Hatef Madani, Joachim Claesson, Per Lundqvist

Transkript

1 Dynamic Heat Pump System with Capacity Control Effsys2 Project Final Report By Hatef Madani, Joachim Claesson, Per Lundqvist Division of Applied Thermodynamics and Refrigeration Department of Energy Technology Royal Institute of Technology Stockholm, Sweden 2010

2

3

4

5 Abstract The aim of the project is to develop a family of models able to represent the dynamics of the real heat pump system, i.e. the typical heat pump installations in single family houses. The models should be used to quantitatively evaluate different techniques and strategies for capacity control of heat pumps. First of all, a comprehensive survey was carried out to find the existing methods and strategies for capacity control of heat pump systems. Then development of a generic model is suggested in order to approach the challenge of capacity control in heat pump systems. Then the generic model including several sub-models such as the heat pump unit, building, ground source, thermal storage tank, auxiliary heater, and climate models are developed. The developed model is used for comparative analysis of different control methods and strategies aiming at improvement of the system seasonal performance. Experimental studies and in-situ field measurements are carried out to validate the models and obtain a better understanding about the system behavior. For example, the loss behavior in the variable speed compressor and frequency inverter are analyzed when the compressor speed varies. The results show that the inverter loss increases as the compressor speed is increased, although the inverter loss as the percentage of the total compressor power decreases. The generic model developed and validated at the earlier stages is applied in a wide range of applications such as Making a comparison between the annual performance of an on/off controlled and variable capacity systems; the results show that the variable capacity heat pump system yields a better performance only when the ambient temperature is below the balance point and the auxiliary heater operates; Evaluation of the effect of the mass flow rate of the secondary fluid flowing in the borehole heat exchanger on the overall efficiency of a ground source heat pump; The results show that there is a certain optimum brine mass flow rate which gives a maximum overall system COP at every compressor speed. However, concerning system COP maximization, a constant but carefully-selected brine mass flow rate can still be an appropriate option for the variable capacity heat pump unit. Evaluation of an innovative heat recovery system to which a capacity controlled heat pump is retrofitted; results from the annual modelling i

6 show that by retrofitting a well-sized variable capacity heat pump unit to the system, there is a potential to increase the amount of heat provided by the system up to more than 50%. ii

7 Svenskt abstrakt Syftet med projektet är att utveckla modeller som kan representera dynamiken i verkliga värmepumpar, det vill säga typisk installation av värmepump i enfamiljshus. Modellerna är tänkta att användas för att kvantitativt utvärdera olika tekniker och strategier för kapacitetskontroll av värmepumpar. Inledningsvis i rapporten genomförs en omfattande undersökning för att identifiera befintliga metoder och strategier för kapacitetskontroll av värmepumpsystem. En generisk modell beskrivs sedan som tar hänsyn till karaktäristiken i kapacitetsreglerade värmepumpar. Flertalet undermodeller interagerande med värmepumpmodulsmodellen som behövs för att simuleras ett kapacitetsreglerat värmepumpsystem beskrivs också, såsom byggnaden, värmekällan (borrhål), ackumulatortank, tillsatsenergikälla och klimatmodell. Modellen används sedan i rapporten för att illustrera jämförelser mellan olika kapacitetsstrategier som avser att höja systemets årsvärmefaktor. Både laboratorie- och fält-mätningar har genomförts med avsikt att validera de framtagna modellerna samt att få en bättre förståelse för systemens dynamik. Bland annat har förlusterna över frekvensomriktaren och dess interaktion med elmotorn experimentellt mätts då varvtalet (frekvensen) varieras. Studien visar att förlusterna ökar i absoluta tal då varvtalet ökar (lasten), medan procentuellt minskar förlusterna. Som en sista del har den framtagna modellen använts för att utvärdera ett antal utvalda olika applikationer: Jämförelse mellan årsvärmefaktorn för en on/off-reglerad och en frekvensstyrd värmepumpenhet. Resultaten visar att den frekvensstyrda värmepumpen är mer effektiv då utetemperaturen är lägre än värmepumpen balanstemperatur och tillsatsenergi måste tillföras den on/off-styrda värmepumpen. Utvärdering av inverkan av massflödet i borrhålet och i förångaren på värmepumpssystemets värmefaktor. Resultaten visar att det finns ett optimalt flöde där högsta värmefaktor erhålls. Utvärdering av ett vätskekopplat ventilationsvärmeåtervinningssystem där en kapacitetsreglerad värmepump installeras. Simuleringarna visar att genom att installera en kapacitetsstyrd värmepump i dessa system kan tillförd energi till till-luften öka upp till 50 %. iii

8 Acknowledgements This research is a part of a national research and development program in Sweden called Effsys2. It is a four year program for applied R&D in Refrigeration and Heat Pump Technology financed by Swedish Energy Agency and several national and international industrial partners. Special thanks to Swedish Energy Agency and the following project partners for providing the opportunity to do this project. Project partners: ETM Kylteknik Thermia Climacheck Swedish Heat Pump Association (SVEP) NIBE CTC Wilo Copeland UPV iv

9 Table of contents 1 Introduction Background Aim of the project Methodology Publications Conference papers Journal papers Internal report Results summary Model development Experimental studies First example of the models applications: Comparative analysis between on/off controlled and variable capacity systems The second example of the models applications: use of a variable speed pump in ground heat source of a heat pump system The third example of the models applications: Retrofitting a variable capacity heat pump to a ventilation heat recovery system: modeling and performance analysis Capacity control in heat pump systems: literature review Capacity control by components: literature review Compressor cycling (on/off control) Rapid cycling technique Bowtie compressor technique Multiple compressors technique Variable speed compressor Variable speed pumps or fans on the secondary fluid sides Prognostic climatic control strategy: literature review References Development of the Model Introduction Terminology Modeling Approach The overall objective Modeling philosophy Required model complexity The qualitative model of the system The quantitative model of the system v

10 4.5.1 A variable speed hermetic scroll compressor sub-model The condenser sub-model The evaporator sub-model The expansion valve sub-model Modeling the HP unit (variable and single speed) Modeling the building Modeling of ground heat source (boreholes) Modeling of the storage tank Modeling of pumps Modeling of climatic condition Annual simulation of the whole system Validating the model with the experimental results through in-situ field measurement Use the model to address the questions Conclusion Experimental studies: an example Introduction Methodology Result The heat pump heat capacity and Coefficient Of Performance (COP) The compressor power before and after inverter The inverter loss versus compressor frequency Compressor modeling The compressor total isentropic efficiency Conclusion Nomenclature REFERENCES The First Example of the Models Applications: Comparative Analysis Between On/off Controlled and Variable Capacity Systems Introduction Methodology Sub-model Description The building The on/off controlled (single speed) GSHP unit equipped with an electrical auxiliary heater The variable speed GSHP unit The ground heat source (borehole) The storage tank The liquid pumps Climatic conditions Results from system modeling vi

11 6.4.1 The unit COP in different temperature ranges The annual performance of the system Conclusion References The second example of the models applications: use of a variable speed pump in ground heat source of a heat pump system Introduction Definitions and system boundary Methodology A variable speed hermetic scroll compressor sub-model The condenser sub-model The evaporator sub-model The Expansion valve sub-model The pumps model Results The effect of brine mass flow rate on the pumping power (obtained from modeling) The effect of brine mass flow rate on the heat pump heating capacity (obtained from modeling) The effect of brine mass flow rate on the overall COP The optimum brine mass flow rate in a variable capacity heat pump system Conclusion Nomenclature References The third example of the models applications: Retrofitting a variable capacity heat pump to a ventilation heat recovery system: modelling and performance analysis Introduction Retrofitting a variable speed heat pump to the system Modeling and simulation Description of the sub-models Climate model Heat exchangers models Pumps models Fans models Heat pump unit model Control unit model in TRNSYS The Control unit model in EES Annual simulation of the whole system Modeling results When the ambient temperature is above 18 C vii

12 8.6.2 When the ambient temperature is below 18 C and above 14 C When the ambient temperature is below +14 C and above+11 C When the ambient temperature is below +11 C and above-6 C When the ambient temperature is below -6 C A comparative analysis between system A and B Conclusion References Appendix: List of formula Compressor modeling Evaporator modeling Condenser modeling viii

13 1 Introduction 1. 1 B a c k g r o u n d In order to reach the full thermodynamic potential of heat pump systems, an appropriate capacity control technique must be developed and adjusted for each unique installation. The dynamic interaction among the components of the heat pump system makes system control complicated and complex; changing one parameter in any single component is likely to affect several other components whose behaviors might be difficult to predict. The heat pump system as a whole is affected simultaneously by climatic conditions, the behavior of the heat distribution system (such as floor heating or radiators), the building (light or heavy), the heat source (air or ground), the storage tank (if any), different user behavior etc. Furthermore, on the cycle level, components enabling the thermodynamic cycle (the heat pumping), such as the compressor, condenser, evaporator, and expansion valve also interact in a fairly complex dynamic manner, however typically with much shorter time constants compared to that of the heat source or the building systems. Some of the boundary conditions such as instantaneous solar insolation, or the behavior of the building inhabitants can impose fast changes in the dynamic behavior. The role of the auxiliary heater which is commonly used in an on/off controlled heat pump can also influence the system performance. Taking all these together, it is essential to take all the components and their interaction into consideration when different control strategies are analyzed or compared together. Despite all the studies and experiments already done in this field, there have been some limitations and difficulties hindering the researchers to draw a clear practical conclusion and pragmatic results, from either experiment or modeling of capacity-controlled systems. It is also very difficult to compare and generalize the results from different studies. Some of the reasons are: 1. Inconsistency between the compared systems components (different type of compressors or heat exchangers), 2. Focus solely on the heat pump unit efficiency and neglecting the behavior of other parts of the system, 1

14 3. Lack of knowledge about the losses induced by variable speed compressor due to complex measurement process required, 4. Neglecting the dynamics of the whole system, including heat source and heat sink (residential building in our case). The inconsistency between the compared systems components is typical in experimental comparative analyses. For instance, the seasonal performance factor of a variable speed heat pump system is compared to one of a single speed on/off controlled unit whereas the heat load and heat sink conditions are completely different or, even more commonly, the compressor or refrigerant types are different in these two cases. Obviously, some limitations and obstacles hinder the researchers to examine different control methods on heat pumps with exactly the same specifications such as the same compressor, refrigerant, or building. So, the most appropriate way seems to be modeling the system and applying available experimental data to validate the model. It is then feasible to compare the performance of the different control strategies on the same systems. Another common phenomenon is a strong focus on the super-efficient heat pump unit forgetting the undeniable role that the other parts of the system such as building dynamics, temperature level of the heat distribution system, or heat source characteristics can play to increase the system COP. A complicated and costly improvement of the heat pump unit may easily be offset by a small change in the load side conditions. This issue demonstrates the necessity to look at the heat pump system in a wider sense covering the building, building inhabitants, heat source and control units all as the parts of the system. A common conclusion in some of the investigations is that heat pump systems with variable speed compressor in theory has a great potential to improve heat pump unit COP or SPF; however, because of other losses imposed by the variable speed compressor, the actual outputs are not as good as expected. The new losses introduced can stem from the frequency inverter or changes in volumetric or total efficiencies of compressors due to frequency changes. Obtaining sufficient knowledge about what exactly happens to the compressor efficiencies when the frequency is varied or where the new inefficiencies come from is required in order to understand the advantages of variable speed and minimize the disadvantages. Another often neglected significant characteristic of the system usually is the dynamic behavior of the system; boundary conditions are time dependant and different sub-systems have different time constants. Certain parameters such as outdoor or indoor temperature of buildings change relatively slowly; whereas, other such as solar insolation and tap water use are more rapid. Going into all these details of heat pump unit control needs a very short time step (a few seconds) through the investigation; whereas, the control issues in the whole system in a 2

15 wider sense can be hourly-based. Heat source and heat load are being dynamically altered by boundary conditions such as user behavior or climatic changes. Thermal inertia of the heating system is another significant issue which was only considered by a few investigations before such as (Sakellari et al., 2006). Consequently, carrying out a research considering all these dynamic aspects of the system is a challenge A i m o f t h e p r o j e c t The aim of the project, as mentioned in the project application, is to develop a family of models able to represent the dynamics of the real system, i.e. the typical heat pump installations in single family houses. The models are verified against real systems through instrumentation, measurement and evaluation of at least three different systems. The models will be used to quantitatively evaluate different techniques and strategies for capacity control of heat pumps M e t h o d o l o g y The work on the project started on May 2007 by surveying the existing methods and strategies for capacity control of heat pump systems. Then the modeling phase started by making simple models of heat pump units, multi-family houses, and ground heat source. The complexity of all the models has gradually increased, as it can be seen from the schematics in figure 1. The size of the boxes shown in figure 1 represents the complexity of the models made during the project. The size of the upper boxes in figure 1 represents the heat pump unit model comprising the compressor, condenser, evaporator, and expansion valve. The first heat pump unit model was a black box model based on test data expressed as polynomials based on EN14511 (CEN 2007) or other test standards. The more complex model of the heat pump unit consists of a few parameters representing the behavior of the evaporator, condenser, compressor, and the expansion valve. However, at the latest stages of the heat pump unit modeling, a very detailed model of the heat exchangers or compressor makes the unit model more complex and more accurate, but dependent on further detailed input data. Furthermore, the size of the lower boxes in figure 1 represent the complexity of the heat sink (building in this case) and heat source (such as ground or outside air). At the beginning of modeling, the building model can be as simple as a single zone box whose heat demand is proportional to the ambient temperature. Similarly, the ground heat source model could be as simple as a constant temperature or an equation describing how the temperature in the ground changes over the year, for example on an hourly basis. However, as it is shown in the lower part of figure 1, the complexity of the models has increased over the time making the 3

16 system model, which is a pair of an upper box and a lower box, more detailed and more reliable. Heat Pump unit models HP system models Test data unit Black box 1 Parameter 2 HP Design 3 HP Design 4 HPC 1 HPC 2 HPC 3 HPC 4 time Figure 1. Moving forward by stepwise increased complex ity. Each line represents a possible model-pair Beside the development of the generic model of heat pump system described in chapter 3, experimental studies and in-situ field measurements are carried out in order to have a better understanding about the system behavior, to facilitate the modeling and also to validate the models. Chapter 4 presents some examples of the experimental studies done in order to validate the models and also evaluate the variable speed compressor and inverter loss behavior at different operating conditions. The generic model of a heat pump system, made and validated at the earlier stages, are used to Compare different control techniques and strategies such as variable speed versus single speed compressor, as it can be seen in chapter 5 Evaluate the annual performance of the system equipped with any specific component such as a variable speed pump in the ground heat source (borehole), as it can be seen in chapter 6 4

17 Evaluate the annual performance of the new energy systems equipped with heat pump systems, such as the building heat recovery system equipped with a capacity controlled heat pump which it is presented in chapter 7. Assess the annual performance of the heat pump system with different heat source characteristics (such as different borehole characteristics) or different heat load characteristics (such as different types and sizes of buildings), which is done based on the requests from the project sponsors and is not included in this report. Compare the annual performance of different system configurations for the heat pump systems such as the system with or without the storage tank, which is done based on the requests from the project sponsors and is not included in this report P u b l i c a t i o n s C o n f e r e n c e p a p e r s 1. Madani H., Claesson J., Lundqvist P Variable capacity heat pump systems, modeling and simulation, published in 9th IEA Heat pump conference, Zurich. 2. Madani H., Wallin J., Claesson J., Lundqvist P Ventilation heat recovery with run around coil: System analysis and a study on efficiency improvement Part I published and presented in ASHRAE Region- At-Large conference, Kuwait. 3. Madani H., Wallin J., Claesson J., Lundqvist P Retrofitting a variable capacity heat pump to a ventilation heat recovery system: modelling and performance analysis published and presented in the International Conference on Applied Energy (ICAE), Singapore. 4. Madani H., Acuna J. et al The ground source heat pump: a system analysis with a particular focus on the U-pipe borehole heat exchanger, published in 14th ASME international heat transfer conference, Washington DC., USA. 5. Madani H., Ahmadi N. et al Experimental analysis of a Variable Capacity Heat Pump System Focusing on the Compressor and Inverter Loss Behavior, published in International Refrigeration and Air Conditioning Conference at Purdue, USA. 6. Wallin J., Madani H., Claesson J Ventilation heat recovery with run around coil: System analysis and a study on efficiency improvement 5

18 Part II published and presented in ASHRAE Region-At-Large conference, Kuwait. 7. Wallin J., Madani H., Claesson J Run around coil ventilation heat recovery system: A comparative study between different system configurations published and presented in the International Conference on Applied Energy (ICAE), Singapore J o u r n a l p a p e r s 1. Madani H., Lundqvist P., Claesson J., 2009 Dynamic heat pump with capacity control: the project description published in Rehva journal (the journal of federation of European HVAC association). 2. Madani H., Claesson J., Lundqvist P Capacity control in ground source heat pump systems, Part I:modeling and simulation, submitted to Journal of International Institute of Refrigeration. 3. Madani H., Claesson J., Lundqvist P Capacity control in ground source heat pump systems, Part II: Comparative analysis between on/off controlled and variable capacity systems, submitted to Journal of International Institute of Refrigeration I n t e r n a l r e p o r t 1. Madani H Capacity control in heat pump systems: methods and strategies, submitted to the project partners. 6

19 2 Results summary 2. 1 M o d e l d e v e l o p m e n t The present report suggests a method to approach the challenge of capacity control in Ground Source Heat Pumps (GSHP). The report describes the development of a model of the system which includes several sub-models such as the heat pump unit, building, ground source, thermal storage tank, auxiliary heater, and climate (see figure 2). The developed computer model can be used for comparative analysis of different control methods and strategies aiming at improvement of the system seasonal performance. With this model, on/off controlled and variable capacity GSHPs, with a single speed or variable speed pumps in the systems, can be evaluated in a wide range of operating conditions and more energy efficient methods of the system control can be found. The computer model is developed in the two environments EES and TRNSYS utilizing so-called cosolving technique. Furthermore, the use of the developed model provides the opportunity to some important questions such as: How does the COP (Coefficient Of Performance) of a variable capacity heat pump system change over a year? What are the conditions under which the variable capacity heat pump units may yield a better performance than on/off controlled ones? How does the auxiliary electrical heater elimination in variable capacity heat pump systems influence the yearly seasonal performance factor, SPF, for the heat pump system? How does the relative sizing of the heat pump unit capacity (the magnitude of the balance point temperature 1 ) affect the seasonal performance of the system? 1 The ambient temperature at which the heat capacity of the heat pump unit is equal to the heat demand of the building. 7

20 How do variations in the mass flow rate of brine in the ground heat exchanger and water in the heating system (floor heating or radiators) affect the performance of the system, etc. Figure 2. The qualitative model of the system in second level of complexity. The number beside each component represents the section number in the present chapter which is devoted to the specific component s model 8

21 Inverter loss(% of total compressor power) 2. 2 E x p e r i m e n t a l s t u d i e s Furthermore, the present report shows an example of how the experimental studies are used to obtain more knowledge about the behavior of the capacity controlled heat pump systems, especially loss behavior of the system components at different operating condition. In the present examples, the loss behavior in the variable speed compressor and frequency inverter are analyzed when the compressor speed varies. The examples provide an understanding how changing the compressor speed influences the frequency inverter losses and the total isentropic efficiency of the compressor; It would then be possible to evaluate the impact of the inverter and compressor loss on the annual performance of the heap pump system. The experimental results show that increasing the compressor speed reduces the heat pump COP up to 30%. The inverter loss increases as the compressor speed is increased, although the inverter loss as the percentage of the total compressor power decreases. Figure 3 shows that the ratio of the inverter loss to the total compressor power (the power measured before the inverter) decreases from 11% to 3% when the frequency changes from 30 to 100 Hz when R134a is used as the refrigerant. The trend for the system with R407C is very similar ºC/26 ºC as the source/load side temps 1.5 ºC/21.5 ºC as the source/load side temps Compressor frequency (Hz) Fig. 3. The measured inverter loss in % of the total compressor power when the frequency varies between 30Hz and 100Hz: the heat pump with R134a 9

22 Total isentropic efficiency of compressor Regarding the total isentropic efficiency of the variable speed compressor, as it is shown in figure 4, it varies about 8% when the compressor frequency changes from 30Hz to 90Hz. Furthermore, figure 4 shows that the highest total isentropic efficiency of compressor occurs when the compressor frequency is close to 50Hz ºC/26 ºC as the source/load side temps 5 ºC/30 ºC as the source/load side temps Compressor frequency (Hz) Figure 4.The variation in the total isentropic efficiency of the compressor based on the changes in the compressor frequency when R407C is used as the refrigerant : experimental results 2. 3 F i r s t e x a m p l e o f t h e m o d e l s a p p l i c a t i o n s : C o m p a r a t i v e a n a l y s i s b e t w e e n o n / o f f c o n t r o l l e d a n d v a r i a b l e c a p a c i t y s y s t e m s The present report compares the annual performance of the on/off controlled and inverter-driven variable capacity heat pump systems, using modeling and simulation via the generic model which was already developed. The liquid pumps used in both systems are constant speed pumps. The results from the annual modeling for both systems (on/off controlled and variable capacity systems) indicate: When the ambient temperature is above the balance point of the single speed HP unit, the on/off controlled system yields a better performance 10

23 than the variable speed unit. This is mainly because the inverter and compressor losses in the variable speed compressor. When the ambient temperature is below the balance point of the single speed HP unit, the variable speed system has higher efficiency than the on/off controlled unit. The poor performance of the electrical auxiliary heater makes the on/off controlled system less efficient in the low ambient temperatures. The energy use by liquid pumps in the variable speed GSHP system can be up to 10% higher compared to the single speed system. Compared to on/off controlled system, the variable speed system operates for a longer time during a year which makes the pump in the ground heat exchanger work for a longer time as well. When the on/off controlled HP system is designed to cover about 90% of the annual energy demand and the auxiliary heater takes care of the rest, the seasonal performance factor of the system may be improved about 10% by switching to a variable speed HP system. However, if the on/off controlled HP system is designed to cover much more than 90% of the annual energy demand, which is common nowadays, the variable speed heat pump loses its advantage and on/off controlled system can be a better alternative (cheaper with more or less the same annual performance with variable capacity one) T h e s e c o n d e x a m p l e o f t h e m o d e l s a p p l i c a t i o n s : u s e o f a v a r i a b l e s p e e d p u m p i n g r o u n d h e a t s o u r c e o f a h e a t p u m p s y s t e m The mass flow rate of the secondary fluid flowing in the borehole heat exchanger of a ground source heat pump is an influential system parameter whose variation can influence the pumping power, efficiency of the pump, heat pump heat capacity, and above all, the system Overall Coefficient Of Performance (COP). The 11

24 Heat pump heat capacity (kw) developed model of the heat pump system is used in order to evaluate these influences. Figure 5 shows an example of the variation of heat pump heating capacity (kw) when the brine mass flow rate (kg/s) and compressor speed change. In the figure 5, the X axis represents the brine mass flow rate (kg/s) and the Y axis shows the heat pump heating capacity (kw). Each curve represents one compressor speed (Hz) which varies from 30 Hz to 75 Hz. The inlet temperature to the evaporator and condenser are kept constant at 0 C and 35 C. As shown in figure 5, at low compressor frequency, increasing the brine mass flow rate yields a slight increase in the heating capacity. For example, when the compressor frequency is 30Hz, increasing the brine mass flow rate from 0.38 kg/s to 0.8 kg/s increases the heating capacity about 9%; whereas, at high compressor frequencies, the brine mass flow rate can significantly influence the heat capacity. For instance, when the compressor frequency is 75 Hz, the heating capacity rises up to 17% when the brine mass flow rate changes from 0.38 kg/sec to 0.8 kg/s. 14 f=30 f=40 f=50 f=60 f= Brine mass flow rate (kg/s) Figure 5. Variation of the heat pump heat capacity (kw) based on the changes in the brine mass flow rate: different colors show different compressor frequency (Hz) Figure 6 shows the variation of overall COP (considering both compressor and pump power) when the brine mass flow rate (kg/s) and compressor speed change. The X axis represents the brine mass flow rate (kg/s) and the Y axis 12

25 Overal COP represents the overall coefficient of performance. Each curve represents one compressor speed (Hz) which varies from 30 Hz to 75 Hz. As shown in the figure 6, for a given compressor frequency, f, there is a certain brine mass flow rate, which yields the maximum overall system COP. The lower the compressor speed, the lower. For example, when the compressor speed varies from 30 Hz to 75 Hz, increases from 0.41 kg/s to 0.59 kg/s. 3.4 f=30 f=40 f=50 f=60 f= Brine mass flow rate (kg/sec) Figure 6. Variation of the overall COP of the system based on the changes in the brine mass flow rate: different colors show different compressor frequency (Hz) Consequently, in order to maximize the overall COP of the system at every operating condition, the brine mass flow rate should be changed proportionally to the compressor speed; however, the results from the present study showed if the brine mass flow rate is selected to be constantly equal to the optimum brine mass flow rate at 50 Hz (close to the average between the maximum and minimum compressor speed), the overall COP at all the operating conditions would be very close to the optimum values. Therefore, a single speed pump but with a very carefully- selected brine mass flow rate would be still an appropriate option for a variable speed heat pump systems, if COP maximization is the main concern. Concerning the heat capacity maximization, it is found that the brine mass flow rate can play a significant role in increasing the heat pump heat capacity whenever 13

26 the compressor operates at high speed; so if the designed compressor is not able to cover the peak heat demand of the building, higher brine mass flow rate can be used to increase the heat capacity of the heat pump system T h e t h i r d e x a m p l e o f t h e m o d e l s a p p l i c a t i o n s : R e t r o f i t t i n g a v a r i a b l e c a p a c i t y h e a t p u m p t o a v e n t i l a t i o n h e a t r e c o v e r y s y s t e m : m o d e l i n g a n d p e r f o r m a n c e a n a l y s i s Heat recovery from the building ventilation system is commonly used in countries with cold climates such as Sweden. There are different methods to perform the heat recovery from the ventilation system and run-around coil system is known as one of the typical systems for this application. Run around coils are often used in buildings where contamination of the supply air is a major issue, for example hospitals. In order to improve the system efficiency, a variable capacity heat pump is retrofitted to a conventional run around coil ventilation heat recovery system. The present study aims at evaluating the annual performance of both existing run around coil system and the system to which a variable capacity heat pump is retrofitted. The capacity controlled heat pump model which was already developed is used for dynamic modeling of both systems over a year. Figure 7 shows an example of the results from annual modeling of both conventional run around coil system and the new system to which a heat pump is retrofitted. Figure 7 shows how much heat could be recovered from the traditional run around coil system and the new system in which the variable speed heat pump retrofitted. As may be seen from figure 7, by retrofitting a variable speed heat pump to the run around coil system, the total amount of energy provided by the heat recovery system can be increased up to 58%. The run around coil system can recover 47% of the total ventilation heat demand for Stockholm s case. As it can be observed from figure 7, this value can increase to 80% just by retrofitting a variable speed heat pump unit to the old system. 14

27 Energy (MWh) Heat provided by run around coil Heat provided by the system with heat pump retrofitted Total ventilation heat demand Berlin Stockholm Figure 7. The total heat provided annually by both conventional run around coil system and the new system to which a heat pump is retrofitted, beside the total ventilation energy demand in Stockholm and Berlin (MWh) Consequently, by considering the energy used by the compressor of the heat pump, it can be concluded that the new system to which the heat pump retrofitted looks promising from the economic point of view and it can lead to save a large amount of energy use and money annually. 15

28 3 Capacity control in heat pump systems: literature review There are several possible methods to provide capacity control of heat pump systems. Capacity control can be divided into three main categories: capacity control by components (variable speed compressor or pumps for example) capacity control by advanced control algorithms (prognostic climate control using building mass for example) cycle design for capacity control (for example changing the system configuration) Some examples of the different capacity control techniques and strategies under these three categories are presented as following: 3. 1 C a p a c i t y c o n t r o l b y c o m p o n e n t s : l i t e r a t u r e r e v i e w C o m p r e s s o r c y c l i n g ( o n / o f f c o n t r o l ) The conventional on/off control is one of the most common methods for capacity control in heat pump systems. In on/off controlled heat pump (HP) units, the compressors are switched on or off by aid of a control algorithm that estimates the discrepancy between building heat demand and heat pump capacity. Figure 8 shows an example of how the heat pump outlet temperature (supply temperature to the building or the storage tank) and compressor work varies in an on/off controlled heat pump unit over about 10 hours (in-situ field measurements, Stockholm, 2007). The x axis represents the measurement time step which is one minute during the on time and 5 minutes during the off time. The left y axis represents the actual and the required output temperature from the heat pump unit and the right y axis also represents the compressor power input. The

29 variation of heat pump outlet temperature (from 35.3 C to 46.4 C) comes from the frequent compressor cycling aiming at matching of the unit heat capacity and the heat demand; whereas, the required unit output temperature which can be calculated based on the ambient temperature variation over this period is almost constant (only varying from 39.4 C to 40.5 C). In this example, the average output temperature of the unit is 4.6K higher than the average of the required output temperature. Fahlen (2004) and Karlsson (2007) suggest that this temperature discrepancy may lead to a 2% COP reduction/k; so system COP is reduced about 9% in this case HP output temp(c) compressor work (kw) time step HP output temp- on/off controlled HP output temp-required compresspr work (kw) 0 Figure 8. Dynamic output and the required temperature of heating water and the compressor work in an on/off controlled heat pump unit. x axis represents the measurement time which is one minute during the on time and 5 minutes during the off time. So off time in the figure should be multiplied by five. Furthermore, frequent switching of a compressor on and off introduces so called "cycling losses". The reason for these losses are the migration of the refrigerant after cut-off and the need to heat up the thermal mass of the system - mainly the heat exchangers and the compressor at the start up after an off period. Henderson (2000) stated that in air/air heat pumps, the losses due to the thermal inertia of the compressor and the refrigerant migration during the off-period accounted for 4-11% reduction in efficiency. Tassou and Votsis (1992) also indicated that cycling losses are responsible for 11% COP reduction in air-to-water heat pumps; however, Karlsson (2003) and Bergman (1985) mentioned that the cycling losses can be negligible due to short duration of start-up compared to the total cycle 17

30 time and decrease in the refrigerant migration by use of thermostatic expansion valve which, if working correctly, will prevent refrigerant migration from condenser to evaporator. Furthermore, according to IEA HPP Annex 28 (Wemhoener and Afjei 2006), the cycling losses of the compressor are neglected in the calculation of the Seasonal Performance Factor (SPF) of the system. If the heat pump system is designed to cover 100% of the annual load and consequently eliminate the supplementary heat in the system, an intermittent controlled heat pump unit will always be oversized and it will have too many on/off cycles increasing the importance of cycling losses. Consequently, in Sweden and the rest of Scandinavia the heat pumps (except exhaust air heat pumps) are generally designed to cover approximately 60% of the load at the DOT. The heat pump will then cover about 90% of the annual heating demand (Forsen 2002). The electrical heater is usually used as the supplementary heating system to take care of the rest of heating demand. Use of electrical supplementary heater in the heat pump system results in two major problems: firstly, these electrical heaters mostly work during the coldest hours of a year when the power plants and electricity production companies are suffering from the annual peak loads. The second problem comes from high price of electricity compared to relatively very low price of other fuels for heating which makes the end-users to pay more. In an on/off controlled heat pump system, the stand-by power consumption is another significant point affecting the system COP. During the off cycle, pumps or fans in the secondary fluid sides in both condenser and evaporator should be turned off. Otherwise, the parasitic energy consumption leads to reduction in system COP. Riveire et al. (2004) shows experimentally how these parasitic energy consumption can deteriorate the system efficiency in an air-to-water heat pump system R a p i d c y c l i n g t e c h n i q u e Rapid cycling is another type of capacity control methods using the compressor cycling. The difference between conventional compressor cycling and rapid cycling is that in rapid cycling, the off period of the cycle is shorter than the time constant 2 of the evaporator (SHERHPA 2005). In this method, cycle period may be on the order of a few seconds to a few minutes. Since the cycles are short, the pressure lift oscillates around a mean that is close to that of variable speed operation. Figure 9 shows schematically the saturation temperatures on the high and 2 The time constant of the evaporator is the time elapsed after the considered parameter, which in this case is the refrigerant mass flow rate through the evaporator, has been risen to 63% or fallen to 37% of the maximum mass flow rate. 18

31 low sides of the system and the on cycle average temperature lifts for each operating condition (M.J. Poort 2006). Figure 9. Temperature lift for different types of operation (Poort 2006) One existing commercial version of rapid cycling concept uses a proprietary mechanism in the compressor to disengage the scrolls while the motor runs unloaded during the off cycle (Hundy 2002). It is called digital scroll compressor. The digital scroll compressor is a compressor capable of modulating capacity from 10% to 100%. A cycle in a digital scroll compressor consists of loaded state and unloaded state provided by the help of a solenoid valve. When the solenoid valve is in its normally closed position, there is no pressure difference across the piston; consequently, the compressor operates at full capacity or in the loaded state (figure 10.a). When the solenoid valve is energized, the piston can be actuated by gas pressure; so the two scroll elements move apart axially about 1 mm and the unloaded state occurs (figure 10.b). 19

32 a) Loaded b) Unloaded Figure 10. The loaded and unloaded state in a digital scroll compressor (Bianchi and Winandy 2007) During the unloaded state which lasts about 10 sec, the compressor motor continues running, but since the scrolls are separated, there is no compression and compressor delivers 0% capacity. The power input drops close to zero during the unloaded period. The compressor is switched between the loaded and unloaded states with a cycle time of typically 20 sec (see Figure 11). The duration of the loaded state within this time determines the capacity. At this capacity modulation method, there must be sufficient thermal inertia allowing almost all systems to experience continuous capacity output from the compressor (Bianchi and Winandy 2007). There is a lack of detailed information about the possible technical problems which might prohibit the small scale heat pump units to take any advantage out of digital scroll technology. 20

33 Figure 11. A simple digital waveform from the controller activates the unloading (Bianchi and Winandy 2007) B o w t i e c o m p r e s s o r t e c h n i q u e A bowtie compressor is a novel refrigeration (or heat pump) compressor which is able to modulate the unit capacity by simple mechanical means. The name, Bowtie, stems from the compressor s two sector-shaped, opposing compression chambers forming a bowtie. It uses a unique off-center-line mechanism so it modulates the unit capacity by changing the piston stroke without changes of the clearance volume for better thermodynamic efficiency (Kim and Groll 2007). The compressor uses an inexpensive fixed speed motor and does not need a rotational change of the crankshaft. Figure 12 shows how a novel-designed bowtie compressor works. As shown in the figure, the crankshaft rotates counter-clockwise, and the piston rod makes the piston reciprocate axially and not linearly. Thus, two compressions take place at the same time. The crankshaft, piston rod, and piston are arranged in such a way that the piston rod is perpendicular to the crankshaft when the piston reaches the top-dead centre (the top schematic of Figure 12). 21

34 22 Figure 12. The diagram of the bowtie compressor design proposed by Kim and Groll 2007 The cylinder is placed in the sliding chamber with mechanical springs and pressurized on both ends by the suction and the discharge gases. The cylinder is able to slide from left to right (or vice versa) by an actuator working based on the difference of the two pressure forces. At compressor start-up, the actuator is off so that the compressor operates with a smaller swept volume. This can lead to a smooth start-up due to the reduced motor torque. When more capacity is needed, some of the discharge gas is bypassed to activate the actuator resulting in movement of the module to the left end, which creates a larger swept volume as depicted in the bottom schematic of figure 12. It is significant to notice that the cylinder movement does not create an extra clearance volume and the clearance volume is kept constant all the times; therefore, the capacity is modulated only by changing the piston stroke (Kim and Groll 2007). For better understanding, it should be mentioned that the suction and discharge valves are positioned here and that inlet and outlet tubes are not shown in the figure 12. By use of the Bowtie compressor, it is possible to modulate the refrigeration or heat pump unit capacity continuously. Although the Bowtie compressor yields lower cost and avoids the inverter losses, it provides a narrower range of capacity modulation (only from 50% to 100%) when compared to a variable speed or digital compressors. Furthermore, it may introduce some new losses such as more leakage due to its more leakage path than the conventional reciprocating compressor. Kim and Groll (2007) investigated the effect of changes of the leakage

35 clearance and they concluded an optimum value for that. They also proposed to use an electronic control valve in order to achieve a continuous capacity modulation which may increase the system complexity, particularly when it is used simultaneously beside an electronic expansion valve in the heat pump system M u l t i p l e c o m p r e s s o r s t e c h n i q u e In this capacity control method, two or more compressors (usually with different capacities) are connected in parallel providing multiple operating points for the system. By alternating between the operating points, the heat output can be adjusted to be as close to the heat demand as possible. It is more conventional to use two compressors, one low capacity and another one high capacity in tandem so having three operating points: one when only low capacity compressor works, one when only high capacity compressor works, and one when both work simultaneously. The operating points can be altered according to control input signal such as the ambient temperature or indoor temperature. By applying this method, the supplementary electrical heater can be avoided since 100% of the annual load can be covered by the unit while all the compressors work simultaneously. Furthermore, two or more compressors in different sizes can provide a variety of options for defrosting in the case of an air-source heat pump system. Defrosting can be performed by reversed cycling with only the smaller compressor, only the bigger compressor or with both together. Flash-Malaspina et al. (2004) made a comparative analysis between two air-towater heat pump systems, one with two dissimilar compressors in tandem and one with a single compressor. They concluded that the system with twin compressor yields higher COP due to decreasing the sleep power consumption and omission of supplementary heater. A multiple compressor system needs a complex oil management system to assure that no single compressor will experience oil shortage V a r i a b l e s p e e d c o m p r e s s o r Generally, in a variable speed heat pump unit, the speed of the inverter-driven compressor is modulated by aid of control algorithm so that the heat output from the heat pump unit offsets the load of the building; consequently, the discrepancy between the heat supply the heat demand (in the building) can be decreased. For example, at periods of low load, the refrigerant mass flow rate reduces to match the load. The lower refrigerant mass-flow-rate leads to lower temperature differences in the heat exchangers and thus lower condensation and higher evaporation pressure. This happens because the mass flow rate has been reduced while the heat-transfer area of the heat exchangers remains constant. This gives more time to the refrigerant to exchange heat with the secondary fluids. Consequently, the temperature difference between the refrigerant and secondary fluids decreases 23

36 forcing the evaporating temperature to rise and the condensing temperature to drop. Capacity control by adding variable speed compressors in heat pump systems has been tackled both theoretically and experimentally by many researchers. For example, (Neale et al. 1981; Lida et al. 1982; Wang et al. 1983; Halozan 1988; Zubair 1989; Sheldrake 1991; Tassou and Qureshi 1996; Vargas & Parise 1995; Qureshi 1998; Claesson & Forsen 2002; Zhao et al. 2003;Karlsson 2007; Parnitzki 2007) analyzed the variable speed heat pump systems. For example, Zhao et al (2003) did a comparison among on/off and variable capacity controls, in addition to two other methods, for ground source heat pumps and concluded that changing the compressor speed is the preferred method. However, Karlsson and Fahlen (2007) concluded from their research that despite improved performance at part load (as was already shown by Tassou and Qureshi,1996), the variable-speed controlled heat pump did not improve the annual efficiency compared to the intermittently operated heat pump. Karlsson (2007) suggested that it is mainly due to the inefficiencies of the inverter and the electrical motor of the compressor and the need for control of pumps used in the heating and ground collector systems (Karlsson 2007). The experimental study by Cuevas and Lebrun (2009) showed that in the variable speed scroll compressor, the inverter efficiency varies between 95% and 98% when the compressor electrical power varies between 1.5 and 6.5 kw. It is also found that the additional electrical motor losses induced by the presence of the inverter are negligible (Cuevas and Lebrun 2009). Koury et al. (2001) presented both a steady state and a transient model of a waterwater type of variable speed refrigeration system. These modeling results showed how the mass flow rate of the refrigerant in the compressor and the thermostatic expansion valve as well as the degree of superheat changes by variation of compressor speed (Figure 13). By increasing the compressor speed (after 60 sec), the evaporation temperature and pressure decreases. The evaporation pressure reduction leads to the increase in the superheat temperature. Then, by 10% increase in the throttling area of the thermostatic expansion valve (after 300 sec), the mass flow rate of the refrigerant increased at both inlet and exit of the evaporator resulting in superheat temperature reduction. 24

37 Figure 13. The dynamic behavior of the mass flow rate and the superheat in response to a step variation of the compressor rpm and of the throttling section area (Koury et al. 2001). Apera et al. (2006) experimentally showed that in a variable speed heat pump unit with a thermostatic expansion valve, the variation in the superheat temperature is lower compared to the unit with electronic one (Figure 14). Aprea et al. (2006) did a comparison in terms of energy consumption between the electronic valve and the thermostatic valve and it is concluded that in a variable speed refrigeration system, the thermostatic expansion valve allows an energy saving of about 8% in comparison with the solenoid electronic valve in the same operating conditions. Choi et al. (2001) and Yoon et al. (2001) tried to develop an appropriate superheat control method for variable speed heat pumps. Furthermore, due to the fact that the compressor frequency and the superheat temperature cannot be controlled independently and they make interfering loops when these two parameters change simultaneously, Li Hua et al. (2008) presented a decoupling model to eliminate the interfering loops. 25

38 Figure 14. S uperheating with TEV and EEV with the set - point of 0 C and fuzzy control (Aprea et al. 2006) V a r i a b l e s p e e d p u m p s o r f a n s o n t h e s e c o n d a r y f l u i d s i d e s Variable speed pumps or fans can be used in the source/sink sides of the heat pump system in order to optimize the system energy consumption and increase its performance. The flow rates of secondary fluid sides of the heat pump have been analyzed by (He 1996; Granryd 2002; Granryd 2007; Karlsson and Fahlen 2008; Finn et al. 2008). For example, Granryd (2007) and Finn et al. (2008) analyzed two control approaches: finding the optimum secondary fluid mass flow rates which yields either maximum COP or maximum heat capacity. Granryd (2007) concluded that for the case giving maximum COP, much less power is used to operate the pumps in the secondary system, compared to the case of maximum heat capacity. Granryd (2007) suggested a rule of thumb in which the flow velocity in the secondary system can be halved to maximize COP compared to the solution for maximum capacity. Finn et al. (2008) showed that when switching from capacity maximization to COP maximization, the heat pump COP can be improved about 12% with an associated 5% reduction in capacity. Moreover, it is found by Finn et al. 26

39 (2008) that a system with the COP maximization control method can yield 11% higher seasonal performance factor than the one with capacity maximization control algorithm P r o g n o s t i c c l i m a t i c c o n t r o l s t r a t e g y : l i t e r a t u r e r e v i e w Prediction of climatic conditions such as the ambient temperature and solar radiation for the next day (or the next few hours) and regulation of the heat pump control algorithm based on this prediction is another way to control the heat pump systems. If a precise prediction of climatic condition can be available, it is feasible to integrate the daily demand; so reducing the compressor cycling or switching the demand from peak hours to off-peak hours. By predicting the daily solar radiation while considering the lag caused by the thermal inertia of the building, it is possible to reduce the heating capacity of the heat pump unit or turning the unit off to avoid over-heating the building. Bianchi (2005) applied a Model Predictive Control (MPC) algorithm 3 in order to optimize the conventional on/off control of the heat pump with single speed compressor; so they take the daily weather forecast into account and calculate the required heat energy over a fixed on/off period (e.g. two hours) by integrating the required heat flow. The heat pump switched on at the beginning of each period and remains on until the required heating energy has been delivered. Bianchi (2005) compared this new method with conventional relay-type controller from different points of view such as the energy costs, the energy consumption, the COP of the heat pump, as well as the mean room temperature. The results of the two-week simulation showed that the new predictive model was able to keep the same mean value and standard deviation of the room temperature as the conventional controller, while requiring nearly the same amount of electrical energy. However, it placed considerably more heat energy in the low-tariff period. The predicted reduction in heating cost by switching the heating load from peak hours and high tariff period to low tariff periods is about 10%. Sakellari et al. (2006) also tried to utilize a daily forecast of solar radiation to avoid over-heating the building during a sunny day. The system s control is set so that the heat pump is turned off when the instantaneous solar radiation to a horizontal surface exceeds one third of the maximum expected radiation for a given day. This means that the heating system is turned off a few hours (she found 2-3 hrs to be the optimum) before solar radiation gets its peak even if the indoor temperature is below the set point temperature. The heat pump is turned on again 3 Model Predictive Control (MPC) refers to a class of computer control algorithm that use an explicit process model to predict the future response of the system. 27

40 when the solar radiation and indoor temperature reach their peaks. They concluded that by applying such a prognostic climatic strategy, it is possible to eliminate the indoor temperature peaks without letting the zone temperatures get very low. At the end, Sakellari et al. highlight the importance of dynamics in controlling functions and the difficulty of incorporating unpredictable factors such as the solar radiation in the model (Sakellari 2005; Sakellari et al. 2006) R e f e r e n c e s Aprea C. et al (2006) Performance of thermostatic and electronic valves controlling the compressor capacity International Journal of Energy Research, Vol. 30, No 15, pp Bahel V. et al. (1989) "An assessment of inverter-driven variable-speed air-conditioners: sample performance comparison with a conventional system" ASHRAE transactions Vol. 95, part 1, pp Bergman, A. (1985) Impact of on/off control on heat pumps-laboratory testing (in Swedish). Byggforskningsrådet R111: P. Stockholm, Sweden Bianchi M. et al. (2005) Comparing new control concepts for heat pump heating systems on a test bench with the capability of house and earth probe emulation Proceeding of 8th IEA heat pump conference, Las Vegas, US. Bianchi, W., E. Winandy (2007) Copeland Digital Scroll for Refrigeration: A simple mechanical system to achieve the broadest step-less modulation range Proceeding of XII Centro Galileo European Conference, Milano, Italy. Choi J.M., et al. (2001) Experimental study on superheat control of a variable speed heat pump Poc. SAREK 13 (4), pp Choi J.M., Y.C. Kim (2003) Capacity modulation of an inverter-driven multi-air conditioner using electronic expansion valves Energy No.28, pp Claesson J., M. Forsen (2002). Capacity control of a domestic heat pump- Part 1- Performance of the heat pump and its components. Zero leakage-minimum charge Proceeding, Stockholm, Sweden. Ehrbar, M, Gusber, et al. (2003) On part-load behaviour of on/off controlled heat pumps, Proceeding of the XXI IIR International congress of Refrigeration, Washington DC, US, ICR0116 Fahlen P. (2004). Heat pumps in Hydronic heating systems-efficient solutions for space heating and domestic hot water heating (in Swedish), effsys. 28

41 Flach-Malaspina, N., J. Lebreton et al. (2004) Performance of a new air-to-water heat pump system with controlled capacity, Proceeding of yje X International Refrigeration and Air-conditioning Conference, Purdue, US. Finn D., F. Murphy et al. (2008) Control algorithms for system integration of heat pumps: steady state and seasonal analysis Proceeding of 9th IEA Heat pump conference, Zurich, Switzerland. Granryd E. (2005) Refrigerating Engineering- Part II, Department of Energy Technology, Royal Institute Of Technology, Stockholm, Sweden. Granryd E., P. Lundqvist, et al. (2005) Refrigerating Engineering- Part I, Department of Energy Technology, Royal Institute Of Technology, Stockholm, Sweden. He X.D. (1996) Dynamic modelling and multivariable control of vapour compression cycles in air conditioning systems Department of Mechanical Engineering, Massachusetts Institute of Technology (MIT), Doctoral thesis. Henderson HI., D. Parker, YJ. Huang (2000) Improved DOE-2 s RESYS Routine: User-defined functions To Provide More Accurate Part Load Energy Use And Humidity Predictions. Proceeding of the ACEEE Summer Study On Energy Efficiency in Buildings, Pacific Grove, CA, US. Hori M., et al. (1985) Seasonal efficiencies of a residential heat pump air conditioner with inverter-driven compressors ASHRAE Transaction Vol. 91, part 2B, pp Hundy G. F. (2002). "Capacity control solutions with scroll compressors." Inst Refrigeration 98: pp Hydeman M., G. Zhou (2007) Optimizing chilled water plant control ASHRAE Journal, June 2007, pp Ilic S.M. et al. (2001). "Effect of Short Cycling Compressor Modulation on Refrigeration System Performance" ACRC CR-43. Ilic S.M. et al. (2002) Experimental Comparison of Continuous vs. Pulsed Flow Modulation in Vapor Compression Systems Proceeding of International Compressor Engineering Conference, Purdue, West Lafayette, US. Jakobsen A. et al. (2000) Development of Energy optimal capacity control in refrigeration systems International Refrigeration conference, pp Purdue, US. 29

42 Jin H., J.D. Spitler (2002) A Parameter Estimation based model of Water-To- Water Heat Pumps for use in Energy Calculation Programs ASHRAE Transactions Vol. 108, part 1:3-17. Jordan and Priester (1956) Refrigeration and Air Conditioning, Prentice-Hall, Inc. Karlsson F. (2003) Integrated control of heat pumps. (Chalmers University of Technology) D203:03 84 p. Göteborg Karlsson F. (2007) Capacity control of residential heat pump heating system. Department of Energy and Environment. Göteborg, Chalmers University of Technology. Doctoral thesis. Karlsson F., P. Fahlen. (2007). "Capacity-controlled ground source heat pumps in hydronic heating systems." International Journal Of Refrigeration no.30, pp Kim J. H., E Groll (2007) Feasibility study of a bowtie compressor with novel capacity modulation International Journal of Refrigeration Vol.30, pp Koury R., L. Machado et al. (2001) Numerical simulation of a variable speed refrigeration system International Journal of Refrigeration no.24, pp Kruse H. et al. (1984) Continuous and discontinuous capacity control for highspeed refrigeration compressors Proceeding of Purdue Compressor Technology Conference, Purdue, US. Li Hua. et al. (1) (2008) An empirical model for independent control of variable speed refrigeration system Applied Thermal Engineering xxx, pp. xxx-xxx. Li Hua et al. (2) (2008) Feedforward control of capacity and superheat for a variable speed refrigeration system Applied Thermal Engineering xxx, pp. xxxxxx. Lida K. (1982) Development of an energy-saving-oriented variable capacity system heat pump ASHRAE Trnasaction, Vol. 88, Part 1, pp Madani H., P. Lundqvist et al. (2008) Variable capacity heat pump systems: modelling and simulation Proceeding of 9th IEA Heat pump conference, Zurich, Switzerland Marquand C., C.J. Tassou et al. (1984) An Economic Comparison of a Fixed Speed, a Two Speed, and a Variable Speed Vapour Compression Heat Pump Applied Energy, No. 16, pp

43 Mei. V.C. (1983) Laboratory tests of a residential low-temperature water-source heat pump ASHRAE Trnasaction, Vol. 89, pp Mikhaeel M. (2007) Performance Evaluation of Constant Speed and Variable Speed Compressors in Brine-Water Heat Pumps for Space Heating: A Modeling Study Department of Energy Technology. Stockholm, Royal institute of technology of Sweden,KTH. Master thesis. Mills M. (1987) Variable speed drives Refrigeration, Air conditioning, and heat recovery Vol.90 No pp Mitsobara K. et al. (1987) The latest compressor technologies for heat pumps in Japan Proceeding of IEA Heat pump conference, Orlando, US. Parken, W.H., Beausoliel, R.W. and Kelly, G.E. 1977, Factors affecting the performance of a residential air-to-air heat pumps, ASHRAE Transactions, vol. 83, part 1, pp Poort M.J. et al. (2006) "Applications and control of air conditioning systems using rapid cycling to modulate capacity" International Journal of Refrigeration vol. 29: Qureshi T.Q., S. A. Tassou (1996) "Variable-speed capacity control in refrigeration systems" Applied thermal engineering 16(2), pp Qureshi S. et al. (1998) "Comparative performance evaluation of positive displacement compressors in variable-speed refrigeration applications" International Journal of Refrigeration Vol. 21. No. I: Riveire, P, Flach-Malaspina, N and Lebreton, J 2004, A new installation for part load testing of Air to water single stage chillers and heat pumps, Proceeding of a X International Refrigeration and Air-conditioning Conference, Purdue, US, R086. Sakellari D., P. Lundqvist (2005). "Modelling and simulation results for a domestic exhaust-air heat pump heating system" International Journal Of Refrigeration No.28, pp Sakellari D., P. Lundqvist et al. (2006) "Investigating control strategies for a domestic low-temperature heat pump heating system" International Journal Of Refrigeration 29, pp Shannon S., D. Finn et al. (2005) SHERHPA WP2 REPORT 1 Department of Mechanical Eng., University college Dublin. 31

44 Sheldrake T. (1991) "Introducing variable-speed drives" Building Services, pp Shimma Y.T et al. (1985). "inverter control systems in the residential heat pump air conditioner." ASHRAE transactions vol. 91, part 2B, pp Stoecker and Jones (1982) Refrigeration and air-conditioning, Mcgraw-Hill college division. Tassou et al. (1988) "Comparison of the performance of capacity-controlled and conventional-controlled heat-pumps" Applied Energy No.14, pp Tassou S.A., P. Votsis (1992) Transient response and cycling losses of air-towater heat pump systems, System Recovery & CHP Vol. 12 (no. 2): pp Toshiba (UK) (1987) Toshiba inverter-aided air conditioner. United Kingdom Umezo K. et al. (1984) Heat pump room air conditioner using a variable-speed capacity air conditioner ASHRAE Transaction Vol. 90, part 1A, pp Vargas. J., J. Parise (1995) Simulation in transient regime of a heat pump with closed-loopand on-off control International Journal Of Refrigeration, vol. 18, No. 4., pp Wang Y., D. Wilson et al. (1983) Heat pump control, IEEE Transactions on control theory and applications, Vol. 130., No 6, pp Wilson D. et al. (1981) The minimization of the power consumption of a thermodynamic heat pump by a microprocessor based control system Commission of the European communities, Luxemberg. Yaqub M. et al (2001) "Capacity control for refrigeration and air conditioning systems: a comparative study" Journal of Energy Resources Technology vol. 123, no. 1: Yoon S.H., H.W. Chang (2001) Empirical modelling and control performance simulation of an inverter heat pump system, Proc. SAREK, pp Yu F.W., K.T. Chan (2007) Part load performance of air-cooled centrifugal chillers with variable speed condenser fan control Building and Environment No.42, pp Zhao L., L. Zhao et al. (2003) Theoretical and Basic experimental analysis on load adjustment of geothermal heat pump systems Energy Conversion and Management, No. 44, pp

45 Zubair S.M et al. (1989) "Compressor capacity modulation schemes" Heating, Piping, Air-conditioning, pp

46 4 Development of the Model 4. 1 I n t r o d u c t i o n Capacity control by adding variable speed compressors in heat pump systems has been tackled both theoretically and experimentally by many researchers. For example, (Neale et al. 1981; Lida et al. 1982; Wang et al. 1983; Halozan 1988; Zubair 1989; Sheldrake 1991;Qureshi and Tassou 1996; Vargas & Parise 1995; Qureshi 1998; Claesson & Forsen 2002; Zhao et al. 2003;Karlsson 2007; Parnitzki 2007; Madani et al. 2008) analyzed the variable speed heat pump systems; Furthermore, the prognostic climatic control strategy for heat pumps is evaluated by (Wimmer and Shafai 1999; Bianchi 2005; Sakellari et al. 2006); Variable speed pumps on the secondary fluid sides of the heat pump have been also analyzed by (e.g. He 1996; Granryd 2007; Karlsson and Fahlen 2008; Finn et al. 2008). Despite all the studies and experiments in this field, there have been some limitations and difficulties preventing the researchers from finding clear answers for the following questions: How does the COP (Coefficient Of Performance) of a variable capacity heat pump system change over a year? What are the conditions under which the variable capacity heat pump units may yield a better performance than on/off controlled ones? How does the auxiliary electrical heater elimination in variable capacity heat pump systems influence the yearly seasonal performance factor, SPF, for the heat pump system? How does the relative sizing of the heat pump unit capacity (the magnitude of the balance point temperature 4 ) affect the seasonal performance of the system? 4 The ambient temperature at which the heat capacity of the heat pump unit is equal to the heat demand of the building. 34

47 How do variations in the mass flow rate of brine in the ground heat exchanger and water in the heating system (floor heating or radiators) affect the performance of the system, etc. The present chapter suggests a structured method to approach the challenge of better understanding the techniques and potential for capacity control in Ground Source Heat Pumps (GSHPs). This chapter describes the development of a generic model of a ground source heat pump system, covering the heat pump unit, the building, the ground as heat source, a thermal storage tank. The model also takes the local climate into consideration. The aim of the model developed in the present chapter is to facilitate a comparative analysis between the annual performance of the systems with a variety of capacity control strategies such as: Variable speed and single speed compressors Variable speed and single speed pumps in the ground source 4. 2 T e r m i n o l o g y Within the present study, the Heat Pump unit is defined as the system facilitating the thermodynamic cycle, i.e. the heat pumping, comprising of an evaporator, a condenser, a compressor, an expansion valve and a working fluid. The Heat Pump system includes several sub-systems such as the heat source (a water-filled borehole for example), the heat sink (for example, a single family house), the heat distribution system (for instance, floor heating or radiator system), the thermal storage tank (which may be or may not be in the system), the liquid pumps and the auxiliary heater. The varying climatic boundary conditions affect the sub-systems either directly (e.g. on the borehole system) or indirectly (e.g. on the heat demand of the building) M o d e l i n g A p p r o a c h T h e o v e r a l l o b j e c t i v e The main objective of the modeling effort is to compare and evaluate some of the methods and strategies used for capacity control of heat pump systems such as using variable speed compressors and variable speed pumps M o d e l i n g p h i l o s o p h y In order to make a fair comparison of the seasonal performance factor of different control strategies for heat pump systems it is essential to provide the same operating and boundary condition for the systems under study. For example, the systems should have exactly the same heating duty (the same building with the 35

48 same user behavior), the same climatic condition, and the same heat source characteristics (the same ground heat exchanger and rock behavior). Our approach is to develop a comprehensive model of the system and then use in-situ field measurements to validate the performance of the model. Consequently, different capacity control techniques can be compared by running the validated computer model over a year R e q u i r e d m o d e l c o m p l e x i t y One key question when modeling complex and interrelated systems is to find the minimum required level of detail so the model is able to capture the necessary behavior of the real system with satisfying accuracy. Figure 15 suggests a road map in order to find the required complexity of the sub-models based on the type of analysis for the system in question. The horizontal axis represents different level of complexity for the heat pump unit model, comprising the condenser, evaporator, compressor, and expansion valve sub-models. The vertical axis represents the level of complexity of the heat source and heat sink systems (such as ground or outside air, and single or multi-family houses). Depending on the type of analysis carried out on the system, the required level of complexity for the components models is different as discussed below: Figure 15. The suggested roadmap for finding the required complexity of the sub-models based on the type of analysis 36

49 In a cause-effect analysis when the influence of an operating parameter on the heat pump unit is investigated and the study focus is on the heat pump unit and the behavior of its components, it is essential to go beyond the typical black box model of the heat pump unit and develop a more complex thermodynamic model. On the contrary, a black box or a fairly detailed model of the heat source/sink would be sufficient for such an analysis (Zone B) if the focus is on the heat pump unit. Oppositely, when the focus is on a cause-effect analysis for the heat source or heat sink systems, more complexity would be needed in the heat source/sink models, allowing for a more simplified heat pump unit model (Zone C). Finally, if the goal of the study is to investigate how the entire system and the components behave for a wide variety of boundary conditions or how a new technology, introduced in one of the sub-systems affect the other systems and the system as a whole, we may have to develop a model where all the models included are fairly (or very) detailed. This gives us a complex model with a large number of input and output parameters (zone D). Such a model is extremely demanding for the computer with long computational times. It is also extremely demanding to validate due to the large number of parameters included. An extremely simple model on the other hand consisting of a set of black box models of all system components will not be able to capture the characteristics of the system in question. The developed computer model in the present study aims at facilitating a reasonable accurate comparative analysis of the seasonal performance of different control strategies (on/off controlled and variable capacity HP system) or different system layouts (for example, a system with or without thermal storage tank) by finding the balance between complexity of models and possible simplifications; so HP unit, heat source and heat sink are all modelled in detail and the models are located in the upper right corner of zone A of figure 15, somewhere close to the crossing of the dashed lines T h e q u a l i t a t i v e m o d e l o f t h e s y s t e m Figure 16 and 17 show two qualitative models of the system in question exemplifying two different levels of complexity. In the first level of complexity (Figure 16), all the component models are black box models inter-connected via certain inputs and outputs, e.g. fluid flow, or energy flow. In the first level of complexity, we do not need to know (or even want to know) what happens inside of any of the system components. Our only concern is how the system behaves due to a change in input/output. A typical black box model of the heat pump system could be based on test data expressed as polynomials based on EN14511 (CEN 2007), EN255-3 (CEN 1997) or other test standards. The ground heat source model could be as simple as a constant temperature or an equation describing how the temperature in the ground changes over the year, for example on an hourly basis. 37

50 The building model can similarly be expressed as the building heating demand as a function of the outdoor temperature. The climate can be described as hourly or so-called binned data which gives the number of hours for each outdoor temperature. Since the binned data is not suited for studying the dynamics of the systems, hourly data may instead be used for a so-called quasi-dynamic analysis of the heat pump system. This type of analysis is today the typical basis for calculation of seasonal performance of heat pump systems. However, such a first level model lacks all ability to predict the outcome of a change of the system and it cannot be used to explain the behavior of the system. Figure 16. The qualitative model of the system in the first level of complexity. Using such a model together with an elaborate calibration with test data may however be quite good at predicting performance of a well-defined system. 38

51 Figure 17. The qualitative model of the system in second level of complexity. The number beside each component represents the section number in the present chapter which is devoted to the specific component s model In the second level of complexity (Figure 17) the black-box sub-models of Figure 16 are opened and the inner processes within the components may be taken into consideration. For example, the ground heat source on the second level of complexity consists of relationships and correlations that describes what happens in the ground source heat exchanger, accounting for the thermal resistance and the rock temperature and how these parameters influences the output of the submodel. Furthermore, the model of the HP unit with the second level of complexity includes new sub-models of the condenser, evaporator, compressor, and expansion valve systems taking the thermo-physical behavior of the working fluid into consideration (these sub-systems are now black boxes on level 2). 39

52 Finally, the level of complexity can move beyond the second level, but an indepth analysis of every single sub-system will make the model full of details increasing the level of complexity considerably. Possible additional levels models will require more measured data for validation but will be able to more in greater detail describe the actual behavior of the component. It is however possible to go beyond the second level of complexity for just a certain component. For example, a change from a constant speed to a variable speed compressor in the HP unit can be analyzed comprehensively with a more elaborated compressor model including the different losses in the compressor, such as mechanical, friction and electrical losses. This technique is used in the present study T h e q u a n t i t a t i v e m o d e l o f t h e s y s t e m At this stage, quantitative models of the system components in focus are developed, using a bottom-up approach. In this approach, semi-empirical sub-models of the HP unit components are developed and the important parameters representing the characteristics of each component are identified. The individual submodels are then linked together (as black boxed systems) to form the model of the heat pump unit, very similar to the physical assembly procedure of the actual HP unit. Besides, the individual model of heat source, heat sink, climate, and thermal storage tank are developed and/or implemented. Finally, all the submodels are then linked together to model the heat pump system as a whole, also considering the boundary conditions. The modeling procedure is explained briefly in the sub-sections to follow A v a r i a b l e s p e e d h e r m e t i c s c r o l l c o m p r e s s o r s u b - m o d e l The variable speed and single speed hermetic scroll compressors are modeled in detail and the focus of the present models is to derive the characteristic parameters and their inter-relations. The models generally represent the behavior of the compressor losses, including motor losses and inverter losses for the variable speed case. The model outputs are as following: Variable and constant part of compressor electromechanical losses Built-in volume ratio Internal mass leakage coefficient Generally, the total electromechanical losses in the compressor can be divided into two main parts: a part of the losses which is almost unchanged in any operating condition and it could be assumed to only depend on the characteristics of some sub-components such as the inverter or the motor; whereas, another part 40

53 of the losses varies based on the operating condition e.g. it changes as the compression pressure ratio or the rotational speed changes (Bourdouxhe 1994;Cuevas et al. 2010). Therefore, the developed model includes these two different kinds of electromechanical losses. The built-in volume ratio of the scroll compressor is another parameter which should be incorporated into the model (Winandy et al. 2002). The built-in volume ratio of a scroll compressor is the ratio of the volume of the trapped gas pocket immediately after closing, to the volume of trapped gas pocket immediately before opening to discharge. This ratio relates to losses according to under- and over-compression of the refrigerant. Furthermore, the internal mass leakage coefficient is another parameter which could be incorporated into the model which contributes to the overall compressor losses. Chen et al. (2000) mention that the mass leakage in a scroll compressor is mostly through the gap between the bottom and the top plate of the scroll (called radial leakage ) and also through a gap between the flanks of the two scrolls (called tangential ). Cuevas et al. (2010) considered the modeling of internal leakage to be complicated because the equivalent leakage area is influenced by the oil circulation, thermal dilatation, internal pressures, compliance on the orbiting scroll, etc. Moreover, the actual model without internal leakage gives reasonable agreement with the experimental measurement (Ibid). Thus, the internal mass leakage is neglected in the modeling of the compressor at the present study. To conclude, the characteristic parameters derived from the compressor semiempirical model made at this stage will be used in the HP unit modeling, see section T h e c o n d e n s e r s u b - m o d e l In the condenser, the condensation heat transfer of refrigerant R134a is modeled assuming a Brazed Plate Heat Exchanger (BPHE). It has been observed for a wide range of operating conditions that the heat transfer coefficient during condensation inside BPHE is either gravity controlled (i.e. classical Nusselt theory) or shear controlled (Claesson, 2005a). In either case the heat transfer coefficient depends on the all-liquid Reynolds number, as complete condensation occurs in these applications under consideration here. However, for relatively large sized condensers (with relatively small temperature differences) as is usual found in domestic heat pumps, the classical Nusselt theory would be appropriate (Claesson, 2005a). To further simplify the model, it is noted from experimental evidence conducted on the modeled heat pump unit, that the condensation heat transfer coefficient could roughly considered to vary in a narrow range (almost constant). The motivation for this simplification is justified due to the fact of the limited range of flow rate in the condenser together with the experimental results in the gravity controlled regime by Claesson (2005a). Therefore, a constant overall heat 41

54 transfer coefficient (UA value) is assumed for the condenser model in the present study in which the water flow rate is constant. This assumption cannot be valid for the case of variable water flow rate in the condenser. The assumed constant UA value in the condenser for each specific heat pump unit is estimated from a calculation of the logarithmic Mean Temperature Difference and the heat capacity achieved from the measurements over a wide range of conditions. To conclude, the constant overall heat transfer coefficient derived from the condenser modeling in this section will be used in the HP unit modeling, section T h e e v a p o r a t o r s u b - m o d e l The evaporation heat transfer of refrigerant R134a is modeled in a BPHE evaporator. Claesson (2005b) shows that although the boiling heat transfer coefficient and the overall heat transfer coefficient are not constant along a BPHE evaporator, the logarithmic mean temperature difference approach may be useful if the boiling heat transfer is governed by the heat flux and the logarithmic mean temperature difference is not too small (>4-5 C). Both evaporation and super-heating portions of the heat transfer surface are considered in the model. Increasing the superheating portion of the heat transfer space leads to a decrease of the overall heat transfer coefficient because of the lower heat transfer coefficient for one-phase flow compared to the coefficient for two-phase flow. The heat flux has a dominating effect on the heat transfer coefficient (Claesson 2005c; Longo and Gasparella 2007). According to e.g. Claesson (2005d), Cooper s pool boiling correlation can be used for estimating the heat transfer coefficient. The overall heat transfer coefficient, as a function of the evaporator heat flux, which is derived from the evaporator modeling in this section will be used in the HP unit modeling, section T h e e x p a n s i o n v a l v e s u b - m o d e l The expansion valve model is assumed to provide a constant superheat temperature through an isenthalpic expansion process providing the appropriate amount of refrigerant to the evaporator. For the variable speed heat pump unit, this assumption is only valid when an electronic expansion valve is used. The experimental studies by the authors also show that it is a fair assumption for a thermostatic expansion valve if the thermostatic valve is well-sized to operate in a wide range of evaporation temperatures. 42

55 M o d e l i n g t h e H P u n i t ( v a r i a b l e a n d s i n g l e s p e e d ) At this stage, the individual sub-models described previously including compressor, evaporator, condenser and expansion valve models are combined together as black box models to form the model of the HP unit (on the second level of complexity). The model is able to process the following input data: Frequency Approach heat sink temperature to condenser (water side) called load side temperature (T_load) Approach heat source temperature to evaporator (brine side) called source side temperature (T_source) The model outputs are the required compressor work and heating and cooling capacity of the HP unit M o d e l i n g t h e b u i l d i n g A single-family house is modeled in TRNBUILD 5 and the heating load of the house is calculated based on the following parameters: The heat losses through the exterior walls, windows, floor and roof Infiltration losses Ventilation load Internal gains Solar gain The building construction materials, thermal inertia, orientation, shading, and occupancy are all possible to consider in the modeling process of the building. 5 TRNBUILD is an interface in TRNSYS in which the building descriptions and its thermal characteristics can be set. 43

56 M o d e l i n g o f g r o u n d h e a t s o u r c e ( b o r e h o l e s ) In this study, a vertical ground heat source is modeled. Heat is exchanged with the bedrock by circulating a fluid through a closed U-pipe loop in a vertical borehole. The secondary refrigerant transports the heat from the rock to the evaporator of the heat pump unit. The so-called Epsilon-NTU method is used for the modeling of the heat source system. According to experimental results by Acuna et al. (2009), the borehole thermal resistance is assumed to be constant over a year. Although the borehole thermal resistance varies to some extent along the depth (Acuna et al. 2009), an average value can be assumed as an overall thermal resistance. The ground temperature is assumed to vary about 3 K over a year, besides varying based on the rate and time of heat extraction. The average rock temperature corresponds to the average ambient temperature over the year for the specific location M o d e l i n g o f t h e s t o r a g e t a n k An insulated vertical cylinder with uniform losses and equipped with two inlets and two outlets is modeled 6. The storage tank is assumed to be a mixed (not stratified) tank which receives the hot water from the condenser as well as the return water from the building and it supplies water to the condenser as well as to the building M o d e l i n g o f p u m p s The pump model estimates the required pump power based on the pressure drops in pipes, connections, and heat exchangers. Constant pump efficiency is assumed for the single speed pumps. However, in the variable speed pump model, the pump efficiency varies with the mass flow rate of the secondary fluid. The correlation for the pump efficiency is obtained from the manufacturer s data (Jonasson 2009) M o d e l i n g o f c l i m a t i c c o n d i t i o n The climate model covers not only typical meteorological data such as temperature and relative humidity but also other important factors such as wind velocity and solar insolation given as hourly data over a year. 6 The TRNSYS pre-built model of storage tank, called type 4b, is used. 44

57 A n n u a l s i m u l a t i o n o f t h e w h o l e s y s t e m At this stage, all the sub-models made in section are combined together to form the models of a single speed and a variable speed heat pump system. TRNSYS (Klein 2005) and EES (Klein 2003) are used as the simulation tools in a co-solving manner (see Figure 18-a). For every single time step of simulation, the heat load of the building determines the outgoing temperature of the heating water from the building. This temperature is fed to the storage tank model in TRNSYS and consequently to the condenser of the heat pump unit model in EES. The rock temperature and the other borehole characteristics in the borehole sub-model determine the source side temperature and this temperature is then supplied as an input to the evaporator sub-model of the heat pump unit model in EES. For the variable speed heat pump system, the required compressor speed is determined based on the required supply temperature (Tsupply in Figure 18-b). Consequently, as it is shown in Figure 18-b, the heat pump unit model in EES processes the input data (see section 3.5.5) and delivers the output to the building, storage tank, and borehole models in TRNSYS and the process continues to the next time step of the simulation. a) b) Figure 18. Co-solving of the model by both TRNSYS and EES 45

58 4. 6 V a l i d a t i n g t h e m o d e l w i t h t h e e x p e r i m e n t a l r e s u l t s t h r o u g h i n - s i t u f i e l d m e a s u r e m e n t At this stage, the usefulness of the model in terms of its ability to predict the real systems behavior is evaluated by making a comparison between modeling results and experimental results which are obtained through in-situ field measurements. The validation process will be explained by details in the next chapter. The validation process for the heat pump unit model is as following: first, the semi-empirical models of HP unit components such as compressor, evaporator, condenser, and expansion valve are developed using a limited set of experimental data. Then, the sub-models are combined together to create a model of HP unit which is validated with another set of experimental data, different from what was used in the development stage of the sub-models. The measurements are done on two variable speed GSHP systems, both located in Stockholm, Sweden. Table 1 presents a brief summary of the systems installed in Stockholm whose data are used in the present study. Basically, three input parameters consisting of load and source side temperatures and compressor speed are varied between the ranges already presented in table 2. Then the results given by the model are compared with the experimental results. So the range of heat capacity and compressor power presented in table 1 stems from different combination of source side temperatures, load side temperatures, and compressor speeds whose ranges are given in table 2. The Root Mean Square errors for selected outputs from the HP unit model are also given in Table 2. As may be seen in Table 2, the HP unit model is able to predict the evaporation and condensation temperatures fairly accurate, as well as the heating capacity, compressor power and COP of the unit. Table 1. The characteristics of the systems installed in Stockholm whose data are used in the present study The variable heat pump unit The range of heat capacity (kw) The range of compressor power (kw) The range of compressor speed (Hz) The type of refrigerant Unit A R134a Unit B R134a 46

59 Table 2. The range of input and output parameters used for the models validations and the root mean square error of the output parameters The input parameters Minimum Maximum RMS error Source side temperature ( C) Load side temperature ( C) Frequency (Hz) The output parameters Evaporation temperature ( C) Condensation temperature ( C) Heating capacity (kw) Compressor power (kw) COP Moreover, it is shown in Figure 19 and 20 that the calculated COP and heating capacity by the developed HP unit model have less than 15% deviation from the measurements. Furthermore, the calculated compressor power has less than 10% deviation from the measurements, as it is presented in figure 21. Figure 21 also shows that the developed compressor model overestimates the compressor efficiency at high frequencies and there is a potential to improve the model to predict more accurately the compressor power at high frequencies. Figure 19. GSHP unit COP: modelling vs. in -situ field measurements. 47

60 Figure 20. GSHP heat capacity (kw): model ling vs. in-situ field measurements. 48 Figure 21. Compressor power input (kw): modelling vs. in-situ field measurements. The building model is the standard component of TRNSYS, called TYPE 56, which has been already validated (Voit et al. 1994). The ground heat source model is a semi-empirical model using certain parameters which were already obtained from in-situ filed measurements (Acuna et al. 2009). Furthermore, the

61 pumps and storage tank models are the standard components of TRNSYS, called type 3 and type 4 respectively, which have been validated in assembly meaning that they were verified as parts of a larger system which was verified (Klein 2005) U s e t h e m o d e l t o a d d r e s s t h e q u e s t i o n s By using the model developed in the present study, it is possible to compare some of methods and strategies used for capacity control of a heat pump system such as using a variable speed compressor and/or variable speed pumps in the HP system C o n c l u s i o n The present chapter suggests a generic method to approach the challenge of further understanding capacity control in a ground source heat pump. The chapter describes the development of a computer model of ground source heat pump, covering the heat pump unit, the building, the heat source, a thermal storage tank, and the climate, in order to analyze existing or new control methods or strategies in the HP system. The conceptual and qualitative models of the system are described and the required level of complexity for the sub-models is determined by aid of the suggested road map. TRNSYS and EES are used as the simulation tools in so-called co-solving manner. The model shows enough robustness for making a comparative analysis between the seasonal performances of the systems with different capacity control methods such as: Variable speed and single speed compressors (the theme of chapter 5) Variable speed and single speed pumps in the ground source (the theme of chapter 6) Furthermore, the model can be used to evaluate some innovative energy systems equipped to heat pump such as a run around coil heat recovery system equipped to a heat pump unit (the theme of chapter 7). 49

62 50

63 5 Experimental studies: an example 5. 1 I n t r o d u c t i o n Experimental studies and in-situ field measurements are carried out mainly to Validate the models Have a better understanding about the system behavior Some examples for validation of the models were already presented in chapter 3 (3.6). The present chapter shows an example of how the experimental studies are used to obtain more knowledge about the behavior of the capacity controlled heat pump systems, especially loss behavior of the system components at different operating condition. In the present examples, the loss behavior in the variable speed compressor and frequency inverter are analyzed when the compressor speed varies. The examples provide an understanding how changing the compressor speed influences the frequency inverter losses and the total isentropic efficiency of the compressor; It would then be possible to evaluate the impact of the inverter and compressor loss on the annual performance of the heap pump system M e t h o d o l o g y The experimental setup built for the measurements consists of a variable speed heat pump unit, liquid pumps, a storage tank, two plate heat exchangers, valves, and a data acquisition system (see figure 22). In the variable speed heat pump unit, two brazed plate heat exchangers are used as the evaporator and condenser. The scroll compressor is equipped with a frequency inverter in order to facilitate various compressor speeds. An electronic expansion valve maintains the superheat temperature of the heat pump unit at 5ºC. When the speed of the compressor is changed, the following parameters are maintained constant: Approach heat sink temperature to condenser called load side temperature 51

64 Approach heat source temperature to evaporator called source side temperature Figure 22. The schematic of the experimental setup Parameters which are measured and used in the present study are: The compressor power before and after the frequency inverter The compressor speed The heating and cooling capacity of the heat pump unit The condensation and evaporation pressure Superheat and sub-cooling temperature Furthermore, the built-in volume ratio of the compressor together with the constant and variable parts of the electromechanical losses, are estimated by a compressor model (section 4.3.4). The built-in volume ratio found from the modeling is used to calculate the compression power in the compressor and to evaluate the loss due to the mismatch between the actual and the built-in pressure ratio. 52

65 Heat pump heat capacity (kw) Heat Pump COP 5. 3 R e s u l t T h e h e a t p u m p h e a t c a p a c i t y a n d C o e f f i c i e n t O f P e r f o r m a n c e ( C O P ) Fig. 23 and 24 show the heat pump heat capacity and COP (defined as the ratio of heat capacity to compressor power) when changing the compressor frequency. R134a and R407C are used as the refrigerant in figure 23 and figure 24, respectively. The source/load side temperatures are held constant at 4.5ºC/26ºC in all the operating points presented in both figures. As may be seen from figure 23, varying compressor speed from 30 Hz to 100 Hz increases the heat capacity from 5 kw to 12 kw and decreases the COP from 4.4 to 2.9. Furthermore, when the R134a is replaced by R407C, the heat pump yields higher heat capacity and lower COP (figure 24). When the compressor speed increases from 30 Hz to 90 Hz, the heat capacity of the system with R407C increases from 6.5kW to 14kW and system COP decreases from 3.9 to 2.7 (see fig.24) R134a Heat capacity (kw) COP Compressor frequency (Hz) Figure 23.. The heat pump heat capacity (kw) and COP versus compressor frequency when R134a is used as the refrigerant : experimental results 53

66 Heat Pump heat capacity (kw) Heat Pump COP R407C heat capacity (kw) COP Compressor frequency (Hz) Figure 24. The heat pump heat capacity (kw) and COP versus compressor frequency when R407C is used as the refrigerant : experimental results T h e c o m p r e s s o r p o w e r b e f o r e a n d a f t e r i n v e r t e r Fig. 25 and 26 present the compressor power measured before and after the inverter when R134a or R407C is used as the refrigerant, respectively. The source/load side temperatures for both cases are held constant at 4.5ºC/26ºC. As may be seen from fig. 25, for the system with R134a, changing the compressor speed from 30Hz to 100Hz increases the compressor power from 1 kw to 4.2 kw. For the system with R407C, the compressor power increases from 1.7kW to 5.1kW by increasing the compressor frequency from 30Hz to 90 Hz (see fig. 26). 54

67 Compressor Power (kw) 5 R134a Power before the inverter power after the inverter Compressor frequency (Hz) Figure 25. Compressor power (kw) measured before and af ter inverter when R134a is used as the refrigerant. R407C Power before the inverter (kw) power after the inverter Compressor power (kw) Compressor frequency (Hz) Fig. 26. Compressor power (kw) measured before and after inverter when R134a (a) or R407C (b) is used as the refrigerant. 55

68 Inverter loss(kw) T h e i n v e r t e r l o s s v e r s u s c o m p r e s s o r f r e q u e n c y Fig show how the inverter loss changes when the compressor frequency increases from 30Hz to 100 Hz for two different set of source/load side temperatures (square and triangle dots). As it may be observed from fig. 27, with R134a the inverter loss varies between 100 W and 200 W with a peak at about 85 Hz. Fig. 28 shows that the ratio of the inverter loss to the total compressor power (the power measured before the inverter) decreases from 11% to 3% when the frequency changes from 30 to 100 Hz. The trend for the system with R407C is very similar, as it can be seen from fig.29. The inverter loss varies between 130 W and 250 W. Furthermore, as it is presented in fig. 30, the higher the compressor speed, the lower the ratio of inverter loss to the total compressor power ºC/26 ºC as the source/load side temps 1.5 ºC/21.5 ºC as the source/load side temps Compressor frequency (Hz) Figure 27. The measured inverter loss in kw when the frequency varies between 30Hz and 100Hz: the heat pump with R134a 56

69 Inverter loss (kw) Inverter loss(% of total compressor power) ºC/26 ºC as the source/load side temps 1.5 ºC/21.5 ºC as the source/load side temps Compressor frequency (Hz) Fig. 28. The measured inverter loss in % of the total compressor power when the frequency varies between 30Hz and 100Hz: the heat pump with R134a Compressor frequency (Hz) Fig. 29. The measured inverter loss in kw when the frequency varies between 30Hz and 90Hz: the heat pump with R407C 57

70 Inverter loss (% of total compressor power) Compressor frequency (Hz) Figure 30. The measured inverter loss in kw (a) and in % of the total compressor power (b) when the frequency varies between 30Hz and 90Hz: the heat pump with R407C C o m p r e s s o r m o d e l i n g A semi-empirical of the compressor is made in order to estimate the built-in volume ratio 7, the compression power of the compressor (at every operating point), and the variable and constant part of the electromechanical losses in the compressor. Winandy (1999) obtained equation (1) to calculate the compression work, assuming an ideal gas: is the built-in volume ratio, i.e. the ratio between the suction and discharge volume. The built-in volume ratio is a compressor characteristic based on the geometry and under idealized condition, it is related to built-in pressure ratio given by equation (2). (1) (2) 7 The built-in volume ratio can also be obtained from the compressor manufacturer s data if it is available and trustworthy. 58

71 The difference between the built-in pressure ratio and the actual pressure ratio (the ratio of condensation pressure to evaporation pressure) can influence the compressor efficiency, as it will be discussed later in this section. Furthermore, the swept volume flow can be determined by the equation (3): n is the compressor speed in rpm. It is possible to estimate the compression power in the compressor by combining Equation (1) and (3). Moreover, as Winandy (1999) and Cuevas et al. (2010) suggested, the total compressor power comprises of the internal compression power and the compressor electromechanical losses which also can be divided into two terms: constant part and variable part which varies by the internal compression power, as given in equation (4): (4) W= W comp +W η cons where is the parameter representing the variable part of electromechanical losses, and is the constant part of electromechanical loss (kw). Consequently,,, and built-in volume ratio are the three parameters which are found by the semi-empirical parameter estimation model (see fig. 31). Table 3 presents the results of the modeling for the analyzed heat pump system with two different refrigerants: R134a and R407C. The difference between the built-in volume ratios for the heat pump with R134a and R407C can be due to the uncertainties in the compressor model (about 6%). (3) The measured compressor power The measured condensation pressure The measured evaporation pressure The measured compressor speed minimization of the error between the calculation and measurment using the generic method Built-in volume ratio η W loss Fig. 31. The flowchart of compressor modeling Table 3. Results of the modeling for the analyzed heat pump system in two different cases: the heat pump uses either R134a or R407C as the refrigerant The estimated parameter The heat pump with R134a The heat pump with R407C Built-in volume ratio

72 ηbuilt_in Fig. 32 shows the influence of the mismatch between the built-in and actual pressure ratio on the isentropic efficiency when the actual pressure ratio changes by varying the compressor speed. Fig. 32 is made based on equation (5) suggested by Granryd (2005) in order to estimate the losses due to the fact that the actual pressure ratio for compression does not match the built-in pressure ratio. As fig. 32 depicts, the compressor tested in the present study is designed in a way that the loss due to the pressure ratio mismatch is very low when the actual pressure ratio is close to 2.7 and it decreases when the pressure ratio increases from 2.7 to 5.8. It is predicted that the (5) 1 R134a R407C Pressure ratio (condensation pressure/evaporation pressure) Fig. 32. The influence of built -in volume ratio on isentropic efficiency of compressor 60

73 Total isentropic efficiency of compressor T h e c o m p r e s s o r t o t a l i s e n t r o p i c e f f i c i e n c y Figures 33 and 34 present the total isentropic efficiency of the compressor versus the compressor frequency for two sets of source/load side temperatures, when R134a (fig. 33) or R407C (fig. 34) is used as the refrigerant. This efficiency, in equation (6) covers all kind of losses including both constant and variable parts of electromechanical losses, presented in equation (3), and also the losses due to the pressure ratio mismatch (shown in figure 32). W= W isen η isen (6) As it is shown in fig. 33, the total isentropic efficiency of the compressor changes about 11% when the compressor frequency changes from 30Hz to 100Hz. Furthermore, fig. 33 shows that the highest total isentropic efficiency occurs when the compressor frequency is between 50Hz and 55Hz for all the presented cases ºC/26 ºC as the source/load side temps 1.5 ºC/21.5 ºC as the source/load side temps Compressor frequency (Hz) Fig. 33. The variation in the total isentropic efficiency of the compressor based on the changes in the compressor frequency when R134a is used as the refrigerant : experimental results 61

74 Total isentropic efficiency of compressor ºC/26 ºC as the source/load side temps 5 ºC/30 ºC as the source/load side temps Compressor frequency (Hz) Figure 34.The variation in the total isentropic efficiency of the compressor based on the changes in the compressor frequency when R407C is used as the refrigerant : experimental results Figure 35 and 36 also present the total isentropic efficiency of the compressor versus the compressor frequency obtained from the compressor modeling when R134a (fig. 30) or R407C (fig. 31) is used as the refrigerant. 62

75 Total isentropic efficiency of compressor ºC/26 ºC as the source/load side temps 1.5 ºC/21.5 ºC as the source/load side temps Compressor frequency (Hz) Figure 35. The variation in the total isentropic effici ency of the compressor based on the changes in the compressor frequency when R134a is used as the refrigerant : modeling results Total isentropic efficiency of compressor ºC/26 ºC as the source/load side temps 5 ºC/30 ºC as the source/load side temps Compressor frequency (Hz) Figure 36. The variation in the total isentropic efficiency of the compressor based on the changes in the compressor frequency when R407C is used as the refrigerant : modeling results 63

76 5. 4 C o n c l u s i o n The present chapter shows an example of the experimental studies done during the project in order to obtain a better understanding about the loss behavior of some components of a capacity controlled heat pump system. The present examples evaluate the loss behavior in the variable speed compressor and frequency inverter when the compressor speed varies. To make the analysis, an experimental setup is made and measurements are done for two different cases: The heat pump uses R134a as the refrigerant The heat pump uses R407C as the refrigerant When the compressor speed is changed, other boundary parameters such as source/load side temperatures are kept constant. The measurement results showed that increasing the compressor speed from 30 Hz to100hz lowers the heat pump COP up to 30% for both refrigerants used. Furthermore, the inverter loss increases quantitatively by increasing the compressor speed. For example, for the heat pump with R407C, increasing the compressor frequency from 30Hz to 90 Hz almost doubles the amount of inverter loss (the inverter loss increases from 130W to 250W) although the inverter loss as the percentage of the total compressor power decreases from 8% to 4%. The compressor loss due to the mismatch between the actual pressure ratio and the built-in pressure ratio is found based on the compressor built-in volume ratio of the compressor. A semi-empirical model of the compressor is made to obtain the built-in volume ratio for the compressor studied in the present study. Then, the loss due to the pressure ratio mismatch is studied and it is found that this loss rises from 1% to 20% when the pressure ratio increases from 2.7 to 5.8. Finally, the total isentropic efficiency of the compressor which represents all kinds of losses in the compressor was analyzed at different compressor speeds. The results showed that the total isentropic efficiency of the compressor shows a maximum value when the compressor frequency is close to 50Hz. This efficiency changes about 11% when the compressor frequency varies between 30Hz and 100Hz N o m e n c l a t u r e 64

77 n Speed (rpm) Subscripts P Pressure (kpa) cons constant V Volume ( ) comp compression Volume flow rate ( ) ex exhaust W Work (J) isen isentropic Power (kw) sup supply Polytropic exponent ( ) Efficiency ( ) Built-in volume ratio (-) Built-in pressure ratio (-) R E F E R E N C E S Cuevas, C., Lebrun, J., Testing and modelling of a variable speed scroll compressor. Applied thermal engineering, vol. 29:p Cuevas, C., Lebrun, J., Lemort V., Winandy, E., Characterization of a scroll compressor under extended operating conditions. Applied thermal engineering, vol. 30:p Granryd, E., Lundqvist, P., et al. 2005, Refrigeration Engineering, Department of Energy technology, Stockholm, Sweden, p Lund, J., Sanner, B. et al., Geothermal (Ground Source) Heat Pumps, A World Overview. Geo-Heat Centre Quarterly Bulletin, Klmath Falls, Oregon: Oregon Institute of Technology, vol. 25, no. 3: p Madani, H., Claesson, J., Lundqvist, P., Capacity control in ground source heat pump systems- Part I:modeling and simulation, International Journal Of Refrigeration xx, xx-xx. Madani, H., Claesson, J., Lundqvist, P., Capacity control in ground source heat pump systems- Part II: Comparative analysis between on/off controlled and variable capacity systems, International Journal Of Refrigeration xx, xx-xx. 65

78 Rybach, L., The Advance of Geothermal Heat Pumps World-wide. Newsletter IEA Heat Pump Centre (23/4). Zhao, L., Zhao, L. et al., Theoretical and Basic experimental analysis on load adjustment of geothermal heat pump systems. Energy Conversion and Management, vol. 44:p Karlsson, F., Fahlen, P., Capacity-controlled ground source heat pumps in hydronic heating systems. International Journal Of Refrigeration, vol. 30:p Karlsson, F., Capacity control of residential heat pump heating system. Department of Energy and Environment. Göteborg, Chalmers University of Technology. Doctoral thesis. Qureshi, T.Q., Tassou, S. A., Variable-speed capacity control in refrigeration systems. Applied thermal engineering, vol. 16, no.2:p Winandy, E., 1999, Contribution to the performance analysis of reciprocating and scroll refrigeration compressors, Mechanical Engineering department, University of Concepcion, Chile. Doctoral thesis. 66

79 67

80 6 The First Example of the Models Applications: Comparative Analysis Between On/off Controlled and Variable Capacity Systems 6. 1 I n t r o d u c t i o n Capacity control in GSHPs is one of the techniques having a potential for efficiency improvement. The conventional on/off control is one of the most common methods for capacity control in GSHPs. In on/off controlled heat pump (HP) units, the compressors are switched on or off by aid of a control algorithm that estimates the discrepancy between building heat demand and heat pump capacity. However, there are some alternative methods suggested by the other researchers which might improve the system efficiency. A variable speed heat pump is one of those methods. Generally, in a variable speed heat pump unit, the speed of the inverter-driven compressor is modulated by aid of control algorithm so that the heat output from the heat pump unit offsets the load of the building; consequently, the discrepancy between the heat supply the heat demand (in the building) can be decreased. However, there are some issues which have not been clarified yet: How does the COP (Coefficient Of Performance) of a variable capacity heat pump system change when the climatic conditions vary over a year? What are the conditions under which the variable capacity heat pump units may yield a better performance than on/off controlled ones? 68

81 How does the elimination of auxiliary electrical heater in variable capacity heat pump systems influence the yearly seasonal performance factor, SPF, for the heat pump system? Some researchers have investigated both on/off controlled and variable capacity systems. For example, Zhao et al (2003) did a comparison among on/off and variable capacity controls, in addition to two other methods, for GSHPs and concluded that changing the compressor speed is the preferred method; however, Karlsson and Fahlen (2007) concluded from their research that despite improved performance at part load (as was already shown by Tassou and Qureshi,1996), the variable-speed controlled heat pump did not improve the annual efficiency compared to the intermittently operated heat pump. Karlsson (2007) suggested that it is mainly due to the inefficiencies of the inverter and the electrical motor of the compressor and the need for control of pumps used in the heating and ground collector systems (Karlsson 2007). The experimental study by Cuevas and Lebrun (2009) showed that in the variable speed scroll compressor, the inverter efficiency varies between 95% and 98% when the compressor electrical power varies between 1.5 and 6.5 kw. It is also found that the additional electrical motor losses induced by the presence of the inverter are negligible (Cuevas and Lebrun 2009). The present chapter aims at making a fair comparison between the annual performance of an on/off and variable capacity GSHP systems, using the generic model developed in chapter 3. The unique features of the current study are as following: Not only heat pump unit model, but also all the other relevant submodels such as the building, the ground heat exchanger, the storage tank, and climate are made in details; so it is feasible to study the influence of any given parameter on the whole system. All the dynamic interactions among the system components are considered, using a quasi dynamic model; Exactly the same boundary condition is established for both the on/off and the variable capacity systems when they are compared to each other. The inverter and motor losses are taken into account in the model of the variable speed GSHP unit; The influence of the compressor speed on the isentropic efficiency of compressor is considered; 69

82 6. 2 M e t h o d o l o g y The annual performance of the two following control strategies is evaluated by modeling both systems over a year: A GSHP system with an on/off controlled compressor and single speed pump in both source and load sides A GSHP system with a variable speed compressor and single speed pumps in both source and load sides Figure 37 shows the schematic of the systems studied in the present chapter. The system not only covers the heat pump unit consisting of compressor, expansion valve, condenser and evaporator, but also building as the heat load, ground as the heat source, storage tank, and pumps. The system also includes the electrical auxiliary heater, when the heat pump unit is not designed to cover the peak heat demand of the building. Figure 37. The schematics of the systems whose annual performances are analyzed in the present study 70

83 The building as the heat load, the ground heat exchanger as the heat source, the heat pump unit components consisting of compressor, condenser, evaporator, and expansion valve, the boundary condition such as the climatic condition, all are established to be exactly the same for both systems; so the systems can be compared in a fair way. The only difference between the two systems is the inverter and the motor losses induced due to the change in compressor speed and capacity S u b - m o d e l D e s c r i p t i o n The general information about the sub-models is already provided in chapter 3. The detailed information about the sub-models made for the cases in the present chapter (two systems mentioned in next sections) are: T h e b u i l d i n g The building is modeled by aid of TRNBUILD, an interface in TRNSYS (Klein 2005) in which the building descriptions and its thermal characteristics can be set. In this case, a large house located in Stockholm, Sweden, with a peak load of 13.2kW and annual heat demand of about kwh is modeled. As presented in chapter 3 (3.5.6), the heating load of the building at every time step is calculated based on the following parameters: The heat losses through the exterior walls, windows, floor and roof Infiltration losses Ventilation load Internal gains Solar gain The construction materials for the walls, roof and floor were selected from the library embedded in TRNBUILD program. In this case, the exterior walls have a wooden lightweight frame construction with mineral wool as insulating material. The windows, added to the exterior walls, are double-glazed with one low-e coating and filled with air. The windows are equipped with internal shading which is applied when the zone temperature exceeds the upper limit zone temperature (20 C). The floor construction is an externally insulated concrete slab on the ground. The position of the wall to the sun, the ambient temperature and wind velocity at every time and date of the year, and also the solar gains of the exterior walls and windows are all considered in calculation of the heat load at each time step. 71

84 Furthermore, the building model provides the opportunity to set the infiltration rate, ventilation rate, internal gains for each zone. In this case, the ventilation and infiltration rates for all the zones are set to 0.5 ACH (Air Change per Hour) and 0.1 ACH, respectively. Moreover, time schedules are applied for occupancy and appliance loads. That makes a variation in internal gain of every zone over a day. A summary of information about the building model can be found in Table 4. Figure 38 also shows variation of the building heat load in different ambient temperature (obtained from modeling of building in TRNSYS). As it can be seen from figure 38, there is a large variation in different building heat demands at the same ambient temperature. This stems from variation of solar radiation, internal gain, and wind speed and also, significantly, from the variation of ambient conditions at the previous day. For example, a sudden change in the ambient temperature does not change the building heat demand instantly due to the high thermal mass of the building. Building description Single family house large size Table 4. A brief description of the buildings modelled in TRNSYS and used in the system modeling U values (W/m2 K) Wall: 0.3 Roof: 0.2 Floor: 0.2 Windows: 1.8 (frame+glass) Infiltration rate (ach) Ventilation rate (ach) Annual heating load (kw/hr) Heating peak load (kw)

85 Figure 38.The building heat demand (kw) in different ambient temperatures, obtained from modeling in TRNSYS T h e o n / o f f c o n t r o l l e d ( s i n g l e s p e e d ) G S H P u n i t e q u i p p e d w i t h a n e l e c t r i c a l a u x i l i a r y h e a t e r A typical on/off controlled heat pump unit currently used in Swedish single family houses has been modeled. As presented in chapter 3, all the sub-models including the single speed compressor, the evaporator, and condenser are modeled individually and connected together to form the HP-unit model. Engineering Equation Solver (Klein 2003) is selected as the simulation tool for the HP unit modeling. The main function of this model is to process the input data given from the building (load side temperatures), borehole models (source side temperatures), and climate (ambient temperatures) by finding the required compressor work, heating and cooling capacity of the unit at every time step of the simulation based on the given input. Usually, HP units with single speed compressors in Sweden are dimensioned to cover about 60% of the peak heat demand i.e. the building heat demand at the Design Outdoor Temperature. The designed HP units typically cover 85-95% of the total heating energy demand over a year; however the recent trends show that 73

86 the heating coverage by heat pump is increasing, going to 70% of the peak demand and about 98% of the annual energy demand (Forsén 2009). In the present study, the unit has the heat capacity of 7.5kW at source/load temperature of 0 C/35 C (about 57% of the building peak demand). According to the heat load of the building located in Stockholm, this unit covers about 90% of the total heating energy demand over a year. Figure 39 presents the relationship between the on/off controlled HP unit heat capacity, used in the present study, and the heat demand of the building, described in 5.3.1, at different ambient temperature. In this figure, the source side temperature is assumed constant at 0 C, contrary to the variable heat source temperature in the actual modeled system. 74 Figure 39. Principal building heat demand and single speed HP heat capacity at different ambient temperature (a rough estimation). Furthermore, the load side temperature varies based on the variation in ambient temperature in order to meet the heating demand of the building at different ambient conditions. At the point where the two curves cross, often denoted as the balance point, the capacity of the HP unit is equal to the heat demand of the building. For this particular building and HP-system combination, the balance point occurs when the ambient temperature is around -2 C. On the right hand side of the balance point (higher ambient temperatures) the capacity of the HP unit is higher than the heat demand and the compressor will turn on and off intermittently (the compressor cycling region). Oppositely, on the left side of the

87 balance point (lower ambient temperatures), the heat demand of the building is higher than the HP-unit can deliver and the electrical auxiliary heater is activated to compensate the deficit. It should be mentioned that in the real case, the building heat demand does not change linearly by variation in the ambient temperature (as was indicated in figure 38) because of the solar insolation, the thermal inertia of the building, and user behavior. Thus figure 39, and also figure 40, only depict an estimation of the building heat demand at different ambient temperatures; whereas, in the modeling process, all the influential parameters on the building heat demand including the solar insolation and thermal inertia are considered. Additional information about the unit characteristic and performance is presented in the table 5. The overall heat transfer coefficient for the evaporator given in table 5 is an average value to give the readers a rough idea about its order of magnitude and it is obtained from making an average of all the different heat transfer coefficients varying by the heat flux. Table 5. Characteristics of the single speed GSHP unit which is mod eled in the present chapter Type of the condenser Type of the evaporator Type of compressor Type of the expansion valve Braze plate heat exchanger (BPHE) Braze plate heat exchanger (BPHE) Hermetic scroll Thermostatic expansion valve The average overall heat transfer coefficient for Condenser (kw/ K) The average overall heat transfer coefficient for Evaporator (kw/ K) The compressor nominal power source side/load side temp of 0/35ºC The set point temperature for superheat ( K) T h e v a r i a b l e s p e e d G S H P u n i t The same GSHP unit as the on/off controlled one (described before) has been modeled; however, a frequency inverter is retrofitted to the unit in order to change the compressor speed and consequently the unit heat capacity. Similarly to the on/off controlled case, all the components including the compressor (but this time, variable speed one), the evaporator, condenser and expansion valve are modeled individually and connected together to form the HP unit model (Madani et al. 2009). 75

88 Figure 40 depicts the heat capacity of the variable speed GSHP unit at minimum and maximum compressor speeds, alongside the heating demand of the building. Although figure 40 shows an auxiliary heater region, there is almost no need for the electrical auxiliary heater because the heating demand of the building in the graph is only an approximation (as previously explained) and as may be seen in figure 38, in spite of having an ambient temperature lower than -12 C, the heat demand of the building is mainly below 12kW. Furthermore, the storage tank adds thermal mass to the heating system, helping to cover almost 100% of the demand by the stand-alone variable speed GSHP unit. Figure 40. The building heat demand and variable speed Hp heat capacity at different ambient temperature (a rough estimation) For the variable speed HP unit, in spite of decreasing the compressor speed down to the minimum frequency (30Hz), there is still compressor cycling region, as shown in figure 40; however, by decreasing the compressor speed from 50Hz (single speed case) to 30Hz (variable speed case), the cycling region is shifted from about -2 C (see figure 39) to about +4 C (figure 40). As it is shown with the solid blue line in figure 40, when the ambient temperature is between -12 C and +4 C, the compressor speed varies between the minimum (30Hz) and maximum (75Hz) in order to provide as much heat as needed for the building. Figure 41 presents the estimated COP of the variable speed GSHP unit in different ambient temperatures when it operates at minimum and maximum speed, 30 Hz and 75 Hz respectively. The different slopes between below and above 0ºC 76

89 ambient temperatures are due to the difference between the required supply temperatures (heating curves) when the ambient temperature is below or above 0 ºC. Furthermore, figure 41 shows that the single speed unit shows higher performance (the auxiliary heater required at lower temperatures is not included) due to the avoidance of inverter losses and losses caused by operating at different speeds. A deviation from the nominal speed might decrease the isentropic efficiency of the compressor and hence reduce the COP of variable speed GSHP unit. Figure 41. The estimated variati on of GSHP unit COP based on ambient temperature in both variable and single speed cases (no auxiliary heater included) T h e g r o u n d h e a t s o u r c e ( b o r e h o l e ) The heat source for the heat pump in the present chapter is the ground (crystalline rock) in which heat is exchanged with the bedrock by circulating a secondary fluid through a closed U-pipe loop in a vertical borehole. The secondary refrigerant transports the heat from the rock to the evaporator of the GSHP-unit. A summary of information about the modeled ground source for both studied systems is given as following: The borehole is 200 meter deep, water filled and equipped with a U-tube heat exchanger. 77

90 An aqueous solution of ethanol (20% ethanol by mass) is used as the secondary fluid. The pipes are made of Polyethylene and have 40 mm external diameter and 2.4 mm thickness. The working fluid volume flow rate is 1.8 m3/hr T h e s t o r a g e t a n k The tank is a pre-built insulated vertical cylinder with uniform thermal losses and equipped with two inlets and two outlets. The tank is modeled using TRNSYS model, called type 4b. The tank volume for both systems studied in the present chapter is 350 liters T h e l i q u i d p u m p s The liquid pump model estimates the required pump power based on the pressure drops in pipes, connections, and heat exchangers. Constant pump efficiency is assumed for the single speed pumps. As it is shown in the system layout (figure 37), there are three liquid pumps in the system, one between the storage tank and the building with the rated power of 30W, one between the storage tank and the HP unit with the rated power of 40W, and one in the borehole side with the rated power of 180W. The first pump is always on during the heating season; however, the second and third pumps are only in operation during the on time of the HP unit C l i m a t i c c o n d i t i o n s Stockholm, representing the Nordic countries climatic condition, was selected as the location of both systems. Therefore, the climatic data for Stockholm, obtained from Meteonorm (Meteotest 2009) database, is used for the annual simulation R e s u l t s f r o m s y s t e m m o d e l i n g At this stage, all the sub-models are combined together to form the models of both variable speed and on/off heat pump system. The detailed information about the system modeling was already presented in chapter 3. The time step for modeling was set to 5 minutes. 78

91 T h e u n i t C O P i n d i f f e r e n t t e m p e r a t u r e r a n g e s In order to make a comparison between the performance of the single speed and variable speed HP units over a year, their performance are analyzed in different ranges of ambient temperature. Before presenting the results for different temperature ranges, it is worthy to clarify some points: The temperature ranges given below are approximate values and strongly depend on the design of both units. Different sizes of the units or buildings may yield completely different ranges of temperatures. The figures are just some examples of the dynamic COP at different temperature ranges and they are just picked up from the results of annual modeling of both systems. The different time periods for the figures also stems from the fact that the ambient temperature shown in that figure were within that range for that period of time. For example, in figure 43, the ambient temperature was between 10ºC and about 4ºC for 40 hours whereas in figure 45, the ambient temperature was below - 6ºC only for 10 hours). The COPs shown in figure 43 to figure 45 are the unit s COPs including the auxiliary heater, but not the power consumed by the liquid pumps; whereas, the calculated seasonal performance factors in the next section includes the pumps energy consumptions. In the calculation of the system COP, the cycling losses of the compressor are neglected according to IEA HPP Annex 28 (Wemhoener and Afjei 2006). The building thermal inertia is always considered in the simulation; however, the heat pump control unit is set to vary the outgoing temperature from the condenser based on the ambient temperature. That is why the heat pump reacts instantly to the ambient temperature. Ambient temperature is above +12 C When the ambient temperature is above 12 C, the system is mostly turned off because there is no considerable heat demand of the building. When the ambient temperature is between 12 C and 20 C, internal and solar gains of the building are usually enough to compensate for the building heat losses. Figure 42 shows how small the heat load of the building is when the ambient temperature exceeds 12 C. The non-zero heat demand of the building at very short time that the ambient temperature gets to 20ºC can be explained as the result of considering the building thermal inertia. Every single time step presented in figure 42 to figure 45 represents 5 minutes of the simulation. 79

92 Figure 42. Building heat demand (kw) when the ambient temperatur e is above 12 C. Ambient temperature is between +4 C and +12 C In this temperature range (+4ºC_+12ºC), both on/off and variable capacity heat pumps work intermittently. That means that the heat demand of the building is still lower than the minimum heat capacity of the variable speed unit; so the variable speed unit switches intermittently between 0 Hz and 30Hz, similarly to what the on/off controlled unit does between 0Hz and 50 Hz. Due to the fact that the variable speed unit at 30Hz yields lower heat capacity than the on/off unit does at 50Hz, the off time period for the variable speed unit is shorter than the one for single speed unit. At this temperature range, the off time percentage (the ratio between the off time and total time) for the variable speed and single speed unit is about 42% and 55%, respectively. As it is shown in figure 43, the COP of the on/off controlled unit is higher than the one of the variable speed unit. That stems from two causes: first, the inverter losses were added to the variable speed unit. Secondly, the variable speed unit has a lower COP when it works at the minimum speed (30Hz) compared to when it works at the nominal speed (50Hz) which is mainly due to the decrease of the compressor isentropic efficiency. 80

93 Figure 43. An example of dynamic COP of both variable speed and single speed GSHP when the ambient temperature is between 12 C and 4 C When the ambient temperature is between -2 C and +4 C In this range (-2ºC_+4ºC), the single speed unit works intermittently with the off time percentage of about 20%; however, the operation of the variable speed unit strongly depends on the other conditions such as the conditions in previous time step, solar radiation, or wind velocity. Sometimes, but rarely, the unit operates intermittently (the percentage of the off time is about 7%) and during the rest of the time, the compressor speed varies between 30 and 55Hz. Similar to the previous temperature range, the COP of the on/off controlled system is still higher than the one for the variable speed unit. When the ambient temperature is between -6 C and -2 C In this temperature range (-6ºC_-2ºC), both on/off controlled and variable capacity HP units work continuously. Furthermore, the electrical auxiliary heater in the on/off controlled unit is turned on and off intermittently. As it is shown in figure 44, in the on/off controlled HP unit, the auxiliary heater has a stepwise operation due to the heat demand of the building. 81

94 82 Figure 44. An example of dynamic COP of both variable speed and single speed GSHP when the ambient temperature is between -2 C and -6 C Moreover, as it can be seen from the figure 44, when the auxiliary heater is shut off, the COP of the single speed system is higher than the one for variable speed system. However, as soon as the auxiliary heater is turned on in the on/off controlled system, the COP of the variable capacity system exceeds the one of the on/off controlled system. The very low COP of the single speed system at the beginning of the chart is due to the fact that the auxiliary heater works at its maximum capacity (6kW). Then, the capacity of the auxiliary heater decreases stepwise (in 1kW steps) down to a point when the auxiliary heater is turned off and consequently, the COP of single speed system exceeds the one of variable speed unit. The slight change in the COP for the variable speed system at the right hand side of figure 44 is due to the activities of the building inhabitants which reduce the heat demand of the building. That causes a decrease of compressor speed and an increase of unit COP. When the ambient temperature is below -6 C Both single speed and variable capacity units work continuously when the temperature is lower than -6 C; the electrical auxiliary heater in the single speed system also works in a stepwise manner; therefore, the single speed system COP would be decreased considerably. In the variable capacity system, the compressor

95 speed varies usually between 40Hz and 75Hz in order to satisfy the heat demand. As may be seen from figure 45, the variable capacity system works more efficiently than the on/off GSHP system at this range of ambient temperature. This efficiency improvement mostly stems from avoiding the electrical auxiliary heater in a variable capacity system by increasing the compressor speed. Figure 45. An example of dynamic COP of both variable speed and single speed GSHP when the ambient temper ature is below -6 C T h e a n n u a l p e r f o r m a n c e o f t h e s y s t e m Figure 46 presents the annual energy use of compressors, pumps, and electrical auxiliary heater for both single speed and variable speed systems. Obviously, the compressor energy use for the variable speed system is higher than the one in the single speed system because the variable speed compressor covers all the heat demand without any need for the electrical auxiliary heater. Since the on-time period of the variable speed unit is higher than the on-time of the single speed one, the brine side pump works for longer time over a year and the pumps energy consumption in the variable speed system is about 10% higher than the single speed system. 83

96 Figure 46. The annual energy use of system components: variable speed and on/off controlled s ystems However, as it can be seen from figure 46, the total energy use of the single speed system is about 10% higher than the one of the variable speed system. The Seasonal Performance Factor (SPF) of the variable speed system is about 10% higher than the single speed system (see figure 47). The calculated SPF is the ratio of the total energy delivered annually to the building to the annual energy use. The saving by frequency control is mainly due to the role of electrical auxiliary heater in the single speed system. The larger use of the electrical auxiliary heater, the higher total energy use of the system would be. This suggests that a bigger size of the single speed heat pump which decrease the role of the auxiliary heater (and lowers the balance point of the system) may yield a better performance and make the SPF of the system closer to SPF of the variable speed system. The present analysis indicate that a larger single speed HP unit with a balance point temperature of about -6 C yields almost the same SPF as the variable speed HP unit does. Clearly, the economic constraints limit the size of a single speed HP unit to some extent. 84

97 Figure 47. Seasonal performance factor: comparison between on/off and variable speed systems The existence of the storage tank in on/off controlled system helped the system to provide the stable temperature for the supplied heating water to the building; otherwise, contrary to the variable capacity system, on/off controlled system would suffer from higher oscillation in the temperature of the supplied heating water to the building and consequently, high indoor temperature swing C o n c l u s i o n The current study compares the annual performance of the on/off controlled and inverter-driven variable capacity heat pump systems, using the generic model developed in chapter 3. The liquid pumps used in both systems are constant speed pumps. Both systems are modeled in details and the results of the modeling for the system components are validated with the experimental data from in-situ field measurements (see chapter 3). The results from the annual modeling for both systems indicate: When the ambient temperature is above the balance point of the single speed HP unit, the on/off controlled system yields a better performance than the variable speed unit. This is mainly because the inverter and compressor losses in the variable speed compressor. 85