A New Concept for CNG Carriers and Floating CNG/Oil Processing and Storage Offshore Platforms

Transcription

1 A New Concept for CNG Carriers and Floating CNG/Oil Processing and Storage Offshore Platforms Regu Ramoo & & Mohan Parthasarathy Altair ProductDesign, Inc E. Big Beaver Road, Troy, MI 48083, USA Prof. Thomas Lamb University of Michigan Ann Arbor, MI USA copyright Altair Engineering, Inc. 2011

2 Abstract The collection and transportation of Natural Gas from small gas fields is the current focus of the Offshore Gas industry. To transport the gas from the field to the land distribution point there are two choices LNG or CNG. For the small fields and/or short transportation distance it appears that the best solution is CNG. The paper describes the application of a new containment tank system that has significant benefits for CNG carriers as well as Floating Oil/CNG Processing and Storage Offshore Platforms (FOCNGPSO) and Floating CNG Processing and Storage Offshore Platforms (FCNGPSO). Because of the shape of the CDTS advanced structural analysis is required to design it. Therefore the paper also presents the application of CAE methods using Altair Engineering s Hyperworks suite of advanced structural design, analysis and optimization tools. State-of-the-art optimization techniques like Topology and Free-Size optimization are showcased in this effort. 1.0 Introduction The transport of LNG by sea started in the late 1960s and has continued to grow, almost constantly, since then. Economic trends suggest that this will continue for the near future with new fields and new consumers entering the market. With the growth in demand the collection and transport of natural gas from small (stranded) gas fields has been investigated for the past 15 years. It has been concluded that the best way to collect and transport gas from a small field is by compressing it. Compressed natural gas (CNG) requires a storage volume approximately twice that of LNG but does not need the expensive refrigeration plant at the source or the gasifying equipment at the receiving end. This can be clearly seen from the results of a study (CMCNG 2006) shown in Table 1. Table 1. Delivery Chain Capital Cost Comparison ($US Millions) The concept of carrying compressed natural gas on ships is not new. It was tested by Columbia Gas in 1965 around New York Harbor (FRIIS 2007). The containment system was of the multiple bottle or cylinder type. It is reputed to have failed due the low trading rate for natural gas at that time as well as the high cost of steel pressure vessels designed to ASTM standards. A number of transportation containment systems have been developed but none yet built. They all consist of thousands of meters of pipe in stacks or coils. Some of the solutions are a combination of refrigeration and pressure to reduce the pressure at which the gas can be compressed such as the VOTRANS approach, shown in Figure 1 (FRIIS 2007), which stores Copyright Altair Engineering, Inc.,

3 CNG at -30 C and 125 bar. All the existing systems are very volume inefficient thus requiring large ships. For example an 23MMscm (800MMscf) CNG Carrier carrying 15,000 t of CNG would be the same size as an 70,000 m3 LNG ship carrying 35,000 t of LNG. The only system that is at ambient temperature is the SEA NG coselle approach as shown in Figure 2, but the proposed ships are quite small in capacity, the largest design so far being for 7.7 MMscm (275 MMscf) or 5000 t. Even though the capacity of natural gas is small, the ship would still be relatively large in length. The problem is that the weight of the gas storage coselles is great, about 35,000 t for a cargo deadweight of only 5,000 t. This results in a very low deadweight ratio and it may be the reason why it is taking so long to get a CNG Carrier built and operating. The paper reports on the application of the CDTS to the marine transportation and storage offshore and shows that the resulting benefits are in both operating and acquisition cost for CNG Carriers and FOCNGFPSO and FCNGPSO offshore production and storage platforms. It will also propose a combined CNG/Oil Carrier. Figure 1. VOTRANS System 20MMsdm (700MMscf) CNG Carrier Copyright Altair Engineering, Inc.,

4 Figure 2. SEA NG 16 Coselle 1.4MMscm (50MMscf) CNG Carrier 2.0 CDTS Description The CDTS was developed for the carriage of LNG. The basis for its design was constructing a self-standing tank surface composed of 12 identical, in form, intersecting cylinders that formed the twelve edges of a cube that would have a significantly better volumetric efficiency than a spherical tank. Where the intersecting cylinders met in the center of each face a closing cap was provided. Figure 3 (from the original patent) shows the form of the tank. In 2005 Regu Ramoo joined Lamb in the continuing development of the CDTS using the ALTAIR Engineering advanced structural analysis and simulation systems. It became immediately clear that the original tank structural objectives could not be attained as proposed but they could be attained by connecting all the side centers together by a cross bracing structure as can be seen in reference paper (LAMB 2009). The cross brace system which performed the desired function for LNG containment dealing with hydrostatic and dynamic loads especially sloshing, did not perform as well for the pressurized tank and a complete intersecting cylinder option proved to be superior. The complete cylinder tank option is shown in Figure 4, which shows the current tank configuration with a portion of the top removed and also a corner of the tank to demonstrate the intersecting cylinders and the corner closing cap. Copyright Altair Engineering, Inc.,

5 Figure 3: Original Geometry of the CDTS Figure 4: Current Geometry of the CDTS 3.0 The Application of CDTS to CNG Carrier It was always the intention to explore the use of the tank for pressures above atmospheric, initially to increase the temperature at which LNG could be contained but recently the application of the CDTS to CNG Carriers and FOCNGPSO and PCNGPSO was examined. Whereas the CDTS size for LNG application had no limitation up to that required for the largest LNG Carriers under development, the CDTS tank for the carriage of CNG will be much smaller and will be a compromise between shell thickness, weight and manufacturability. Copyright Altair Engineering, Inc.,

6 A solution that has superior volumetric and weight usage is proposed that utilizes a new CNG containment tank system, namely the Cubic Doughnut Tank System (CDTS). It has a hold volumetric efficiency of 0.33 (VOTRANS 0.18 and SEA NG 0.20) a ship volumetric efficiency of 0.14 (VOTRANS 0.09 and SEA NG 0.09) and the ships would have a cargo deadweight coefficient of (VOTRANS 0.12 and SEA NG 0.09); a significant improvement over other proposed designs. Another important advantage of the CDTS is the very significant reduction in surface area of the containment system. This is important as it impacts the heat transfer into or out of the contained CNG and this in turn increases the CNG pressure due to increasing in gas temperature. The result of its many benefits is significant acquisition and life cycle cost savings compared to the other proposed designs. The CDTS offers the following benefits compared to other proposed systems for CNG Carriers and offshore platforms: significant reduction in ship or platform length, significant reduction in Gross Tonnage significant reduction in tank surface area and thus CNG gain of heat, smaller propulsion power for CNG Carriers, savings in fuel for CNG Carriers, the in service maintenance benefit in that the tank structure can be inspected, and significantly reduced number of tank manifolds While the CDTS offers benefits just from the tank design, construction, and installation in the ship, it also offers unique benefits in the design of the ship/platform including shorter length, which has construction benefits in less work content for the same capacity ship compared with any other system. The CDTS has been applied to LNG Carrier design (Lamb 2009) and offers significant acquisition and life cycle cost savings compared to other tank containment systems. It has been found to offer similar cost savings for CNG Marine applications. A range of CDTS size was explored in the preliminary structural analysis to determine tank volume and average shell thickness, and is presented in Table 2. The 10 m CDTS tank was selected to demonstrate its application to CNG Carriers and FPSOs, as it was the best compromise between shell thickness, weight and other construction limits. Copyright Altair Engineering, Inc.,

7 Table 2. CDTS Tank Characteristics As a means to demonstrate these benefits a comparison of CDTS CNG Carrier designs with the VOTRANS and SEA NG designs is offered. Table 3 shows the Principal Characteristics for VOTRANS and a much smaller SEA NG Carriers as well as a CDTS Carrier of the same capacity as the VOTRANS Carrier. Figure 5 shows the three containment system ships of the same capacity. Note the significant reduction in ships size for the CDTS CNG Carrier. The VOTRANS ship is a published design by EnerSea Transport, the SEA NG design was developed by Lamb based on available published data and the CDTS Carrier is also developed by Lamb. Table 3. Comparison of Current Designs with CDTS Designs Copyright Altair Engineering, Inc.,

8 Figure 5. Comparison of 23MMscm (800MMscf) CNG Carriers - All to same Scale The Cost Index is based on ship differences and excludes the cost of manufacturing and installing the CNG tanks/modules. All cost estimates were made using a Preliminary Design Cost Estimating Model and were for Average World Prices for Labor and Material. They were converted to cost indices as it was only the magnitude of the differences in cost that was necessary for comparison. The model is a parametric weight-based approach at the US Navy Ship Work Breakdown Structure (SWBS) Summary (single digit) level with special factors to take into account ship size and complexity (LAMB 1998, LAMB 1999). It was first developed over 40 years ago but has been continuously refined by use over all the years since. It was further developed for the PODAC Cost Model developed for the US Navy in 1998 (ELLIS 1998). This approach (or methodology) has been found over time to predict shipbuilding cost within plus or minus 10% with very few outliers. The estimated ship acquisition cost, excluding tank/module material, fabrication and installation, for the CDTS CNG Carrier would be 12% less than the VOTRANS Carrier and 29% less than a comparable capacity SEA NG Carrier. Required SHP for the CDTS would be 6% and 37% less than that for the VOTRANS and SEANG Carriers respectively, thus providing a life cycle cost benefit from the fuel savings. The Gross tonnage Index is based on estimates of the Gross Tonnage for each CNG Carrier. The Gross Tonnage is an important life-cycle cost driver as many of the fees are based on it. The GT Index is 15 % less than VOTRANS and 60% smaller than SEA NG. Copyright Altair Engineering, Inc.,

9 A series of 11 CNG Carrier designs utilizing the CDTS tanks is presented, and shown in Figure 7, based on number of tanks ranging from 4 to 108 and 1.1 to 28.4 MMscm, at 125 bar pressure and -30 C temperature. The tank size selected was 10 meters length, breadth and height based on a workable thickness of shell plating and lifting weight. Characteristics for the Series CNG Carriers is given in Table 4. The Cargo Deadweight Coefficient for the CDTS Carrier is which is 11% better that the VOTRANS and 48% better than the SEA NG Carriers. The Ship Volumetric Coefficient is given by dividing the Carrier CNG Capacity in m3 by the LBP, Beam and Depth to Top of Cover. Again the CDTS CNG Carrier is 14, which is 40% better than the VOTRANS and the SEA NG Carriers. The Hold Volumetric Efficiency is given by dividing the Carrier CNG Capacity in m3 by the length, breadth and height of the surrounding bulkheads. The CDTS CNG Carrier is 33, which is 83% better than the VOTRANS and 135% better than the SEA NG Carriers. Figure 6. CDTS CNG Carrier Series Copyright Altair Engineering, Inc.,

10 Table 4. CDTS CNG Carrier Series It is possible to design a combination CNG/OIL Carrier by simply increasing the beam and locating the Oil Tanks in wing tanks Port and Starboard similar to that shown in Figure 7 for the FOCNGPSO. The figure shows a 40 tank platform but obviously the concept can apply to any tank number from 4 to 108. A detailed design of the tank and the derivation of production data is still to be done but as an interim check, basic production information such as number of plates and joint weld length, was estimated for the VOTRANS and the CDTS 800MMscf CNG Carriers and is presented in Table 5. The significant reduction in tank/pipe surface area should be noted. Copyright Altair Engineering, Inc.,

11 Table 5. Production Parameters for 800 MMscf CNG Carrier 4.0 FOCNGPSO Tank Arrangement Before the natural gas can be transported by ships it must first be collected. Unfortunately many of the gas fields are small compared to the large oil fields. Thus it has not been economically viable to recover the gas from them up until now. However with increasing demand, and a decreasing supply of easily recovered energy it is becoming necessary to investigate how to change the situation. The first Floating Liquefied Natural Gas (FLNGPSO) platform is in operation. Figure 7 shows a concept design for a 10.5 MMscm/200,000 Bbl FOCNGPSO and Figure 8 shows the midship structural arrangement. Other smaller and larger combinations are given in Table 6. As already stated CNG has a density slightly less than QUARTER that of oil. Thus the CNG tank space dominates the design. To date there are only designs for FOCNGPSOs. Copyright Altair Engineering, Inc.,

12 Table 6. CDTS FOCNGPSO Series Figure 7. CDTS FOCNGPSO Arrangements Copyright Altair Engineering, Inc.,

13 Figure 8. FOCNGPSO Midship Section 5.0 Structural Analysis 5.1 Introduction ALTAIR Engineering is a leading global provider of technology that strengthens and empowers client innovation and decision-making through technology that optimizes the analysis, management and visualization of business and engineering information. Recent focus has been on exploring the application of its software to the marine design sector. Altair Engineering s Hyperworks is a computer-aided engineering (CAE) simulation software platform that allows businesses to create superior, market-leading products efficiently and cost effectively. The Hyperworks platform offers: modeling & visualization as well as analysis & optimization solutions. Further information is available on ALTAIR S URL: To demonstrate a typical application, the software was used to design a new tank system for cargoes such as LNG and CNG. The basis for the CDTS design was constructing a selfstanding tank composed of 12 identical intersecting cylinders that formed the edges of a cube. A closing cap was provided at the center of each face and at the 8 corners where the cylinders intersected as can be seen in Figure 1 (from the original patent), and Figure 2 which shows the configuration of the tank. The tank is of a complex shape and as such does not lend itself to simple analysis. An advanced structural analysis approach is required. Starting from 2005, Altair Engineering applied the Hyperworks suite of advanced structural design, analysis and optimization tools in order to improve the design to meet the structural objectives which could not otherwise be attained by the proposed original design. This involved connecting the center of all faces by an internal cross brace also shown in Figure 2. The finite element analysis and optimization was performed using Altair OPTISTRUCT, which is a linear finite element solver available in Altair Engineering s Hyperworks. Only the forces due to hydrostatic pressure were considered for the initial analyses and optimization. Copyright Altair Engineering, Inc.,

14 The CDTS was originally intended for LNG applications and was designed to withstand the hydrostatic and dynamic loads, including sloshing. The CDTS structure for LNG application has been extensively analyzed. A CNG tank will see none of those loads. Instead it has to be designed as a pressure vessel to meet ASTM and Classification Society Rules. For the CNG application, the design is driven by internal pressure and must meet ASTM and Classification Society Rules for pressure vessels. A brief overview of the different optimization techniques that are available in Altair Optistruct is presented in the next section. Results of the analyses and optimization of the CDTS tank under internal pressure are discussed in the subsequent sections. 5.2 Optimization Techniques The mathematical statement of any structural optimization problem can be posed as Minimize f(x) = f(x1,x2, Xn) Subject to gj(x) 0 j = 1,2, m Where f(x) is the objective function, X1,X2, Xn are the design variables and gj(x) are the constraints. Typically the objective function is the compliance of the structure for the given loading and boundary conditions and the constraint is on the mass, volume fraction of the material in the design space or any response like displacement, stress, etc. When there are multiple load cases, a weighted compliance is used as the objective. The weighted compliance is given by Cw = Σ wici, where Ci and wi are the compliance and weight associated with each load case respectively Topology Optimization Topology Optimization is a mathematical technique that produces an optimized material distribution/shape of the structure within a given package space. As in the size and free-size optimization, the objective function is typically the weighted compliance of the structure for the given load cases. The design variable is the material density of each element in the finite element model of the design space and it varies continuously between 0 and 1 which represent the states of void and solid respectively. A distinction should be made between this density and the physical mass density of the material of the structure. The goal of any topology optimization is to achieve a value of either 0 or 1 for the density variable. Since the density variable is continuously varying, many intermediate values are possible though not desirable. In order to avoid intermediate values for the density variable, a penalization technique is use and is given by K (ρ) = ρp K Where K is the actual element stiffness matrix (the real density of the material is used to compute the actual element stiffness matrix), K is the penalized element stiffness matrix, ρ is the material density or the design variable and p is the penalization constant which varies between 2 and 4. Using a value of p greater than 1 gives a small value for the stiffness and thus penalization is achieved when the optimization problem is posed as minimization of Copyright Altair Engineering, Inc.,

15 compliance (or maximization of stiffness). For details of the different optimization techniques mentioned above the reader is directed to the Optistruct user manual and the references given in [2] Free-Size Optimization In free-size optimization, the thickness of each element in the finite element model of the design space is treated as a design variable. This is the fundamental difference between freesize and conventional size/gage optimization. Unlike conventional size optimization, free-size optimization results in continuously variable shell thickness in the design space, between the given lower and upper bounds of the thickness. A part with variable thickness is typically far more expensive to manufacture and may not be a viable choice at first glance. It should be emphasized that the results of free-size optimization should not be considered as a final design. Based on this result, the design space should be subdivided into smaller zones and a conventional size optimization could then be performed to fine tune the thickness of the different zones. The design variables for this size optimization would be the thickness of various zones Size/Gage Optimization Conventional finite-element based size optimization techniques require the use of engineering judgment or intuition to make a priori decisions as to how the design space should be discretized using different design variables. Based on how the design variables are defined, the optimization algorithm then iteratively explores the combination of design variable levels that minimizes the objective function subject to the constraints that were imposed. The number of design variables is typically limited to about 50 to 300 due to computational cost and effectiveness of computational search algorithms. Any size parameter in the finite element model of the design space like the thickness of a shell component, the moment of inertia of a beam component etc. could be used as a design variable. 5.3 Preliminary Results and Discussion Baseline Design The baseline design of the CDTS is shown in Figure 9. The CDTS is made of 12 identical cylinders of diameter about 4.7m which intersect at the four corners to spherical caps. The size of the cube circumscribing the CDTS (excluding the base) is 10m. A uniform shell thickness of 100mm was initially assumed. The material used for the tank is manganesemolybdenum steel alloy with a modulus of 210,000 MPa and Poisson s ratio of 0.3. The mass of the baseline design is 873 t. An internal pressure of 2000 psi was applied on the walls of the tank. Due to symmetry, a quarter model of the tank was considered for the finite element analysis. The base was constrained in vertical displacement and symmetry boundary conditions were applied to the two planes of symmetry. The contour plot of von Mises stress is shown in Figure 10. The average value of the ultimate strength of manganese-molybdenum steel alloy is about 800 MPa. The desired stress level was set as 400 MPa which is about 50% of the average value of the ultimate strength. As can be seen in Figure 10, a significant portion of the tank is over the desired stress level of 400 MPa. Copyright Altair Engineering, Inc.,

16 Figure 9: Baseline Design of the CDTS Figure 10: Von Mises stress in MPa Baseline Topology Optimization Topology optimization was performed on the baseline design in order to determine the optimal material distribution. The design space used for the topology optimization is shown in Figure 11. Figure 11: Design Space used for Topology Optimization Copyright Altair Engineering, Inc.,

17 The objective of the topology optimization was minimization of the compliance with a constraint on the volume fraction of the material as 30%. The design space was filled with first order tetrahedral elements. The load path or the optimal material distribution obtained from the topology optimization is shown in Figures 12 and 13. Figure 12: Load Path from Topology Optimization Figure 13: Load path from Topology Optimization Free-Size Optimization Using the load path of the topology optimization as a guideline, internal bulkheads were added as shown in Figure 14 to 15. As topology optimization is a design tool used to provide critical insight on the structural load path, manufacturability and fabrication considerations must be taken into account when interpreting these results. Copyright Altair Engineering, Inc.,

18 Figure 14: Bulkheads Incorporated based on the Load Path from Topology Optimization Figure 15: Bulkheads Incorporated based on the Load Path from Topology Optimization A free size optimization was performed on the modified design in order to determine an optimal thickness distribution that reduces the mass and yet maintains a lower stress level. The free-size optimization was posed as minimization of compliance due to the 2000 psi internal pressure with a stress constraint of 400 MPa and mass constraint of 500 t. The thickness of the various components of the CDTS was allowed to vary from 15mm to 120mm. The continuously variable thickness distribution obtained from the free-size optimization is shown in Figures 16 and 17. Copyright Altair Engineering, Inc.,

19 Figure 16: Thickness from Free Size optimization Figure 17: Thickness from Free Size optimization Based on these results, discrete thickness values were assigned to different parts of CDTS so as to minimize the number of regions with disparate thicknesses. The mass of the tank is about 594 t. This discrete thickness map is shown in Figure 18. The resulting von Mises stress distribution is shown in Figure 19. With this thickness distribution the maximum stress is close to the endurance limit (400MPa). Figure 18: Discrete Thickness Map Figure 19: Contour Plot of Von Mises Stress (MPa) Copyright Altair Engineering, Inc.,

20 5.3.3 Design Refinement Considering the stress contours of Figure 19 and factoring manufacturability considerations, the internal bulkheads were trimmed. The critical load path contours from the earlier topology runs also indicated a sparser material distribution on the bulkheads adjacent to the spherical caps. Additionally, limiting the welding of the bulkheads to the seams of the intersecting cylinders and cap instead of the center of the cap will significantly reduce construction complexities and the need to weld the bulkheads to the spherical caps. High stress concentration seen at the corners of the bulkheads can be addressed by designing in generous fillets in these regions. Figure 20: Bulk head with cut out Figure 21: Contour Plot of Von Mises Stress (MPa) Validation of Shell Model A solid model with the same thickness as the shell model was created using hexahedral and pentahedral elements and the analysis was performed using Radioss Block which is part of Altair Hyperworks. Unlike the shell model, the pressure load was applied to the bulkhead including the edges around the cut outs. Figures 22 to 25 compare the results obtained using the shell and solid models. Copyright Altair Engineering, Inc.,

21 Figure 22: Von Mises stress - Shell Model Figure 23: Von Mises stress - Solid Model Figure 24: Von Mises stress - Shell Model Figure 25: Von Mises stress - Solid Model The results obtained are in good agreement. Though this was expected since transverse shear flexibility is considered for the shell elements used in Optistruct, this comparison adds credibility to the design and analysis approach using the shell model. For future work only the shell model will be used as it is easier to implement design changes to a shell model than a solid model. Copyright Altair Engineering, Inc.,

22 6.0 Conclusions The paper has shown the benefits of a new Tank Containment System, namely the CDTS, for the carriage of CNG in ships and floating production and storage platforms, that compared to other existing designs: significantly reduced the size (length, displacement and SHP) compared to the other CNG systems currently being developed, reduced the estimated acquisition cost of carriers by 5% to 20%, reduces the Gross Tonnage and therefore many operating costs by 15% to 60%, reduces surface area and thus heat transfer by a factor of 5 compared to VOTRANS and 30 compared to SEA NG, all combining to offer a technical cost effective solution for both CNG Carriers and FOCNGPSO/FCNGPSO. It also presented the results of the preliminary structural analysis validating the adequacy of the design while demonstrating the use of ALTAIR Engineering's Hyperworks suite of software. Structural simulation studies evaluating trade-offs between material and fabricating cost with containment pressures and temperatures are currently ongoing. 7.0 Acknowledgements The authors would like to acknowledge with thanks the support of ALTAIR Engineering and their vision of a future for the CDTS. The application of their advanced CAE tools to a practical idea has enabled this concept to be further developed and proven its feasibility. 8.0 References CCMCNG -Center for Marine CNG, Inc, Marine CNG: Viability of Supply, Presentation to National Energy Board, April 25, 2006 FABER, F, BLIAULT, A.E.J., RESWEBER, L.R., and Jones, P. S., "Floating LNG Solutions from the Drawing Board to Reality," Paper 14100, OTC 2002 FRIIS, D., and Abdi, A., "Marine Transportation of Compressed Natural Gas - Overview of Existing Technologies and Standard Implications," CCMCNG, 2007 LAMB, T, and Ramoo, R, "The Application of a New Tank Containment System to ULTRA- Large LNG Carriers," Paper, OTC 2009 LOTHE, P., "Pressurized Natural Gas-An Efficient and Reliable CNG Solution for Offshore Gas Transportation," Paper 17231, OTC 2005 RADIOSS/Optistruct 9.0 User s Guide, Altair Engineering Inc., 2008 Copyright Altair Engineering, Inc.,

23 RAMOO, R. and LAMB, T., "The Use of Advanced Structural Analysis and Simulation Tools to Validate a New Independent LNG Containment System," ICCAS 2009, Shanghai, China Copyright Altair Engineering, Inc.,