Poster PO-39

NATURAL GAS LIQUEFACTION PROCESSES COMPARISON COMPARAISON ENTRE PROCEDES DE LIQUEFACTION DE GAZ NATUREL
Pierre-Yves Martin Jérôme Pigourier Axens (France) www.axens.fr Béatrice Fischer IFP (France)

ABSTRACT This paper presents Axens efforts to compare LiquefinTM with the competing process, specially the reputed most efficient ones: C3/MR, C3/MR followed by nitrogen cycle, dual mixed refrigerant process with spiral wound exchangers. To compare properly one process to another, it must be done with the same gas, with the same site conditions, with the same gas turbines and with the same cooling medium temperature (air or water). That done, to compare processes like for like is still not that easy. For instance, it seems fair to take the same efficiencies for the compressors, however, axial and centrifugal compressors do have different efficiencies. Similarly, equal basis leads to have the same temperature approach for the air-cooler (or water coolers), however between mixed refrigerant and propane, the heat exchange area will be much lower for mixed refrigerant if the approach is kept identical. The end flash vapour quantity has also a big influence on the process efficiency, but each process has a different fuel gas consumption. Those added factors may lead to wide differences. Axens has calculated the effect of all those parameters on efficiency. The equipment characteristics play also an important role in the comparison: the limitations of axial compressors, of centrifugal compressors (Mach number) and possibly spiral-wound exchangers maximum size do have to be taken into account. Even the gas turbines or alternative drivers chosen can be well adapted to one process, but not to the other. Another important parameter is the LPG recovery: a large LPG recovery will decrease the efficiency of any process, but not to the same extent. For Liquefin, the efficiency decrease is not very big. This paper shows the detailed results of those comparison studies, and the effect of several parameters on Liquefin efficiency. RESUME Cet article présente le travail d’Axens pour comparer son procédé LiquefinTM aux autres procédés avec lesquels il est en compétition, et spécialement ceux réputés les plus efficaces : C3/MR, C3/MR suivi d’un cycle azote, DMR avec échangeur bobiné. Pour comparer sérieusement un procédé avec un autre, on doit considérer le même gaz, les mêmes conditions ambiantes, les même turbines à gaz , et le même moyen de refroidissement

PO-39.1
SESSIONS CONTENTS

Poster PO-39

(air ou eau). Ceci fixé, il n’est toujours pas si aisé d’avoir une comparaison honnête. Par exemple, il semble normal de prendre la même efficacité pour les compresseurs, cependant les compresseurs axiaux et centrifuges n’ont pas la même efficacité. De même, on aurait tendance à prendre la même approche thermique entre la température côté procédé et la température de l’air ambiant (ou de l’eau de mer), cependant si on fait cela, la surface d’échange sera beaucoup plus petite avec un mélange réfrigérant qu’avec du propane. La quantité vaporisée dans le flash final est très importante en terme d’efficacité du schéma, cependant chaque procédé a une consommation de fuel gaz différente. Tous ces facteurs combinés peuvent conduire à des différences importantes. Axens a calculé l’effet de chacun de ces paramètres sur l’efficacité. Les caractéristiques des équipements jouent également un rôle important dans la comparaison : les limitations des compresseurs axiaux et centrifuges (le nombre de Mach), ainsi que la taille maximum possible pour un échangeur bobiné doivent être prises en compte. Même les turbines à gaz ou les autres moyens d’entraînement des compresseurs peuvent être bien adaptés à un procédé, mais pas à l’autre. Un autre paramètre important est la récupération de GPL : une récupération poussée de GPL va faire chuter l’efficacité de tous les procédés, mais pas dans la même mesure. Liquefin a une baisse d’efficacité qui n’est pas trop importante dans ce cas. L’article présente les résultats détaillés de ces études de comparaison, et l’effet de différents paramètres sur l’efficacité de Liquefin. INTRODUCTION For many years, there was absolutely no problem to choose the process of a new liquefaction plant: C3/MR was the only choice. The same process was implemented again and again, with small improvements, sometimes bigger gas turbines, and anyway bigger capacities along the years. However, this process is now reaching the technology limits: maximum mach number on the propane compressor, spiral wound exchanger becoming enormous. So now many new processes are appearing: APCI has launched the APX process (C3/MR/N2 cycles), SHELL a DMR process, LINDE a process with three mixed refrigerant cycles, and IFP/Axens another DMR with plate-fin heat exchangers. The old cascade process has come back in Trinidad, with a new concept. Nowadays, the new projects consider capacities of 5, 6, sometimes 8 MTPA, whereas the biggest unit in operation is below 4 MTPA. Deciding of the process to be used for a given project is now much more difficult, and many factors must be considered to make a proper comparison. OLD AND NEW PROCESSES All the natural gas liquefaction baseload plants built during the last twenty years or so are C3/MR units, to the exception of Trinidad. The C3/MR process is well known (see figure 1): the MR and natural gas are pre-cooled with propane, at 3 or now 4 levels of pressure. The mixed refrigerant is only partially condensed, and separated before entering the large spiralwound exchanger.

PO-39.2
SESSIONS CONTENTS

Poster PO-39

Trinidad plant is built with a cascade process (propane/ethylene/methane), with also several levels of pressure on each cycle . The process (see figure 2) is arranged so as to have the same power on the three cycles. Another interesting feature is to install parallel lines of compression, with Frame 5 - variable speed gas turbine - so as to have a high availability and easier operation: no compressor trip will shut down completely the unit, and the restart of the compressor can be done without loss of refrigerant.

Propane cycle
CW

CW

MR cycle
CW

LNG

Feed Gas

Figure 1: C3/MR process simplified scheme

CW

CW

C1 C2 2 Frame 5

2 Frame 5

CW

C3 2 Frame 5

PFHE Feed gas

PFHE

PFHE LNG

Figure 2: Cascade process

PO-39.3
SESSIONS CONTENTS

Poster PO-39

APCI APX process [1] This process (see figure 3) is a three cycle process: propane, mixed refrigerant, nitrogen. The exchangers used are kettles for the propane, spiral wound for the mixed refrigerant, another spiral wound and plate-fin exchanger for the nitrogen cycle. Compared to the C3/MR process, the new third cycle allows to decrease the propane and MR flow-rates, so as to achieve with existing equipment much higher capacities (7 – 8 MTPA).
Propane cycle MR cycle
CW

N2 Cycle
LNG

CW CW

CW

Feed Gas

Figure 3: APX process simplified scheme

LNG

Natural gas
First mixed refrigerant cycle

Figure 4: Shell DMR process simplified scheme

PO-39.4
SESSIONS CONTENTS

Poster PO-39

SHELL DMR Process [2] This process (see figure 4) is a dual mixed refrigerant process, with different power on the two cycles, and with two spiral-wound exchangers. Having mixed refrigerant on the first cycle allows to have a smaller condenser, and also to remove the propane compressor bottleneck: For propane compressors, the compressor size, thus the capacity of the unit is limited by the mach number at the tip of the blades. Using a mixed refrigerant, with a lower molecular weight, allows to push further this limit as the mach number is lower with this gas. (see figure 7) LINDE process [3]

Liquefaction MR cycle Pre-cooling MR cycle Sub-cooling MR cycle

Natural gas

Plate-fin Exchangers

Spiral-wound heat-exchanger

Spiral-wound heat-exchanger

LNG

Figure 5: LINDE process simplified scheme This process is a three cycle process, like the cascade process, but with mixed refrigerant on all cycles (see figure 5). Compared to the cascade, the efficiency is better, as mixed refrigerants allow to have a closer approach. However, the power is not the same on all three cycles, unlike the new cascade. Plate-fin exchangers are used on the first cycle, and spiralwound exchangers on the two colder cycles. IFP/Axens Liquefin Process [4] [5] This process (see figure 6) is a dual mixed refrigerant process, with the same power on both mixed refrigerant cycles. Plate-fin heat exchanger are used for the whole exchange line. As for all processes with mixed refrigerant on the first cycle, the main condenser is smaller (see figure 11) and the compressor of the first cycle has a lower mach number (see figure 7). The lower amount of mixed refrigerant on the cold cycle allows to reach much higher LNG capacities with the existing axial compressors.

PO-39.5
SESSIONS CONTENTS

Poster PO-39

CW CW

Feed gas
Heavy mixed refrigerant compression line

Scrubber

Hot Oil
CW

Main exchange line

Light mixed refrigerant compression line

LNG

Figure 6 : Simplified Liquefin process scheme All these processes were developed to overcome the technology limits reached by the C3/MR process. The main limitations being the propane compressor as explained already , but also to some extend the axial compressor and the spiral-wound exchanger.

Biggest propane compressors
1.25 1.20 1.15 1.10 1.05 1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 0.45 0.40 0.35 0.30 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16

Peripheral Mach number (Mu)

Flow Coefficient (φ1)

Tip relative mach number (inlet)

Flow Coefficient (φ1)

Liquefin MR1 compressors (4.8MTPA)
Figure 7 : Mach number vs flow coefficient for the first stage compressor (propane or MR)

PO-39.6
SESSIONS CONTENTS

Poster PO-39

To require proven compressors imposes to fit the light refrigerant flow-rate and pressure ratio to one existing axial compressor. This is another constraint making difficult to increase very much the LNG production with a proven process. In case the process could not accommodate an axial compressor, there will be an efficiency penalty : one consider usually a polytropic efficiency of 86% for the axial compressors instead of 82% for the centrifugal compressors. This axial compressor will be used on the coolest stage of the refrigeration. On LIQUEFIN, the simulations give a difference of about 1.5 % on the LNG production. The use of axial compressor is also beneficial for operation: the inlet vanes possible angle variation will be very useful for control. The size of the spiral wound exchanger, already huge, cannot be increased forever. So for very high capacities, it would be necessary to increase the LMDT of this exchanger to stay within a feasible size, but with an efficiency penalty. Another possibility would be to have 2 spiral-wound exchangers in parallel, but it would increase the cost and the delivery time. SIMULATION PARAMETERS To compare one process with another, it is necessary to be very careful about several parameters, which can change hugely the result: end flash quantity, compressor efficiencies, condenser temperature approach, LPG recovery. End Flash Quantity. The quantity of end flash usually corresponds to the plant fuel gas consumption (minus some margin). If it is possible to increase this quantity (fuel gas export to other plants, recycle, etc), the cold end temperature of the main exchange line will increase, and the efficiency of the plant, thus the quantity of LNG produced will increase.
LNG Production vs End Flash Quantity
106 105 LNG production 104 103 102 101 100 100 110 120 End flash quantity 130 140

Figure 8 : Effect on LNG production of end flash quantity If no end flash is wanted for any reason, this is a large decrease of LNG production (but a simplification of the scheme: removing of the end flash compressor). In many cases however, the quantity of fuel gas cannot be decreased below a certain quantity because of the nitrogen content of the feed gas. The figure 8 shows the production variation with the quantity of end flash. If the end flash can be increased by 40%, the LNG production will be increased by 5%

PO-39.7
SESSIONS CONTENTS

Poster PO-39

with the same power on the refrigeration compressors (but of course more power on the fuel gas compressors). This also must be checked carefully for process comparison. Compressor Efficiency. Depending upon the compressor efficiency considered, the LNG production can vary a lot: the LNG production is increased by nearly 10% when the polytropic efficiency is changed from 79 to 85% (see figure 9). This has to be checked carefully when making comparisons.
LNG Production vs Compressor Polytropic Efficiency
105 104 103

LNG Production

102 101 100 99 98 97 96 95 79 80 81 82 83 84 85

Compressors Polytropic Efficiency

Figure 9 : Effect on LNG production of compressor efficiency Temperature Approach on The Main Condenser Whatever the process, a large condenser on the first refrigerant cycle must evacuate the heat produced by the refrigeration compressors. As the outlet of this condenser is at bubble point, to modify the outlet temperature of this condenser will change the discharge pressure of the corresponding compressor, so the power of this compressor, and thus the overall efficiency. The temperature approach of the other coolers will also have an impact on the power, but less important than this one. We have plotted for a Liquefin case the capacity versus the temperature approach of the condenser (see figure 10). The closer the temperature approach, the larger the LNG production. However, in air-cooling case, the size of this condenser can be a problem, as it governs more or less the size of the plant area, so a part of the cost. The size of the condenser will depend upon whether the refrigerant of the first cycle is a pure component (propane as in the C3/MR and cascade), or a mixed refrigerant (Liquefin and Linde process). With a pure component, the condensation is done at a fixed temperature (the dew point temperature is the same as the bubble point temperature), whereas with a mixed refrigerant, the temperature varies linearly between the dew point temperature and the bubble point temperature. (see figure 11). Either the condenser will be much smaller with a mixed refrigerant in the first cycle, or inversely with the same condenser size, the LNG production will be increased.

PO-39.8
SESSIONS CONTENTS

Poster PO-39

Plant capacity vs MR1 Condenser Outlet temperature
103.0%

102.5%

Plant Capacity (%)

102.0%

101.5%

101.0%

100.5%

100.0% 39 40 41 42 °C 43 44 45

Figure 10: Effect on LNG production of condenser temperature approach
Inlet temperature 60

MR Condensation
55 50 Temperature °C 45 40 35 30 Air or Water 25 20 0 50 100 150 200 250 DUTY MW 300 350 Larger LMDT (+35%) CONDENSER 35% Smaller Propane Condensation outlet temperature

Figure 11: Effect of using propane or mixed refrigerant on the size of the condenser LPG Recovery. To recover LPG from the gas can help sometimes to make the project economically sound. However, this recovery will increase the power for LNG liquefaction, and not with the same amount for all processes, nor for all gas compositions. We have simulated for LIQUEFIN the effect on efficiency of different C3 recovery ratios with a very lean gas (1.2% C3 only).

PO-39.9
SESSIONS CONTENTS

Poster PO-39

LNG production vs LPG recovery 100 99 LNG production (%) 98 97 96 95 94 93 92 91 30 40 50 60 70 80 90 C3 recovery (% of C3 in feed)

Figure 12 : LNG production decrease with increased LPG recovery CONCLUSION To solve the capacity increase problem, many new processes have proposed solutions. The comparison of these processes for a specific case must be done with care, taking into account all parameters including site conditions, end flash quantity, compressor efficiency, temperature approach and LPG recovery. The equipment availability and risk factor must also be taken into account. In this respect, the flexibility of Liquefin could make it the right choice in many cases. REFERENCES CITED 1. M.J. Roberts, J.C. Bronfenbrenner, Yu-Nan Liu, J.M. Petrowski - Large Capacity Single Train AP-X Hybrid LNG Process - Gastech 2002, Qatar 2. R. Nibbelke, S. Kauffman, B. Pek - Liquefaction Process Comparison of C3MR and DMR for Tropical Conditions - GPA 81st annual convention, 2002 3. H. Bauer - A Novel Concept for Large LNG Baseload Plants - AICHE Spring National Meeting, 2001 4. M. Khakoo , B.Fischer, J.C.Raillard - The Next Generation of LNG plants - LNG13, Seoul, Korea, 2001 5. B.Fischer - A New Process To Reduce LNG Cost - AICHE Spring National Meeting, 2002

PO-39.10
SESSIONS CONTENTS