Printer Friendly

Experimental Study and Simulation of a Thermosiphon Defrosting Technique for Air-Source Heat Pumps.


Air-source heat pumps are easy to install and offer low running costs for space heating. The main drawback of this type of heat pump is frost formation at the air evaporator (Xia et al. 2006) (Huang et al. 2009) (Shao et al. 2009) during winter period. Under low ambient temperatures, water vapor contained in the humid air turns to frost on the fins of the air coil. Frost increases thermal resistance and pressure drops on the air side flow. Defrosting is usually carried out either by reversing the heat pump cycle or injecting hot gas from the discharge line to the frosted evaporator. Because these classic defrosting techniques provoke a break in the heat production and heat losses, the seasonal system performance strongly decreases when heating demands are at their highest.

A two-phase thermosiphon defrosting technique is proposed through the design of a heat pump for simultaneous heating and cooling (HPS) (Byrne et al. 2009, Byrne et al. 2011a and Byrne et al. 2011b). A previous paper assesses the impact of the defrosting technique on the heat pump performance during a heating sequence (Byrne et al. 2011c). Coefficient of performance was increased by 12% and exergetic efficiency, by 18%. The HPS can carry out space heating, space cooling and domestic hot water (DHW) production for hotels and small office or residential buildings. On the low pressure side of refrigeration plants using several evaporators, thermosiphons are usually cancelled out by non-return valves because they represent risks of refrigerant trapping in the circuit. The HPS takes advantage of this phenomenon to run an "automatic" defrosting without stopping the heat production. Thermosiphons are known to be efficient means of heat transfer (Lee et al. 2009) (Hakeem et al. 2008) (Dobson 1998). This article first presents the HPS concept and its defrosting technique. Then the thermosiphon is tested using a metrology of temperature and pressure sensors and infrared video recording. Finally a simulation study is carried out using a numerical model developed thanks to the experimental results. This simulation study gives paths to optimize the two-phase thermosiphon for defrosting purpose.


HPS concept

The HPS sectors are hotels or glass-fronted buildings in which the proportion of simultaneous demands in heating and cooling is high. This situation occurs in mid-season (spring and autumn) for north-south oriented buildings in which rooms facing north need heating and rooms facing south need cooling. Another situation is encountered in summer when cooling and domestic hot water demands are simultaneous. To satisfy these fluctuating thermal demands, the HPS can operate under three modes:

- heating mode: corresponding to a classic air-source heat pump operation with an air evaporator and a water condenser,

- cooling mode: corresponding to a chiller operation with a water evaporator and an air condenser,

- simultaneous mode: producing hot and chilled water simultaneously.

The HPS prototype (figure 1) produces hot and chilled water using plate heat exchangers. A balancing air coil works either as a condenser for heat rejection in a cooling mode or as an evaporator for heat suction in a heating mode. The air evaporator and the air condenser are never used at the same time. These functions have been assembled in the same three-fluid air coil (air, high pressure refrigerant and low pressure refrigerant) in order to decrease the finned surface area compared to separate air condenser and evaporator. When the tubes of the air evaporator are used the surface of the fins near the tubes of the air condenser are also used and vice versa. A subcooler is connected to the cold water loop to carry out a short-time heat storage during winter sequences. Depending on the operating mode, the electric components (compressor, fan and electronic valves named Evr) are managed automatically by a programmable controller or manually by the operator. The thermostatic expansion valves are named TEV1 (connected to the water evaporator) and TEV2 (connected to the air evaporator). Non-return valves named Nrv1 and Nrv2 are placed at the outlets of the air and water condensers to avoid refrigerant trapping in the condensers.

A pressure control system ensures that condensation is completed in the condenser (and does not finish in the subcooler). Moreover it is able to control the condensation temperature and thus the heating capacity. A liquid receiver is placed on the liquid line. It is connected to the compressor discharge line and the inlet of the air evaporator by copper tubes of smaller diameter on which electronic on-off valves are placed (EvrHP and EvrLP in figure 1). The high pressure control system indirectly controls the volume of liquid in the receiver. The volume of liquid in the different condensers depends on the operating mode. If the high pressure is below the set point, the electronic valve EvrHP is opened by the controller. The receiver is filled up with gas coming from the compressor discharge line at a pressure higher than the pressure in the receiver until the set pressure is reached. The gas entering the receiver drives the liquid towards the evaporator. The non-return valve closes because pressure becomes higher at the outlet than at the inlet. The subcooling heat exchanger and the bottom part of the water condenser are filled up with more liquid until the appropriate level of liquid is reached. If however the chosen mode is the cooling mode, the condenser becomes the air heat exchanger. The set point for pressure is then the lowest possible. The pressure is reduced by driving the vapor out of the top part of the receiver towards the inlet of the air evaporator. The refrigerant in a liquid phase is sucked out of the water condenser and the subcooler and enters the receiver. The operation of the control system depends upon a special liquid receiver being designed sufficiently high and narrow with the main objective to enhance temperature stratification and to limit as much as possible thermal transfer between the gas and the liquid. When the gas is injected, part of it condenses. When gas is rejected to the low pressure, part of the liquid evaporates. Although these phenomena can reduce the efficiency of the liquid variation in the receiver, it stabilizes the control system. The receiver is also thermally insulated to reduce the heat transfer towards the ambience. The chosen refrigerant is R407C.

Thermosiphon defrosting

The thermosiphon circuit involves a water evaporator linked to a heat source, a frosted air evaporator and the compressor. Firstly some energy from the air-source heat pump cycle has to be stored. The principle is here to recover energy by subcooling the refrigerant after condensation using a supplementary heat exchanger and to store this energy in a water tank. The energy recovery is carried out during a classic air-source heat pump operation involving a water condenser and an air evaporator. Under low ambient temperatures, the air evaporator gets frosted. Before reaching a dramatic level of frosting, a second operating mode is engaged. Evaporation now takes place in the water evaporation. The mode change is done by opening Evr1 and closing Evr2 simultaneously. The water tank, being also connected to the water evaporator, is now the heat source of the heat pump cycle and the air evaporator is free for defrosting. Refrigerant vapor coming out from the water evaporator is hotter than 0 [degrees]C whereas the frost on the air evaporator fins imposes the vapor temperature inside the air evaporator tubes. The temperature difference provokes a vapor pressure difference and a thermosiphon forms. Some amount of refrigerant vapor migrates from the water evaporator to the air evaporator. This amount of vapor condenses inside the air evaporator tubes. The condensation energy is transferred to the frost layer that melts. The refrigerant liquid returns back to the suction accumulator and then to the compressor by gravity.


The objective of the experimental study is to better understand and characterize the thermosiphon defrosting technique. The protocol consists in running the HPS in a heating mode during 30 minutes (step 1) then switch to evaporation on water in a simultaneous mode (step 2). During step 1, the heat source is air and the heat sink is water, energy is also stored in a water tank by recovering the subcooling energy. During step 2, the thermosiphon starts to operate and defrosts the air evaporator.

Experimental set up

The experimental apparatus is shown in figure 2. The thermosiphon circuit is divided into two lines connected to two air heat exchangers in parallel. This configuration was preferred to a one-heat exchanger configuration to improve compactness of the machine. The two lines are symmetric. Two temperatures and one pressure measurements are recorded during the defrosting procedure with a time step of 0.1 s. T1 and T2 correspond to the temperatures measured by type K thermocouples at the bottom and the top of the tube respectively. Pressure is measured using a pressure transducer having an accuracy of [+ or -]1% for linearity, hysteresis, repeatability and offset. Water resulting from condensation of frost is collected in a vase and weighed to obtain a gross evaluation of frost mass.

Experimental measurements

The experimental measurements are reported in figure 3. Temperatures T1 and T2 and the pressure are represented versus time.

The experiment starts at the very end of the heating mode, just before switching to the simultaneous mode. During the first 6 seconds of the graph, the heating mode operates with a low vapor temperature and a low pressure. Then Evr2 (that feeds the air evaporator) closes and Evr1 (that feeds the water evaporator) opens. The heat source is now water from the cold water tank at a temperature higher than ambient air. Therefore pressure increases up to between 4.4 and 5.1 bar in an oscillatory regime. The oscillations are due to the pressure control system and the effect of the thermostatic expansion valve. In the meantime temperature increases inside the thermosiphon pipes up to 15 [degrees]C. After 140-145 s, pressure and temperatures do not increase again. A visual observation allows affirming that the air coil is completely defrosted. Then at time 155 s, T1 decreases drastically down to nearly 0 [degrees]C whereas the evolution of temperature T2 shows a less steep decrease. The temperature difference between the top and the bottom of the tube indicates that liquid is flowing in a free surface flow pattern. At 240 s, the experiment is stopped. The mass of water resulting from condensation of frost is measured 15 minutes later is 382 g.

Infrared observation

The first study can be correlated to infrared observation of the tubes involved in the thermosiphon. The following images (figure 4) were taken from the 240-second video corresponding to the experimental results shown in figure 3. Figures 4a, 4b, 4c and 4d correspond to times 0 s, 100 s, 160 s and 200 s respectively. The tube surface being painted in black, the emissivity parameter of the camera was chosen equal to 0.98. The reflected temperature was set to 20 [degrees]C. The image scale ranged between -10 and 20 [degrees]C. Image (a) shows vapor at a low temperature during a heating mode. When evaporation occurs in the water evaporator, some amount of vapor at a higher temperature flows up the thermosiphon tube to condense and carry out defrosting of the air evaporator (image (b)). At time 160 s (image (c)), a colder zone is observed at the bottom of the tube. The tube temperature becomes progressively homogeneous to reach the stage of image (d). This infrared observation provides us with the following important information.

1. Frost on the air evaporator fins creates a cold region in the refrigeration circuit where vapor condenses and where vapor pressure is low. The thermosiphon is effectively created.

2. In the observed part of the circuit, refrigerant vapor flows up, defrosting occurs and subsequently refrigerant liquid flows down by gravity.

3. Refrigerant liquid does not flow down inside one of the two thermosiphon tubes. The condensed refrigerant must be trapped inside this air evaporator.

Thermosiphon defrosting limitations

The main limitation is liquid trapping inside the air evaporator tubes. The air coil of the HPS prototype is classic technology of finned-tube heat exchangers. The horizontality of the tubes does not act favorably on the liquid flow driven by gravity. Indeed the liquid downstream is even not observed in one of the thermosiphon tubes.

Other experiments with higher levels of frosting were carried out. Air temperature was 2 [degrees]C and air relative humidity 98%. After 2 hours of operation, under this level of frosting the thermosiphon was not able to defrost the air evaporators. The major cause envisaged is refrigerant liquid trapping inside the air evaporator. Other studies have to be carried out to confirm this assumption and to redesign the thermosiphon circuit so that defrosting could be achieved under such levels of frosting.


The aim of the simulation study is to propose improvements regarding the thermosiphon configuration. Following the experimental results, the thermosiphon defrosting seems to be decomposed into two successive phenomena. For the configuration of this prototype, refrigerant vapor firstly flows up and condenses inside the air evaporator then refrigerant liquid subsequently flows down back to the suction line. It is as if the liquid could not return back by gravity because of high vapor velocity. The following numerical model focusses on the vapor upstream. Finally some paths for optimization of the thermosiphon are given.

Vapor upstream

In the first part of the defrosting procedure, vapor flows up the thermosiphon tubes. The phenomenon can be described using equation 1 based on the conservation of momentum. Parameter [beta] is the momentum factor equal to 1 in the present case of turbulent flow. Friction factor [f.sub.v] is given by equation 2 (Bonjour et al. 1997). Within the thermosiphon, the vapor pressure gradient is assumed to be the difference between the experimental measurement of evaporating pressure and the saturation pressure at a dew temperature of 0 [degrees]C for condensation. The volume conservation is also applied to the system. Equation 3 uses the cross section area of the vapor flow [S.sub.v]. The hydraulic diameter of the vapor flow [d.sub.v] is the adjustment parameter of the model to the experimental results. It actually corresponds to an average diameter of the thermosiphon tube having different section areas in proportion to pressure drops. Once the refrigerant vapor has reached the frosted air evaporator, the refrigerant condenses and energy is supposed to be conserved following equation 4. The remaining frost mass is calculated using equation 5.

[partial derivative][P.sub.v]/[partial derivative]x = -[beta] x [[rho].sub.v] x [partial derivative][u.sub.v.sup.2]/[partial derivative]x - 2[f.sub.v] x [[rho].sub.v] x [u.sub.v.sup.2]/[d.sub.v] (1)

[f.sub.v] = 0.078 [Re.sub.v.sup.-0.25] (2)

[m.sub.v] = [u.sub.v] x [S.sub.v] (3)

[delta][n.sub.f] x [L.sub.f] = [m.sub.v] x [L.sub.v] x dt (4)

[m.sub.f] = [m.sub.f0] - [[].integral over (0)][delta][m.sub.f] x dt (5)

Figure 5 shows the evolution of vapor velocity and frost mass depending on the pressure difference between the evaporating pressure and the vapor saturation pressure at the assumed temperature of the frost layer of 0 [degrees]C. When the evaporating temperature is superior to the saturation temperature, the vapor velocity in the upwards direction inside the thermosiphon evolves. When the pressure difference is negative, the vapor velocity is supposed to be zero. Frost mass decreases from the initial value of 382 g to zero within 140 s, which corresponds to the visual observation of what happens on the air coil. The hydraulic vapor diameter [d.sub.v] calculated to adapt the fhermosiphon model to the visual and infrared observations is equal to 0.0313 m.

Paths for optimization

The thermosiphon model is used to identify paths for optimization of the thermosiphon effect. Figure 6 shows the impact of the change in diameter on the mass of frost that could be melted if the other parameters of the system were kept constant. Figure 7 shows the impact of the thermosiphon length on frost mass. An increase of the diameter and a decrease of the tube length could be envisaged to enhance vapor migration towards the frosted air evaporator and the efficiency of the defrosting technique. The increase of the diameter would act immediately. The effect of shortening the thermosiphon tube would be relevant for an important length reduction.

Other paths of optimization regard the liquid flow such as inclining the air coil to prevent refrigerant trapping or implementing a porous media inside the tube surface to improve the refrigerant liquid return. These options could enable the defrosting to be completed for higher levels of frosting such as presented in figure 5. However, they require more development in terms of design and engineering or numerical modeling.


This article presents a heat pump for simultaneous heating and cooling in which a thermosiphon forms to defrost the air evaporator. The thermosiphon was observed visually and by infrared thermography. The phenomenon starts by a migration of vapor driven by a vapor pressure gradient, then the vapor condenses on the cold point of the circuit (frosted evaporator) and finally liquid returns back to the suction line by gravity. In the configuration of our prototype, the thermosiphon does not seem to reach a steady state regime in which the two phases of vapor upstream and liquid downstream would coexist in the tube. More detailed numerical studies have to be carried out to entirely characterize the phenomenon, to optimize the thermosiphon effect for more efficient defrosting and to adapt this defrosting technique to newly designed air-source heat pumps.

[beta]      =  momentum factor
d           =  diameter
l           =  length
f           =  friction factor
m           =  mass
m           =  mass flow rate
P           =  pressure
[rho]       =  density
Re          =  Reynolds number
S           =  cross section area
u           =  velocity
[L.sub.v]   =  latent heat of refrigerant vaporization
[L.sub.f]   =  latent heat of frost liquefaction


f           =  frost
TS          =  thermosiphon
v           =  vapor


Bonjour, J., Lefevre, F., Sartre, V., Bertin, Y., Romestant, C, Ayel, V., Platel, V., Systemes Diphasiques De Controle Thermique, Thermosiphons Et Caloducs, Editions T.I. BE 9 545, Mise a jour de l'article [B 9 545] Caloducs, 1997.

Byrne, P., Miriel, J., Lenat, Y. 2009. Design And Simulation Of A Heat Pump For Simultaneous Heating And Cooling Using HFC Or C02 As A Working Fluid, International journal of Refrigeration, 32: 1711-1723.

Byrne, P, Miriel, J, Lenat, Y. 2011a. Experimental Study Of An Air-Source Heat Pump For Simultaneous Heating And Cooling--Part 1: Basic Concepts And Performance Verification. Applied Energy 88: 1841-1847.

Byrne, P, Miriel, J, Lenat, Y. 2011b. Experimental Study Of An Air-Source Heat Pump For Simultaneous Heating And Cooling--Part 2: Dynamic Behaviour And Two-Phase Thermosiphon Defrosting Technique. Applied Energy 88: 3072-3078

Byrne, P, Miriel,J, Lenat, Y. 2011c. A Two-Phase Thermosiphon Defrosting Technique For Air-Source Heat Pumps, IEA Heat Pump Centre Newsletter, 29(3): 72-82.

Dobson RT. 1998. Simulation Of The Two-Phase Flow In A Thermosyphon Using An Inclined Transparent Tube With The Lower-End Closed And The Upper-End Open, Revue Generale de Thermique, 37: 968-972.

Hakeem M.A., Kamil M., Arman I. 2008. Prediction Of Temperature Profiles Using Artificial Neural Networks In A Vertical Thermosiphon Reboiler, Applied Thermal Engineering, 28: 1572-1579.

Huang, D., Li, Q., Yuan, X. 2009. Comparison Between Hot-Gas Bypass Defrosting And Reverse-Cycle Defrosting Methods On An Air-To-Water Heat Pump, Applied Energy, 86: 1697-1703.

Lee J., Ko J., Kim Y., Jeong S., Sung T, Han Y., Lee J.-P, Jung S. 2009. Experimental Study On The Double-Evaporator Thermosiphon For Cooling HTS (High Temperature Superconductor) System, Cryogenics, Vol. 49, pp. 390-397.

Shao, L.-L., Yang, L., Zhang, C.-L. 2010. Comparison Of Heat Pump Performance Using Fin-And-Tube And MicroChannel Heat Exchangers Under Frost Conditions, Applied Energy, 87: 1187-1197.

Xia, Y., Zhong, Y., Hrnjak, P.S., Jacobi, A.M. 2006. Frost, Defrost, And Refrost And Its Impact On The Air-Side Thermal-Hydraulic Performance Of Louvered-Fin, Flat-Tube Heat Exchangers, International journal of Refrigeration, 29: 1066-1079.

Paul Byrne, PhD

Laurent Serres, PhD

Redouane Ghoubali

Jacques Miriel, PE

Paul Byrne and Laurent Serres are assistant professors, Redouane Ghoubali is PhD student and Jacques Miriel is professor at the University of Rennes1, France--Corresponding author:
COPYRIGHT 2013 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc. (ASHRAE)
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Byrne, Paul; Serres, Laurent; Ghoubali, Redouane; Miriel, Jacques
Publication:ASHRAE Conference Papers
Date:Dec 22, 2013
Previous Article:Design of a Vertical, Closed-Loop, Geothermal Condenser for Domestic Air Conditioning.
Next Article:Development of a Residential Ground-Source Integrated Heat Pump.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters