Air Grilles Designed to Prevent Backflows in Natural Ventilation Stacks – Experimental Investigation

Natural ventilation is the main type of ventilation in residential buildings in most countries in the world [1]. But since it relies on natural conditions to create the pressure difference required to move the air, its performance strongly depends on the weather. While most of the time natural ventilation may work as intended, in unfavorable weather conditions airflows may be too high or too low [2], and in extreme cases it may even lead to backflows in ventilation stacks. In Poland this effect is increased by the fact that at the beginning of the 21^st century in order to save energy people massively changed old wooden windows in their homes for sealed PCV windows. Nowadays air inlets in the windows are obligatory, but they can only be found in new and recently retrofitted buildings. The lack of airflow through window gaps disturbed the intended operation of natural ventilation, which combined with the unfavorable atmospheric conditions may lead to sucking the air through ventilation stacks. Such backflow may be dangerous if the building has chimneys emitting fumes containing carbon oxide on the roof, and it usually significantly decreases temperature in the room. It may also transport odors from the kitchen or toilet [3] to other rooms. To counter this phenomenon various methods are available – from complex diagnostics of the ventilation to simply adding window air inlets or chimney cowls, or installing fans that change the system from natural to mechanical ventilation [4]. But from the author’s observations the most common method used by the inhabitants is to simply seal the air vent through which cold air goes into the room, further decreasing the efficiency of ventilation (also described in [5]).

To counter the backflows, various manufacturers produced special air grilles meant to prevent the reverse air flow the through ventilation stack. In this paper two such grilles were experimentally tested – one with moving flaps, and one shaped in a way that should greatly increase its flow resistance in case of back-flow. For each of them, pressure loss during airflow in both directions was measured for various air speeds. Also for comparison, the ordinary air grill was tested the same way.

There are several systems of natural ventilation used in the world: single-sided [6] or cross-ventilation [7] through the openings in the wall (usually windows), stack ventilation with individual channels [5] or with a stairwell with the outlet at the top acting as a ventilation stack, and other solutions [8]. But all these installations rely on two phenomena to transport the air: wind pressure and thermal buoyancy.

Wind pressure can either create higher or lower pressure inside the building compared to outside, depending on the wind direction [3]. On the roof where the wind usually blows in a direction parallel to the surface, it creates low pressure in the openings sucking the air from the inside. In the case of natural ventilation with outlets on the roof, this effect can be further increased by the use of a chimney cowl. On the walls the wind will usually create positive pressure on the wall it blows on, and negative pressure in all the other directions. This effect however strongly depends on the shape of the building, and exact values of pressure are very hard to predict. Even if several identical building are placed next to each other, wind pressure distribution on each of them can vary a lot [3]. The wind is an unstable phenomenon, and its speed can vary in a wide range, from standstill to a gale, making the design process of natural ventilation very hard. The automatic regulation is rarely used in natural ventilation, so the amount of supplied air and the direction of the airflow inside the building strongly depend on the wind speed. The literature study shows that in the natural ventilation wind pressure is usually the dominant force causing the airflow [9, 10].

Thermal buoyancy occurs usually in cold seasons when the air inside the building is significantly warmer than the outside air. The difference in density causes the warm air to rise, increasing the airflow through ventilation stacks leading up to the roof. In the lower part of the building thermal buoyancy causes pressure to decrease, and in the upper part it increases [3]. The presence of the staircase leading through all the heights of the building makes it possible for air to travel between flats. In the lower floors air is sucked from flats to the staircase, and in the upper part, it is pushed from the staircase to the flats [5]. When there are ventilation stacks in the building, thermal buoyancy always increases the airflow causing the air in the stacks to rise. This effect is, however, weaker than the influence of the wind and can be easily countered by it in unfavorable weather conditions [5].

Backflows in the ventilation stack usually happen when there is no other way for the air to get into the building, and the natural conditions (usually the wind) cause the pressure in the building to drop very low. It usually happens in a situation described in the introduction, when the windows are airtight and there is no other way for the air to get into the building. It may also happen when the air intakes are equipped with automatic flaps controlled by humidity. The obvious solution is to make some way for the air to get in, but users often seal the windows on purpose to save energy or because of a lack of knowledge of how natural ventilation works. There are also more complicated ways like adding a chimney cap [11] or solar chimney [12].

One of the solutions for this problem is installing the air grilles that should prevent the air from flowing in the opposite direction. In construction of the air grilles, there are two ways of preventing the airflow in the undesirable direction: mechanical elements like flaps, or aerodynamic shape that increases the pressure losses during backflow.

The first solution with the moving parts is relatively simple. Its main advantage, besides its simplicity, is the fact that flap may completely shut off the reverse flow. Its disadvantages are the facts that flap may require some initial force to move, and with low pressure difference it may not open. Also closing of the flap may cause noise, although it is not very important in rooms like toilet or bathroom were ventilation stacks are usually placed.

The second solution is more complex, because it is difficult to design a shape that has low flow resistance in one direction and much higher in the other. The most well – known of such devices is Tesla valve [13] (Fig. 1a). The device made of many channels in this shapes called the “fluid diode” is being currently tested [13, 14]. But the complexity makes it difficult to manufacture it in small size, and those devices usually take shape of large plates placed in window opening. Other, easier to make shape is simple diffuser, where smaller air duct expands into larger diameter (Fig. 1b). For example as described in [15], round diffuser with angle of 50° where diameter of one duct is two times as big as the other has pressure loss coefficient x equal to 0.2 when the air flows from the wide side, and 0.9 when it comes from the narrow side. The exact effect in the diffuser may be slightly stronger or weaker than inside the air duct, but the principles of the phenomena will not change. The advantage of such solution is lack of moving parts, which makes the device more durable and, when the airflow does-n’t create noise, quiet. The main disadvantage is that it cannot block the backflow completely.

In the paper air grilles using both of those principles were tested.

Figure 1.

The physical model of a fragment of a room with a ventilation stack was created for this study (Fig. 2, 3). The model consists of two chambers representing the upper part of the room (1 in Fig. 2) with a lower part of the ventilation stack (2 in Fig. 2). The tested air grill is placed in the channel with a diameter of 100 mm placed between those two spaces (5 in Fig. 2). The whole model is airtight, with connections between parts additionally sealed with silicone (Fig. 3). The air may enter or exit the model with two pipes (one in each chamber), with exhaust fan (4 in Fig. 2) at the end of one of them. The fan is equipped with a voltage regulator to change its output, and may be moved to the other pipe to reverse the flow direction. It creates negative pressure inside the chamber it is connected to, causing the air to flow through the whole model. The measured values are air speed in the middle of the pipe cross-section, and the pressure difference between two chambers. The airspeed is used to calculate the mass flow of air.

Figure 2.

Figure 3.

Because the expected air speed was very low, especially in the case of reverse flow, thermoanemometers AirDistSys 5000 [16] were used to measure the air speed (Table 1). Two thermoanemometers were used in the setup at the ends of both pipes (3 in Fig. 2). During measurements only data from one thermoanemometer placed on the inlet pipe (without the fan) was used, to avoid flow disturbances from the fan to influence the readings. Even though the fan is placed upstream of the anemometer, initial tests showed that it disturbs the airspeed measurements. The data was recorded and stored on the computer.

The pressure difference between the two chambers was measured using a difference micromanometer CMR-10 (Fig. 3d, Table 1). The pressure difference between those two parts of the model is equal to the pressure drop on the air grill.

Three different air grilles, all intended to be installed on a 100 mm air duct, were tested: an ordinary air grill for comparison, grill with moving flaps that close during reverse flow, and grill shaped to decrease the reverse flow by its higher hydraulic losses in the reverse direction.

The ordinary air grill (Fig. 4) is a round plastic grill with fixed air vanes and mesh to prevent insects and larger solid particles to get into the room. It is commonly used for natural ventilation in homes and public buildings.

Figure 4.

The grill with moving flaps (Fig. 5) is also made mainly of plastic. It consists of a frame and four vertical flaps, each attached to the frame with a hinge in its upper part. When there is no airflow, gravity causes the flaps to stay in the closed position. If the pressure behind the flaps is lower than in the room, they open and allow the air to flow. The flaps cannot move in the opposite direction, so if the pressure in the ventilation stack is higher than in the room they stay closed.

Figure 5.

The grill with a shape that prevents backflow (Fig. 6) is made of plastic and styrofoam. It is shaped like a diffuser, similar to the one presented in Fig. 1b. The diameter of the narrow part of the diffuser is 28 mm. There is a mesh in the narrow part, but the manufacturer does not inform if it is just to stop insects or if has any aerodynamic purpose.

Figure 6.

Total of six measurement series were made, two for each of the grilles: one for forward, and one for reverse flow. In each series, 15 measurements of pressure drop on the air grill were made for different air speeds measured in the pipe.

To calculate the relationship between the mass flow of air and the air speed in the measurement point, a CFD simulation of the pipe inlet was made (Fig. 7). The simulation was made using the Ansys CFX software. The geometrical model includes the pipe inlet with a measurement point, and a box representing open space around the pipe. In the physical model, both inlets have the same dimensions, including sensor placement, so only one numerical model was created. The boundary conditions were: opening with relative pressure of 0 Pa on the walls of the box representing the outside space, and outlet with the defined mass flow at the end of the pipe (green in Fig. 7a). The small volume of the model allowed to use a very fine mesh, which increased the accuracy of the calculation. Five simulations were made, for mass flow of 15, 20, 25, 30 and 35 kg/h. For each simulation air speed at the measurement point was noted. With this data, it was possible to formulate an equation showing the dependence of mass flow ṁ from measured air speed w (equation 1). (1) m˙[kgh]=25.9⋅w[ms]+0.027 \dot m[{{kg} \over h}] = 25.9 \cdot w[{m \over s}] + 0.027

Figure 7.

The air temperature during experiments measured by thermoanemometers was about 18°C with small fluctuations. Figure 8 shows the measured relationship between the mass flow of air and pressure drop on the tested air grill. For each of the measurement series this relationship was approximated using the parabola (dotted lines in Fig. 8). Figure 9 shows the comparison between airflows through all the tested grilles for a pressure difference of 15 Pa.

Figure 8.

Figure 9.

The results show that the ordinary grill has lower hydraulic resistance when the air travels in the reverse direction, towards the room. But this difference is small, and would not make much difference in performance of the natural ventilation.

The grill with moving flaps works as intended. Since the area of the openings is smaller than in the ordinary grill, its hydraulic resistance is slightly higher. The airflow in the forward direction for the same pressure difference is about 20% lower for the grill with flaps. In the reverse direction, the flaps stay closed strongly affecting the airflow. They don’t cut off the flow completely as the air can still travel through the gaps between the flaps and the frame, but the airflow drops to about 18% of the flow in forward direction with the same pressure difference.

The diffuser–shaped grill has very high hydraulic resistance in both directions, almost as high as the previously described diffuser with closed flaps. The airflow in the forward direction is 81% lower than through the ordinary grill with the same pressure difference. The reason may be the fact that the channel in the narrow part of the diffuser is very small, with the mesh further increasing its flow resistance. The effect of the shape designed to increase its hydraulic resistance during reverse flow is barely noticeable. It may be due to the fact that on the rear part of the diffuser where the air enters the ventilation stack, there is a similar shape, but turned in the other direction (Fig. 10). It is not shaped like a diffuser, but like a rapid increase in a radius of the air duct. According to [15], this shape has similar properties to the diffuser – when the air travels from the narrow channel to open space, the hydraulic losses are much higher than in the other direction. Research [13, 14] shows that it is possible to design a part with hydraulic losses strongly depending on flow direction, but this particular air grill doesn’t work as intended.

Figure 10.

The hydraulic loss coefficients were calculated for all the air grilles in both directions (Table 2), using formula 2 [15], where Δp is pressure drop, ρ is air density (0.19 kg/m³) and v is average airspeed. The average air speed was calculated for the air duct diameter of 100 mm.

The pressure loss coefficient for the diffuser with flaps is higher than for the ordinary diffuser. The reason for it is a noticeably smaller area of openings, even if the flaps are in the opened position. Despite this, the airflows within the range of pressures common for natural ventilation are only slightly lower. Probably the air grill could be improved by adding two additional flaps on the sides since there is enough room for them. The pressure loss coefficient in the reverse direction is very high, essentially almost blocking the flow, as the flaps are intended to work. The diffuser-shaped air grill has a very high-pressure loss coefficient, almost the same in both directions.

Two air grills blocking the reverse flow in the ventilation duct were tested, working on different principles. The air grill with moving flaps turned out to be working as intended. The airflow in the forward direction was only slightly lower (about 20%) than for the ordinary air grill, while the reverse flow was blocked almost completely (18% of the forward flow with the same pressure difference). The diffuser-shaped air grill had very high hydraulic resistance, blocking the flow in both directions. The flow is 81% lower than for the ordinary diffuser with the same pressure difference. The effect of increasing the flow resistance through the grill in the reverse direction was barely noticeable.

Even though the use of the air grill blocking the back-flow isn’t the best way to fix the malfunctioning natural ventilation, it is much better than sealing the air vents. The air grill with moving flaps may help to prevent the inflow of cold air into the room through the ventilation stack.