NEW DESIGN CONSIDERATIONS OF CRYOGENIC FUEL PUMPS FOR LNG CARRIERS USING DUAL FUEL PROPULSION SYSTEMS Jinkook Lee, Ph.D. Director, Advanced Technology Carter Cryogenic Products Division Argo-Tech Corporation Cleveland, Ohio, USA leej@argo-tech.com www.argo-tech.com ABSTRACT The purpose of this paper is to address the considerations of solving practical problems in designing fuel pumps for LNG Carriers particularly using dual fuel propulsion system. For decades, steam turbine propulsion for LNG Carriers has been the rule. However, it was the obvious choice in order to dispose of the unavoidable natural boil of gas. Therefore, alternatives to steam turbines were developed such as Diesel Gas Electric (DGE) propulsion with engines operated in a lean burn gas combustion as the best choice. In dual fuel propulsion systems, pumping liquid gas (LG) during voyage of high wave can be a problem with particularly low fuel level because the pump will experience sudden loss of priming by sucking air due to an exposed inlet. Therefore, it is increasing that the ever-growing demand of a high performance cryogenic pump be able to handle severe two-phase flow for continuous operation. Discussed in this paper, a novel method of solving continuous fuel pumping problem of LNG Carriers at high sea. The hope is that the definition of problems and the outline of general solutions described in this paper will open meaningful discussion and application to other fluid engineering applications. PO-35.1
INTRODUCTION The recent development of dual-fuel operated diesel and gas engines has made it possible to use the boil-off gas efficiently and therefore propulsion based on diesel engines is a strong option for modern LNG carriers today. Although slow or medium speed diesel engines have been selected for some of the recent LNG carriers, owners are also considering and specifying the dual fuel installation options that uses both gas boil-off and ordinary bunker C fuel oil. Variations of the dual fuel arrangements include: diesel engine or gas turbine driven generators with one propulsion shafting system and a liquefaction plant; and diesel engine or gas turbine driven generators with two azimuth thrusters and a liquefaction plant. To date, slow speed diesel with re-liquefaction plant as well as a gas combustion unit and medium speed dual fuel diesel with gas combustion units, are the preferred options for the new large LNG carriers recently ordered. In modern LNG carriers, the amount of natural boil-off is decreasing due to advances in tank insulation technology and design. Therefore the energy in the natural boil-off gas is far from sufficient to produce the propulsion power needed for the relatively high operating speed. A natural boil-off rate of 0.15% per day is typically considered as the design point. However, in recent days these values as low as 0.10% per day have been reported in modern vessels. Therefore it is expected that the energy content of gas is not constant due to decreasing of nitrogen contents of the boil-off gas during the voyage. Based on electric propulsion where the prime movers in the power plant are fourstroke low-pressure dual-fuel gas engines, the typical 135,000 m 3 ship needs approximately 30 MW total engine output power. Recent studies suggest that the most beneficial solution to top up the need for additional energy is to use forced boil-off instead of fuel oil. This solution in combination with dual-fuel engine electric propulsion is economically superior both in installation cost and operation. At the engine inlet, the dual-fuel engine operating gas pressure is in between four and five bar. In dual fuel propulsion system, pumping liquid gas (LG) at rough sea could be a problem, particularly when fuel level is low. The fuel pump will experience sudden loss of priming by sucking air due to exposure of pump inlet into air during voyage through high waves. The resulting fuel supply to the engine may be interrupted momentarily when the ship needs more propulsion power. In this paper, a novel fuel pump designed specially for LNG carriers and its performance test data is introduced to demonstrate continuous fuel supply to the engine operating in rough sea voyage. DESIGN CONCEPT The basic design concept of fuel pump for dual fuel LNG transport ships was conceived to supply liquid fuel to the engine continuously even at the very rough sea PO-35.2
voyages. A novel fuel pump was design to overcome the high fuel lift in case of fuel level lower than pump inlet up to 0.3 meter or 12 inches and reprime the pump within a reasonably short period time for smooth and continuous engine operation. DESIGN REQUIREMENTS However, this fuel pumping system is a functional unit tailored to respond to engine demand requirements. Therefore, the engine operating characteristics define the limits of a pump s performance. It is significant in quantitative view point to note that the typical engine fuel pump requires relatively low flow rate such as less than 20 m 3 /hr and very high head rise of over 200 meter. As previously mentioned, two other significant requirements at rough sea voyage exist compared to the pumps on ground, high two-phase flow or high vapor to liquid mixture (V/L) flow pumping capacity and self-prime (dry-lift) capability. Both are associated with, and caused by, conditions upstream of the engine pump. GENERAL DESIGN OBJECTIVES To supply continuous liquid fuel to the engine even at the time of high sea. To be designed to meet and exceed all performance requirements. To be designed for safety and environmental soundness. To be designed for high quality, high reliability and using proven technology. To be designed for long service life and give more value to the customer. DESIGN CONSIDERATIONS Two-phase flow occurs when boil-off gas due to external heat transfer to the liquid enters to the pump inlet or when liquid gets agitated by turbulence (sloshing effect) due to rough sea. Under these circumstances the liquid fuel has to be sucked through the plumbing by the engine pump itself. Because of turns and change in elevation of the fuel line, a small amount of pressure drop of the fuel occurs by the time when it arrives at the engine pump. This might cause some fuel vaporization. The result is a mixture, not necessarily homogeneous of gas and liquid. However, this fuel pump has to deliver fuel in 100% liquid form to the engine. This is the function of the engine fuel pump component design consideration. Under the dry-lift condition due to exposure of pump inlet to the air at rough sea, the pump is required to evacuate the air from a dry inlet (no liquid fuel) line, of a given volume and at a given conditions of line elevation above fuel level to lift fuel through the line to the pump and to deliver the fuel to the engine at some required pressure within a specified time interval. Thus, in final selection of pump performance parameters, the most restrictive combinations of pressure, flow and temperature conditions should be selected. PO-35.3
TEST SET UP The fuel pump test set up for dry lift is illustrated in Figure 1. It is intended to test the capability of ingestion of air trapped in the inlet fuel line due to exposure of bellmouth of pump inlet to the air in the event of high sea. For the worst case scenario, the center line of pump is located 12 inch (0.3 cm) above fuel surface in the tank to measure the dry lift capability and duration of repriming the pump. It is also equipped with sight glass for observation of the fuel level. Vacuum leak test of the inlet line is also recommended if fuel tank pressure becomes lower than atmospheric pressure. Figure 1. Schematic of Dry Lift Test Set Up Actual pump test set up is shown in Figure 2. The pump inlet line is equipped with a clear tube to visualize the two phase flow coming to the pump inlet. The clear tube makes it easy to see the air and liquid flowing into the pump and by having this clear tube it is much easier to measure the duration of dry lift and repriming. PO-35.4
Figure 2. Dry Lift Pumping Test Set Up TEST PROCEDURE The pump is installed subject to the requirements and the test schematic of Figure 1. The test will be conducted in the following order: a). The test liquid level in the tank is adjusted up to 12 inches below the center line of pump inlet and specific gravity of test liquid is around 0.74. b). The pump and suction line shall be primed as specified. c). The pump shall be operated at a specified pump speed to establish test fluid circulation and pump will operate at low and high flow rates. d). While maintaining fuel circulation and control of tank pressure and temperature, fuel samples shall be drawn for measuring fuel properties. e). Record the time from the start of pump to start of liquid sucking into the pump. RECOMMENDED FACTORY PERFORMANCE TEST IN LNG Hydraulic Performance at rated speed Pump down and NPSH Vibration and rotor stability (monitor) Starting current test (monitor) PO-35.5
Post test disassembly and inspection New set of bearings installed in each pump after performance testing METHOD OF TEST AND RESULTS The pump is run with no fluid supplied at the mouth of the inlet tube for both short and long periods of time. With the mouth of inlet tube now submerged in liquid, the pump is tested to see if it can reprime itself and deliver rated flow rate in a very short period of time. The centrifugal pump which has an impeller equipped with reprime capability was tested to measure the pump performance at severe two-phase flow conditions which can happen in the very high sea voyages. To test the capability of rapid pump repriming, the specimen pump was located at an elevated position from the fuel level. The inlet line which was made of transparent tube was used to visualize the two-phase flow flowing into the pump. Air entrained in inlet pipe between pump inlet and fuel surface was calculated and reprime time was measured to know how quickly liquid fuel was pumping into the engine. Each time, short and long dry run tests were also made prior to reprime test. Tear down pump visual inspection was made after testing. Pump components such as rotor and shaft assembly, impeller, bearings, volute housing, reprime pump exhaust port, o-rings, were examined and inspection results were recorded for book keeping. REPRIMING The reprime capability of the pump was tested while the pump was running at various speeds at sea level and the test results are depicted in Figure 3. The first test was conducted at various pump speeds and 170 second of reprime time was recorded at 5% pump speed and 5 second at 100% pump speed, respectively. It was also conducted at constant pump speed with two different flow rates of 12,000 and 3,000 pound per hour after two different dry run times of 1 minute and 11 minutes, respectively. It was found that there were no major differences in reprime time with respect to the duration of dry run. PO-35.6
Reprime Time Test Rep rim e T im e ( sec) 180 160 140 120 100 80 60 40 20 0 0 0.2 0.4 0.6 0.8 1 1.2 Normalized Pump Speed Figure 3. Pump reprime time test versus normalized pump speed DRY LIFT The pump must be able to prime itself from a fuel supply located below the pump for all anticipated operating conditions. However, dry lift is seen to be adversely affected by inlet line length, excessive diameter, pump leakage, back pressure and carry over or trapped volume of air. Large displacement rates for the pump are helpful if either power or weight conservation is not a requirement for the design. Dry lift capability was demonstrated for several pump speeds while fuel specific gravity was maintained at 0.74. The pump and fuel were conditioned by circulating the fuel. The pump was stopped and the fuel within the pump and inlet line drained by means of system vent valves. The pump was then operated with pump discharge vented to the tank ambient and the time to reprime and establish flow and adjust PO-35.7
pressure recorded for each of the pump operating speeds. The results of dry lift test data are plotted in Figure 4. CONCLUSIONS Figure 4. Dry lift test data at constant flow rate The test pump was capable of priming itself when operated at specified conditions and showed dry lift capability of up to 14 inches. It was also found that faster the pump runs shorter the reprime time gets at a fixed inlet pipe volume. ACKNOWLEDGMENTS The author is grateful for the permission and valuable support by Argo-Tech Corporation of supplying test pump, test apparatus, conducting various tests, and test data which makes this paper possible. BIBLIOGRAPHY Cooper, Paul C., 1963, Advanced Inducer Study, NAS 8-4005, ER-5288, TRW Electromechanical Division, May. PO-35.8
Gorla, R. S. R. and Khan, A. A., 2003, Turbomachinery-Design and Theory, First Edition, Marcel Dekker, Inc. Jacobsen, J. K. and Keller, R. B., 1971, Liquid Rocket Engine Turbopump Inducers, NASA SP-8052m May. Lee, Jinkook, 2005, Considerations of High Performance Inducer (HPI) Design at Low Cavitation Breakdown for Applications in the LNG Pump Industries, GASTECH2005, Bilbao, Spain, 14-17 March. Lee, Jinkook and et al., 1997, Aircraft Gas Turbine Engine Fuel Pumping Systems in the 21 st Century, Journal of Engineering for Gas Turbine and Power, July, Vol. 119, pp. 591-597. Sayers, A. T., 1998, Hydraulic and Compressible Flow Turbomachines, McGraw-Hill Book Company. Stepanoff, A. J., 1957, Centrifugal and Axial Flow Pumps, Theory, Design and Application, John Wiley, Second Edition. Stripling, L. B. and Acosta, A. J., 1962, Cavitation in Turbopumps Part1, ASME Journal of Basic Engineering, September, pp. 326-338. Turton R. K., 1994, Rotodynamic Pump Design, Cambridge University Press. Tuzson, John, 2000, Centrifugal Pump Design, John Wiley & Sons, Inc. PO-35.9