Simulation of performance of intermediate fluid vaporizer under wide operation conditions

Abstract

The intermediate fluid vaporizer (IFV) is a typical vaporizer of liquefied natural gas (LNG), which in general consists of three shell-and-tube heat exchangers (an evaporator, a condenser, and a thermolator). LNG is heated by seawater and the intermediate fluid in these heat exchangers. A one-dimensional heat transfer model for IFV is established in this paper in order to investigate the influences of structure and operation parameters on the heat transfer performance. In the rated condition, it is suggested to reduce tube diameters appropriately to get a large total heat transfer coefficient and increase the tube number to ensure the sufficient heat transfer area. According to simulation results, although the IFV capacity is much larger than the simplified-IFV (SIFV) capacity, the mode of SIFV could be recommended in some low-load cases as well. In some cases at high loads exceeding the capacity of a single IFV, it is better to add an AAV or an SCV operating to the IFV than just to increase the mass flow rate of seawater in the IFV in LNG receiving terminals.

Keywords

liquefied natural gas / intermediate fluid vapo-rizer / heat transfer performance / numerical simulation / extreme condition

Introduction

At present, natural gas (NG) has become one of the most popular energy sources because of its advantages of cleanness and safety. Liquefied natural gas (LNG) is convenient for transportation. However, LNG should be vaporized and heated before it is conducted into the public domain network. There are several types of vaporizers and heat exchangers employed in LNG receiving terminals, for example, ambient air vaporizer (AAV), open-rack vaporizer (ORV), submerged-combustion vaporizer (SCV), and intermediate fluid vaporizer (IFV), etc.

IFV is of complicated structure, but has some advantages as well, i.e., compact volume, no icing, low requirement for seawater, and low operation cost [¹]. Therefore, IFVs have been increasingly employed in LNG receiving terminals at present.

An IFV usually consists of three shell-and-tube heat exchangers, namely, the evaporator, the condenser, and the thermolator respectively [²,³]; the LNG is heated by intermediate fluid rather than by seawater directly, as shown in Fig. 1. In actuality, warm water or air [⁴] can be used as the alternative heat source in some cases where seawater is used restrictedly due to environmental consideration.

Intermediate fluid plays an important part in IFVs. Xu et al. [⁵] discussed several refrigerants as intermediate fluids and compared the required heat transfer areas of IFVs with propylene, propane, isobutene, butane, and dimethylether respectively by numerical calculation. They found that propylene and dimethylether were feasible refrigerants for an IFV system besides the widely reported propane. Solberg [⁶] made a comparison between propane and ethylene glycol as intermediate fluids in IFVs and thought that ethylene glycol required less heat transfer area because of its higher specific heat. However, glycol is hard to be utilized when the temperature decreases significantly because of its higher freezing point (–33°C).

In addition, the flow and heat transfer characteristics inside the IFV were also taken into consideration. Pu et al. [⁷] established a lumped model to analyze the thermal performance of an IFV and discussed the influences of LNG pressure, LNG mass flow rate, seawater inlet temperature, and seawater mass flow rate. Xu et al. [⁸,⁹] established a thermal design method to investigate the subcritical liquefaction process for LNG and to predict the heat transfer characteristics among different sections and flow boiling zones. Such characteristics include the increment of LNG pressure, the transition from one flow boiling zone to another, and the profitless increment of the required heat transfer area of the condenser.

Additionally, Wang et al. [¹⁰] conducted CFD simulations on flow and heat transfer in an IFV and showed heat transfer deterioration occurred in some cases at a low mass rate.

To deeply analyze the IFV performance, especially when LNG operates in the pseudo critical zone, IFV structure and operation parameters should be combined together on optimization discussion. Although some researchers have discussed the optimizing structure, including plate-fin configuration [¹¹] or spiral-wound tube instead of smooth tubes [¹²], the discussion on heat exchanger was not sufficient. Moreover, at a low supplying load, when it is suitable for the vaporizer to operate in the simplified IFV (SIFV) mode, the simulation of the SIFV is helpful in evaluating its feasibility in actual application. Therefore, the present paper intends to discuss the heat transfer performance of IFV affected by some structure parameters, and to propose corresponding alternative modes under extreme conditions.

Methods for calculation

Physical problem description

Figure 2 shows a typical schematic diagram of an IFV. The LNG flows from the entrance (point 1) to the exit (point 2) through the U-tubes in the condenser, then enters the thermolator and flows through the inside tubes of the thermolator for extra heating. Seawater flows from the entrance (point 5) to the exit (point 6) to heat NG in the thermolator. Then the cooling seawater enters the evaporator for further heat exchange. The intermediate fluid condenses on the outside of the tubes in the condenser and boils on the outside of the tubes in the evaporator.

Therefore, there are three heat transfer processes in an IFV: the heat transfer in the condenser, the heat transfer in the thermolator, and the heat transfer in the evaporator. In the condenser, the supercritical LNG is fed into the U-tubes and heated by the propane vapor, and the intermediate fluid condenses on the outside of tubes, whose saturation temperature and pressure are T_po and P_po; In the thermolator, seawater is fed into the exchanger to further heat the NG coming from the condenser. On the other side, after being heated by intermediate fluid in the condenser, the NG flows across the tubes to exchange heat with seawater and reaches the temperature beyond 0°C. In the evaporator, seawater heats the propane; the intermediate fluid boils on the outside of tubes.

Thermal properties of working medium

In this calculation, propane is selected as intermediate fluid, while LNG and seawater are simplified as pure methane and pure water respectively, and their thermal properties are obtained from REFPROP.

Figure 3 depicts the variation of density, viscosity, specific heat, and thermal conductivity of methane with temperature and pressure. The specific heats have violent variations near supercritical pressure. Moreover, specific heat reaches the maximum near special temperatures, which are called pseudo-critical temperatures and vary with pressure. As critical pressures get closer, the variation amplitudes of specific heat increases. Other properties have similar trend as temperature increases at the supercritical pressures.

Heat transfer correlations

The heat transfer processes in the three parts of IFV are quite different. The related correlations from Ref. [⁷] are adopted in this paper except for the supercritical heat transfer process and the condensate of propane outside the condenser. Since the research on heat transfer about supercritical LNG is insufficient, Jackson and Hall’s correlation [¹³] has been considered as one of the feasible correlations for the supercritical heat transfer in tubes with water, carbon dioxide, and some hydrocarbons [¹⁴]. Wang et al. [¹⁵] modified the correlation to well predict the methane process in similar cases. Compared with Bae and Kim’s correlation utilized in Ref. [⁷], it is used to calculate such heat transfer process in the present condition. On the other hand, in the case of a horizontal tube bundle, the drained condensate from upper tubes results in a stronger thermal resistance of the condensate layer on lower tubes, which reduces the condensation heat transfer. It is proposed that this bundle effect in horizontal condensers be discussed with a correction factor of N_{\mathrm{eff}}{}^{-\mathrm{1}/\mathrm{6}} to obtain the average condensation heat transfer coefficient [¹⁶]. The total heat transfer correlations are listed in Table 1.

Mathematical model

The process in an IFV could be simulated, which leads to a deep understanding of the supercritical heat transfer on LNG. To evaluate the IFV performance, it is assumed that the IFV is working steadily; the LNG mass flow is distributed evenly in each tube; and the pressure drop of LNG in the tubes is neglected since the operation pressure is much higher than the pressure drop.

The default structure parameters of the evaporator, the condenser, and the thermolator are listed in Table 2.

The energy balance equations in the evaporator, the condenser, and the thermolator are expressed as

(1)Q_{\mathrm{S}}=m_{\mathrm{S}}\left(h_{\mathrm{S}\mathrm{1}}-h_{\mathrm{S}\mathrm{3}}\right)=Q_{\mathrm{L}}=m_{\mathrm{L}}\left(h_{\mathrm{L}\mathrm{3}}-h_{\mathrm{L}1}\right),

(2)Q_{\mathrm{S-th}}=m_{\mathrm{S}}\left(h_{\mathrm{S}\mathrm{1}}-h_{\mathrm{S}\mathrm{2}}\right)=Q_{\mathrm{L-th}}=m_{\mathrm{L}}\left(h_{\mathrm{L}\mathrm{3}}-h_{\mathrm{L}\mathrm{2}}\right),

(3)Q_{\mathrm{S-ev}}=m_{\mathrm{S}}\left(h_{\mathrm{S}\mathrm{2}}-h_{\mathrm{S}\mathrm{3}}\right)=Q_{\mathrm{L-con}}=m_{\mathrm{L}}\left(h_{\mathrm{L}\mathrm{2}}-h_{\mathrm{L}\mathrm{1}}\right),

where Q_s, Q_S-th, and Q_S-ev represent the heating heat transfer rates of the seawater in the whole IFV system, the thermolator, and the evaporator, respectively; Q_L, Q_L-th, and Q_L-con represent the cooling heat transfer rates of LNG/NG in the whole IFV system, the thermolator, and the condenser, respectively; h_S1, h_S2, and h_S3 represent the seawater enthalpies at the inlet, the outlet of the thermolator, and the outlet of IFV, respectively, while h_L1, h_L2, and h_L3 represent the LNG/NG enthalpies at the inlet, the outlet of the condenser, and the outlet of IFV, respectively.

LNG presents complex heat transfer characteristics near pseudo-critical point on account of the violent variations of the properties. The heat transfer coefficient varies significantly along the tube. Therefore, in order to get more detailed results, a one-dimensional model is employed, and the evaporator, condenser, and thermolator are divided into 180, 90, and 90 elements, respectively, each element being 0.1 m in the condenser and the evaporator, and the element being 0.04 m in the thermolator.

In each calculation element of each heat exchanger, the energy balance equation is expressed as

(4)Q\left(i\right)={\alpha }_{\mathrm{in}}\left(i\right)A_{\mathrm{in}}\left(T_{\mathrm{in}}\left(i\right)-T_{\mathrm{w}}\left(i\right)\right)={\alpha }_{\mathrm{out}}\left(i\right)A_{\mathrm{out}}\left(T_{\mathrm{w}}\left(i\right)-T_{\mathrm{out}}\right),

where α_in(i) and α_out(i) are the heat transfer coefficients inside and outside the tubes, respectively.

The next element temperature can be calculated by using

(5)T_{\mathrm{in}}(i+\mathrm{1})=T_{\mathrm{in}}\left(i\right)+Q\left(i\right)/\left(c_{\mathrm{p}}\right(i\left)m\right).

The discrete elements in the three heat exchangers are depicted in Fig. 4.

Numerical method

The solution procedure is as follows.

(1) Assume the propane temperature T_po. Then the NG temperature out the condenser T_L2 and the heat transfer rate Q_L-con in condenser can be calculated by using Eqs. (4) and (5).

(2) Assume the seawater temperature T_S2. Then the seawater inlet temperature {T^{\prime }}_{\mathrm{S}\mathrm{1}} and the NG outlet temperature T_L3 can be calculated by using Eqs. (4) and (5).

(3) The relative error between {T^{\prime }}_{\mathrm{S}\mathrm{1}} and the real inlet temperature T_S1 is checked. If the relative error is less than a certain value, such as 0.001, the iteration ends; otherwise the value of T_S2 will be modified. Step (2) will be repeated until a converged solution is obtained through this inner iteration.

(4) Based on T_S2 and T_po, T_S3 and heat transfer rate Q_S-ev can be obtained by using Eqs. (4) and (5).

(5) The relative error between Q_L-con and Q_S-ev is checked. If the relative error is less than 0.001, the iteration is completed; otherwise, T_po will be updated. Steps (1) to (5) will be repeated until a converged solution is obtained.

The logical diagram of the program is illustrated in Fig. 5.

Validation

First, the grid-independence study is performed. The temperature profiles in the condenser, the evaporator, and the thermolator are displayed in Fig. 6 with different numbers of grid. The results indicate that the calculations with a grid number of 180, 90, and 90 for the condenser, the evaporator, and the thermolator are acceptable.

The one-dimensional calculation was validated with some practicing data before conducting the analysis of supercritical heat transfer in IFVs. Some operation data about IFV were collected from the LNG terminal in Ningbo, China. There were four operation cases where the pressures and the mass flow rates of LNG are tabulated in Table 3. Other operation parameters were the same as the default values presented in Table 4.

Figure 7 exhibits the comparison between the numerical calculation and the operation data of the LNG/NG temperature and the seawater temperature, whose values agree with each other. The outlet temperatures of LNG/NG are well predicted, while the LNG/NG temperatures outside the condenser have certain deviations compared with the operation data. There is a temperature difference of about 8 K and a relative error of about 6%. The seawater temperatures out of the thermolator have similar deviations between numerical calculation and operation data, and the errors are acceptable as well.

Results and discussion

Performance on the rated condition and optimization for structure parameters

The performance of the IFV is discussed with some structure parameters. The structure variations are combined with four tube diameters and four tube numbers in the condenser, which are discussed and compared with the typical structure parameters in Table 2. The inlet operation parameters are the rated values as given in Table 4. Additionally, an analysis of the thermolator is also performed.

Tube diameters

Since the LNG thermal properties vary dramatically near the critical region, there are violent variations in local heat transfer coefficients accordingly. In Fig. 8(a), the heat transfer coefficient profile reaches the peak at the position about 3 m away from the entrance in each case, since the violent variations of the LNG properties and the LNG reach the pseudo-critical temperature at this position. The local heat transfer coefficient increases with the tube diameter reducing, the average values being 2784 W/(m²·K), 2025 W/(m²·K), 1539 W/(m²·K), and 1211 W/(m²·K) with the tube diameter of 12 mm, 14 mm, 16 mm, and 18 mm, respectively, since the tube diameter affects the velocity and the Reynolds number, and the heat transfer coefficient is dominated with Re number. The variations of inner heat transfer coefficients with different tube diameters have similar profiles, while the pseudo-critical peak value waves considerably. As the tube diameter decreases, the peak value increases.

In contrast, the outer surface heat transfer coefficient is less affected by the tube diameter, as shown in Fig. 8(b). Comparing the inner and the outer heat transfer coefficients, the major thermal resistance is on the inside of the tube in cases with large tube diameters. The major thermal resistance locates at the outside of the tube in cases with small tube diameters, as shown in Fig. 9. It is well known that reducing tube diameter results in the increase in pressure drop. Therefore, the tube diameter should not be reduced excessively.

Tube number

Figure 10 shows the influence of the tube number on the condenser. Both of the tube diameter and the number are related to the mass flux and the velocity in the tubes. Therefore decreasing the tube number would lead to the rise in Reynolds number, as shown in Fig. 11, and, in turn, the heat transfer coefficient. Moreover, the Reynolds number vary with similar profiles.

The outer heat transfer coefficient is not sensitive to the tube number, in comparison with the tube diameter. In general, increasing tube number slightly weakens the outer heat transferability. Comparing the inner and outer heat transfer coefficients, the thermal resistances at the two sides of the condenser tube are roughly close to each other.

In fact, the heat transfer area is usually determined preferentially in the design, and the interaction between tube diameter and tube number should not be neglected. Figure 12 shows the heat transfer coefficient with four different tube numbers where their total heat transfer areas are the same. Obviously, with large tube numbers, the performances are better. Of course, the tube number should not be increased excessively, since increasing tube numbers with the same total heat transfer areas implies to decrease the tube diameters, and leads to the increase in pressure drop.

Thermolator

The simplified IFV (SIFV) comes from the traditional IFV, which operates without thermolator and is also employed in some specific conditions. Figure 13 shows the capacities of IFV and SIFV at different seawater inlet temperatures and mass flow rates, with the other operation parameters in Table 4. When the seawater inlet temperatures are 4°C, 6°C, 8°C, and 10°C respectively, the heat transfer capacities of IFV are higher than those of SIFV by 233%, 146%, 118%, and 107%. When the seawater mass flow rates are 4000 t/h, 6000 t/h, 8000 t/h, and 10000 t/h, respectively, the heat transfer capacities of IFV are higher than those of SIFV by 133%, 125%, 118%, and 113%. Obviously, the thermolator significantly improves the performance when the IFV operates at low local seawater temperatures and low seawater mass flow rates, where the heat transfer capacity should be maintained with sufficient heat transfer areas.

The influences of the heat transfer area of the thermolator are shown in Fig. 14. The capacity of the system can be improved significantly by increasing the thermolator heat transfer area. The high inlet temperature of seawater and the large mass flow rate of seawater are helpful for the IFV to provide a large heat transfer capacity as well. The heat transfer capacity could be increased by about 0.16 t/h by increasing 1m² of the heat transfer area of the thermolator. Obviously, the thermolator is appropriate and necessary when the IFV operates in most common cases.

Discussion of extreme condition

Sometimes, the vaporizers in a LNG terminal properly operate at low loads or high loads. For example, IFV properly faces a rated load of about 40% in actual situation. If so, the heat transfer areas designed under the rated load are redundant unavoidably.

Low-load conditions

In low-load situations, the NG temperature could be higher than 0°C just at the condenser exit. Thus, the heat transfer capacity of thermolator is unnecessary, and further the thermolator operation would still bring a certain pressure drops on two fluid sides. As summarized in Table 5, even though the NG temperature at the condenser exit is not higher than 0°C at an LNG load of 100 t/h, the proportion of heat exchanged in the thermolator is less than 3% because of the small heat transfer temperature difference, but the pressure loss proportion in the thermolator is more than 40%.

In addition, the seawater may probably freeze in IFVs under extreme conditions. For example, at a seawater inlet temperature of 4°C, a LNG mass flow rate of 140 t/h, and a seawater mass flow rate of 6300 t/h, the NG outlet temperature is 0.23°C, the seawater outlet temperature would be –0.01°C in Fig. 15, and the temperature value is close to the freezing point of seawater. In actual operation, the SIFV could relieve this problem, the seawater and the LNG/NG could avoid exchanging heat deeply in the thermolator with the SIFV mode at low loads, and the seawater outlet temperature could be controlled above the icing point.

High-load conditions

In high-load situations or when the seawater inlet temperatures are low enough, the NG outlet temperature would be less than 0°C, as shown in Fig. 16, which could not satisfy the operation reliability. At an LNG mass flow rate of 250 t/h, the NG outlet temperature is about –2°C, unless the seawater mass flow rate is increased by 50%. The NG outlet temperature would be lower than 0°C. Moreover, some extra pumping cost should be taken for seawater cycling.

In actuality, there are various vaporizers configured in a LNG receiving terminal. It should be a better choice to arrange several AAVs or SCVs to operate in combination with an IFV. For example, just by changing the LNG mass flow rates to 250 t/h, 300 t/h, and 350 t/h, the NG temperatures out of the IFV are –2.12°C, –7.39°C, and –13.11°C, and there leave 621.22 kW, 2022.23 kW, and 4010.67 kW extra heating requirements for heating the NGs over 0°C, respectively. If the AAVs are employed at the ambient temperature of 20°C with a heat transfer area of more than 94.12 m², 249.66 m², and 405.11 m², the NG outlet temperatures could reach the common requirement. If an SCV is employed to balance the extra heating load, it would consume 0.011 kg/s, 0.036 kg/s, and 0.071 kg/s LNG as fuel to compensate the heating loads. Therefore, it is necessary to flexibly employ various vaporizers to operate in combination in an LNG terminal to satisfy the peak requirements of NG supply.

Conclusions

In summary, the performance of a typical IFV is simulated and analyzed, and the structure optimization is discussed. Besides, the operation of IFV in some extreme conditions is also investigated, and several methods are proposed.

There are violent waves on the LNG heat transfer and flow performances in an IFV once the LNGs are near the peso-critical points. The violent variation of the inner heat transfer coefficient near the critical region could be weakened by increasing the tube diameter. Adjusting the tube diameter and the tube number could change the inner heat transfer resistance to a certain extent. The major thermal resistance is located inside the tube in cases at large tube diameters, and the outside thermal resistance dominates at small tube diameters.

The thermolator could significantly improve the IFV performance, especially at low local seawater temperatures and low seawater mass flow rates, in most common cases.

The SIFV could be adopted in the cases at a low seawater inlet temperature and a low LNG mass flow rate. In some cases at low LNG mass flow rates, the operation in the SIFV mode could avoid freezing in the heat exchanger, which can save some pumping energy in the thermolator.

In some high-load situations, an IFV could operate in combination with several AAVs or SCVs to balance the extra heating load, rather than just reluctantly increasing the mass flow rate of seawater in an IFV.

References

[1] Dendy T, Nanda R. Utilization of atmospheric heat exchangers in LNG vaporization processes: a comparison of systems and methods. In: AIChE Spring Meeting and 235th ACS National Meeting, New Orleans, LA, 2008

[2] Iwasaki M, Egashira S, Oda T, Asada K, Sugino K. Intermediate fluid type vaporizer, and natural gas supply method using the vaporizer. US Patent: 6164247, 2000

[3] Iwasaki M, Asada K. Intermediate fluid type vaporizer. US Patent: 6367429, 2002

[4] Groothuis C K, Tanaeva I. Method and apparatus for vaporizing a liquid stream. US Patent: 9103498, 2015

[5] Xu S, Cheng Q, Zhuang L, Tang B, Ren Q, Zhang X. LNG vaporizers using various refrigerants as intermediate fluid: comparison of the required heat transfer area. Journal of Natural Gas Science and Engineering, 2015, 25: 1–9

[6] Solberg E L. A comparative analysis of propane and ethylene glycol as intermediate fluid in a LNG regasification system. Norwegian University of Science and Technology, 2015

[7] Pu L, Qu Z, Bai Y, Qi D, Song K, Yi P. Thermal performance analysis of intermediate fluid vaporizer for liquefied natural gas. Applied Thermal Engineering, 2014, 65(1–2): 564–574

[8] Xu S, Chen X, Fan Z. Thermal design of intermediate fluid vaporizer for subcritical liquefied natural gas. Journal of Natural Gas Science and Engineering, 2016, 32: 10–19

[9] Xu S, Chen X, Fan Z. Design of intermediate fluid vaporizer for liquefied natural gas: a review. Chemical Engineering & Techno-logy, 2017, 40(3): 428–438

[10] Wang B J, Kuang Y W, Qi C, Wang W, Xu J W, Huang Y. Analysis of heat transfer to supercritical LNG in intermediate fluid vaporizer. Journal of Chemical Industry and Engineering (China), 2015, 66(S2): 220–225 (in Chinese)

[11] Kim D Y, Sung T H, Kim K C. Application of metal foam heat exchangers for a high-performance liquefied natural gas regasification system. Energy, 2016, 105: 57–69

[12] Liu F, Dai Y, Wei W. Feasibility of intermediate fluid vaporizer with spiral wound tubes. China Petroleum Processing and Petrochemical Technology, 2013, 15(1): 73–77

[13] Jackson J D. Consideration of the heat transfer properties of supercritical pressure water in connection with the cooling of advanced nuclear reactors. In: Proceedings of the 13th Pacific Basin Nuclear Conference, Shenzhen, China, 2002: 21–25

[14] Pioro I L, Khartabil H F, Duffey R B. Heat transfer to supercritical fluids flowing in channels—empirical correlations (survey). Nuclear Engineering and Design, 2004, 230(1–3): 69–91

[15] Wang Y Z, Hua Y X, Meng H.Numerical studies of supercritical turbulent convective heat transfer of cryogenic-propellant methane. Journal of thermophysics and heat transfer, 2010, 24(3): 490–500

[16] Nikitin K, Kato Y, Ishizuka T. Steam condensing-liquid CO₂ boiling heat transfer in a steam condenser for a new heat recovery system. International Journal of Heat and Mass Transfer, 2008, 51(17–18): 4544–4550

[17] Dittus F, Boelter L. Heat transfer in automobile radiators of the tubular type. International Journal of Heat and Mass Transfer, 1930, 2: 443–461

[18] Cooper M G. Saturation nucleate pool boiling—a simple correlation. In: 1st UK National Conference on Heat Transfer, 1984, 86: 785–793

[19] Jung D, Chae S, Bae D, Oho S. Condensation heat transfer coefficients of flammable refrigerants. International Journal of Refrigeration, 2004, 27(3): 314–317

[20] Žukauskas A. Heat transfer from tubes in crossflow. Advances in Heat Transfer, 1972, 8: 93–160

RIGHTS & PERMISSIONS

Higher Education Press