Influence of Control Strategy on Heat Recovery Efficiency in a Single-Duct Periodic Ventilation Device

Abstract

The subject of the research was a single-duct, decentralised periodic ventilation unit, using accumulative heat exchanger for heat recovery (also called single-core fixed-bed regenerator). It can achieve high efficiency of heat recovery but is vulnerable to pressure differences between the interior of the building and the outside. To counter this, two control strategies were proposed: adjustment of the fan speed based on an air flow sensor and adjustment of the working cycle length based on temperature sensors. The strategies were tested experimentally in actual working conditions. Due to the use of cheap and simple sensors, it was possible to retain the low price of the device. Both control strategies proved to be successful in equalising the amount of supplied and removed air in a single cycle. Moreover, the heat recovery efficiency increased by more than 10% compared to the default working mode.

1. Introduction

As the heat recovery in ventilation becomes not just a luxury but a common feature in ventilation systems, the need for small, compact devices for use in a single room arises. Commonly used ventilation systems with air ducts and air-handling units are sufficient for most new buildings, but in the existing structures, especially multi-family residential buildings, classic installations may be difficult to build. Moreover, the amount of energy used by fans to move the air through the ducts may in some cases be greater than the savings from heat recovery [1]. One of solutions of this problem is the use of decentralised ventilation systems. Literature studies show that properly designed decentralised ventilation can increase the thermal comfort of occupants and decrease the system energy demand [2,3] and may be advantageous in European climates [2]. Individual HVAC controls for each room can further increase the comfort of occupants and lower the energy demand by adjusting the airflow rate according to actual needs [4]. That is why single-duct periodic ventilation installations with accumulative heat exchangers were introduced by the manufacturers.

These devices are cheap, simple in terms of construction and installation, and, according to manufacturers, feature high efficiency of heat recovery. The device (Figure 1) works in two cycles: At first, warm air is exhausted from a room by a fan and passes through an accumulative heat exchanger, thus increasing the exchanger’s temperature. Then, in the second cycle, the direction of airflow is reversed by the fan, and cold air from the outside is pumped into the room. As it passes through the heat exchanger, its temperature increases. Next, the cycle repeats. The process is controlled automatically, and a single cycle lasts about a minute.

Since those devices are relatively new on the market, they have not been thoroughly tested yet. Theoretical analyses of such a device, confirming its high potential in heat recovery, were carried out by Nizovtsev et al. in 2016 [5]. The tests executed on a commercially available unit [6] showed that both the air flow and the heat regeneration efficiency in perfect conditions are within the values declared by the manufacturers. However, the air flow and, as a result, the heat recovery efficiency depend strongly on the pressure difference inside and outside the room, especially when the fan operates at a low speed. This problem especially concerns decentralised ventilation devices [7,8] but can also affect traditional systems [9]. This pressure difference may be caused by the stack effect and wind, especially in tall buildings with a staircase [10,11]. In most cases, pressure is lower inside the building, with a typical value of about 10 Pa and a maximum value of 30 Pa [12,13]. For example, during the measurements [6] on the second fan speed, at the pressure difference of 7 Pa, the supplied-airflow rate was 22 m³/h, and the exhausted-airflow rate was equal to 6 m³/h, while the declared value was 15 m³/h. Similar results were obtained during measurements by Zemitis and Bogdanovics [13]. These conclusions were confirmed by the research of Mikola et al. [14], who made the simulations and laboratory tests of single-duct heat recovery units installed in a block of flats, and by Pekdogan et al. [15] in the laboratory tests.

The solution to this problem may be the utilisation of an automatic controller equipped with a sensor that measures the airflow rate and adjusts the fan speed accordingly. Other approaches include using temperature sensors on both sides of the exchanger and adjustment of the exhaust and supply time periods. In the existing literature, it is difficult to find studies depicting the internal controls of the ventilation device. Most of the works related to control strategies in ventilation depict utilisation of various sensors located in the room or algorithms that predict the user behaviours [16]. However, Gama et al. [17] describe control of the fan motor using readouts from the airflow sensor. The study focuses more on controlling the DC motor and reading two encoders simultaneously, but it shows that the Arduino microcontroller can be used for this task. Ni et al. [18] describe the ventilation system with the turbine flowmeter. Its readings are not used as input for regulation, but the authors state that such a sensor can be reliable, even after years of use.

The aim of this paper was to test if additional control mechanisms based on airflow or temperature measurements would equalise the amounts of exhausted and supplied air, increasing the heat recovery efficiency of a single-duct ventilation device in actual working conditions. To do this, a custom single-duct ventilation unit with a capacitive heat exchanger was built and placed in a testing station. The device was equipped with additional sensors and programmable microcontroller to gather data during the working cycles. Three different control strategies were tested in real working conditions.

2. Materials and Methods

2.1. Selection of an Airflow Sensor

Since all commercially available single-duct periodic ventilation units with heat recovery are equipped with both an automatic controller and a fan speed regulator, the only physical part that should be added is an airflow rate sensor. The sensor must meet three major requirements. First of all, it has to be cheap in order to maintain one of the greatest advantages of the device—its low price. Secondly, it cannot cause high pressure losses that would require use of a fan of higher output that would generate more noise. Moreover, it should also enable operation at low air speeds—lower than 1 m/s [6].

The initially selected thermo-anemometric sensor was too expensive. Due to their price, sensors that require very accurate measurement of pressure, such as orifice plate or Pitot tube, could not be used. Moreover, the orifice plate generates large pressure loss, and the Pitot tube is less accurate at low airspeed. As a result, a turbine metre was chosen. With proper implementation, it meets all the requirements. However, its disadvantage is relatively slow reaction to the airflow change. But, excluding the cycle change, during which measurements are not necessary, the airflow through the device should not change rapidly.

Additionally, two improvements were made in the device to increase accuracy of the turbine sensor. Instead of one reversible fan, two fans were used on both ends of the device to ensure that the sensor always remains on the suction side of the working fan. This way, high turbulence and irregularity of the speed profile created by the discharge from the fan [6] do not affect the sensor. Furthermore, the honeycomb-shaped flow equaliser was installed in front of the sensor on the external side [19]. On the other side, this role is fulfilled by an accumulative heat exchanger.

The turbine (Figure 2) was made of plastic (PLA) using a 3D printer. It is symmetric to allow measurements of the air flow in both directions. It is mounted on two bearings of a high-performance RC car so that the turbine could move with very low resistance. To measure the rotation speed of the turbine, an infrared sensor was used, and the turbine blades were crossing between the transmitter and the receiver, thus cutting the IR beam off.

2.2. Testing Station

The testing station was built using a 125 mm ventilation duct (Figure 3). There are fans with built-in speed regulators, controlled by a PWM (pulse-width modulation) signal, on both sides of the ducts. The PWM signal can take values from 0 to 255, corresponding to the fan speed. Inside the duct, there is a ceramic regenerative heat exchanger taken from a commercially available ventilation device, RHU-WALL, manufactured by Alnor (05-552 Wola Mrokowska, Poland). The next part is a turbine flow sensor, followed by a honeycomb flow equaliser. There are also temperature sensors: thermocouples at both ends of the ducts and two DS18B20 digital thermometers manufactured by Maxim Dallas (San Jose, CA, USA) outside the duct, in an area not affected by air plume discharged from the device. Since thermocouples have small thermal capacity, they were chosen to register rapid changes in temperature inside the device. The fans, airflow metre, and all the temperature sensors were connected to Arduino Uno microcontroller. For connecting the thermocouples, Adafruit (New York, NY, USA) MAX31856 thermocouple amplifiers with an accuracy of ±0.7 °C were used. Moreover, the Arduino board was equipped with a memory card reader to save the results of the measurements to a file.

To calibrate the turbine flowmeter, SwemaFlow 236 airflow hood manufactured by Swema (Hökarängen, Sweden) was used (Figure 4). It utilises a thermo-anemometric sensor to measure airflow, and its accuracy is ±3.6 m³/h. The calibration was made separately for flows in both directions. As a result, equations depicting the dependence of the turbine’s rotational speed on the airflow rate were formulated.

The entire measuring station was placed in a window opening in an office room located on the ground floor in a building of the Silesian University of Technology. It was mounted in place using a wooden board with thermal insulation and rubber seals along the edges (Figure 5). Moreover, on the board, there was a connector opened towards the outside environment, the purpose of which was to measure the pressure difference between the room and the outside. Since the ventilation device could affect the pressure difference, the measurements were made with the device removed and with the opening closed (Figure 5). To measure the pressure difference, a CMR-10 micromanometer with an accuracy of ±0.005Δp + 0.5 Pa was used.

2.3. The Control Strategy

Three different control strategies were used: one with constant fan rotational speed for comparison and two proposed strategies using input from sensors.

• Strategy 1—constant rotational speed

This control strategy (Figure 6) is used in case of commercially available devices. The fan speed is constant, set to achieve desired airflow through the device with no pressure difference between inside and outside environment. The airflow rate can be selected by the user from among predefined values.

During the measurements, the fan speed was set so that the airflow in both directions was equal to 25 m³/h, calibrated when the device was not mounted in the wall, and the pressure on both ends was identical. The direction change time was set to 60 s. These parameters are identical as in the RHU-WALL unit from which the accumulative heat exchanger was taken.

• Strategy 2—fan speed control based on airflow rate

This control strategy (Figure 7) makes the fan speed dependent on the readings of an airflow rate sensor. At first, the fan is started at the speed necessary to achieve the desired airflow with no pressure difference. Then, every 5 s, the regulator checks if the airflow measured by the turbine sensor is correct and slightly adjusts the fan speed. The working of this regulator is similar to the integral part of a PID (proportional–integral–derivative) controller. At the end of the supply and exhaust cycle, the current value of the signal controlling the fan is saved to memory. When the next cycle starts, the value of the signal at the end of the previous cycle of the same type is loaded. For example, at the start of the supply cycle, the fan speed is identical to the speed at the end of the previous supply cycle. Since the flowmeter turbine needs some time to stop and start rotating in the opposite direction, the first ten seconds after the airflow direction changes are not taken into account in the process of changing of the fan speed.

• Strategy 3—cycle length control based on temperature readings

In this control strategy (Figure 8), the fan speed is constant, just like in strategy 1. But, the length of the cycle is based on the readings of the temperature of air when it leaves the heat exchanger. The idea is that even if the mass flow rate of the air is slightly different in both cycles, it is possible to shorten the cycle where the airflow rate is higher and extend the cycle where it is lower, so the total mass of the air passing the exchanger in a single cycle will be closer to equal. The cycle shifts when the air that passed the exchanger reaches the set temperature, which means that the exchanger absorbed or gave away a certain amount of energy. The advantage of this method may be the fact that it requires a simpler and thus cheaper type of sensors compared to strategy 2. Based on the temperature changes during normal operation of the device [6], the temperature at which the phase changes was set as arithmetic average of inside and outside air temperatures.

2.4. Values Calculated on the Basis of the Measurements

To facilitate comparison of the performance of the device between various control strategies, their working parameters were calculated. In these calculations, the two first exhaust–supply cycles were excluded, as the device could not yet reach its steady-state working parameters.

The following parameters were calculated:

• The average supplied temperature during the supply cycle, averaged directly based on the sensor readouts;
• The average supplied and removed airflow rate, averaged directly based on the sensor readout;
• The volume of air supplied and removed in a single cycle, summarised based on the sensor readouts;
• The supply heat recovery efficiency [6,15], which shows the amount of energy obtained from the accumulative heat exchanger in relation to total amount of energy required to heat up the air to the room temperature. It is calculated using the following Formula (1):

  $$\eta ={\displaystyle \frac{Q_r}{Q_n}}={\displaystyle \frac{t{e}^{\prime }-te}{ti-te}}$$

  (1)

  where:
  Q_r—energy recovered by the device;
  Q_n—energy needed to heat up the air from external temperature to room temperature;
  te—external temperature;
  te′—temperature after the heat recovery device—in this case supply air temperature;
  ti—room temperature.
• The removed-heat-recovery efficiency [6,15], which shows how much energy was stored in the accumulative heat exchanger in an exhaust cycle in relation to the maximum theoretical amount. The maximum amount of energy would be stored if the air leaving the exchanger would be cooled down to outside temperature. It was calculated using the following Formula (2):

  $$\eta ={\displaystyle \frac{Q_r}{Q_l}}={\displaystyle \frac{ti-tx}{ti-te}}$$

  (2)

  where:
  Q_l—maximum amount of energy that can be recovered in exhaust cycle if the air was cooled down in the exchanger to external temperature;
  tx—temperature of air removed from the device to atmosphere.

3. Results

The temperature courses (Figure 9, Figure 10 and Figure 11) were measured at an outside temperature of approximately 5 °C and an inside temperature of about 23 °C (exact values are given in Figure 9, Figure 10 and Figure 11). The pressure difference between the inside and outside with no airflow through the device (Figure 5) was fluctuating, but the mean value from 10 min of measurements was equal to 7.4 Pa. The measurements for all the control strategies were made subsequently to achieve measurements conditions as similar as possible.

The working parameters of the device in all three control strategies as described in Section 2.4. are collected in

Table 1. Working parameters of the device for three control strategies.

Control Strategy	1. Constant Fan Speed	2. Fan Speed Based on Airflow Rate	3. Cycle Length Based on Temperature
Average supplied air temperature, °C	13.9	16.4	16.8
Average supplied air flow rate, m3/h	29.6	24.3	28.0
Average removed air flow rate, m3/h	17.1	24.2	17.0
Volume of air supplied in a single cycle, m3	0.493	0.406	0.380
Volume of air removed in a single cycle, m3	0.285	0.403	0.340
Supply heat recovery efficiency	0.52	0.63	0.65
Exhaust heat recovery efficiency	0.76	0.61	0.66

.

4. Discussion

As Table 1 and Figure 9, Figure 10 and Figure 11 show, in the default working mode with a constant fan speed, the difference between the amount of supplied and removed air is noticeable. Only 58% of supplied air is removed through the device, and the rest is removed by natural ventilation systems in other rooms. This indicates that the ventilation system does not work as intended and confirms the results from other studies [3,6,7,9,13]. The exhaust heat recovery efficiency is very high due to low airflow rate of removed air. But the supply heat recovery efficiency, which represents the actual amount of energy saved, is low, barely exceeding 50%. That is far below the values provided by manufacturers of the devices. Those two disadvantages indicate the need for improvements.

For the second strategy with the fan speed regulated by indication of airflow sensor, the supplied and removed airflow rates in a single cycle were almost identical. This indicates that the automatic regulation works as intended, and its main criterion was met. Supply heat efficiency was increased by 11% compared to the default working mode, and the supplied-air temperature increased by 2.5 K. The increase in heat recovery efficiency is possible due to the fact that, compared to the default working mode, the amount of air removed in a single cycle is higher. This allows for heating the exchanger to higher temperatures, using its full capacity to store the thermal energy. With the utilisation of the airflow rate sensor, the actual airflow rates were close to the designed values.

For the third control strategy with the temperature sensors regulating the cycle length, the amount of air removed and supplied in a single cycle was almost equal to 90% of the supplied air removed through the device. The total volume of air supplied and removed in a single cycle was slightly lower than for control strategy 2, but still, there is much improvement over the default control strategy. Thanks to the fact that the operation of the device is controlled by temperature, it is possible to achieve the highest supply heat recovery efficiency, equal to 65%, which means 14% improvement over the default working mode. This mode allows more precise estimation of amount of energy stored in the exchanger, resulting in better use of its full heat capacity.

Considering all the data, we can conclude that both control strategies that use sensors noticeably increase the performance of the device in two ways. First, the supplied and removed airflow rates are closer to declared values, ensuring that the required amount of air is supplied to and removed from the room. Second, due to better control over the amount of exhausted and supplied air, it is possible to achieve higher heat recovery efficiency and higher temperature of supplied air in cold seasons. This allows for saving energy and assures that the ventilation system works as intended, regardless of weather conditions influencing the pressure difference between the room and the outside.

The comparison of the two control methods using sensors shows that there is not much difference in terms of the performance of the devices. In both cases, the main criterion of the automatic regulation is met—the amounts of removed and supplied air are close to equal. However, for the strategy 2, where the regulator directly measures the airflow rate, the amount of supplied air was closer to the designed value of 25 m³/h. On the other hand, for the control strategy 2, fan speed had to be greater in the exhaust cycle compared to the default working mode, which may generate noise in case of large pressure differences between the room and the outside.

The changes in operation of the sensor over time should be the subject to future research. It can be expected that the characteristics of the temperature sensors used in control strategy 3 would not change in time. However, the turbine flowmeter may require recalibration after few years of operation due to wear of the bearings or accumulation of ash.

The acoustic measurements were not part of this research, but it was noticeable that the test station was quieter than the commercially available device working with the same airflow rate. This may be caused by the fact that two fans were used instead of one reversible fan. In a fan that works in one way, the blades can be asymmetrically curved, which results in better performance and lower noise emission, which is an important factor in a decentralised ventilation unit. This should be the topic of a future research.

This research was made in real working conditions with no control over external temperature and pressure difference between the inside and outside of the buildings. While it provides valuable data to confirm that automatic regulation improves performance of the device, in the future, it would be useful to test it under fully controlled laboratory conditions. This would allow us to determine how the device reacts to more extreme conditions, such as very low temperatures or strong winds.

5. Conclusions

• It is possible to increase the performance of a single-duct periodic ventilation device with accumulative heat exchangers using control strategies utilising simple and cheap sensors.
• For both control strategies using sensors, the heat recovery efficiency increased by more than 10% compared to the default operating mode, and the amounts of air supplied and exhausted in a single cycle were much more balanced. The average supplied air temperature was increased by about 3 K compared to the default working mode.
• Both control strategies using sensors produced similar results. The average supplied air temperature was higher by 0.4 K for strategy 3 compared to strategy 2. The difference in heat recovery efficiency was equal to 5%. The largest difference was observed in case of the volume of supplied and removed air. For strategy 2, these values were equal to 0.4 m³ per cycle. In case of strategy 3, slightly less air was removed and supplied—the amount of supplied air was 0.34 m³ per cycle.

6. Patents

Polish patent PL243707—Urządzenie do pulsacyjnej wentylacji pomieszczeń z ciągłą regulacją strumienia objętości powietrza (Device for pulsatile ventilation of rooms with continuous regulation of the air volume flow).