Category Archives: EuroSun2008-11

Optimization of Polymeric Solar Thermal Collectors by Fluid Dynamic Simulations

Steffen Jack1, Michael Kohl1, Axel Mkller2, Karl-Anders Weiss1*

1 Fraunhofer Institute for Solar Energy Systems ISE, Heidenhofstrasse 2, 79110 Freiburg, Germany
2 Dr. Axel Muller — HTCO, Postfach 700203, 79056 Freiburg, Germany
* Corresponding Author, Karl-Anders. Weiss@ise. fraunhofer. de


In order to develop a preferably efficient and durable polymer collector the use of computational fluid dynamics (CFD) and FEM-Simulations opens up a fast and efficient way to determine key parameters and problems. The low intrinsic thermal conductivity of polymeric materials and their limited temperature stability can be partly compensated by the optimization of the design of the collectors. Here numerical simulation tools are utilised to analyse and compare different geometries of absorbers or absorber tubes.

Computational fluid dynamics is used to optimize the collector geometries by Dr. Axel Muller — HTCO. The aim is to develop a layout which assures a homogeneous flow, a maximized contact area between the absorber and the heat transfer fluid and an optimised heat transfer into the fluid. The simulation of different collector and absorber layer geometries discussed in this paper point out the quantifying parameters which allow an efficiency optimisation.

Mechanical stresses in the collectors and channels which are caused by temperature gradients or external loads are calculated with FEM-tools. This topic is important in particular if low-cost polymers like polyolefins shall be used. The material stresses are in acceptable range for the analyzed steady state situations. But the combination of different polymers in an integrated collector with bonded absorber and frame can lead to high deformations due to thermal expansion.

Keywords: Solar Thermal Collectors, Polymers, Fluid Dynamics, FEM-Simulation

1. Introduction

The scarcity of fossil fuels is beyond question — one way to save these resources is to make use of solar thermal energy for domestic hot water. So far, solar thermal collectors mainly consist of glass and metal parts. Not simply substituting materials in existing systems but developing a funda­mentally new design for a polymeric collector is the objective of the research at the Fraunhofer ISE and Dr. Axel Muller — HTCO in the framework of Task 39 of the Solar Heating and Cooling Programme of the IEA.

Since the economic viability of solar collectors is strongly linked to the costs of the system, a decrease of the system costs would lead to a higher market penetration. Key advantages of polymers are abundance, weight reduction and more freedom in design, along with the benefits and cost savings associated with well established manufacturing processes and improved fastening, reduced number of parts, and overall assembly refinements. However, also the probably changed system performance is an important element and may not be forgotten.

The Fraunhofer ISE is currently working on the concept of a fully polymeric collector, in order to consider these elements in an integrated way, as only then the full potential of polymeric materials can be used. Important parameters are of course the absorption of solar radiation and — in comparison to metals — the usually lower thermal conductivity and heat capacity. On the other hand one has to consider the intrinsic stress factors like UV-radiation, high temperatures and mechanical loads because the systems have to reach a service life-time of more than 20 years.

In this early development phase numerical simulation tools are used to analyse these topics. The aim is to develop a collector layout considering production and material boundary conditions with a best possible degree of efficiency. With the help of fluid dynamic simulations, one can calculate e. g. the heat transfer from the absorber layer into the heat transfer fluid and optimize it. So one can realise cost and time savings for engineering and prototype production. Parameter sensitivity studies for absorber channels designs are presented as well as efficiency calculations of entire collectors. Another important topic is the simulation of the behaviour of collectors or parts of them during stagnation conditions. Here the temperature distribution and the maximum temperatures are most interesting in reference to thermo-mechanical stresses which occur due to thermal expansion.

Characterisation of the Optical Properties of Solar Collectors by Ray-tracing Simulations

P. Di Lauro*, S. HeB, S. Rose, M. Rommel

Fraunhofer Institute for Solar Energy Systems (ISE), Department Thermal Systems and Buildings,
Heidenhofstrasse 2, 79110 Freiburg, Germany

* Corresponding Author, paolo. di. lauro@ise. fraunhofer. de


The optical efficiency n0 and the Incidence Angle Modifier (IAM) are very important parameters in terms of the characterisation of the thermal performance of solar collectors. These two parameters give the significant distinguishing information about the optical performance of a solar collector.


The research group “Thermal Collectors and Applications” at the Fraunhofer ISE uses the simulation program OptiCAD for investigations on the optical properties of collectors. OptiCAD is a so called “forward ray-tracer”. Flat-plate collectors, evacuated tubular collectors and also collectors with reflectors may be modelled in OptiCAD. After parameterisation of different components and materials of the collector (cover glass, absorber, reflector etc.) the optical efficiency and the 3D-IAM (Incidence Angle Modifier in three dimensions) can be calculated by ray-tracing. The principle followed is that every simulated ray starting from the light source is calculated, followed during all processes (transmission, reflection, absorption) until it is finally absorbed.

The properties of the collector with respect to diffuse radiation can be calculated from the simulated 3D-IAM assuming an isotropic distribution of diffuse radiation.

Results and Conclusion

Different collectors which were tested at the testing laboratory of the Fraunhofer ISE have been modelled in OptiCAD and their optical parameters have been determined by simulation. In the poster presentation it is shown that the measured and the calculated parameters fit very well. It is discussed how the ray-tracing investigations are a valuable and helpful tool for the development and improvement of solar thermal collectors.

Keywords: solar thermal collector, 3D-IAM, diffuse radiation, ray-tracing

1. Introduction

The optical efficiency no[11] and the Incidence Angle Modifier (IAM) play an important role for the efficient utilisation of solar radiation by means of thermal collectors. Those parameters often are determined at an outdoor test facility, in most cases by using a tracker, or at an indoor solar simulator. But these characteristics now can also be calculated by modelling the collector in a simulation software. For this process the optical properties of the different component’s materials that the collector is composed of are to be displayed. The correct description of the actual optical and geometrical parameters of all of the collector components is important for the quality of the results of such a ray-tracing simulation. Therefore, it makes sense to measure the optical properties like transmission, reflection and absorption of each individual element and accordingly implement them in the ray-tracing program.

The result of a simulation of an emulated collector is the effective transmittance-absorptance products (та)е^ under a particular direction and incidence angle of sun rays. This characteristic

can be simulated for all directions (3D) of incident rays. However, it has to be considered that this factor is not equal to the mathematical product of the two optical properties (transmission and absorption). It additionally takes into account general geometrical effects of the complete collector and the interaction of the components like multiple reflections between them.

The IAM then presents the relation of (та)during solar radiation of a certain direction to the one

under perpendicular solar incidence. In order to finally validate those simulated data by comparing with the measured “optical efficiencies” an assumption for the collector efficiency factor F’ has to be made which is assumed to be constant.

Moreover, it is possible to give a proposition about the acceptance for diffuse radiation of the collector by using the simulated IAM-values for direct radiation and supposition of an isotropic distribution of diffuse irradiation on the collector from the virtual hemisphere.

Ray-tracing simulations of the optical parameters of solar thermal collectors offer the following advantages:

• Determination of the optical properties of a collector that is not yet build

• Variation and improvement of the design of a collector during the construction phase

• Predictions of solar gains by simulation of the IAM for direct solar incidence under any incidence angle

• Identification of the acceptance for diffuse radiation of the modelled collector by deriving a diffuse IAM from the IAM-values for direct radiation (assumption of isotropic distributed diffuse radiation)

The Mathematical Modeling of a Solar Collector’s Absorber

P. Shipkovs1*, M. Vanags1, KXebedeva1, G. Kashkarova1, J. Shipkovs1, V. Barkans2

institute of Physical Energetics, Aizkraukles Street 21, Riga LV-1006, Latvia
2 Latvian Maritime Academy, Kronvalda Boulevard 6, Riga, LV-1206, Latvia
* Corresponding Author, shipkovs@edi. lv


The paper gives the stationary heat conduction process is studied theoretically for flat plate surfaces of the absorber of a solar collector. Based on the mathematical model of a stationary heat conduction process in flat surfaces of a solar collector’s absorber described by the authors in previous works, the geometry of such an absorber is analysed from the viewpoint of the most optimal temperature field in it. Special attention is given to simplification of the mathematical model and to creation of a computer model that would reflect the two-dimensional temperature field in the collector’s absorber. Introducing into the computer model the solar radiation density received from an external data source/storage medium the authors obtain the temperature field in the absorber and show that it is possible to define its optimal geometry if the maximum heat power is found from the cross-sections area, since the main parameters in the mathematically described heat conduction process are geometrical sizes of the absorber. The developed computer model will form a basis for creation of new software intended for this particular new innovative idea relating to the design and making technology of flat plate solar collectors without application of expensive experimental materials. The main conclusion is that the carried out mathematical description can help to find the optimal sizes for the absorber, which, taken for the whole collector, would provide its maximum efficiency.

Keywords: heat flow, solar collector’s absorber, mathematical modeling

1. Introduction

The heat flow in the absorber of a solar collector, which occupies there a definite space and in which this heat spreads by conduction, is now being studied intensively [1-3]. The process can be considered mathematically if we know the temperature at any time and any point of this space. There are cases when it is sufficient to measure the temperature at separate points and to tabulate the data obtained. However for a solar collector it is advantageous to preliminarily analyze the heat conduction thus considering the problem theoretically. Besides, the temperature not always can be measured at all points. The challenge is therefore to obtain the temperature function theoretically, depending on the time and spatial coordinates. Having obtained such a function we can use it further for mathematical modeling. This function will definitely contain some solar collector’s parameters, such as, e. g., the thickness of the absorber’s plate and the distance between the tubes (which are ideally connected by soldering to the absorber), the tube diameter, etc.; by varying these parameters in our mathematical model we can calculate the most optimal temperature condition in the absorber, making it possible to abandon a huge time-consuming and expensive experimental work for determination of the heat transfer from the absorber’s plate to the heat carrier flowing through the tube [1].

The paper presents the mathematical description for the heat conduction on the plane surface of a solar collector’s absorber (further in the text absorber). The collector plate’s cross-section is considered that is perpendicular to the axis of a tube with liquid. The cross-section is conditionally divided into three parts. Having chosen one of them, we will attach to it a corresponding coordinate system.

The temperature [K] on the plate (the OiABC section) is designated with T(x, z), the
temperature in the cylindrical coating — with Т2(г, ф), and the temperature of the liquid — with T3(r,


Measurements of hot water in two multi-family houses

Energy use is since 2005 measured in the apartments in two multi-family houses in Vasteras, Sweden, by the housing company. The momentary power required for hot water is continuously registered in every apartment and the hourly data, which the analysis is based on, are averages of a number of such momentary measurements. Measured data from two years, 2005 and 2006, has been processed and evaluated so far and is presented in this paper, but has also been reported in [7]. Both hot water and electricity demand are measured on an hourly basis.

The two buildings comprise 24 apartments of a total area of 1894 m2, occupied by totally 40 persons (19 and 21 residents respectively). The buildings were constructed in 2001 and are provided with energy efficient facilities. Energy costs for utilities, such as heating, electricity, cold and hot water, are included in the total rent per apartment, which is the most common solution in Sweden. Therefore, the tenants are not encouraged to save energy to the same extent as households in single-family houses or apartments with individual billing. There are further plans to install individual measuring equipment for hot water and electricity in 1 300 apartments. Another 300 apartments have already been measured by the residential company, but this data has not yet been processed.

3. Results and analysis

The results have been analysed in different ways; the modelled and measured hot water demand have been compared on an annual and hourly basis to investigate the agreement and the validity of the model. Furthermore, the variation in measured hot water use on a monthly, weekly and daily basis has been investigated to enable further improvements of the model.

Comparison with a producer design software

The model performance with 2 different airside correlations (Wang et al. [12] for wavy fins and Wang et al. [10] for plain fins) was compared with Type1223new and the producer design software Guntner Product Calculator (GPC) of the company Guntner GmbH. Two staggered tube lay outs available in GPC (HX 1 and HX 2, heat exchanger length of 1.25 m and height 1 m, 10 passes) with wavy fins (corrugation angle = 15°) were considered. For both geometries the heat transfer rate, calculated by the model with the correlation for wavy fins, is about 10% for HX 1 and less than 5% for HX 2 lower than that of the GPC, (Fig. 4). It has to be noted, that transverse tube pitch Pt in HX 1 is out of the validity range of the correlation for wavy fins, extrapolation of the correlation is in general not recommended. The heat transfer rate, calculated by the model for plain fins, is lower than the GPC heat transfer rate too (Fig. 4), which is plausible. However, HX 1 is also out of the validity range of the plain fin correlation, the deviation of this correlation to the GPC for HX 1 is even smaller than that of the wavy fin correlation, probably because of the less complex structure. Unlike these correlations, Type1223new with the Elmahdy and Biggs [5] correlation for plain fins gives higher heat transfer rate values than the GPC. It is, however, unfeasible that plain fins have higher heat transfer coefficient than wavy fins (compare with [20]).

Подпись: Fig. 4 Heat transfer rate calculated by the model (as plain and wavy fins), Type1223new and GPC for different air flow rates and two geometries (left for HX 1, right for HX 2).

The airside pressure drop, determined with the correlations for plain and wavy fins, is significantly lower than that, calculated by the GPC, Fig. 5. Whereas liquid-side pressure drop calculation slightly overestimate the pressure drop calculated by the GPC and is in agreement with the GPC when calculated as smooth tubes.


Fig. 5 Airside pressure drop calculated by the model (as plain and wavy fins) and GPC for different air flow
rates and two geometries (left for HX 1, right for HX 2).

2. Conclusion

A model for fin-and-tube heat exchangers, which is based on empirical heat transfer and flow friction correlations, is presented here. The selected correlations are developed with larger data base and have complete description of the reduction method than the one used in Type1223new. In general one needs to be careful with empirical correlations, especially with complex ones, and one needs to prove simulation results. It is not recommended to extrapolate the correlations for configuration outside of the validity range. If configuration outside of the validity shall be simulated (e. g. for optimization of heat exchanger configuration) it appears to be sensible to use less complex correlations, e. g. Wang et al. [10] instead of Wang et al. [12].


The authors would like to express their gratitude to the Volkswagen Foundation, Germany for the financial support.




minimum flow area



heat transfer rate



frontal area


Reynolds number



tube inside surface area






total airside surface area



heat transfer coefficient



heat capacity rate



fin thickness



collar diameter


heat exchanger effectiveness



tube inner diameter


fin efficiency



tube outer diameter



fin corrugation angle


friction factor






air mass flux based on


ratio of minimum flow area

minimum flow area

to face area




Colburn factor



thermal conductivity


airside inlet


number of tube rows


airside outlet



longitudinal tube pitch


tube inner side



transverse tube pitch




Prandtl number


minimum value


[1] E. Frank, K. Vajen, A. Obozov, V. Borodin (2006): Preheating for a District Heating Net with a Multicomponent Solar Thermal System, Proc. EuroSun 2006, Glasgow

[2] Guntner GmbH (2007), personal communications.

[3] Brandemuehl, M. J., HVAC 2 Toolkit: A Toolkit for Secondary HVAC System Energy Calculations, ASHRAE 629-RP, Joint Center for Energy Management, University of Colorado, 1993

[4] Chillar, R. J, Liesen, R. J., Improvement of the ASHRAE secondary HVAC toolkit simple cooling coil model for simulation, Proceedings of the 1st SimBuild Conference, International Building Performance Simulation Association, 2004

[5] Elmahdy, A. H and Biggs, R. C., Finned tube heat exchanger: correlation of dry surface heat transfer, ASHRAE Transactions Vol. 85, Part 2, 1979

[6] Haaf, S., Warmeubergang in Luftkuhlern, pp. 435-491, in Plank, R., Handbuch der Kaltetechnik,

Springer Verlag, Berlin, 1988

[7] Wang, C. C., Chang, C. T. (1998), Heat and Mass Transfer for Plate Fin-and-Tube Heat Exchangers, with and without Hydrophilic Coating, Int. J. of Heat and Mass Transfer 41, 3109-3120

[8] Wang, C. C., Hsich, Y. C., Chang, C. T and Lin, Y. T. (1997), Performance of Finned Tube Heat Exchangers under Dehumidifying Conditions, J. Heat Transfer 119, 109-117

[9] Wang, C. C., Chi, K.-Y. (2000), Heat transfer and friction characteristics of plain fin-and-tube heat exchangers, Part I: new experimantal data, Int. J. of Heat and Mass Transfer 43 (2000), 2681-2691

[10] Wang, C. C., Chi, K.-Y., Chang, C.-J. (2000), Heat transfer and friction characteristics of plain fin-and — tube heat exchangers, Part II: Correlations, Int. J. of Heat and Mass Transfer 43 (2000), 2693-2700

[11] Wang, C. C., Lin, Y.-T., Lee, C. J. (2000), An airside correlation for plain fin-and-tube heat exchangers in wet conditions, Int. J. of Heat and Mass Transfer 43 (2000), 1869-1872

[12] Wang, C. C., Hwang, Y.-M., Lin, Y.-T. (2002), Empirical correlations for heat transfer and flow friction characteristics of herringbone wavy fin-and-tube heat exchangers, Int. J. of Refrigeration 25, 673-680

[13] Pirompugd, W., Wongwises, S., Wang, C. C. (2006), Simultaneous heat and mass transfer characteristics for wavy fin-and-tube heat exchangers under dehumidifying conditions, Int. J. of Heat and Mass Transfer 49, 132-143

[14] Jacobi, A. M., Park, Y., Tafti, D., Zhang, X. (2001), As assessment of the state of the art and potential design improvements for flat-tube heat exchangers in air conditioning and refrigeration applications — Phase I, Final Report. ARTI 21-CR Program Contract No. 605-20020

[15] Schmidt, Th. E. (1949), Heat transfer calculations for extended surfaces, Refrigerating Engineering, April 1949, 351-357

[16] Perrotin, T. and Clodic, D. (2003), Fin efficiency calculation in enhanced fin-and-tube heat exchangers in dry conditions, Proc. Int. Congress of Refrigeration 2003, Washington, D. C.

[17] Engineering Science Data Unit 86018 with Amendment A, ESDU International plc, London, 1991, pp. 92-107

[18] Shah, R. K. and Seculic’, D. P. (2003), Fundamentals of heat exchanger design, John Wiley & Sons Inc.

[19] Wagner, W. (2001), Stromung und Druckverlust, 5. Auflage, Vogel Buchverlag

[20] Gomini, G, Nonino, C. and Savino, S. (2003), Effect of space ratio and corrugation angle on convection enhancement in wavy channels, Int. J. of Num. Methods for Heat&Fluid Flow Vol. 13, No. 4, 500-519


The simulation tool allows the user to choose different parameters related with the simulation process: time-step, accuracy, etc. The smaller the time-step, the longer the simulation time but the more precise are the results concerning the PCM interaction. This statement has to be taken into account when a PCM simulation is carried out. It was checked in previous simulations that bigger time-steps (to perform faster simulations) hid the operation of the PCM.

The length of the simulation can be chosen by the user going from daily simulations to two years simulation. Four different climates can be selected representing the four most common climates in Europe: Stockholm (Sweden) for a moderate northern climate, Zurich (Switzerland) for a moderate central climate, Barcelona (Spain) for a coast Mediterranean climate with high humidity and temperature in summer but a mild winter, and Madrid (Spain) with a continental Mediterranean climate with high temperature but low humidity in summer. The climate chosen for the simulations carried out was Madrid.

The building simulated is a two 70 m2 storey single family house facing south. Four different buildings can be selected and their differences depend on the energy demand of the building, existing an energy demand of 15, 30, 60 o 100 kWh/m2. Even a 100 kWh/m2 without shading can be chosen in order to simulate cooling demand. The 60 kWh/m2 was the selected house for the simulations performed.

The auxiliary system had a nominal power of 10 kW and the water introduced into the store was heated up to 63 °C (set point). The collectors field is 10 m2 of flat plate collectors facing south with a slope of 45 °C. The storage tank volume is 800 L following the recommendations of W. Weiss [6].

Two different types of simulations were carried out. In the first one, the aim was to observe the influence of some parameters in the result of the simulations. The parameters changed were the different inlets and outlets of the store and the position of the temperature sensors that operates the auxiliary system. Annual simulations were carried out and the results were evaluated for every month
and every week. Temperatures into the store and in every inlet and outlet of the tank as well as mass flow rates were the values checked to observe the influence of the PCM in the system behaviour. Fig. 2 shows the relative position of each sensor into the store.

image010 image011 image012 Подпись: Variable Description Tfluid4 Fluid temperature at sensor 4 Tfluid3 Fluid temperature at sensor 3 Tfluid2 Fluid temperature at sensor 2 Tfluid1 Fluid temperature at sensor 1 Tpcm PCM temperature Tpcm PCM temperature TSA Water temperature from store to auxiliaiy TAS Water temperature from auxiliary to store TSB Water temperature from store to space heating TBS Water temperature from space heating to store TSD Water temperature from store to DHW TXdS Water temperature from DHW heat exchanger to store TSC Water temperature from store to solar collectors TXsS Water temperature from solar collectors heat exchanger to store Tssa1 First on/off auxiliary temperature sensor Tssa Second on/off auxiliary temperature sensor

The second set of simulations was mainly focused on the influence of the PCM. Three different values of the Parea ratio were simulated: 0.25, 0.5 and 0.75. Also the geometry of the module was changed, different lengths of the PCM modules and different diameters were simulated. For a given PCM mass, the thinner the module was, the higher the number of modules into the storage. Also two different kind of PCM were simulated into the store focused in the two different demands. The sodium acetate — graphite compound for the DHW demand and placed at the top of the tank and the RT48 paraffin for the space heating demand and placed in the middle of the tank.


Подпись: TXdS

Fig. 2. Position of the monitored sensors inside the storage tank

More results about case 2

Cases 2A and 2B deal with photovoltaic panels, table 4 shows field production on the whole year. PV field feeds grid network, so electricity generation can be done when vapour compression chiller is not operating. Conservative shading assumptions is taken for all other results. For case 2A, PV field produces more than the whole vapour compression chiller electricity consumption, thus the auxiliaries consumption will be decreased respectively by 2.04, 3.68 and 6.31 kWh/(m2 year) for Paris, Stockholm and Lisbon.


(linear shading)













Table 4. PV energy production [kWh/(net_panel_area year)]

2.3. More results about case 3.

Simulations run for case 3 have been optimised to minimise boiler gas consumption. For each location it results a panel slope, a storage tank volume. A summary of results dedicated to case 3 is presented in table 5. Figures given are only about heating and cooling production consumption.

Gas cons







Solar energy



m2 coll by kWcold





case 3A










coll_net_area/ kWc nominal

































case 3B































Table 5. Case 3 summary

Some details should be given on this table: solar energy is the heat collected by the solar field and feeding storage tank; solar fraction is computed following equation 1. Table 6 proposes primary energy savings by net collector area for case 3A. It can be put in relation with table 4 given for case 2 (table 4 values must be multiplied by 2.5 to have primary energy).

Подпись: (equ. 1)Solar energy

Solar fraction =

(Heating load + Chiller hot water consumption)



Primary energy savings by area of collector





[kWh/(net_coll_area year)]



Table 6. Primary energy savings by collector area

3. Conclusion

In this work, design and energy performance of different kinds of air conditioning system were analysed. A complete building simulation model was developed with parameters found in IEA ECBCS Annex 48 research project. Heating and cooling emission and distribution systems were also defined as well as heat/cold production devices. This simulation has been run in three different locations. The comparison between three cases gives the potential energy savings of two solar technologies in relation to classical air-conditioning. A first analysis of auxiliaries showed that they have a huge weight in the primary energy balance. For case 1, it varies from 60% in Stockholm to 80% in Lisbon. Therefore it shows an important energy savings potential. When using classical vapour compression chiller and no solar energy, cooling cost less primary energy than heating due to COP higher than 2.5 (converting net energy to primary energy for Belgium). If solar energy conversion technology is implemented, a key point is available area for installing panels. In each case (A or B), primary energy savings are higher using PV panels instead of thermally driven chiller (assisted by solar thermal panels). It is another important result of this study. PV panels are directly connected to grid, then solar energy is use even if the building has low needs (e. g. during the weekend). For case 3.B (12 floors), solar fraction is very low, operating this system consumes more energy than classical air-conditioning. A better control can reduce this trend but a much higher solar fraction is required to save primary energy.

Systems costs have not been approached in this study. If this analysis is performed, the comparison should be done between solar and classical air-conditioning. An important point for PV is that, nowadays, electrical energy can be sold by the producer at a very interesting price. Moreover in some countries, such Belgium, kWh photovoltaic are paid directly at the production, so money can be saved even if electricity is not used in the building. Solar thermal energy has no such high financial incentive.


[1] W. Pridasawas (2006). Solar-Driven Refrigeration Systems with Focus on the Ejector Cycle, Doctoral

Thesis, Royal Institute of Technology, KTH, Sweden

[2] TRNSYS simulation studio, Version 16.00.0038 Licensed to Universite of Liege

[3] Stabat P. (2007). IEA48 — Description of Type 1c air-conditioned office buildings for simulation test,.IEA — ECBCS Annex 48 working document

[4] ALESSANDRINI J. M. et al. (2006) Impact de la gestion de l’eclairage et des protections solaires sur la consommation d’energie de batiments de bureaux climatises, Climamed, Lyon, France, 2006

[5] Project PVGIS : PV Estimation Utility tool : http://re. jrc. ec. europa. eu/pvgis/index. htm PVGIS © European Communities, 2001-2008

[6] Water Fired Chiller/Chiller-Heater WFC-S Series: 10, 20 and 30 RT Cooling, http://www. yazakienergy. com/waterfiredperformance. htm

[7] Henning, H.-M. (2008). Solar Cooling Components and Systems — an Overview, proceedings Solar Air­Conditioning international seminar, 11th June 2008, Munich, Germany

[8] U. Speicher (2008). Demand and market development, proceedings Solar Air-Conditioning international seminar, 11th June 2008, Munich, Germany

[9] Henning, H.-M. (2007). Solar-Assisted Air-Conditioning in Buildings, A Handbook for

Planners (Second Revised Edition), Springer-Verlag/Wien.

Simulation of Thermosiphon Solar Hot Water Systems. UsingMatlab/Simulink and Carnot

S. Brandmayr1*, M. Konrad1, W. Zorner1 and V. Hanby[2]

1 Ingolstadt University of Applied Sciences — KompetenzzentrumSolartechnik, Esplanade 10,

85049 Ingolstadt, Germany

2 De Monfort University, Institute of Energy and Sustainable Development, The Gateway,
Leicester LE1 9BH, United Kingdom
* Corresponding Author, sebastian. brandmayr@fh-ingolstadt. de


This paper describes with the R&D activities at Ingolstadt University of Applied Sciences in terms of thermosiphon solar hot water systems. The simulation tool Matlab/Simulink and CARNOT was enhanced by a double mantle heat exchanger storage in order to be able to investigate the behaviour of all kinds of thermosiphon systems in theory. Taking data measured at the university’s thermosiphon testing rig into the simulation models, provides the possibility to improve systems by all relevant parameters without performing additional outdoor tests.

Keywords: simulation, Matlab/Simulink, thermosiphon system, storage, development

1. Introduction

Thermosiphon solar hot water systems have been a subject to R&D activities of the Kompetenzzentrum Solartechnik (Centre of Excellence for Solar Engineering) at Ingolstadt University of Applied Sciences since 2004. After building up a test rig, several thermosiphon systems were tested according to the specifications given in ISO 9459-2 [1]. In addition to that, tests according to methods and procedures developed at Ingolstadt University have been carried out in order to learn about the system’s behaviour under special conditions, e. g. its stagnation behaviour.

In the end of 2007, a R&D project was started which aims at the development of an optimised thermosiphon system based on scientific results. A market analysis carried out beforehand showed that most thermosiphon systems are still developed through trial and error [2]. This project, however, aims at demonstrating a closed development cycle. This cycle includes the analysis of thermosiphon systems in theory, the transfer of the mathematical model into simulation, the design of a prototype based on the simulation results and, eventually, the testing of the prototype in order to maximize the system performance and to achieve validation of the computer model. This validated system model is going to offer the project partner, a manufacturer of solar thermal applications, the possibility of adapting their thermosiphon systems to the customers’ and climatic conditions.

blockset (Conventional And Renewable eNergy systems Optimization Toolbox [4]), which is a tool for the calculation and simulation of the thermal components of heating systems with regard to conventional and regenerative elements, was used. It provides models for heat sources, storage systems, hydraulics and fundamental material calculation as well as the possibility of integrating further models. The models used, except the developed double mantle heat exchanger storage, were validated by Hafner et al [4].

Existing district heating grid

The existing district heating grid with a total grid length of approx. 34 km is powered by a gas-fired power plant (Fig.1) and has an annual heat demand of 156 GWh, the peak heat load is in the amount of 72 MW. The power plant is joined with a water-based thermal storage tank with a total storage volume of 5000 m3 and can store a heat quantity of 250 MWh in the operable temperature range.

Подпись: 250 MWh120’C

ізоот district heating grid

Fig. 1: Existing district heating with a combined gas and steam cogeneration plant

Effect of evaporation heat loss coefficient at the swimming pool

By running the model of the system for swimming pool operation with inputs from meteorological models, we can compare the calculated pool temperatures determined by different recommendations for the evaporation heat loss coefficient using different sources from relevant literatures.

Evaporation causes the dominating part of heat losses at swimming pools. It normally accounts for more than 60% of the total energy losses [6]. The mostly used equation for determining the evaporation heat loss power at the water surface is the following:

Подпись:Подпись: (1)Q = A. h

z—eva pool /leva


Подпись: Qeva A ^pool heva Pv, sat {Tpool} Pv,amb evaporation heat loss power at the swimming pool, W, swimming pool surface area, m2, evaporation heat loss coefficient, W/(m2Pa),

vapour pressure of saturated air directly at swimming pool temperature, Pa, vapour pressure of ambient air, Pa.

The evaporation heat loss coefficient can be determined as

heva = a + b • W>t, (2)


w is the wind speed at the swimming pool, m/s, a, b and n are constants.

For the value of a , b and n different references [10], [11], [12] and [13] contain recommendations (see in Table 1).

Table 1. Recommendations with adequate sources for the constants of evaporative heat loss coefficient.





Height of relevant

wind speed in meters




Richter, 1969 [10]





ISO TC 180/ SC 4 N 140 [11]





Rowher, 1931 [12]





HVAC Handbook, 1987 [13]

0 — on ground level

According to the descriptions in this paper the model have been run using meteorological models. Specifications for the calculations are as follows:

The modelled day is a clear day [3] with number 163 (12 June).

Irradiance, ambient temperature, wind velocity is determined by the model.

Air humidity is fixed to be constant, ф=0,65.

There is no auxiliary heating.

The initial swimming pool temperature is 25°C.


Fig. 6. Comparison of the swimming pool temperatures affected by the evaporation heat loss coefficient.

The biggest difference of the calculated swimming pool temperatures is in the cases of using the recommendations by Richter and by the HVAC Handbook. Namely the difference has a mean value of

0, 76°C, maximum value of 1,25°C, residual value, at the end of the modelled day, of 0,9°C.


[1] I. Farkas, Z. Rendik, International Journal of Ambient Energy, 14, 2 (1993) 59-68.

[2] J. A.Duffie, W. A.Beckman, (1991). Solar Engineering of Thermal Processes, John Wiley and Sons, New York.

[3] Gy. Szabo, Zs. Tarkanyi, (1969). Solar radiation data for the planning in building industry, Institute for Building Sciences, Budapest (in Hungarian).

[4] J. Buzas, I. Farkas, A. Biro, R. Nemeth, Mathematics and Computers in Simulation, 48, 1 (1998) 33-46.

[5] J. Buzas, I. Farkas, The 3rd ISES-Europe Solar Congress (EuroSun 2000), Copenhagen, Denmark, June 19­22, 2000, CD-ROM Proceedings, 9.

[6] E. Hahne, R. Kubler, Solar Energy, 53, 1 (1994) 9-19.

[7] B. Molineaux, B. Lachal, O. Guisan, Solar Energy, 53, 1 (1994) 21-26.

[8] I. Farkas, I. Vajk, Energy and the Environment, I /ed. by B. Frankovic/, Croatian Solar Energy Association, Opatija, October 23-25, 2002., 91-99.

[9] B. Bourges, (1991). European simplified methods for active solar system design. Kluwer Academic Publishers for CEC.

[10] D. Richter (1969). Ein Beitrag zur Bestimmung der Verdunstung von freien Wasserflachen dargestellt am Beispiel des Stechlinsees, Abhandlung des Meteorologischen Dienstes der DDR Nr. 88 (Band XI), Akademie- Verlag, Berlin.

[11] ISO/TC 180/SC 4 N 140, Solar Energy — Heating Systems for Swimming Pools — Design and Installations.

[12] Rowher, United States Department of Agriculture, Tech. Bulletin 271 (1931, December).

[13] HVAC Handbook (1987). Section 20.8.