Robust control of run-around coil heat recovery systems

Run-around coil heat recovery systems are nowadays often used in combination with demand-controlled ventilation. Many existing systems experiences low heat recovery efficiencies due to a poor control of the liquid flow (water and glycol mixture). In this article a more robust control method avoiding the problems with the liquid control is presented.

Keywords: Run-around coil, heat recovery, ventilation, control, glycol

Background

Run-around heat recovery systems have been used for many years and are commonly used in large buildings, e.g. hospitals, shopping centres and office buildings. The heat recovery is based on a liquid circuit connected to coils in the exhaust and supply air streams, see Figure 1.

Figure 1. Run-around heat recovery system.

A run-around heat recovery system allows separately located air streams. This means that contaminants, e.g. particles or gases, in the exhaust air cannot find their way into the supply air. Another advantage is that the air ducts can be placed separated from each other, e.g. on different floors which often is valuable in situations with space limitations. It is also possible to connect more than one exhaust air systems to one supply air system. The system allows also that complementary heat exchangers in the liquid circuit can be used for heating or cooling of the supply air, instead of placing heat exchangers (coils) in the supply air which increases pressure drops and fan energy.

 The system design and the flexibility has also disadvantages. Instead of one step of heat exchange between exhaust and supply air (as in a plate heat exchanger), there are two steps heat exchange, i.e. from exhaust air to liquid and from liquid to supply air. This impairs the overall heat transfer effectiveness. The possibility to add complementary heat exchangers in the liquid circuit and utilize a local heat (or cold) source can be positive (e.g. from an economic point of view) but makes the control more complex and it will reduce the overall heat transfer effectiveness. The liquid flow rate is a crucial parameter of the system. Its importance has grown due to designing for better thermal performance (higher heat transfer effectiveness) and as demand-controlled ventilation has become more common. The liquid flow rate is commonly determined by measuring a pressure drop and controlled by a variable speed pump. The liquid needs to have frost protection (e.g. using glycol) which influences the properties of the liquid and needs to be considered when determining the flow rate.

All together several Swedish facility managers experience low heat transfer effectiveness, often below 50%, in their run-around heat recovery systems. Besides the complexity of the systems, it is common that the design and/or the control of the liquid circuit has some faults. In the following an alternative to the traditional control of a basic run-around heat recovery system without additional heat exchangers in the liquid circuit is presented.

Experimental setup

To enable detailed investigations, we have installed a system designed for 68% heat transfer effectiveness at an air volume flow of 1 m³/s (1000 l/s) in our laboratory. The system is further equipped with a modern control system. The run-around heat recovery system used for experiments in this study is presented in Figure 2.

Figure 2. Schematic drawing of the experimental setup.

Temperature sensors are in the outdoor air, supply air, extract air, exhaust air, the cool liquid side and up- and downstream of the pump on the warm liquid side. The liquid flow rate is determined by a TA Smart 2-way control valve with an integrated ultrasonic flow meter with an uncertainty of ±3 % when used with a water glycol mixture. The system is equipped with VED-driven fans, the airflow rates can be varied from 300 l/s to 1200 l/s, allowing the simulation of a DCV application. Air flow rates are measured with fan integrated pressure drop sensors.

Accordance of the air flow sensors was ensured by short-circuiting the supply and extract air while checking that the sensors measured equal air flow rates. Integrated air and liquid temperature sensors were calibrated against a Dostmann P600 temperature sensor with an uncertainty of ±0.1°C. The exhaust and supply air temperatures represent the temperature upstream the fans (in contrast to the location of the sensors). The main reason to locate the sensors downstream of the fans is to obtain conditions closer to fully mixed. The fan heat dissipation was measured and used to determine the temperatures upstream the fans.

The experiments were carried out with water with 30% glycol as the liquid transferring heat from the extract to the supply air. Supply and exhaust air flows were equal from 300 to 1200 l/s. Outdoor air temperature was 7.0 – 12.4°C, extract air temperature was 14.2 – 26.5°C and liquid flow rate 0.1 – 0.5 l/s.

System control (Balanced air flows)

The traditional way to design a run-around heat recovery system for high efficiency is to have a liquid heat capacity rate equal to the air heat capacity rate, usually for the supply air coil, making sure that Xv in equation 1 equals 1. If the system is operated with a lower air flow than designed for the liquid flow must be decreased to keep the high efficiency, but the tolerance for deviation narrows when aiming for optimal flow.

                                                                                                                 (1)

This means that a system with varying air flow rates needs more accurate liquid flow control which requires more accurate liquid volume flow measurements considering glycol content and type. In the case the liquid flow measurement is based on pressure measurements, e.g. the pressure drops over the supply air coil, it will be wrong unless it is compensated for the properties of glycol.

An alternative control method is based on the traditional control method but instead of measuring flow rate of air and liquid, the alternative control method is based on measuring temperatures and aiming at temperature rise of the air equal to the temperature drop of the liquid, implying that Xt of equation 2 equals 1.

                                                                                                                  (2)

The alternative control is thus based on four temperature sensors and without detailed knowledge about the liquid properties. The same sensors can also be used for supervision. The theory behind the alternative control is presented more in detail in a scientific article [Filipsson et al, 2025].

Experimental results

The control of the system in our laboratory is designed to handle both traditional heat capacity balance control (Xv) and the alterative temperature control (Xt). To verify that the alternative control method gives the same result as the traditional method a number of laboratory tests have been carried out and presented in a master thesis [Lorensu, 2025].

The results presented in the following are related to the simulated conditions in a lecture hall where the airflow changes due to different numbers of persons during different lectures. Here the air flow rate is changed manually in the control which results in a much faster change than in reality. The heat transfer effectiveness (ƞt) is defined according to equation 3.

                                                                                                                    (3)

Figure 3 shows the result using the traditional method (Xv). It shows how the liquid flow is changed related to the changes in air flow (750 – 300 – 1200 - 300 l/s) in order to maintain the heat recovery efficiency.

Figure 3. DCV example with traditional control method (Xv).

Figure 4 shows the result using the alternative temperature control (Xt). It shows how the liquid flow is changed related to the changes in air flow in order to maintain the heat recovery efficiency in a similar way as the traditional control.

Figure 4. DCV example with the alternative control method (Xt).

During the airflow rate change from 300 l/s to 1200 l/s, the liquid flow rate using Xt shows a spike during the first few minutes compared to the liquid flow rate with Xv, primarily due to the manually selected P value in the controller.

Discussion and future research

In this comparison, both methods appear equally effective. However, the key point is that the traditional method requires detailed knowledge of the thermophysical properties of the fluid. While such data is available in our laboratory environment, it is typically not well known in practical, real-world applications.

In future research work, the new controller should be configured to automatically select appropriate PI values to enhance the system’s stability in a fast and smooth manner.

The alternative control method will be implemented and evaluated in existing systems. The influence of unbalanced air flows and complementary heat exchangers in the liquid circuit will be evaluated.

References

Filipsson, P., et al, 2025. Liquid Flow Rate Control in Run-around Heat Recovery Systems. Journal of Science and Technology for the Built Environment.

Lorensu, S., P., 2025. Novel Control of Run-around Heat Recovery System. Master thesis Chalmers University of Technology.