乘风破浪

作者: Daniel Fonseca de Carvalho e Silva,Petrobras研究工程师,巴西里约热内卢

大多数新油田位于海上。许多深水项目在海况不佳的远海位置中使用浮式生产储存卸货装置(FPSO)。在过去,极为耗时且成本高昂的物理实验是确保船舶能够在最有可能发生的海况下不受损坏的唯一方案。Petrobras使用ANSYS仿真来减少所需的实验数量,并取得更详尽的载荷数据。

Save PDF

盐下层是大陆架上的地质构造,在Gondwana超大陆分裂成我们如今所了解的大陆的过程中,盐下层在位于它上面的盐层之前形成。过去几十年来,我们在巴西大陆架上的盐下层里有众多发现,估计有500亿桶石油,是巴西以往储量的五倍。这些储藏带来了巨大的钻井挑战,因为它们位于3,000米的海水、2,000米的岩石和2,000米的盐层之下。并且,由于它们所在的深水区距离海岸线有数百公里之遥,往往是天气恶劣、海况复杂,对于负责从油气井中接收这些碳氢化合物,然后加工、存储并卸载到邮轮或输油气管线的FPSO船舶来说,将石油天然气提升到海面是一项特殊的挑战。

最恶劣的情况被称为绿水,发生在连续波浪翻过FPSO的甲板时。绿水不会威胁到船舶的完整性,但会破坏其表面的关键设备,例如控制阀、电缆槽、消防设备等等。在最恶劣的情况下,可能会进行成本高昂的停产维修。企业可能每天损失数十万美元的收入。目前,石油公司主要使用比例模型实验来评估绿水条件下的载荷,但由于在模型比例实验中很难监测非常拥堵的顶部区域的载荷,这种方法颇受局限。工程师难以提前预测何处会产生最高载荷,因此通常没有将传感器放在正确的位置。Petrobras通过采用 ANSYS Fluent 计算流体动力学(CFD) 软件,能够比物理测试更精确地预测甲板结构上的力,从而克服了这些挑战。

ANSYS Meshing
CFD simulation of green-water loading throughout hull and deck structures of an FPSO

“ANSYS CFD软件能够比物理测试更精确地预测甲板结构上的力。”

DECK STRUCTURE DESIGN CHALLENGES

FPSO deck structures must be designed to withstand a 100-year maximum significant wave height of 12 meters in pre-salt fields gearing up for production, compared to 9 meters or less in post-salt fields. This upgrade may require the addition of features such as structural barriers and local reinforcements to on-deck equipment. These features add weight to the structure, which increase its cost and reduce the vessel's storage capacity. The goal is to quantify the loads as accurately as possible so that structures can be designed with a sufficient margin of safety but not overdesigned.

ANSYS Meshing
CFD mesh with refined area shown in black

A problem with scale-model experiments is that the model is so small that only a few force measurements can be performed on the deck. The forces on other structures must be estimated, and these estimates must be high to account for uncertainty. Scale-model experiments also take about three months to plan and run, and are very expensive. One-dimensional hydraulic codes are sometimes applied to this problem, but since they don't account for the geometry of the structures, they also can only provide estimates of the relevant loading.

"Physical experiments take about three months; simulation can be completed in 10 to 50 days."

SIMULATING GREEN-WATER LOADING

Petrobras engineers recently set out to apply ANSYS CFD to this problem, starting with a simulation of a scale-model experiment so that the simulation results could easily be validated. The model test condition was carefully chosen to intensify the green-water effects; it does not represent a true operational configuration. In the physical experiment, loads were measured at six locations and water elevation at 38 locations. Petrobras engineers captured a time series of the wave from the experiment, selected the most critical part and wrote a MATLAB program that runs a fast Fourier transform to represent the irregular wave interest interval as a combination of linear wave components. They wrote a Fluent user subroutine to impose the wave combination as a boundary condition on the CFD simulation. Engineers imposed the movement of the vessel as measured in the lab on the CFD simulation as another boundary condition, using a dynamic mesh to accommodate vessel movement. They used the Fluent volume of fluid (VOF) model to track the interface between the surface of the water and the air. The shear stress transport (SST) turbulence model solved a turbulence/frequency-based model (k–ω) at the wall and k-ε in the bulk flow.

A problem with scale-model experiments is that the model is so small that only a few force measurements can be performed on the deck. The forces on other structures must be estimated, and these estimates must be high to account for uncertainty. Scale-model experiments also take about three months to plan and run, and are very expensive. One-dimensional hydraulic codes are sometimes applied to this problem, but since they don't account for the geometry of the structures, they also can only provide estimates of the relevant loading.

A mesh refinement study was performed for a 2-D model to select the appropriate mesh refinement for wave propagation. Prior wave impact studies were used to define the surrounding vessel mesh. This combination leads to a mesh of about 40 million cells. Petrobras engineers validated the simulation by comparing the simulation sea state to the experimental results. After an initial transient phase of up to 14 seconds, the simulation data matched the physical testing results.

Engineers synchronized video images from simulation animations and experimental data to qualitatively compare the location, time and intensity of wave breaking effects. Additionally, fluid forces from wave impact were measured in a few locations and compared with the simulation results. Simulation compared well with experimental data except in the hard-to-model breaking wave regions. Even in these cases, simulation values were higher than measured values, demonstrating that they could safely be used to design the deck structures.

loading
Validation of loading results
Validation of wave propagation results
Validation of wave propagation results
green water
FPSO in green water
green water event
green water
 
Simulation using different perspectives of a green-water wave event
Green-water wave event from different perspectives
Simulation provided a good match to experimental results.

"The simulation results provided considerable information that cannot be measured with physical experiments."

SIMULATION PROVIDES ADDITIONAL DATA

The simulation results provided considerable information that cannot be measured with physical experiments. For example, simulation determined loading at every point in the deck structures and hull. Also, simulation provided, for the first time, sufficient measurements to determine the physical mechanisms involved in wave interaction with the hull, especially the velocity field.

Simulation will not replace physical experiments, but will save time and money by reducing the number of experiments that are needed. Setting up and running a physical experiment takes about three months, but additional time is often needed due to test basin availability. Simulation can be completed in about 10 to 50 days depending on problem complexity. This time frame can be reduced in the future by harnessing additional computing resources. Most important, simulation has improved Petrobras' ability to ensure that FPSOs are able to withstand rough seas in lifting oil and gas in pre-salt deposits by providing more detailed loading predictions and other information that cannot be measured by physical testing.

Petrobras is supported by ESSS, an ANSYS Elite Channel partner.

Reference

Silva, D.F.C.; Esperança, P.T.T.; Coutinho, A.L.G.A. Green water loads on FPSOs exposed to beam and quartering seas, Part II: CFD simulations. Ocean Engineering, 2016.

点击此处按钮和ANSYS进行对话

联系我们
联系我们

本网站使用Cookie来改善用户体验,您访问我们的网站即表示您同意接受所有符合我们的cookie政策的cookies。