August 2022

Heat Transfer

Use CFD analysis to evaluate the impact of hot air recirculation on air-cooled heat exchangers

Air-cooled heat exchangers are widely used in the petroleum, petrochemical, natural gas and other industries for cooling purposes. In contrast to a water-cooled system, which consists of a traditional shell-and-tube heat exchanger and a cooling tower, an air-cooled heat exchanger uses ambient air as the main cooling medium.

Chen, C-C., CTCI, Corp.

Air-cooled heat exchangers are widely used in the petroleum, petrochemical, natural gas and other industries for cooling purposes. In contrast to a water-cooled system, which consists of a traditional shell-and-tube heat exchanger and a cooling tower, an air-cooled heat exchanger uses ambient air as the main cooling medium. The air is blown from the exterior of the bundle to cool the higher-temperature process fluid inside the tube to the desired operating temperature. Since this procedure does not require a large amount of industrial water, air-cooled heat exchangers are suitable in areas that face geographical restrictions, as well as regions that lack water resources or are required to save water. However, as hot air discharged from an air cooler recirculates, a backflow phenomenon occurs, causing the temperature of the air used as the cooling medium to increase, leading to decreased operational performance of the air cooler.

Several factors play a role in hot air recirculation. Among them, ambient air temperature, external wind direction/speed and the height/distribution of the surrounding buildings are the most important factors. While ambient wind direction and speed are natural forces beyond human control, computational fluid dynamics (CFD) software can be used to conduct flow field analysis and simulate air circulation of air coolers under various environment conditions to avoid adverse effects from hot air recirculation. The most suitable configuration and improved measures can be determined to ensure the heat-exchange efficiency of the whole system meets operational requirements, regardless of wind direction and speed in different seasons.

CFD model formulation

In an overall field area, building location, building height, wall size, direction of entry and exit access, and relative position of each building determine the environmental boundary conditions, which in turn affect the operational performance of the air-cooled heat exchanger. Since hot air moves vertically due to thermal buoyancy and pressure, air discharged from an air-cooled heat exchanger will flow upward and be blown away by the crosswind. Therefore, it is also important to consider the external flow wind direction in the overall impact of boundary conditions.

This article will discuss these aspects based on a refinery engineering, procurement and construction (EPC) project in Southeast Asia. As shown in FIG. 1, two major buildings (Side Building A and Side Building B) are on the west side of the main building where an air-cooled heat exchanger is located, as well as a few smaller buildings. A preliminary flow field result suggests that these two major buildings have a greater impact on the flow field than the other smaller buildings because they have solid walls that obstruct air from flowing freely. Therefore, the smaller buildings can be ignored due to their negligible effect on flow field—the focus should instead be placed on simulating the flow field of these three buildings. This also has the benefit of lowering resource loading during simulation. Moreover, all irrelevant factors that may affect rapid convergence of the overall simulation are deleted from the analysis model.

FIG. 1. Schematic plant layout.
FIG. 1. Schematic plant layout.

Reference index: Flow field (velocity field and temperature field)

Once the boundary condition of the steady flow field is set and the turbulence module calculation is complete, the flow field can be observed through simulation.

To quantify the results of CFD numerical analysis, the now-known flow field phenomenon was combined with various energy condition inputs to produce a temperature distribution map around the building, as shown in FIG. 2. Comparing the gradient of the temperature field and flow field diagram shows that temperatures rise more significantly where the flow field is disturbed and where turbulences and vortices occur. In particular, temperatures rise significantly in places where accumulated heat is unable to disperse easily due to slower average flow velocity.

FIG. 2. The temperature gradient of air inlet elevation with various interval lengths between bays: plan view (control group) (A); and the temperature gradient of air inlet elevation with various interval lengths between bays: plan view (experimental group) (B).
FIG. 2. The temperature gradient of air inlet elevation with various interval lengths between bays: plan view (control group) (A); and the temperature gradient of air inlet elevation with various interval lengths between bays: plan view (experimental group) (B).

Simulation result 1: Impact from the angle between the equipment setting direction and crosswind direction

The construction site’s wind rose diagram (showing the percentage distribution of local wind direction/wind speed) indicates that the prevailing winds are north-northeast (NNE), north (N) and south (S). Although the north wind, which is the second strongest wind, has little effect on the outflow air volume of the air cooler, the streamline diagram in FIG. 3 shows that air coming out of the air cooler outlet is clearly affected by the crosswind, as the air discharged from the upstream air outlet is sucked by the downstream fan, causing hot air recirculation. This phenomenon is absent in the case of the strongest wind, the north-northeast wind. The takeaway from this analysis is to avoid displaying air coolers in parallel to the prevailing wind direction to reduce the chance of hot air recirculation.

FIG. 3. Streamline diagram showing the effects of north wind.
FIG. 3. Streamline diagram showing the effects of north wind.

Simulation result 2: Impact from wind speeds

How hot air recirculated at various crosswind wind speeds when blown by the north wind was also calculated. FIG. 4 illustrates the changes in temperature as affected by wind speeds. At low wind speed, temperatures at the inlet/outlet accesses rise gradually. As wind speed increases, the trend of temperature rise is reversed and tends to remain stable. Conversely, in terms of the air-cooled heat exchanger’s accumulated heat-dissipation performance, small fluctuations were seen within a certain range (FIG. 5).

FIG. 4. Temperature rise ratios of air-cooled heat exchanger inlet/outlet in various crosswind speed scenarios.
FIG. 4. Temperature rise ratios of air-cooled heat exchanger inlet/outlet in various crosswind speed scenarios.
FIG. 5. Accumulated heat dissipation rate of air-cooled heat exchanger in various crosswind speed scenarios.
FIG. 5. Accumulated heat dissipation rate of air-cooled heat exchanger in various crosswind speed scenarios.

The faster the ambient crosswind, the higher the temperature at the air cooler outlet/inlet (the air cooler outlet temperature peaks at a wind speed of 7 m/sec). However, the temperature begins to drop as the wind speed exceeds 7 m/sec, until finally beginning to fluctuate at 15 m/sec and heading toward dynamic balance.

In a conventional CFD analysis report on air coolers/heaters, a system is considered free from hot air recirculation and, therefore, efficient as long as the temperature at the air inlet is close or equal to the ambient temperature. The challenge, however, is that the design wind speed is generally higher than the actual ambient wind speed and does not consider lower wind speeds. From the simulation results, it was determined that the worst case in terms of higher temperature actually occurs in a low wind speed scenario. Therefore, traditional verification methods may need to be re-examined in this case.

A further examination of the relationship between crosswind speed and an air cooler’s heat dissipation capability reveals that the trend of average heat dissipation does not necessarily drop or increase as the crosswind speed increases. This means that the heat dissipation capability does not necessarily decrease as wind increases. This may be due to the fact that when the wind is strong, a large proportion of the heat will be dissipated from the wall of the air cooler, causing less accumulated heat and more overall heat dissipation.

Simulation result 3: Impact from the interval length between bays

In a windless flow field, the air exiting the air cooler outlet should flow regularly upward. Observing the flow vector from a structure elevation view (FIG. 6) indicates that air emitted from the outlet fans becomes diffused. Closer to the exit, the outlet air speed is faster and the cross-sectional area is smaller. Conversely, the farther away from the exit, the outlet air speed becomes slower and the cross-sectional area becomes larger. When the distance between two fans becomes too close, the scattered air flow lines are affected, causing turbulent flow. As the crosswind passes, it will interrupt the original flow field between the bays, causing interference against each other and a resulting change in temperature distribution.

FIG. 6. A flow vector diagram with various interval lengths between bays: elevation view  (control group) (A); and a low vector diagram with various interval lengths between bays: elevation view (experimental group) (B).
FIG. 6. A flow vector diagram with various interval lengths between bays: elevation view (control group) (A); and a low vector diagram with various interval lengths between bays: elevation view (experimental group) (B).

A further comparison can examine how spacing between adjacent bays of the air cooler would impact air flow and heat dissipation by looking at the differences between a control group and an experiment group. Despite the flow vector diagram, in a control group (FIG. 6A), the air coming from the outlet of the air cooler flows upward regularly, and adjacent bays in the central area have a greater impact on each other in terms of air flow, as outlet air is obstructed due to the adjacent bays’ outlet air pressure. When the temperature plan view of the air inlet is observed, it can be seen that in the control group (FIG. 2A), air discharged from the upstream air outlet is sucked by the downstream fan. Such hot air recirculation causes the temperature superposition effect. Coupled with air obstruction at the air outlet due to outlet air pressure from adjacent bays, heat is accumulated.

In contrast, the experimental group with larger interval lengths between bays (FIG. 6B) shows a more uniform upward flow of outlet air from the air cooler, as well as a lower mutual impact between bays. The outlet air speed increases, while hot air recirculation becomes less frequent. The inlet temperature tends to be consistent, while the temperature superposition effect is gentle (FIG. 2B). With less obstruction at the outlet from the adjacent bays’ outlet air pressures, more heat can be removed.

Flow field simulation shows that the impact of spacing between two adjacent bays on air outlet and heat dissipation will vary nonlinearly with interval lengths. Therefore, the optimal air cooler spacing can be determined through tests and simulations. The design of air coolers is often such that bays stand next to each other—when the angle between the display direction of the air coolers and the ambient wind direction sits at 45°, the largest interval lengths between bays are available. Echoing conclusions from a previous assessment on the angle between equipment setting direction and crosswind direction, the best angle should indeed be 45°.

Takeaway

The most effective approach to reduce the impact from hot air recirculation on air cooler performance is to avoid placing the air cooler in parallel to the prevailing wind direction. The author’s company’s analysis and simulation indicate that a larger temperature rise ratio of the air cooler outlet/inlet occurs at lower wind speeds rather than at higher wind speeds. The analysis and simulation also show that, contrary to the assumption on the heat dissipation efficiency of air coolers, increased ambient wind speeds do not necessarily mean worse heat dissipation capacities. The efficiency depends on the overall cumulative effect of the heat conduction and heat convection. HP

ACKNOWLEDGEMENT

Research results discussed in this publication come from industry-academia cooperation between CTCI Corp. and the National Taiwan University of Science and Technology.

The Author

Related Articles

From the Archive

Comments

Comments

{{ error }}
{{ comment.comment.Name }} • {{ comment.timeAgo }}
{{ comment.comment.Text }}