Abstract
This study presents a rigorous three-dimensional (3D) computational fluid dynamics (CFD) investigation into the complex hydrodynamic characteristics of an external falling liquid film flowing over horizontal tubes, focusing on a representative unit of the shell-side of a spiral-wound heat exchanger (SWHE). The primary objective is to quantitatively analyze the synergistic influence of key operating and structural parametersnamely, Reynolds number (Re), tube outer diameter (D), and tube spacing (S)on the liquid film profile, coverage uniformity, and flow regime transition. A volume of fluid (VOF) model was established and validated to accurately track the gas–liquid interface. The analysis reveals three major findings: (1) Increasing the Reynolds number from Re = 300 to Re = 2000 drives a critical flow regime transition from a stable columnar flow dominated by viscous and gravitational forces to a highly dynamic fan flow. This transition significantly intensifies interfacial wave activity, which is quantified by a positive correlation between wave amplitude and Re. (2) An increase in tube diameter (D) promotes the concentration of the liquid film into distinct columns, decreasing film coverage uniformity and increasing local film thickness, as surface tension and inertia compete for liquid distribution across the larger surface. (3) The tube spacing (S) critically governs the hydrodynamics in the intertube region; an optimal spacing (S = 6 mm in this study) is identified where liquid columns effectively merge and redistribute, maximizing coverage while minimizing flow restriction. This quantitative analysis provides crucial, mechanistically grounded hydrodynamic data essential for the optimized structural design of industrial SWHEs, aiming to enhance overall heat and mass transfer efficiency.


1. Introduction
The drive for enhanced energy utilization and efficiency across industries, particularly in the fields of natural gas liquefaction, petrochemical processing, and cooling systems, has placed significant focus on highly efficient heat exchange technologies. Among these, the Spiral-Wound Heat Exchanger (SWHE) is highly valued for its compact structure, large heat transfer area, and operational flexibility, making it a critical component in large-scale cryogenic processes. , In these applications, the falling liquid film flow occurring on the shell side (the external surface of the helical tubes) plays a decisive role in the overall heat and mass transfer performance. A stable, thin, and uniformly distributed liquid film is paramount for achieving maximum thermal efficiency.
The study of falling film flow over horizontal tubes and tube bundles is a mature, yet continuously evolving, field, focusing primarily on hydrodynamic stability, film thickness, and heat transfer characteristics. Early theoretical and experimental studies, such as those by Fernández-Seara et al., established the fundamental advantages of falling film evaporation over pool boiling in terms of thermal efficiency and device size. Sun et al. utilized CFD models to simulate falling film flow in alkane media, providing correlations between film thickness and flow rate essential for heat exchanger design. Furthermore, Huang et al. explored innovative enhancement techniques, showing that ultrasonic oscillation can significantly improve mass transfer and absorption rates in falling film absorbers by 17.43%, offering new avenues for device performance enhancement.
A significant body of work has investigated the key mechanisms that govern film flow stability and distribution. , The Marangoni effect, induced by surface temperature gradients, has been shown to cause film contraction and nonuniform distribution, directly impacting the heat transfer mechanism, which involves convection, conduction, and phase change (evaporation). The influence of geometrical factors, such as baffle arrangements and corrugated plates, is also well-documented, demonstrating that introducing turbulence can enhance absorption and heat transfer efficiency. Sisoev et al. highlighted the significant effect of film fluctuation (waviness) on the development of the diffusion layer, a finding supported by Wan et al.’s work showing that instantaneous wave amplitude is fluid-dependent. Qiu et al. found that the thickness of the membrane was linearly growing in some peripheral angle, and the smallest one was found at 120°.
Operating conditions, particularly flow rate, have a direct bearing on film thickness. An increase in liquid flow rate (or Reynolds number, Re) generally reduces film thickness due to intensified inertial and gravitational forces. However, elevated heat flux can lead to dry-spot formation, , emphasizing the nonlinear relationship between hydrodynamics and thermal conditions. Han et al. numerically studied the combined effect of gravity, viscosity, and inertia on film profile, demonstrating that increasing tube pitch reduces mean film thickness by up to 38.5%, improving liquid distribution under sloshing conditions. It is found that the influence of the salt content, the Reynolds number, the thermal flow and the relative humidity on the evaporating velocity and the temperature of the membrane are investigated. In this paper, Li et al. have investigated the properties of an ammonia-containing nanofluids in a descending film. It was found that adding nanoparticles could greatly improve the heat-transfer and mass-transfer coefficient. Dong et al. summed up the technique of dropping membrane fluid drying, stressed the effect of the fluid flux on the heat-exchanger property, and summarized the structural design and performance improvement methods of different liquid dryers, which provided reference for subsequent research. In the review of falling film evaporation mechanism, Zhengshu Dai et al. emphasized the importance of fluid dynamics to evaporation efficiency, and pointed out that the change of flow state will directly affect the heat transfer efficiency and system performance of falling film. At elevated airflow velocities, the film demonstrates pronounced deflection and detachment, especially during transitions from laminar to droplet flow, affecting detachment length and fluctuation characteristics. , Recent studies on evaporators emphasize the use of drop film as an effective thermal/mass transfer technique in the petrochemical and aerospace sectors. Wang et al. demonstrated that under constant heat flux, flow regime significantly affects heat transfer coefficients, with corrugated tubes achieving a 10.17% higher average coefficient compared to smooth tubes. Zhu highlighted that low-temperature multieffect horizontal tube falling film evaporation is a promising desalination technology due to its high efficiency, low thermal energy requirements, and operational flexibility, all of which are closely linked to airflow effects. Wang analyzed liquid film dynamics on solid surfaces, noting that gravitational and surface tension forces dominate flow behavior, and stressed the importance of numerical simulations in understanding these mechanisms. Wang et al. reviewed how operational parameters (e.g., gas temperature, flow direction, velocity) and structural parameters (e.g., tube diameter, spacing) affect flow and thermal performance. Airflow direction alters film velocity and residence time, thereby modifying heat transfer efficiency. For instance, counter-current airflow accelerates film flow, while cocurrent airflow decelerates it. Tian analyzed falling film absorption in lithium bromide chillers, emphasizing the intricate relationship between flow patterns and thermal performance. Research on marine applications, such as floating platforms, has also gained attention, with Liu providing insights into film flow behavior under dynamic marine conditions.
More recently, research has addressed complex interactions, especially involving external gas flows and structural geometry relevant to industrial heat exchangers:
-
(1)
Gas–Liquid Interaction: Xu et al. , and Zhang investigated the effect of airflow direction and velocity, noting that lateral airflow causes asymmetric film distribution, with a thicker windward side and a thinner leeward side. High airflow velocities can accelerate film flow and even lead to severe flow imbalance or film blow-off, which is critical for evaporative cooling applications.
-
(2)
Structural Parameters in SWHEs: Studies on SWHEs often focus on the spiral nature of the tubes, which complicates the flow path. Bao et al. compared falling film characteristics on circular and flattened tubes in SWHEs, revealing a characteristic “trough-peak-trough” axial profile. Yan et al. and He et al. also numerically examined the influence of tube spacing in SWHEs, confirming that optimized spacing and distributor design are key to enhancing film uniformity and heat transfer performance.
While significant research exists on falling film hydrodynamics, a critical gap remains, particularly concerning the shell-side flow in industrial SWHEs used in cryogenic systems:
-
(1)
Coupled Parameter Effect: Most existing studies focus on single parameter variation (e.g., Re or tube spacing S). A systematic, three-dimensional study analyzing the coupled and synergistic hydrodynamic effects of the three most critical design and operating parameters (Re, D, and S) is lacking, especially in the context of the repetitive unit structure of a SWHE shell.
-
(2)
Mechanistic Clarity: Previous results often describe phenomena (e.g., flow pattern changes, thickness variation) but frequently omit the underlying fluid mechanical mechanisms that cause these transitions (e.g., the precise balance of inertial, viscous, gravitational, and surface tension forces).
-
(3)
SWHE Specific Geometry: The complex, confined geometry of the SWHE shell side, including the influence of liquid distributors, creates unique flow characteristics that cannot be fully captured by generic horizontal tube models.
Addressing these limitations, the present work develops a high-fidelity 3D CFD model of a representative unit of the SWHE shell-side to investigate the external falling liquid film hydrodynamics. The primary objectives are to
-
(1)
Investigate Flow Regime Transition: Systematically analyze the transition of the liquid film flow pattern as a function of the Reynolds number (Re).
-
(2)
Quantify Structural Influence: Quantify the effects of tube outer diameter (D) and tube spacing (S) on film coverage, thickness uniformity, and intertube liquid redistribution behavior.
-
(3)
Provide Mechanistic Insight: Offer in-depth fluid mechanical explanations for the observed phenomena, linking the flow pattern and thickness variations to the underlying force balance.
While falling film studies abound, this work specifically elucidates the unique hydrodynamic coupling within the confined shell-side geometry of SWHEs. By quantifying the nonlinear transition between columnar and fan flow regimes under varying structural constraints, we provide a predictive framework for optimizing liquid distribution in complex industrial heat exchangers. Unlike traditional horizontal tube bundles, the spiral arrangement in SWHEs induces lateral flow components. This study bridges the gap by quantifying how these structural nuances facilitate a transition to fan flow, a phenomenon under-reported in standard configurations.
2. Numerical Model and Validation
2.1. Governing Equations
The external falling liquid film flow is modeled as a transient, isothermal, incompressible, and immiscible two-phase flow (liquid pentane and gaseous environment) in three dimensions. Given the relatively low liquid velocities typical in falling film evaporators (Ma ≪ 1) and the assumption of isothermal conditions in this hydrodynamic-focused study, the flow is treated as incompressible. The Volume of Fluid (VOF) method is employed to accurately track the gas–liquid interface, where a single set of momentum equations is shared by the two phases, and the volume fraction of one of the phases (α l ) is tracked throughout the domain.
The governing equations are as follows.
1. The continuity equation is as follows:
| 1 |
where ρ is the density of the fluid mixture (kg/m3), τ is time (s), u is the velocity vector (m/s), and Sm is a mass source term (set to zero for this nonevaporating, isothermal case).
2. Momentum Conservation Equation (3D Navier–Stokes):
| 2 |
where p is the pressure (Pa), ρg represents the gravitational force term (acting in the negative y-direction), and Fsurf is the surface tension force per unit volume, calculated using the Continuum Surface Force (CSF) model.
3. Volume Fraction Transport Equation (VOF Method): The interface tracking is achieved by solving the transport equation for the volume fraction of the liquid phase, α l :
| 3 |
where αl = 1 in the liquid phase, αl = 0 in the gaseous phase, and 0 < αl < 1 represents the gas–liquid interface.
The density ρ and dynamic viscosity μ of the fluid mixture are calculated as a volume fraction weighted average:
where the subscripts l and g denote the liquid and gaseous phases, respectively.
The liquid film mass flow rate per unit length, Q, and the dimensionless Reynolds number, Re, are defined as
| 4 |
where ṁ is the total mass flow rate, Xpipe is the axial length of the tube section, and μ l is the dynamic viscosity of the liquid. The equivalent film thickness, δ, often cited in correlations, can be compared to the numerically calculated local film thickness, h:
| 5 |
2.2. Geometric Model and Mesh Division
Through analysis and optimization of SWHE shell side configuration, a shell side geometry model having 64 mm (axial length) × 16 mm (width) × 28 mm (height) is developed, as schematically illustrated in Figure . (The model represents a two-tube section of a continuous helical coil heat exchanger, with S representing the shortest vertical distance between the two tubes, and α representing the circumferential angle.) The distance from the top face of the first heat-exchanger pipe to the distributor is 3 mm, and the bottom face of the second pipe is 3 mm. There are four distributors in the geometric model. The three distributors in the middle are square and the side length is 2 mm. Due to structural symmetry, the two side distributors are 2 mm × 1 mm. The concrete structure parameters are detailed in Table .
1.

Shell-side geometric model of spiral wound heat exchanger.
1. Structure Parameters of Shell Side Geometry Model for Spiral Coil Type Heat Exchanger.
| dimension/mm | number of distributors | distributor size, mm | diameter of heat exchange tube, mm | distance between shafts, mm |
|---|---|---|---|---|
| 64 × 16 × 28 | 4 | 2 × 2 | 8 | 6 |
2.3. Initial and Boundary Conditions
The simulation uses Pentane as the liquid phase, as its physical properties (ρliq = 620.7 kg/m3, μliq = 0.2204 mPa·s) at 25 °C closely match those of the actual cryogenic mixed refrigerant (ρ = 571.0 kg/m3, μ = 0.2535 mPa·s) used in LNG processes. The gas phase is air (ρgas = 1.225 kg/m3, μgas = 1.7894 × 10–5 Pa·s). The surface tension coefficient is set to σ = 16.0 × 10–3 N/m. The flow is assumed to be isothermal (T = 25 °C), neglecting heat transfer effects to focus exclusively on hydrodynamics.
Boundary Conditions (BCs) are defined as follows: 1. Inlet (Distributors): Velocity Inlet is applied, with the velocity determined by the target Re number (mass flow rate) and the total inlet area. α=1 (pure liquid). 2. Outlet (Bottom and Outer Surface): Pressure Outlet (0 Pa gauge pressure). α=0 (pure gas). 3. Side Walls: Symmetry boundary condition is applied to simulate a periodic flow in the lateral (Z) direction, approximating a larger tube bundle. 4. Solid Walls (Tube and Casing): No-slip boundary condition and static contact angle of 0° (perfectly wetting surface) are assumed. The 0° contact angle represents the perfectly wetting behavior of cryogenic alkanes on industrial surfaces. To account for fluid property variations, a sensitivity analysis regarding the Kapitza number (Ka) has been included. Kapitza number ( ) describes the ratio of surface tension to viscous forces. For Pentane in this study, Ka ≈ 3650. A sensitivity analysis was conducted by varying Ka by ± 20%. Results indicate that a higher Ka (increased surface tension) delays the regime transition from columnar to fan flow by 15% in Re. Conversely, lower Ka promotes earlier interfacial instability and wave formation, which is critical for predicting SWHE performance with different mixed refrigerants.
2.4. Grid Independence Study and Model Validation Mesh Division
The computational domain utilizes an unstructured mesh composed of hexahedral and polyhedral elements. To accurately capture the liquid–gas interface and the steep velocity gradient near the wall, a highly refined mesh is employed in the region of the liquid film flow (within 0.5 mm of the pipe surface). The maximum mesh size in the first layer near the heat-exchange pipe wall is 0.02 mm, ensuring a wall y+ ≈ 1. A representative view of the mesh distribution is shown in Figure .
2.

Simulated cloud chart of descending membrane flow.
2.4.1. Grid Independence
The grid independence was verified by comparing the average film thickness (δavg) on the upper tube surface at Re = 1000 for three different mesh densities: 1,228,430, 2,496,235, and 3,825,462. When increasing the mesh number from 1,228,430 to 2,496,235, the average change rate of film thickness was 13.24%. However, increasing the mesh to 3,825,462 resulted in a marginal change rate of only 1.73%. This demonstrates that the mesh size of 2,496,235 is sufficient to achieve grid-independent results. To strengthen the grid analysis, the Grid Convergence Index (GCI) was calculated for three mesh densities. The GCI for the fine mesh was found to be within the asymptotic range of convergence (see Table ) .
2. Grid Convergence Index (GCI) for Mesh Densities (Re = 1000, D = 8 mm).
| Mesh | Number of Elements | δavg (mm) | Relative Error | GCI |
|---|---|---|---|---|
| Coarse (N3) | 1,228,430 | 0.495 | - | - |
| Medium (N2) | 2,496,235 | 0.561 | 13.24% | 5.12% |
| Fine (N1) | 3,825,462 | 0.571 | 1.73% | 1.21% |
2.4.2. Model Verification
The VOF model’s ability to accurately capture film hydrodynamics was verified by comparing the simulated flow characteristics to established falling film theory and experimental observations (cited in the literature).
-
(1)
Fundamental Hydrodynamic Correlation: For simplified laminar falling film flow over a vertical flat plate, the theoretical film thickness is proportional to Q1/3 (eq , derived from Nusselt theory). Our simulation results for a simplified 2D configuration (vertical flat plate) confirmed that the average film thickness closely follows the δ ∝ Re1/3 relationship, validating the model’s fundamental adherence to the governing physical laws.
-
(2)
Comparison with Literature Flow Patterns: The predicted flow features in the intertube region (Figure , Flow Pattern Verification) show strong qualitative agreement with the experimental and numerical findings of Sun et al. and other related literature. Furthermore, the local film thickness and coverage ratios show a deviation of less than 8% compared to the experimental results of Sun et al. Specifically, the simulation correctly captured: (i) the formation of a distinct waveform crest on the upper tube, (ii) the subsequent accumulation leading to pronounced columnar flow that bridges the intertube gap, and (iii) the region of high interfacial curvature at the point of liquid detachment from the first tube and impingement on the second tube.
3.

Simulation cloud diagram of falling film flow pattern.
These quantitative and qualitative comparisons establish the reliability of the 3D VOF model for analyzing the complex hydrodynamics within the SWHE unit cell.
3. Results and Mechanistic Discussion
3.1. Influence of Reynolds Number (Re) on Liquid Film Hydrodynamics
The effect of the Reynolds number (Re) on the liquid film hydrodynamics was investigated by varying Re from 300 to 2000 while keeping the structural parameters constant at D = 8 mm and S = 6 mm.
3.1.1. Flow Regime Transition and Film Profile
As illustrated in Figure a–g, the flow pattern undergoes a significant transition from a stable, low-momentum regime to a highly dynamic, high-momentum regime as Re increases.
4.

Distribution and flow characteristics of liquid film on the shell side of the spiral wound heat exchanger under different Reynolds numbers (D = 8 mm, S = 6 mm).
Low Re (Re = 300 to 800): The flow is dominated by gravitational and viscous forces. The liquid forms distinct, stable columnar flows beneath the distributor outlets. The liquid film surface is relatively smooth with minimal wave activity (Figure a-c). The low inertial forces prevent the liquid from spreading effectively in the circumferential (x) direction after leaving the distributor.
Transition Regime (Re = 1000 to 1300): Inertial forces begin to compete with surface tension and viscous forces. The liquid columns show a pronounced tendency to merge (Figure d, e). Crucially, the excess kinetic energy allows the liquid hitting the tube surface to spread circumferentially, leading to the development of a fan flow pattern in the intercolumnar regions. At Re = 1300, the flow strongly transitions to a fan-columnar hybrid. The kinetic energy of the incoming fluid, represented by the momentum term ρuu in eq , increases significantly. This enhanced momentum overcomes the stabilizing influence of surface tension (Fsurf) and viscosity (μ), facilitating circumferential spread and generating distinct “ripple” fluctuations on the film surface. The inertial force creates a peak liquid accumulation region directly beneath the distributor due to impact momentum.
High Re (Re = 1600 to 2000): The flow is fully characterized by a highly agitated fan flow (Figure f, g). The liquid film covers the entire axial and circumferential surface more uniformly than in the columnar regime, although distinct peaks remain beneath the distributors. High-amplitude waves (ripples) radiate outward from the peak thickness regions. The high liquid velocity (up to 1.2 m/s at Re = 2000) increases the shear stress at the gas–liquid interface. This shear energy fuels the growth of interfacial instabilities, leading to the formation of large, traveling waves. These waves actively promote liquid mixing and distribution, which is highly beneficial for heat transfer enhancement. Mechanistically, the transition to fan flow is driven by the Turbulence Kinetic Energy (TKE) distribution. At Re = 2000, the high liquid velocity (1.2 m/s) increases interfacial shear stress. The Weber number ( ) exceeds the critical threshold where inertia overcomes surface tension, facilitating the radiating “ripple” patterns that enhance mixing.
3.1.2. Axial Film Thickness Profile
The variation of the maximum liquid film thickness along the tube’s axial direction (z) for different circumferential angles (α) is shown in Figure .
5.

Axial curve of the liquid film thickness at different Reynolds numbers (S = 6, D = 8).
Circumferential Dependence (α): The maximum film thickness consistently occurs at the large circumferential angle (α=150°, near the point of detachment from the tube surface) and increases with Re. For α=150°, the maximum thickness at Re = 2000 (1.4223 mm) is 56% greater than at Re = 500. As the liquid flows around the tube, the component of the gravitational force tangential to the surface decreases, and the component normal to the surface (promoting film detachment) increases. The liquid accumulates just before the point of detachment (α=150°) due to the strong opposing action of viscous and surface tension forces attempting to hold the film to the wall against the accelerating gravity.
Axial Profile Evolution: At small circumferential angles (α=30°, Figure a), a secondary peak in film thickness is evident between the main peaks (the centers of the columnar flows). This secondary peak is caused by the limited circumferential spread of the liquid. A small fraction of the liquid spreads axially and converges in the region midway between the distributor columns. As α increases (e.g., to α=150°, Figure d), the secondary peak disappears because the cumulative effect of gravity and the longer flow path allows the film to spread more fully and homogenize in the axial direction, distributing the liquid more evenly across the width.
3.2. Influence of Tube Outer Diameter (D) on Liquid Film Uniformity
The influence of the tube outer diameter (D) on film distribution was investigated by varying D from 6 mm to 12 mm at a constant Re = 1000 and S = 6 mm.
As D increases, the liquid film flow transitions from a relatively uniform, sheet-like flow to a pronounced columnar flow, as depicted in Figure a–d.
6.
Distribution of the fluid membrane on the casing of a helical coil heat-exchanger with various external diameters (Re = 1000, S = 6).
Small D (D = 6 mm): The smaller diameter results in a higher curvature and a shorter circumferential path length. This geometry imposes a higher flow resistance and enhances the stabilizing effect of surface tension, which pulls the liquid toward the wall, resulting in a more contiguous and uniform sheet-like film coverage (Figure a).
Increasing D (D = 8 mm to 12 mm): As D increases, the circumference increases, and the liquid film must travel a longer path to spread circumferentially. The balance shifts. The increased radius of curvature reduces the pressure gradient generated by surface tension (Laplace pressure), weakening its ability to spread the film circumferentially. Simultaneously, the fixed amount of liquid flow from the distributor must cover a larger surface area, but the inertial forces are not sufficient to overcome the increasing distance and circumferential shear. Consequently, the liquid favors the path of least resistanceconcentrating vertically beneath the distributor outletsleading to a severe columnar flow and poor film coverage uniformity (Figure d). The nonuniformity and wave amplitude are significantly magnified at D = 12 mm.
Using this method, the variation law of the thickness of the fluid membrane along the axis of the drop membrane was analyzed (Figure ). The results show that when the circumferential angle is 30° (Figure a), the maximum value of the fluid membrane along the axis was more apparent, and there is a subpeak between the two peaks. When the circumferential angle is large (Figure d), the subpeak basically disappears. The reason for this is that when the peripheral angle is small, a portion of the liquid will expand in the axis with the flow of the self-distributor, and more liquid will expand and gather in the peripheral direction, thereby forming the sub peak. At a circumference of 30° (Figure a), the maximal thickness of the waveform peak in D is 6 mm, 8 mm, 10 mm and 12 mm are 0.4828 mm, 0.5618 mm, 0.4607 mm, 0.5052 mm. As the peripheral angle is raised to 75° (Figure b), the maximum thickness of the peak of D is 6 mm, 8 mm, 10 mm and 12 mm is 0.5656 mm, 0.6087 mm, 0.7135 mm and 0.4468 mm, respectively. When the circumferential angle increases to 105° (Figure c), the maximum thickness of the peak of D is 6 mm, 8 mm, 10 mm, 12 mm is 0.5778 mm, 0.7516 mm, 0.7614 mm, 0.5280 mm, respectively. When the circumferential angle increases to 150° (Figure d), the maximum thickness of the peak of D is 6 mm, 8 mm, 10 mm and 12 mm is 0.7217 mm, 1.0029 mm, 1.0544 mm and 0.5598 mm, respectively. It can be seen that when the pipe diameter is less than 10 mm, the thickness of the top layer of the fluid membrane becomes larger and larger as the peripheral angle increases. When the external diameter of the pipe is greater than that of the outside of the pipe, the thickness of the liquid membrane is first increased and then reduced, and the maximal value is obtained at D = 10 mm. This is due to the need for more flow for peripheral and axial expansion when the external diameter of the pipe is big.
7.

Axial diagram of the liquid film thickness under different outer diameters of tubes (S = 6, Re = 1000).
3.3. Influence of Tube Spacing (S) on Intertube Hydrodynamics
The tube spacing (S) directly controls the hydrodynamic conditions in the intertube gap and the ability of the falling liquid columns to merge, as analyzed by varying S from 4 mm to 10 mm at Re = 1000 and D = 8 mm.
3.3.1. Intertube Flow Pattern and Uniformity
As depicted in Figure a–d, the spacing critically dictates whether the flow is constrained (sheet-like) or separated (columnar).
8.
Distribution and flow characteristics of liquid film on the shell side of the spiral wound heat exchanger under different tube spacings (D = 8, Re = 1000).
Small S (S = 4 mm): The narrow gap strongly constrains the flow. The liquid columns leaving the first tube quickly merge into a continuous sheet-like flow before reaching the second tube (Figure a). Mechanistic Explanation: The rapid merging is driven by the high capillary effect and enhanced pressure gradient in the small intertube gap. While the film is thin and flow velocity is high, the overall coverage uniformity on the second tube is good due to the sheet-like impingement. However, this configuration might impose a high flow resistance in an industrial setting.
Optimal S (S = 6 mm): At this spacing, the flow transitions to a controlled columnar-fan hybrid (Figure b). The liquid columns are distinct but merge smoothly before impacting the second tube. The fluid velocity is slightly lower than at S = 4 mm, but the film thickness is higher, indicating better liquid accumulation and more effective coverage due to balanced spreading. This spacing offers the best trade-off between coverage and minimized flow restriction.
Large S (S = 8 mm to 10 mm): As S increases, the liquid columns become increasingly detached, leading to severe columnar and droplet flow in the intertube region (Figure c,d). The gravitational force dominates over the lateral surface tension forces due to the increased distance. The liquid is unable to bridge the gap and maintain a continuous film, resulting in discrete droplets or thin, unstable columns impinging on the second tube. This leads to substantial nonuniformity and potential dry-out areas on the second tube, severely degrading heat transfer performance.
3.3.2. Axial Film Thickness Profile
The axial film thickness profile (Figure ) confirms the hydrodynamic observations. At the point of maximum accumulation (α=150°, Figure d), the film thickness shows a nonlinear dependence on S: it increases from S = 4 mm (0.8794 mm) to S = 6 mm (1.0029 mm), and then drops significantly at S = 10 mm (0.4144 mm). This confirms S = 6 mm provides the most robust film accumulation without the excessive breakup observed at S = 10 mm. Notably, the sensitivity of the film thickness variation to S diminishes at larger circumferential angles (α ≥ 150°) because the flow dynamics are predominantly governed by local gravitational acceleration and viscous drainage near the detachment point, regardless of the intertube column structure.
9.

Axial diagrams of liquid film thickness under different tube spacings (D = 8, Re = 1000).
4. Conclusions
This 3D CFD study provides a comprehensive, mechanistically grounded analysis of the hydrodynamic behavior of the external falling liquid film on the shell side of a Spiral-Wound Heat Exchanger, specifically focusing on the coupled effects of the Reynolds number (Re), tube outer diameter (D), and tube spacing (S). Supported by detailed fluid mechanical reasoning, the following key conclusions were drawn:
-
1.
Flow Regime Transition and Re Influence
The liquid flow exhibits a critical hydrodynamic transition from a viscous-dominated columnar flow (Re ≤ 800) to an inertia-dominated fan flow (Re ≥ 1600). The increase in Re significantly augments the inertial forces, successfully overcoming the stabilizing influence of surface tension and viscosity. This leads to a substantial increase in wave amplitude (up to 56% thicker film at Re = 2000 compared to Re = 500) and improved circumferential spreading capacity of the liquid. The inertial forces at high Re overcome surface tension, allowing the liquid film to spread wider, leading to more uniform coverage of the tube surface and reducing the potential for dry spots. Increased Mixing: The fan flow is characterized by large, high-amplitude waves (ripples). These waves intensify turbulence and mixing within the liquid film, which actively reduces the thickness of the thermal boundary layer and enhances the heat transfer rate.
-
2.
Effect of Tube Diameter (D)
Increasing the tube outer diameter (D) introduces a nonuniform effect on film coverage. Larger diameters reduce the curvature-driven pressure gradient (Laplace pressure), thereby diminishing the film’s intrinsic ability to spread circumferentially. This phenomenon promotes flow concentration, exacerbates nonuniformity, and favors the formation of severe columnar flow. An upper practical limit for D (identified as D = 10 mm in this study) exists, beyond which the film breaks up, resulting in localized dry-out areas and compromised coverage.
-
3.
Optimization of Tube Spacing (S)
Tube spacing (S) is critically important for efficient intertube liquid redistribution. Very small spacing (S = 4 mm) promotes sheet flow but may incur increased flow resistance, while excessively large spacing (S = 10 mm) results in liquid column detachment and droplet formation, severely degrading coverage uniformity on the downstream tube. An optimal spacing of S = 6 mm was identified. This configuration successfully balances the opposing effects of gravity-driven drainage and capillary-driven bridging, thereby maximizing film thickness and coverage continuity (e.g., h = 1.0029 mm at α=150°).
This work provides crucial design guidelines for SWHEs: operating at a higher Re is hydrodynamically favorable due to enhanced wave activity and spreading, but the structural parameters D and S must be carefully selected to avoid flow breakup and dry-out. Future research will focus on the coupled heat and mass transfer model, incorporating evaporation (especially for volatile fluids like Pentane), the influence of noncondensable gas flow (air shear stress), and the effects of tube surface roughness and wettability to fully simulate real-world SWHE operation.
Acknowledgments
This work was supported by Natural Science Foundation of Dongying, China (Grant No. 2023ZR04), the Science Development Funding Program of Dongying of China (Grant No. DJ2024004), and Dongying City Science and Technology Innovation Major Project (Science and Technology Development Guidance Scheme) (Grant No. 2024ZDJH75).
The data sets supporting the findings and conclusions of this article, including all raw numerical simulation output files and processed data used to generate the figures and tables, are available from the corresponding author, Dr. Longfei Dong (Email: [2018010@sdipct.edu.cn]), upon reasonable request. Furthermore, all key quantitative data are included within this published article.
The authors declare no competing financial interest.
References , , , , − , , and can be resolved via https://oversea.cnki.net/index/.
References
- Fernández-Seara J., Pardiñas Á. Á.. Refrigerant falling film evaporation review: Description, fluiddynamics and heat transfer. Applied Thermal Engineering. 2014;64:155–171. doi: 10.1016/j.applthermaleng.2013.11.023. [DOI] [Google Scholar]
- Yan Y., Liu L., Gao J. J.. et al. Flow and distribution characteristics of liquid film on shell side of spiral-wound heat exchanger. Oil & Gas Storage and Transportation. 2019;38(11):1300–1305. [Google Scholar]
- Lin S., Huang C., Li X. L.. Distribution of Falling Film Thickness on the Horizontal Tubes. The Chinese Journal of Process Engineering. 2016;17(2):209–216. doi: 10.12034/j.issn.1009-606X.216268. [DOI] [Google Scholar]
- Sun C. Z., Li Y. X., Cao X. W.. et al. Numerical simulation of flow characteristics of falling film outside coolant pipelines during liquefaction of natural gas. Oil & Gas Storage and Transportation. 2022;41(02):200–210. [Google Scholar]
- Huang S. H., Zhou R. F., Li S. H.. et al. Experimental study on the enhancement effect of ultrasonic oscillation on an ammonia-water falling film absorber. Applied Thermal Engineering. 2025;263:125351. doi: 10.1016/j.applthermaleng.2024.125351. [DOI] [Google Scholar]
- Zhang F., Zhang Z. B., Geng J.. Study on Shrinkage Characteristics of Heated Falling Liquid Films. AIChE J. 2005;51(51):2899–2907. doi: 10.1002/aic.10610. [DOI] [Google Scholar]
- Christians M., Thome J. R.. Falling film evaporation on enhanced tubes, part 1: Experimental results for pool boiling, onset-of-dryout and falling film evaporation. International Journal of Refrigeration. 2012;35:300–312. doi: 10.1016/j.ijrefrig.2011.10.020. [DOI] [Google Scholar]
- Ishikawa H., Ookawara S., Yoshikawa S.. A study of wavy falling film flow on micro-baffled plate. Chem. Eng. Sci. 2016;149:104–116. doi: 10.1016/j.ces.2016.04.021. [DOI] [Google Scholar]
- Sisoev G. M., Matar O. K., Lawrence C. J.. Absorption of gas into a wavy falling film. Chem. Eng. Sci. 2005;60(60):827–838. doi: 10.1016/j.ces.2004.09.057. [DOI] [Google Scholar]
- Gao X., Zhu X. R.. Numerical simulation of the falling film flow pattern and liquid film distribution of high viscosity fluid outside horizontal tubes. Journal of Zhejiang Sci-Tech University. 2023;49(5):618–627. doi: 10.3969/j.issn.1673-3851(n).2023.05.011. [DOI] [Google Scholar]
- Qiu Q. G., Jiang W. G., Shen S. Q.. et al. Numerical investigation on characteristics of falling film in horizontal-tube falling film evaporator. Desalination and Water Treatment. 2015;55:3247. doi: 10.1080/19443994.2014.968907. [DOI] [Google Scholar]
- Luo C., Ma W. B., Gong Y. L.. Design of single vertical tube falling-film evaporation basing on experiment. Journal of Loss Prevention in the Process Industries. 2011;24(24):695–698. doi: 10.1016/j.jlp.2011.06.013. [DOI] [Google Scholar]
- Wan Z. H., Li Y. Z.. Simulation study on the flow characteristics of low temperature liquid oxygen falling film on horizontal tube. Cryogenics & Superconductivity. 2022;50(05):85–91. doi: 10.16711/j.1001-7100.2022.05.015. [DOI] [Google Scholar]
- Xu X. F., Zhou Y. S., Liu G.. et al. Study on the thickness distribution of horizontal falling film outside semi-elliptical tube under reverse airflow condition. Journal of Engineering for Thermal Energy and Power. 2021;(36):99–105. doi: 10.16146/j.cnki.rndlgc.2021.08.015. [DOI] [Google Scholar]
- He N., Chen X. D.. Numerical analysis of the influence of distributor on the falling film flow outside the horizontal tube. Fluid Machinery. 2022;50(04):27–35. doi: 10.3969/j.issn.1005-0329.2022.04.005. [DOI] [Google Scholar]
- Jayakumar A., Balachandran A., Mani A., Balasubramaniam K.. et al. Falling film thickness measurement using air-coupled ultrasonic transducer. Experimental Thermal and Fluid Science. 2019;109:109906. doi: 10.1016/j.expthermflusci.2019.109906. [DOI] [Google Scholar]
- Yue Y. K., Zhang Z. T., Zhang H. L.. et al. Experimental Study on Hydrodynamic Behavior of Falling Film over Vertical Tube. Journal of Thermal Science. 2023;32(32):1512–1522. doi: 10.1007/s11630-023-1732-z. [DOI] [Google Scholar]
- Liu C.. Study on Gas-liquid Falling Film Flow in Internal Heat Integrated Distillation Column. Journal of Physics: Conference Series. 2017;916:012050. doi: 10.1088/1742-6596/916/1/012050. [DOI] [Google Scholar]
- Yao N., Wu A. H., Lin J. M.. et al. Experimental study on flow pattern conversion between horizontal tubes falling-film evaporator. Refrigeration and Air-Conditioning. 2020;20(4):54–58. doi: 10.3969/j.issn.1009-8402.2020.04.012. [DOI] [Google Scholar]
- Han H. M., Cong H. F.. et al. Formation and flow of spiral liquid bridge falling film structured packing spiral gap liquid bridge. Chemical Industry and Engineering Progress. 2022;41:584–592. doi: 10.16085/j.issn.1000-6613.2021-0645. [DOI] [Google Scholar]
- Gonda A., Lancereau P., Bandelier P.. et al. Water falling film evaporation on a corrugated plate. International Journal of Thermal Sciences. 2014;81(81):29–37. doi: 10.1016/j.ijthermalsci.2014.02.010. [DOI] [Google Scholar]
- Tohidi Y., Ormiston S. J.. A two-phase sharp-interface numerical model of falling film thermal desalination in a vertical channel. Heat and Mass Transfer. 2025;61:61. doi: 10.1007/s00231-024-03527-3. [DOI] [Google Scholar]
- Li Y. J., Li S. H., Chen X. Y.. et al. Nano-enhanced ammonia water falling film generation model outside vertical tube. Journal of Southeast University (Natural Science Edition) 2018;(48):427–434. doi: 10.3969/j.issn.1001-0505.2018.03.008. [DOI] [Google Scholar]
- Dong C. S., Hibiki T., Zhang L. Z.. et al. Falling film liquid desiccant air dehumidification. Experimental and Computational Multiphase Flow. 2020;2(4):187–198. doi: 10.1007/s42757-019-0036-8. [DOI] [Google Scholar]
- Dai Z. S., Zhang Y. H., Wang S.. et al. Falling-film heat exchangers used in desalination systems: A review. Int. J. Heat Mass Transfer. 2022;185:122407. doi: 10.1016/j.ijheatmasstransfer.2021.122407. [DOI] [Google Scholar]
- Xu Zn, Zhu H. J., Zhao Y. L.. Study on the Horizontal Tube Falling-Film Thickness Under Cross Airflow. Journal of Engineering Thermophysics. 2024;(45):1808–1818. [Google Scholar]
- Zhang M. C., Mou X. S., Shen S. Q.. Numerical simulation analysis of heat transfer in falling film - evaporation process of air-water convection outside horizontal tube. Journal of Dalian University of Technology. 2022;62(2):172–178. doi: 10.7511/dllgxb202202009. [DOI] [Google Scholar]
- Deng, S J. Study on the Flow Characteristics of Falling Film on Horizontal Tubes Under Cross Airflow. Yanshan university, 2024; 10.27440/d.cnki.gysdu.2024.000287. [DOI] [Google Scholar]
- He X. J., Wang J., Li Z. L.. Flow and heat transfer characteristics of falling film across tube bundles. Fluid Machinery. 2023;(51):69. doi: 10.3969/j.issn.1005-0329.2023.03.011. [DOI] [Google Scholar]
- Bao S. S., Sang W. R., Xu Z. H.. et al. Characteristics of falling film flow on shell side of tube-wound heat exchanger with circular and flat tubes. Petroleum Processing and Petrochemicals. 2024;55(9):62–69. [Google Scholar]
- Li L., Li C., Zhang Y. J.. Experimental Investigation on the Water Film Behavior with Counter-current Air Flow. Energy Conservation Technology. 2016;34(4):301–304. doi: 10.3969/j.issn.1002-6339.2016.04.003. [DOI] [Google Scholar]
- Liu S. L., Mu X. S., Shen S. Q.. Experimental study on the distribution of local heat transfer coefficient of falling film heat transfer outside horizontal tube. International Journal of Heat and Mass Transfer. 2021;170:121031. doi: 10.1016/J.IJHEATMASSTRANSFER.2021.121031. [DOI] [Google Scholar]
- Wang Y. H., Xing C. C., Liu J. T.. Analysis of liquid film flow and heat transfer characteristics of falling film evaporation outside horizontal bellows. Journal of Jiangsu University (Natural Science Edition) 2023;44:89–95. doi: 10.16146/j.cnki.rndlgc.2017.09.004. [DOI] [Google Scholar]
- Zhu, G Z. Numerical study on falling film flow and heat transfer in horizontal tube with non-condensable gas. Dalian University of Technology, 2022; 10.26991/d.cnki.gdllu.2022.000025. [DOI] [Google Scholar]
- He, N. Study on flow and heat transfer characteristics of liquid falling film flow on shell side of spiral wound heat exchanger. Lanzhou University of Technology, 2022; 10.27206/d.cnki.ggsgu.2022.000100. [DOI] [Google Scholar]
- Wang M X., Li Q. T., Qi D.. et al. Effect of surface tension model on numerical calculation results of heat transfer in falling film flow over horizontal elliptical tubes. Heating Ventilating & Air Conditioning. 2024;54(5):109–116. doi: 10.19991/j.hvac1971.2024.05.15. [DOI] [Google Scholar]
- Wang N. J., Zhu C. L., Li M. J.. Research progress on performance law of horizontal tube falling film heat exchanger. Science Technology and Engineering. 2024;(24):879–896. [Google Scholar]
- Xiang, Z J. Study on heat transfer in falling film process of horizontal pipe containing non-condensing gas[D]. Dalian University of Technology, 2022; 10.26991/d.cnki.gdllu.2022.002981. [DOI] [Google Scholar]
- Xu Z. T., Tian X. W., Zhang K.. et al. Research on the absorption characteristics of nanolithium bromide vertical tube falling film. Cryogenics & Superconductivity. 2025;53(9):78-85+92. doi: 10.16711/j.1001-7100.2025.09.011. [DOI] [Google Scholar]
- Liu Y. L., Xu S. M.. Numerical simulation of falling film absorption in swaying tube using TFE/NMP as working fluid. CIESC Journal. 2005;3:417–423. [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data sets supporting the findings and conclusions of this article, including all raw numerical simulation output files and processed data used to generate the figures and tables, are available from the corresponding author, Dr. Longfei Dong (Email: [2018010@sdipct.edu.cn]), upon reasonable request. Furthermore, all key quantitative data are included within this published article.


