Experimental investigation of oil leakage from damaged ships due to collision and grounding
船舶碰撞和搁浅导致的油泄漏实验研究

Received 24 May 2010  收到 2010 年 5 月 24 日
Accepted 10 September 2011
接受于 2011 年 9 月 10 日
Editor-in-Chief: A.I. Incecik
主编：A.I. Incecik
Available online 3 November 2011
2011 年 11 月 3 日在线可用

Experimental test  实验测试
Oil spill  石油泄漏
Damaged ship  受损船只
Collision  碰撞
Grounding  接地
Tank design  坦克设计

© 2011 Elsevier Ltd. All rights reserved.
© 2011 埃尔塞维尔有限公司。版权所有。

An important part of protecting the environment is to ensure that there are as few spills as possible. Accidental oil leakages sometimes occur and require a quick and adequate response in order to reduce the environmental consequences. Both the government and industry are working constantly to reduce the risk of oil spills by introducing strict new legislation and stringent operating codes (Fingas (2001)).
保护环境的重要部分是确保尽可能减少泄漏。意外石油泄漏有时会发生，需要迅速和充分的应对措施以减少环境影响。政府和行业都在不断努力，通过引入严格的新立法和严格的操作规范来降低石油泄漏的风险（Fingas（2001））。

Most tankers are loaded such that the internal pressure at the tank is larger than the external sea pressure. Thus, if the tank is damaged, cargo flows out. If the tanker carries substantially less cargo such that the hydrostatic balance is established at-or several meters above-the tank bottom, water tends to enter the ship through the hole in the hull as long as the highest point of damage is below the hydrostatic balance level (National Research Council (1991)). The industry has invoked new operating and maintenance procedures to reduce the number of accidents that lead to spills. Based on historical statistics, grounding is the leading cause of oil spills from vessels (26%), followed by collisions at $22\mathrm{\%}$$22\mathrm{\%}$22%22 \%. Some other accidental causes of oil spills are explosion/fire (9%), ramming (9%) and sinking (7%); human error (5%) and mechanical failure ( $2\mathrm{\%}$$2\mathrm{\%}$2%2 \% ) cause the least number of spills
大多数油轮装载时，罐内压力大于外部海压。因此，如果罐体损坏，货物会流出。如果油轮装载的货物量显著减少，以至于静水平衡在罐底或以上几米处建立，只要损坏的最高点低于静水平衡水平，水就会通过船体上的孔进入船内（国家研究委员会，1991 年）。该行业已采取新的操作和维护程序，以减少导致泄漏的事故数量。根据历史统计数据，搁浅是船舶发生油污泄漏的主要原因（26%），其次是碰撞（ $22\mathrm{\%}$$22\mathrm{\%}$22%22 \% ）。其他一些导致油污泄漏的意外原因是爆炸/火灾（9%）、冲撞（9%）和沉没（7%）；人为错误（5%）和机械故障（ $2\mathrm{\%}$$2\mathrm{\%}$2%2 \% ）导致的事故数量最少。

(Keisha, 2005). The complete analysis of ship grounding and collision is an extensive process and can be divided into several approaches. A systematic design procedure for grounding was presented by Amdahl et al. and is shown in Fig. 1 (Amdahl et al., 1995). The starting point is to characterize the ship dimensions, structural scantlings, forward speed and cargo arrangement. The sea floor conditions are very important and may vary from sharp rocks to hard shoals or soft clay/sand banks. Step two describes the external mechanics during grounding and collision, i.e., rigid body motions and the hull girder loads (Tavakoli et al., 2007). During this stage, the hydrodynamic loads interact with the grounding and collision loads. In collision and grounding, when the hull of a loaded tanker is ruptured, some of the oil flows into the sea. The damage to the hull girder is determined in step three. This step describes the internal mechanics and is closely related to step two in the sense that the evolution of the hull damage must be determined by taking into account rigid body motions that affect the degree of structural damage (Tavakoli et al., 2007; Alsos et al., 2007). The next step deals with the residual strength of the ship in its damaged condition. If the ship rests on the sea floor, further damage and additional hull girder loads may be caused by ebb tide and waves. After the hull damage is determined, the consequences of the grounding and collision incident may be estimated in terms of oil spill, water flooding and damage stability, which is done in step five. The present paper focuses on this step. Finally, in step six, the resulting damage/oil spill can be evaluated against some acceptance criteria. These criteria may be expressed as limits to oil.
(Keisha, 2005)。船舶搁浅和碰撞的全面分析是一个复杂的过程，可以分为几个方法。Amdahl 等人提出了一种搁浅的系统设计程序，如图 1 所示（Amdahl 等人，1995）。起点是对船舶尺寸、结构细部、前进速度和货物布置进行表征。海底条件非常重要，可能从尖锐的岩石到坚硬的浅滩或软泥/沙坝。第二步描述了搁浅和碰撞过程中的外部力学，即刚体运动和船体梁载荷（Tavakoli 等人，2007）。在此阶段，水动力载荷与搁浅和碰撞载荷相互作用。在碰撞和搁浅时，当装载油轮的船体破裂时，部分油会流入海中。第三步确定船体梁的损伤。这一步描述了内部力学，并且与第二步密切相关，因为必须通过考虑影响结构损伤程度的刚体运动来确定船体损伤的发展（Tavakoli 等人，2007；Alsos 等人，2007）。 下一步处理的是船舶在损坏状态下的残余强度。如果船舶停靠在海底，潮汐和波浪可能会导致进一步的损坏和额外的船体梁载荷。在确定船体损坏后，可以根据漏油、水淹和损害稳定性来估计搁浅和碰撞事件的影响，这将在第五步进行。本文重点关注这一步骤。最后，在第六步，可以对产生的损坏/漏油与某些验收标准进行比较。这些标准可以表示为油的限制。

Fig. 1. The process of grounding analysis.
图 1. 接地分析过程。

Generally, an oil spill involves the motion of two immiscible liquids in complex geometries and at a wide range of length scales. The oil loss includes both instantaneous loss and subsequent loss due to environmental effects such as tide, current and wave action. The instantaneous oil loss depends on the tank design, ship speed and the size of the hole at the damaged location. In the grounding scenario, if the internal hydrostatic pressure is higher above the opening, outflow occurs and produces a gravity current. The hydrostatic pressure is the key factor in analyzing the leak rate. The flow continues at an ever-decreasing rate until the inside and outside pressures are equalized. In
一般来说，油污泄漏涉及两种不相溶的液体在复杂几何形状和广泛长度尺度上的运动。油损失包括瞬时损失和由于潮汐、水流和波浪等环境效应引起的后续损失。瞬时油损失取决于罐体设计、船速和受损位置的孔径大小。在搁浅场景中，如果开口上方的内部静水压力较高，就会发生流出并产生重力流。静水压力是分析泄漏率的关键因素。流量以不断减小的速率持续，直到内外压力平衡。在
order to calculate the theoretical outflow rate and spill volume, a model based on Bernoulli’s principle is proposed (Tavakoli et al., 2008) and developed to cover the hydrostatic changes due to an oil spill (Tavakoli et al., 2009). Karafiath and Bell (1992a,b) performed the model tests of oil outflow for mid-deck and double-hull tanker configurations.
为计算理论排放速率和溢出体积，提出并开发了一个基于伯努利原理的模型（Tavakoli 等，2008），以覆盖由于油污造成的静压变化（Tavakoli 等，2009）。Karafiath 和 Bell（1992a,b）对中甲板和双层壳油轮配置的油排放模型进行了测试。

Analytical methods were applied for the analysis of oil-water flow for tanks with damage in the form of openings, and the total spill and loss rate were established. The effects of various configurations of hull designs were studied. The analytical model was also extended to cover collision Scenario and CFD simulation
分析方法被应用于分析具有开口损伤的油水流动的储罐，并建立了总泄漏和损失率。研究了各种船体设计配置的影响。分析模型也被扩展以涵盖碰撞场景和 CFD 模拟
with FLUENT (FLUENT Inc., 2003) software was carried out to examine the model (Tavakoli et al. 2010). Model tests for leaking heavy oil from single hull in collision scenario are performed by Debra et al. (2001). In the collision scenario, even if the average pressures are equal across the interface between oil and water, there is still a local imbalance, and the flow will cease only when the lower lip of the hole is covered by water. Accordingly, the side leakage process is divided into two phases (Fannelop, 1994).
使用 FLUENT（FLUENT Inc., 2003）软件进行了模型（Tavakoli 等人，2010 年）的检验。Debra 等人（2001 年）进行了单壳体碰撞场景下泄漏重油的模型测试。在碰撞场景中，即使油水界面处的平均压力相等，仍然存在局部不平衡，只有当孔的下唇被水覆盖时，流动才会停止。因此，侧泄漏过程分为两个阶段（Fannelop，1994 年）。

In order to investigate the fluid dynamics of accidental oil spillage due to grounding and collision for various designs and to judge the performances of the proposed models and CFD simulations, some experimental tests were designed and performed at NTNU. In this paper, the results of these model tests are presented. The purpose of these tests is to provide quantitative information on the behavior of an oil spill from a damaged tank and validate the results of analytical and numerical models. The model tests are conducted to study the oil release from tanks with:
为了研究因搁浅和碰撞导致的意外油泄漏的流体动力学，以及评估所提出模型和 CFD 模拟的性能，挪威科技大学（NTNU）设计并进行了某些实验测试。本文展示了这些模型测试的结果。这些测试的目的是提供有关受损油罐油泄漏行为的定量信息，并验证分析和数值模型的结果。模型测试旨在研究从油罐释放油的行为：

A puncture in the bottom corresponds to the grounding scenarios and a puncture in the side represents the collision scenarios. The double-side designs and double-bottom hulls are intended to provide protection for all collision and grounding scenarios, respectively, except those associated with the highest energies. The double-hull configuration is a combination of the double-bottom and the double-side designs. The advantage of the double-hull tanker with respect to the oil spill is the structural protection afforded by the inner hull and the claimed capability to capture some oil in the space between the inner and outer hull in the case of grounding and collision events that are severe enough to cause a rupture of the inner hull. None of the models and configurations are representative of the much more complex geometry that would occur in a real accident. However, they give an idea of the basic physical processes that will occur.
底部穿孔对应接地场景，侧面穿孔代表碰撞场景。双面设计和双层船体旨在为所有碰撞和接地场景提供保护，除了与最高能量相关的那些。双层船体配置是双层船体和双面设计的组合。与油污相比，双层油轮的优势在于内壳提供的结构保护，以及在接地和碰撞事件中，如果内壳破裂，声称能够捕获内壳和外壳之间的部分油。没有任何模型和配置能代表实际事故中发生的复杂得多几何形状。然而，它们给出了基本物理过程发生的一个概念。

The model tank was built at $1/30$$1/30$1//301 / 30 scale of an existing FPSO, which is studied in collisions with stern and bow of a large supply vessel in a previous paper by the authors (Tavakoli et al., 2007). Test tanks must be at a small scale; hence, it is important to verify that the cargo properties and release conditions are modeled in a reasonable way. In the event of a punctured tank, a variety of factors will determine the oil outflow. The most important factors are the oil density, initial oil level, the tank and puncture dimensions and the location of the puncture relative to the waterline. However, there are also other factors that can affect the discharge, some of which are subtle. Their significance depends on the specific conditions of the tank (Debra et al. (2001)). For example, many vessels carrying petroleum products have a fixed piping system to blow inert gas into the cargo tanks (ISGOTT, 1996); this gas not only renders the tank atmosphere non-flammable but also adds a slight pressure head to the cargo, which may increase the oil outflow (Dodge and Bowles, 1982). Furthermore, water levels outside the tank may fluctuate, increasing or decreasing the release rate. Finally, accidents may cause other complications. Vessel instability may create free surface waves inside the tanks, causing pressure fluc-tuations-particularly if the hole is exposed to the sea. These factors are important with respect to oil outflow. However, it is extremely
模型罐以现有 FPSO 的 $1/30$$1/30$1//301 / 30 比例建造，该 FPSO 在作者之前的一篇论文中研究了与大型供应船艉部和艏部的碰撞情况（Tavakoli 等人，2007 年）。测试罐必须是小规模；因此，验证货物特性和释放条件以合理的方式进行建模非常重要。在罐体被刺破的情况下，多种因素将决定油料流出。最重要的因素是油密度、初始油位、罐体和穿孔尺寸以及穿孔相对于水线的位置。然而，还有其他可能影响排放的因素，其中一些因素较为微妙。它们的重要性取决于罐体的具体条件（Debra 等人，2001 年）。例如，许多运输石油产品的船只都有一个固定的管道系统，将惰性气体吹入货物罐（ISGOTT，1996 年）；这种气体不仅使罐体气氛不可燃，而且也给货物增加了一点点压力头，这可能会增加油料流出（Dodge 和 Bowles，1982 年）。此外，罐体外部的液位可能会波动，增加或减少释放速率。 最终，事故可能导致其他并发症。船舶不稳定可能在舱内产生自由表面波，导致压力波动——尤其是如果漏洞暴露在海水中。这些因素与油泄漏有关。然而，这极为
difficult to scale these conditions so that smaller, experimental releases in the laboratory are applicable to much larger releases. For this reason, the experiments are simplified to consider a small, stationary, vented tank. It is essential to ensure that the oil loss at the model scale is governed by the same hydrodynamic physical phenomena that cause oil loss at the full scale. The different forces that act on a fluid particle are static pressure forces, inertial forces and viscous forces. In order to facilitate the direct transfer of results from model tests to full scale conditions, the geometry, kinematic and dynamic similarities should be fulfilled.
难以将这些条件进行扩展，以便在实验室中进行的小型实验性释放适用于更大规模的释放。因此，实验被简化为考虑一个小型、静止、通风的储罐。确保在模型尺度上的油损失受与全尺度上引起油损失的相同流体动力学物理现象的控制是至关重要的。作用于流体粒子的不同力包括静压力力、惯性力和粘性力。为了便于将模型试验结果直接转移到全尺度条件下，应满足几何、运动学和动力学相似性。

Geometric similarity occurs when the model has the same shape as the prototype such that all linear dimensions of the model are related to the corresponding dimensions of the prototype by a constant scale factor $\lambda$$\lambda$lambda\lambda. To achieve geometric similarity, a model tank is built at $1/30$$1/30$1//301 / 30 scale with the dimension of $100\times 100\times 50{\mathrm{cm}}^3$$100\times 100\times 50{\mathrm{cm}}^3$100 xx100 xx50cm^(3)100 \times 100 \times 50 \mathrm{~cm}^{3}. The horizontal area of the cargo tank is $5000{\mathrm{cm}}^2$$5000{\mathrm{cm}}^2$5000cm^(2)5000 \mathrm{~cm}^{2}. It is, however, impossible to apply the correct scale factor to the plate thickness, which could have an effect on the discharge rate.
几何相似性发生在模型与原型具有相同形状时，即模型的全部线性尺寸都通过一个常数比例因子 $\lambda$$\lambda$lambda\lambda 与原型的相应尺寸相关。为了实现几何相似性，以 $1/30$$1/30$1//301 / 30 的比例建造了一个模型舱，尺寸为 $100\times 100\times 50{\mathrm{cm}}^3$$100\times 100\times 50{\mathrm{cm}}^3$100 xx100 xx50cm^(3)100 \times 100 \times 50 \mathrm{~cm}^{3} 。货舱的水平面积为 $5000{\mathrm{cm}}^2$$5000{\mathrm{cm}}^2$5000cm^(2)5000 \mathrm{~cm}^{2} 。然而，无法将正确的比例因子应用于板厚，这可能会影响排放速率。

Kinematic similarity is obtained when the velocities at corresponding points of the model and the prototype act in the same direction and are scaled by a constant factor ${\lambda }_{\mathrm{v}}$${\lambda }_{\mathrm{v}}$lambda_(v)\lambda_{\mathrm{v}}, which implies that the flows must have similar streamline patterns. The flow regimes must also be the same. For gravity flow, this yields
运动相似性是在模型和原型对应点的速度在同一方向上作用并按一个常数因子 ${\lambda }_{\mathrm{v}}$${\lambda }_{\mathrm{v}}$lambda_(v)\lambda_{\mathrm{v}} 缩放时获得的，这意味着流动必须具有相似的流线模式。流动状态也必须相同。对于重力流动，这导致
${\lambda }_v=\frac{v_p}{v_m}=\frac{\sqrt{2g{h}_p-\frac{2{\rho }_w}{{\rho }_o}g{H}_{w_p}}}{\sqrt{2g{h}_m-\frac{2{\rho }_w}{{\rho }_o}g{H}_{w_m}}}=\frac{\sqrt{2g\lambda h_m-\frac{2{\rho }_w}{{\rho }_o}g\lambda H_{w_m}}}{\sqrt{2g{h}_m-\frac{2{\rho }_w}{{\rho }_o}g{H}_{w_m}}}=\sqrt{\lambda }$${\lambda }_v=\frac{v_p}{v_m}=\frac{\sqrt{2g{h}_p-\frac{2{\rho }_w}{{\rho }_o}g{H}_{w_p}}}{\sqrt{2g{h}_m-\frac{2{\rho }_w}{{\rho }_o}g{H}_{w_m}}}=\frac{\sqrt{2g\lambda h_m-\frac{2{\rho }_w}{{\rho }_o}g\lambda H_{w_m}}}{\sqrt{2g{h}_m-\frac{2{\rho }_w}{{\rho }_o}g{H}_{w_m}}}=\sqrt{\lambda }$lambda_(v)=(v_(p))/(v_(m))=(sqrt(2gh_(p)-(2rho_(w))/(rho_(o))gH_(w_(p))))/(sqrt(2gh_(m)-(2rho_(w))/(rho_(o))gH_(w_(m))))=(sqrt(2g lambdah_(m)-(2rho_(w))/(rho_(o))g lambdaH_(w_(m))))/(sqrt(2gh_(m)-(2rho_(w))/(rho_(o))gH_(w_(m))))=sqrtlambda\lambda_{v}=\frac{v_{p}}{v_{m}}=\frac{\sqrt{2 g h_{p}-\frac{2 \rho_{w}}{\rho_{o}} g H_{w_{p}}}}{\sqrt{2 g h_{m}-\frac{2 \rho_{w}}{\rho_{o}} g H_{w_{m}}}}=\frac{\sqrt{2 g \lambda h_{m}-\frac{2 \rho_{w}}{\rho_{o}} g \lambda H_{w_{m}}}}{\sqrt{2 g h_{m}-\frac{2 \rho_{w}}{\rho_{o}} g H_{w_{m}}}}=\sqrt{\lambda}
where $v$$v$vv is the outflow velocity, $g$$g$gg is the acceleration due to gravity, $h$$h$hh is the oil level, $H_w$$H_w$H_(w)H_{w} is the water level, ${\rho }_{\mathrm{o}}$${\rho }_{\mathrm{o}}$rho_(o)\rho_{\mathrm{o}} is the density of oil, ${\rho }_{\mathrm{w}}$${\rho }_{\mathrm{w}}$rho_(w)\rho_{\mathrm{w}} is the density of water, the subscripts $p$$p$pp and $m$$m$mm refer, respectively, to the prototype and the model and $\lambda$$\lambda$lambda\lambda is the geometric scaling factor.
$v$$v$vv 是流出速度， $g$$g$gg 是重力加速度， $h$$h$hh 是油位， $H_w$$H_w$H_(w)H_{w} 是水位， ${\rho }_{\mathrm{o}}$${\rho }_{\mathrm{o}}$rho_(o)\rho_{\mathrm{o}} 是油的密度， ${\rho }_{\mathrm{w}}$${\rho }_{\mathrm{w}}$rho_(w)\rho_{\mathrm{w}} 是水的密度，下标 $p$$p$pp 和 $m$$m$mm 分别指原型和模型， $\lambda$$\lambda$lambda\lambda 是几何比例系数。

Dynamic similarity is basically met if the model and prototype forces differ by a constant scale factor at similar points. The ratio of all forces acting on corresponding fluid particles and boundary surfaces in the two systems is constant. The non-dimensional parameters that are useful for scaling these forces in fluid flows are the Froude number and Reynolds number.
动态相似性基本上在模型和原型力在相似点相差一个常数比例因子时得到满足。两个系统中对应流体粒子和边界表面上所有力的比值是恒定的。在流体流动中，用于缩放这些力的无量纲参数是弗劳德数和雷诺数。

The Froude number, Fr, relates the inertia term to the gravity term:
弗劳德数，Fr，将惯性项与重力项相关联：
$\mathrm{Fr}={\left(\frac{v}{\sqrt{gL}}\right)}_p={\left(\frac{v}{\sqrt{gL}}\right)}_m$$\mathrm{Fr}={\left(\frac{v}{\sqrt{gL}}\right)}_p={\left(\frac{v}{\sqrt{gL}}\right)}_m$Fr=((v)/(sqrt(gL)))_(p)=((v)/(sqrt(gL)))_(m)\operatorname{Fr}=\left(\frac{v}{\sqrt{g L}}\right)_{p}=\left(\frac{v}{\sqrt{g L}}\right)_{m}
where $v$$v$vv is the outflow velocity, $g$$g$gg is the acceleration due to gravity, $L$$L$LL is the oil level and water level and the subscripts $p$$p$pp and $m$$m$mm refer, respectively, to the prototype and the model. Because velocity in both the prototype and the model is only due to
$v$$v$vv 是出流速度， $g$$g$gg 是重力加速度， $L$$L$LL 是油位和水位，下标 $p$$p$pp 和 $m$$m$mm 分别指原型和模型。因为原型和模型中的速度仅由