship behavior in quartering waves - X _u surge acceleration coefficient - X _u X _vr surge velocity coefficients - Y _v ,Y _r sway acceleration coefficients - Y _v ,Y _r ,Y _vv ,Y _rr ,Y _vr sway velocity coefficients European Union-nominated Fellow of the Science and Technology Agency of Japan, Visiting Researcher, National Research Institute of Fisheries Engineering of Japan     

Tuning of transfer functions for analysis of wave–ship interactions

Transfer functions are often used together with a wave spectrum for analysis of wave–ship interactions, where one application addresses the prediction of wave-induced motions or other types of global responses. This paper presents a simple and practical method which can be used to tune the transfer function of such responses to facilitate improved prediction capability. The input to the method consists of a measured response, i.e. time series sequences from a given sensor, the 2D wave spectrum characterising the seaway in which the measurements are taken, and an initial estimate of the transfer function for the response in study. The paper presents results obtained using data from an in-service container ship. The 2D wave spectra are taken from the ERA5 database, while the transfer function is computed by a simple closed-form expression. The paper shows that the application of the tuned transfer function leads to predictions which are significantly improved compared to using the transfer function without tuning.