OIL SPILL SIMULATION: A CASE STUDY IN THE STRAIT OF ISTANBUL

Transcription

1 OIL SPILL SIMULATION: A CASE STUDY IN THE STRAIT OF ISTANBUL Sevilay Can 1, Fahri Celik 2, Huseyin Yilmaz 2 and Osman Askin Bak 3 1 Istanbul Technical University, Maritime Faculty, Tuzla, Istanbul, Turkey 2 Department of Naval Architecture, Yildiz Technical University, Besiktas, Istanbul-Turkey. 3 Trabzon Port management Inc.Trabzon, Turkey SUMMARY Safe ship operation in Turkish straits, the most important routes of oil transportation, is a fundamental issue. Turkish Straits connect the Black Sea and the Aegean Sea, and they enable direct transport of the oil produced in the northern Black Sea to Europe. There, about 15,000 tankers per year are going up continuously. This fact has inevitably increased the risk of catastrophic accidents and consequent oil spills, especially in the Istanbul strait. Large accidents occurring beside the mega city Istanbul (about thirteen millions of inhabitants) can be a great disaster directly attacking the life in this city. Therefore, in the event of an accident, efforts should be made to contain the spilled oil and minimize damage. This paper describes the numerical simulation of oil spill for two different points in Istanbul strait enabling us to estimate the effects in the case of a tanker accident. The output of the numerical simulation can be used to define and specify the proper methodology, and procedures for the Oil Spill Contingency Plan of Istanbul strait. KEYWORDS: Oil spill, CFD, simulation, tanker accident, fluent, Istanbul. INTRODUCTION Large amounts of oil have been produced in the north of the Black Sea, and most of it in Europe. The Turkish straits are one of the most important routes for oil transportation in the world. They connect the Black Sea and the Mediterranean, especially enabling easy oil transport to Europe. According to OSPRI (Oil spill Preparedness Regional Initiative), the region s oil is about 155 million metric tons to be transported through the Turkish Straits in 2010 [1]. The Turkish straits consist of two parts, and the northern half is called the Istanbul strait. It passes through a mega-city, Istanbul, where thirteen million people live. Disasters based on tanker accidents have happened several times in this strait, directly attacking the life of this city. One of the essential problems is to analyze the tanker accidents and oil spill problems, and to consider the safe ship operation in the Turkish straits. In this paper, the basic features of Istanbul strait are described, and a numerical simulation of oil spill originating from a tanker accident is presented. The main objective of the computer modeling is the prediction of the oil slicks behavior, which is the key element for oil pollution control in case of an accident. This provides the opportunity to take necessary immediate precautions. Once the model grids are developed, this study enables to develop scenarios for different environmental conditions such as current speed and direction, boundary conditions, type and amount of the spill etc. Some researchers use their own CFD codes and methods, instead of commercial ones. Reed et al. [2] gives brief information about many models used in studies between One of them is the dispersion model based on the experimental work of Delvigne and Sweeney [3], which has now become a standard. This method estimates the entrained oil mass per unit area as well as unit time, and is used by plentiful researchers. This basic methodology is used in the ADIOS model, the SINTEF oil weathering, OSCAR and OILMAP [4-9]. Nakata et al. [10] used the model, developed by Yapa et al. [11], who presented the equations for a two-layer, two-dimensional system for a Nakhodka tanker oil spill. Sugioka et al. [12] simulated the fate of spilled oil, caused by a Japanese tanker (Diamond Grace) accident using the Lagrangian discrete-parcel method. Skognes and Johansen [13] simulated oil spills from a given location for a large number of weather situations extracted from a historical wind database using StatMap model. A three-dimensional 1517

2 particle-tracking model is used to simulate instantaneous releases from the location shown in Liverpool Bay [14-16]. Hansen and Ditlevsen [17] simulated the oil outflow caused by a tanker accident using the GRACAT simulation computer program (Grounding and Collision Analysis Toolbox). The GRACAT program is an integrated software package developed at the Technical University of Denmark from 1998 to The software consists of a frequency analysis module, a damage analysis module, a consequence analysis module, and a risk mitigation module. In the risk mitigation module, risk profiles for calculated consequences can be calculated. This paper describes the numerical simulation of an oil spill originating from tanker accidents for two different areas, which are called Büyükliman & Rumelifeneri and Büyükdere & Tarabya in Istanbul Strait. It enables us to estimate the effect of oil spill if a tanker accident occurs. The numerical simulation of oil spill has been carried out using a commercial CFD code fluent. Unlike the previous studies, the cases have been analyzed using the complete model of Istanbul Strait; and also the wind effects have been included. This study enables us to develop scenarios for different environmental conditions, such as current speed and direction, boundary conditions, type and amount of the spill in the Strait. The results of the numerical simulations can be used for protection planning and spill response operations for the Oil Spill Contingency Plan of Istanbul strait. The Istanbul Strait has a very high risk of catastrophic accidents, and disasters originating from tanker accidents occurred several times. As these accidents are very close to city life, it requires quick response as soon as possible. Fig.2 shows the comparison of collision risk areas during two periods [19]. MATERIALS AND METHODS The Geometrical Feature of Istanbul Strait Istanbul strait is a narrow and windy route. Total length of Istanbul strait is 32 km, and the narrowest point of the strait is about 800 m wide. The coastline of the Istanbul strait is shown in Fig. 1. The Black Sea lies on the north side of the strait, and the southern area is called Marmara Sea. Istanbul lies on both sides of the strait, and the center of Istanbul city is located at the south end of the strait. Both sides of the strait are steep, and complicated bottom configuration is observed. The bottom configuration contributes to make complex flow field together with high speed current. The strait s current is mostly based on the difference of water density between the Black Sea and the Mediterranean. The maximum current speed in the strait is over 1 m/s, and the following current acting on a fullloaded tanker makes its maneuverability unsteady. As the main factor of current is the difference in water density, the counter current exists at the bottom. Then a shear flow exists in the middle of the section, and there are possibilities that the counter flow and strong shear layer may attack the full-loaded tanker. Numerical Method FIGURE 1 - Coastline of Istanbul strait [18]. The geometrical data, supplied by the Turkish Navy Department of Navigation, Hydrography and Oceanography, were used for the grid generation. The finite volume method was employed here to obtain the solution of Navier-Stokes equation. From the perspective of the physics modeling of the process, the flow was assumed to be incompressible and turbulent. Due to the nature of processing, the simulation was performed in unsteady mode. The time marching method was the fully implicit method, and the solution algorithm was the PISO solver. As a solver, a commercial CFD code Fluent 6.2 was used. Detailed descriptions of the numerical method can be found in the Fluent 6.2 User s Guide [20]. The Realizable k-ε is chosen as the turbulence model [21]. Viscous model constants are taken as defaults. The segregated solver is used for multiphase calculations (oil and water phases), and implicit formulation is set as default. 1518

3 FIGURE 2 - Comparison of collision risk areas during two periods. TABLE 1 - The case properties for oil spill simulation In Istanbul Strait. CASE Station Risk ratio of collision (%) Water Inlet Velocity (m/s) Wind Velocity (m/s) Oil Inlet Velocity (m/s) Oil Inlet Time (s) Total Oil (m 3 ) 1 Büyükliman & Rumelifeneri 5.5 V= V= W= Büyükdere & Tarabya 5 V= V= W= Numerical Conditions The simulation was performed in transient mode utilizing the extended version of the PISO algorithm. The time step size was 10 seconds, and 1500 time-steps were employed. The solution was considered to be converged, when the continuity residual is lower than 10-5 and velocity residuals are lower than All calculations were performed using Fluent 6.2 code. Table 1 shows case properties including locations, water velocity, wind velocity, oil velocity, oil inlet time and total oil. Gambit, which is a pre-processor of Fluent, is used for grid generation of the model. The number of total cells in volume mesh is 324,780. The coastline of the model consists of vertical walls. The bottom of the strait was assumed as flat surface, since the initial aim of the present simulation was related to the surface flow. The non-slip condition was applied to the sidewalls, and the wall-slip condition was used for the water surface including wind stress. The wall-no slip condition was applied to the bottom boundary by considering the limitation of resolution, which means that the present simulation simulates the quasi 2D-flow field. As one-way current exists on the Istanbul strait s surface, the inlet condition was applied to the north section, and the pressure outlet was used for the southern end section. An oil spill was given from the small inlet on the water surface. The inlet speed of oil is shown as a velocity/ time relation (Table 2), being 1/10 of the main current inlet. TABLE 2 - Oil inlet velocity and time. T (s) U (m/s) RESULTS AND DISCUSSIONS Transient calculation was carried out to obtain the behavior of spilled oil. The density ratio of oil/water was 0.87, and the inlet speed of oil was 1/10 of main current inlet. The oil flow is constant in this stage (Table 2), but then the total amount of oil in the calculation domain keeps increasing. In Fig. 3, the results of simulation for Case 1 are given. In this simulation, which started in Büyükliman-Rumelifeneri, it is seen that between s the oil was collected in Büyükdere area, and up to s, it flowed to- 1519

4 a) Case 1 -time-step 1000 (range is between ). c) Case 1 -time-step 7000 (range is between ). b) Case 1 - time-step 4000 (range is between ). d) Case 1 -time-step (range is between ). FIGURE 3 - Calculated oil distribution at different time steps 1000, 4000, 7000, and s, for Case 1. wards the south end of the strait due to the wind and current coming from the north. This case showed that after an accident occurred at the left side of the strait, near Büyükdere, the oil is collected in Büyükdere area due to the counter current, and leaves the area lately. Also in this case, in the narrowest part of the strait, especially in the Emirgan-Kanlıca area, both sides are polluted. In Fig. 4, the results of the simulation for Case 2 are given. Around 3000 s, some intensity is observed on the Tarabya and Rumelihisarı coasts. Similar intensity is also observed in Kanlıca, the right coast of the strait. At 4000 s, some intensity is also observed on the offshore side of Emirgan, but it is also observed that both coasts of the strait are polluted, even if the pollution is more intense in Bebek-Ortaköy and Kandilli. At 5000 s, intensity is observed in the Emirgan area. The presented case studies could be considered as an initial stage for preparing the Contingency Plan of the Istanbul Strait. Accordingly, similar oil spill simulations for probable accident scenarios should be carried out at different locations of the strait. The simulations can be used to provide preventive measures in assessing the risks and damage of the natural resources from the actual and potential spills, and also assist the emergency staff in the development of oil spill planning strategies and response. 1520

5 a) Case 2 - time-step 1000 (range is between ). c) Case 2 - time-step 4000 (range is between ). b) Case 2 - time-step 3000 (range is between ). d) Case 2 - time-step 5000 (range is between ). FIGURE 4 - Calculated oil distribution at different time steps 1000, 3000, 4000, 5000 s, for Case 2. CONCLUSIONS The safety of Turkish Straits is very important, because by a great tanker accident the whole economy can be affected considerably. In this study, two different cases have been analyzed using the complete model of Istanbul Strait. Unlike the previous studies, the entire domain is simulated and wind effects have been included. It is possible to pinpoint polluted areas based on simulation results. The simulation results have facilitated contingency planning. The results show that a tanker accident inside the Istanbul Strait could extensively threaten the life in the city, and to control an oil spill, quick responses will be required. Such simulation models can be used as a basis for an Oil Spill Contingency Plan of Istanbul Strait. This application of numerical simulation can be considered as a preliminary study. The simulations should be performed in the regions where tanker accidents have happened before, or in the risky parts of the Strait, in terms of a disaster. This numerical study is pioneering for a detailed simulation of the Istanbul Strait. REFERENCES [1] Taylor, P. (2005) Oil Spill Preparedness Regional Initiative. A Briefing Paper about OSPRI, [2] Reed, M., Johansen, O., Brandvik, P.J., Daling, P., Lewis, A., Fiocco, R., Mackay, D. and Prentki, R. (1999) Oil spill modeling towards the close of the 20th Century: Overview of the state of the art. Spill Science & Technology Bulletin 5(1), [3] Delvigne, G.A.L. and Sweeney, C.E. (1988) Natural dispersion of oil. Oil and Chemical Pollution 4, [4] NOAA (1994) ADIOS, Automated Data Inquiry for Oil Spills, User's Manual. NOAA/Hazardous Materials Response and Assessment Division, Seattle, Washington. 1521

6 [5] Aamo, O.M., Reed, M. and Daling, P.S. (1993) A laboratory based weathering model: PC version for coupling to transport models. In: Proceedings of the 16th Arctic and Marine Oil Spill Program Technical Seminar, Environment Canada, pp [21] Shih, T.H., Liou, W.W., Shabbir, A., Yang, Z. and Zhu., J. (1995) A new k-ε eddy-viscosity model for high Reynolds number turbulent flows-model development and validation. Computers Fluids, 24(3): pp [6] Daling, P.S., Aamo, O.M., Lewis, A. and Strùm-Kristiansen, T. (1997) SINTEF/IKU oil-weathering model: Predicting oil properties at sea. In: Proceedings Oil Spill Conference. API Publication No. 4651, Washington DC, pp [7] Reed, M., Aamo, O.M. and Daling, P.S. (1995) Quantitative analysis of alternate oil spill response strategies using OSCAR. Spill Science and Technology Bulletin 2 (1), [8] Aamo, O.M., Reed, M. and Downing, K. (1997) Oil spill contingency and response (OSCAR) model system: sensitivity studies. In Proc International Oil Spill Conference, Ft. Lauderdale, FL, pp [9] Spaulding, M.L., Odulo, A. and Kolluru, V.S. (1992) A hybrid model to predict the entrainment and subsurface transport of oil. In: Proceedings of the 15th Arctic and Marine Oil Spill Program Technical Seminar, Environment Canada, pp [10] Nakata, N., Sugioka, S. and Hosaka, T. (1997) Hindcast of a Japan Sea oil spill. Spill Science & Technology Bulletin 4 (4), [11] Yapa, P.D., Shen, H.T. and Angammana, K. (1994) Modeling oil spill in a river-lake system. J. Marine Systems 7, [12] Sugioka, s., Kojima, T., Nakata, K. and Horiguchi, F. (1999) A Numerical Simulation of an Oil Spill in Tokyo Bay. Spill Science & Technology Bulletin 5 (1), [13] Skognes, K. and Johansen, O. (2004) Statmap a 3-dimensional model for oil spill risk assessment. Environmental Modeling & Software 19 (7-8), [14] Elliott, A.J. (1991) EUROSPILL: Oceanographic processes and NW European shelf databases. Marine Pollution Bulletin 22, [15] Elliott, A.J. and Jones, B. (2000). The need for operational forecasting during oil spill response. Marine Pollution Bulletin 40, [16] Elliott, A.J. (2004) A probabilistic description of the wind over Liverpool Bay with application to oil spill simulations. Estuarine, Coastal and Shelf Science 61 (4), [17] Hansen, P.F. and Ditlevsen, O. (2003) Nature preservation acceptance model applied to tanker oil spill simulations. Structural Safety 25 (1), [18] Örs, H. (2003) Oil Transport in the Turkish Straits System: Simulation of the Contamination in the Istanbul Strait. Energy Sources 25 (11), [19] Yurtoren, C. (2004) Maritime Traffic Management in Istanbul Strait. Department of Maritime Transportation Systems, Graduate School of Science and Technology, Kobe University, PhD Thesis (in Japanese). [20] Fluent Inc (2005) Fluent 6.2 Documentation. Commercial Code Package Users Guide, Germany Received: March 08, 2007 Revised: July 05, 2007 Accepted: August 30, 2007 CORRESPONDING AUTHOR Sevilay Can Istanbul Technical University Maritime Faculty Tuzla, Istanbul TURKEY sevcan@itu.edu.tr FEB/ Vol 16/ No 11b/ 2007 pages