Ships must draw in external air and discharge indoor air to the outside for indoor temperature control and ventilation. In particular, in the engine compartment where various machines, including the engine, are installed, a significant amount of air must be circulated for engine and equipment operation. Thus, an air-conditioning system composed of air-intake louvers, fans, and ducts is installed. An air-intake louver comprises multiple blades, as shown in Fig. 1. It is installed on the ship’s outer plating to draw in external air. Water separators that remove water from the air are applied to louver blades to inhibit the entrance of water, including on snowy or rainy days, as shown in Fig. 2.

However, for air-intake louvers of ships that operate in polar waters, separated water can freeze at the moisture-separation section, which may affect the safety of the ship and its crew by obstructing the air inlet. Therefore, an anti-icing structure that uses electric heat tracing or a thermal-transfer fluid should be applied to air-intake louvers to prevent icing.

Electric heat tracing, which converts electrical energy into thermal energy, is commonly used in ships operating in low-temperature environments. In electric heat tracing, the associated equipment can be easily installed and maintained, and precise temperature control is enabled through sensors. However, its application is limited due to restrictions on heat capacity. Therefore, polar ships generally use thermal-transfer fluids to respond to icing. In this study, Polyester’s Termolan H, which is applicable at both high and low temperatures and safe for human health, was used as a thermal-transfer fluid. The physicochemical and thermophysical properties of Termolan H are listed in Tables 1 and 2, respectively.

To effectively separate water and prevent the freezing of separated water, the air-intake louver blade was designed by implementing moisture-separation sections and thermal-transfer fluid piping, as shown in Fig. 3. Two moisture-separation sections were applied to each blade, around which thermal-transfer fluid piping was installed.

The air-intake louver featured a width of 1 m and a height of 1 m, and 17 blades were applied to it, as shown in Fig. 4. Thermal-transfer fluid was introduced through the upper piping. It flowed into each blade and was discharged through the lower piping. Except for the structure for fixation on the hull, the entire piping of the thermal-transfer fluid and the blades were modeled using Ansys Inc.’s DesignModeler (Ver. 2022), as shown in Fig. 5.

Heating was performed through the thermal-transfer fluid to prevent icing in the blades of the air-intake louver. Moreover, a design that supplies uniform heat to all blades was required to prevent losses due to excessive energy supply and avoid icing even in the least heated blade. In this study, three-dimensional (3D) numerical analysis was conducted to evaluate the temperature distribution and uniformity of the thermal-transfer fluid.

The geometric model used for the analysis was constructed using Ansys DesignModeler (Ver. 2022), and the grid for the inside of the thermal-transfer fluid piping was generated through the Ansys Meshing module. The grid system was constructed based on unstructured irregular hexahedrons, and Ansys Fluent (Ver. 2022)―a commercial software program based on the finite-volume method―was used for flow and heat-transfer analysis. Numerical analysis was conducted based on the semi-implicit method for pressure linked equations algorithm, which is a pressure–velocity coupling technique.

The analysis target was assumed to be a single-phase incompressible flow, and governing equations based on mass, momentum, and energy conservation laws were applied for the overall analysis. Among them, the transport equation for the generalized scalar variable ∅_j is expressed as follows (Ansys, 2009):

where ρ is the density, u_i the i-direction velocity component, x_i the spatial coordinate, Γ_k the diffusion coefficient, S _∅j the source term, and ∅_j the general scalar physical quantity (e.g., energy and temperature). Additionally, j is an index variable that indicates one of several scalar physical quantities (e.g., energy, concentration, and turbulence variables).

Under incompressible flow, energy conservation is expressed as follows:

where e is the internal energy per unit mass, k the thermal conductivity, T the temperature, and r the heat quantity per unit mass.

In this study, the flow and temperature distribution of the thermal-transfer fluid were analyzed under steady-state conditions. Thus, the time derivative terms in Eqs. (1) and (2) were omitted from analysis.

To select a grid model for numerical analysis, four models with different maximum element sizes were verified. The maximum element size and the number of elements for each model are listed in Table 3. Fig. 6 shows the computational domain and boundary conditions. The thermal-transfer fluid was supplied to the louver through piping with an inner diameter of 16.4 mm and then separated into each blade of the louver to transfer heat. In each blade, piping with an inner diameter of 10.5 mm was installed in two moisture-separation sections. Subsequently, the thermal-transfer fluid was discharged to the outside through piping with an inner diameter of 16.4 mm. Velocity inlet and pressure outlet boundary conditions were applied to the inlet and outlet, respectively, while no-slip boundary conditions were applied to the wall. Termolan H, which is the thermal-transfer fluid, was supplied through the inlet of the louver at a temperature of 20 °C. Additionally, a heat-transfer coefficient of 26 W/m²K was applied to the wall considering an external air temperature of −52 °C and a wind speed of 2.7 m/s, which are the design conditions of polar ships, as shown in Fig. 7 (Osczevski, 1995). For grid verification, the average temperatures at the center of each blade (as shown in Fig. 6) were compared under an inlet flow velocity V_i of 0.9 m/s. The analysis results are shown in Fig. 8. The G2 model and the maximum element size were selected to ensure accuracy and efficiency, as shown in Fig. 9.

In this study, the uniformity of the heat supplied into each blade by the thermal-transfer fluid was numerically evaluated, and the average temperatures of the thermal-transfer fluid separated into each blade, which were measured 400 mm below the center of the louver (as shown in Fig. 6) were compared. The difference between the maximum temperature (T_MAX ) and minimum temperature (T_min) of the thermal-transfer fluid separated into each blade at the corresponding position is denoted as ΔT_b.

Numerical analysis was conducted for the case involving the same inner piping diameter (10.5 mm) for the thermal-transfer fluid separated into each blade. The analysis results for V_i = 0.5, 0.6, 0.7, 0.8, 0.9, and 1.0 m/s are shown in Table 4 and Fig. 10. The total pressure loss of the thermal-transfer fluid, ΔP, is included as well.

Although ΔT_b was 49.2 K when the inlet flow velocity was 0.5 m/s, the ΔT_b of each blade decreased gradually and the uneven heat supply resolved gradually as the flow velocity increased. However, ΔP increased significantly with the flow velocity. Fig. 11 shows the temperature distribution of the thermal-transfer fluid under V_i = 0.6 m/s. The thermal-transfer fluid supplied through the upper piping was distributed into the piping of each blade. Relatively high temperatures were maintained at the measurement points of each blade close to the inlet owing to the efficient supply of the thermal-transfer fluid. However, as the distance from the inlet increased, the temperature decreased gradually owing to the reduced flow rate of the supplied fluid caused by pressure loss.

To ensure a homogeneous supply of heat to each blade even under moderate/low flow velocities, the size of the piping installed in the No. 17 blade, which exhibited the lowest temperature, was adjusted. Considering the production of the air-intake louver, nominal pipe sizes were applied in this study. Specifically, sizes of 11.8, 12.5, 13.4, and 13.8 mm were considered in addition to the original size of 10.5 mm as the piping diameter of the No. 17 blade. The analysis results for V_i = 0.5, 0.6, and 0.7 m/s are shown in Tables 5,–7 and Figs. 12,–14, respectively. At V_i = 0.5 m/s, ΔT_b was 49.2K for the original louver with a pipe diameter of 10.5 mm. As the piping size of the No. 17 blade increased, however, ΔT_b decreased gradually and reached 2.6 K at an inner diameter of 13.4 mm. At a pipe diameter of 13.8 mm, however, ΔT_b increased compared with that at a pipe diameter of 13.4 mm owing to a significant increase in the flow rate of the thermal-transfer fluid supplied to the No. 17 blade. Additionally, ΔP decreased gradually as the pipe diameter of the No. 17 blade increased due to the improved flow of the thermal-transfer fluid. At V_i = 0.6 and 0.7 m/s, ΔT_b decreased gradually as the piping size of the No. 17 blade increased, thus resulting in the least difference at an inner diameter of 13.4 mm. At a pipe diameter of 13.8 mm, ΔT_b was greater than that at an inner diameter of 13.4 mm. Additionally, as in the case of V_i = 0.5 m/s, ΔP decreased slightly as the pipe diameter of the No. 17 blade increased. Fig. 15 shows the temperature distribution of the thermal-transfer fluid under a pipe diameter of 13.44 mm for the No. 17 blade and V_i = 0.6 m/s.

The pipe diameter of the No. 17 blade was determined to be 13.4 mm. To supply heat more evenly, the thermal-transfer fluid temperatures of each blade were compared within the inlet flow velocity range of 0.5–0.7 m/s, as shown in Fig. 16. The pipe diameters of the No. 2, 3, and 16 blades, which exhibited relatively low temperatures, were adjusted considering nominal pipe sizes.

The pipe diameters of the No. 2, 3, and 16 blades were changed from the original size of 10.5 mm to 10.8 and 11.8 mm. The analysis results for V_i = 0.5, 0.6, and 0.7 m/s are shown in Tables 8,–10 and Figs. 17,–19, respectively. At V_i = 0.5 m/s, ΔT_b was 2.6 K for the original louver with a pipe diameter of 10.5 mm; however, it decreased to 2.0 K as the pipe diameter of the No. 2, 3, and 16 blades increased to 10.8 mm. At a pipe diameter of 11.8 mm, however, ΔT_b increased to 4.3 K owing to the significant increase in the flow rates of the blades. Meanwhile, ΔP decreased as the pipe diameter increased owing to the improved flow of the thermal-transfer fluid inside the louver. At V_i = 0.6 and 0.7 m/s, the least temperature difference was also observed when the pipe diameter of the No. 2, 3, and 16 blades was 10.8 mm, whereas the temperature difference increased at a pipe diameter of 11.8 mm. In the case of the air intake louver with a pipe diameter of 13.4 mm for the No. 17 blade and 10.8 mm for the No. 2, 3, and 16 blades, the temperature distribution of the thermal-transfer fluid at V_i = 0.6 m/s is shown in Fig. 20.

To ensure an even supply of heat to each blade of the air-intake louver, the pipe diameter of the No. 17 blade was set to 13.4 mm and that of the No. 2, 3, and 16 blades to 10.8 mm. Table 11 shows the temperature of the thermal-transfer fluid at each temperature-measurement point under V_i = 0.6 m/s. The fluid temperature was the lowest (10.0 °C) in the No. 3 blade.

Numerical analysis was conducted to evaluate whether icing occurred in the No. 3 blade, which exhibited the lowest temperature (10.0 °C) for the thermal-transfer fluid. The condensation temperature of moisture is 0 °C; however; a moisture-separation section temperature of at least 5 °C was targeted in this study, considering a safety margin of 5 °C. Two-dimensional (2D) modeling of the blade cross-section was generated using Ansys Inc.’s Mechanical (Ver. 2022), as shown in Fig. 21. The number of grid elements was 5,200. A heat-transfer coefficient of 26 W/(m²·K) was applied to the external wall of the blade considering a wind speed of 2.7 m/s, and −52 °C was applied as the external air temperature. A temperature boundary condition of 10.0 °C was applied to the walls of the two pipes inside the blade for analysis.

Fig. 22 shows the steady-state thermal analysis results for the blade cross-section. The temperature of the moisture-separation sections exceeded 5.4 °C owing to heat transferred from the thermal-transfer fluid. This indicates that the separated moisture will not condense.