Solar assisted ground source heat pump system

(1)

Master of Science Thesis

KTH School of Industrial Engineering and Management Energy Technology EGI-2015-038MSC

Division of Applied Thermodynamics and Refrigeration SE-100 44 STOCKHOLM

–modelling and simulation

Mattias Ericsson

(2)

(3)

Master of Science Thesis EGI 2015-038MSC

Solar Assisted Ground Source Heat Pump System –modelling and simulation

Mattias Ericsson

Approved

2015-09-14

Examiner

Hatef Madani

Supervisor

Hatef Madani

Commissioner

Värmex Konsult AB

Contact person

Emma Myrén

(4)

Preface

This thesis is written as a degree project in Sustainable Energy Engineering with the focus on energy utilization and heat pump/refrigeration systems. For me the project concludes over eight years of study at university level where the last ve years have been spent on rstly a Bachelors degree in Energy and Environment and then this Masters degree in Sustainable Energy Engineer- ing. For the achievement of writing this thesis and completing all these years of study I would like to acknowledge those who helped me through.

Specically for this thesis I would like to thank my supervisor Hatef Madani for your valuable feedback. Peter Lembke at Rydell & Lembke Kyl & Värme- teknik AB for giving me access to your lab and heat pump paradise. Värmex Konsult AB for bringing this project to me and for the economic support of investing in the much needed software's used in this project.

Passing through these last years of study some persons have been extra in- spiring to me. I would like to thank José Acuña for helping me put some of my newly found knowledge into practice. Samer Sawalha for being one of the best and respectful teachers I've met. Your lectures made it easier to go to school!

I would like to thank my friends for sharing all those late night evenings at school. You know who you are!

I wish you all happy reading!

Stockholm September 2015 Mattias Ericsson

(5)

Sammanfattning

Inverkan av strategier för styrning och ackumuleringsvolymer på systemprestandan hos en solkollektorassisterad bergvärmeinstallation har undersökts. Det undersökta systemet är i projekteringsstadiet och kommer att byggas i projektet Slottsholmen i Västervik under 2015.

Genom att använda simuleringsmjukvaran TRNSYS har systemet mo- dellerats i sin helhet och systemets respons på olika styrstrategier och konguration av ackumulatortankar har undersökts. Systemet är de- signat med två ackumuleringstankar för solkollektorkretsen där solvär- me antingen kan användas för direkt beredning av varmvatten(en varm tank) eller som värmekälla för systemets värmepumpar(en kall tank) med syftet att då höja värmepumparnas förångningstemperatur.

Fyra olika styrstrategier har undersökts. Två strategier där antingen den varma eller den kalla tanken är prioriterad, en strategi där båda tankarna är i serie och värme kan lämnas vid båda temperaturnivåer samtidigt samt en fjärde strategi där den kalla tanken alltid förbigås och solvärmen endast används för direkt beredning av varmvatten. För varje styrstrategi har en rad olika kongurationer på ackumuleringstankarna testats.

Resultatet visar att inverkan av styrstrategier dominerar över den eekt som olika ackumuleringsvolymer har. Andelen av systemets vär- melast som betjänas av solvärme varierar mellan 0.10 och 0.13 mellan olika styrstrategier medan variation mellan olika ackumuleringsvolymer är nära försumbar. Elanvändningen i systemet har jämförts mot ett referenssystem där solkollektorerna är avstängda. Resultaten visar att besparingen i elektricitet relativt referenssystemet varierar mellan 6.6

% och 9.9 % mellan olika styrstrategier.

Intressant är att elbesparingen är högre för fall med lägre andel solvärme. För styrstrategier som prioriterar varmvattenberedning ökar temperaturnivån i solkollektorkretsen vilket leder till lägre verknings- grad för solkollektorerna och därmed lägre andel solvärme som förs in i systemet. Dock visas att solvärme som används direkt för varmvattenberedning leder till högre elbesparing än solvärme som används som källa för värmepumparna vilket förklarar den lägre elanvändningen vid lägre andel solvärme.

Ett försök att kvantiera värdet av den skördade solvärmen har utförts genom att introducera ett nyckeltal kallat Solbesparingsverk- ningsgrad (Solar Savings Eciency, ηSS). Nyckeltalet är denierat som kvoten av elbesparingen för en viss strategi/konguration jämfört med referenssystemet och total mängd solenergi som skördats. Solbesparings- verkningsgraden varierar mellan 0.23 och 0.46 med det högre värdet för strategier som prioriterar direkt varmvattenberedning.

(6)

Abstract

The inuence of control strategies and storage tank sizes on the system performance of a solar thermal assisted ground source heat pump(SAGSHP) installation has been investigated. The system investigated is in the design stage and will be implemented in the project Slottsholmen in Västervik, Sweden during 2015. Using the simulation software TRNSYS the suggested system has been modelled in its entirety and the response of the system for dierent control strategies and storage tank size congurations have been investigated.The system is designed with a dual tank conguration where solar heat can either be used for direct domestic hot water(DHW) production(in a high grade tank) or utilized as additional source for the heat pumps(in a low grade tank) with the purpose of increasing evaporation temperatures of the heat pumps.

Four dierent control strategies have been investigated. Two strategies where either tank is prioritized, one where the two tanks are run in series and heat can be delivered at two temperature levels simulta- neously and one strategy where the low grade storage tank is by-passed and heat is only utilized directly for DHW production. For each control strategy a series of dierent tank size congurations have been tested.

Results show that the inuence of control strategies dominate the eect of dierent storage tank size congurations. Solar fraction for the system varies between 0.10 and 0.13 between control strategies while variations between storage tank sizes are close to negligible. The electricity use of the SAGHSP system has been compared to a reference system where the solar collectors are switched o. The results show that fractional energy savings of the SAGSHP system ranges from 0.066 to 0.099 between control strategies.

Interestingly the fractional energy savings increases for cases with lower solar fraction. For control strategies which prioritize DHW production the temperature level in the solar collector loop increased thus leading to lower solar collector eciency and less collected heat. How- ever, solar heat used directly for DHW production leads to a higher electricity savings than using the heat as source for the heat pumps which explains the decoupling of fractional energy savings from solar fraction.

An attempt to quantify the value of the harvested solar collected heat is done by introducing a performance gure named Solar Savings Eciency, ηSS which is the ratio of the electricity savings compared to the reference system to the collected solar heat. The Solar Savings Eciency ranges from 0.23 to 0.46 with the higher value registered for strategies which prioritize DHW production.

(7)

List of Figures

1.1 Buildings at Slottsholmen. Picture extracted from IDA ICE-

model. . . 2

2.1 System ow chart . . . 6

2.2 Solar collector loop . . . 8

2.3 Kitchen refrigeration . . . 9

2.4 DHW system . . . 10

2.5 Heat pump source system . . . 11

2.6 Control strategy for summer/winter case . . . 24

2.7 Forward temperatures for space heating systems . . . 25

3.1 Solar Fraction . . . 30

3.2 Fractional energy savings [-] . . . 31

3.3 Solar savings eciency . . . 32

3.4 Total collected heat at solar collectors . . . 33

3.5 Hourly mean temperature at solar collector outlet . . . 33

3.6 Evaporator inlet temperatures . . . 34

3.7 Annual specic pump energy relative reference system . . . 35

3.8 Specic pump energy divided by circulation pump relative the reference system . . . 36

3.9 Specic pump energy for Prio AK02 [kWh/m²] . . . 36

3.10 Specic pump energy for Only AK02[kWh/m²] . . . 37

3.11 Outlet temperature of solar collectors and AK02 for strategy Prio AK02. VAK01= 3m³and VAK02= 1m³ . . . 38

3.12 Outlet temperature of solar collectors and AK02 for strategy Prio AK02. VAK01= 3m³and VAK02= 6m³ . . . 39

3.13 Specic electricity savings [kWh/m²] . . . 40

3.14 Specic electricity savings for Prio AK02 [kWh/m²] . . . 40

3.15 Specic electricity savings for AK01-AK02 Series [kWh/m²] . . 41

3.16 Pay o times for adding storage to the solar collector loop for strategy Prio AK02 [years] . . . 42

(8)

List of Tables

2.1 Building loads and key gures . . . 5 2.2 Types used in TRNSYS simulation . . . 14 2.3 Solar collector data and coecients for use with equation 2.1 . 15 2.4 Rated heat pump conditions for the Rydell & Lembke Ec-

Multi903p. Rated conditions accordning to EN14511(i.e. 30/35^◦C at condenser side and 0/-3^◦C at evaporator side) . . . 19 2.5 Heat pump test cases for development of input to Type 927

data les . . . 19 2.6 Estimated pressure drops . . . 21 A.1 Flow chart symbols . . . 53

(9)

1 Introduction

Heat pump systems have over the last decades emerged as a natural replace- ment of fossil fuel use in the built environment. The electricity used to run the heat pumps however comes at a fairly high price and heat pump performance is a major point of focus to minimize the bought electricity for the customer thus keeping costs as low as possible. Solar collectors oer a way to enhance the heat pump system performance either by replacing some of the heat that else would have been produced by the heat pump or by enhanc- ing the performance of the heat pump itself by improving running conditions.

Systems however tend to get more and more complex as solar collectors are integrated at several temperature levels in the system. This raises the need for thoroughly developed control strategies in order to make the most of the collected solar heat in the heat pump system.

1.1 Solar assisted ground source heat pumps

Solar assisted ground source heat pumps(SAGSHP) denes a class of heat pump systems where solar thermal collectors are integrated into a ground source heat pump system with the ground acting as main source of thermal energy brought into the system. There are several approaches when integrat- ing solar thermal collectors into ground source heat pump systems. When considering systems with combined solar collectors and heat pumps, whatever the source of the heat pump, there are mainly two approaches. Either the solar collectors act as an assisting source for the heat pumps with the main objective to enhance the heat pump performance since the solar collectors often oer more high grade energy than e.g. a borehole, or the solar thermal energy is used directly for space heating or domestic hot water(DHW) production and therefore relieves load from the heat pump. Which one of these two approaches that is chosen is often decided by whether the designer has a solar collector or heat pump background which naturally places one of the components in main focus (Elimar et al., 2010).

1.2 Project description and objectives

This thesis investigates the eect of control strategies and storage tank con-

gurations(sizes) of a SAGSHP installation. The system investigated is in the design stage and will be constructed at Slottsholmen in the city of Västervik in Sweden during 2015. The project is conducted together with Värmex Kon- sult AB whom are responsible for the design of the energy system supporting

(12)

Figure 1.1: Buildings at Slottsholmen. Picture extracted from IDA ICE- model.

the buildings with space heating/cooling, DHW and ventilation. The energy performance of the buildings have previously been investigated by Värmex using the building simulation software IDA ICE and demand curves from the simulations have been extracted and used as inputs to the simulations conducted in this thesis project. The system is predicted to have an electricity performance of approximately 30 kWh/m² for the heat pumps and auxiliary HVAC equipment. Figure 1.1 shows the geometric model of the buildings in IDA ICE.

The system utilizes solar collectors in a dual tank conguration where heat from the solar collectors can either be used for direct DHW production or as a source for the space heating heat pumps. Even though some general conclusions can be drawn from this thesis the study and it's results should be considered case specic as the boundary conditions concerning system layout as well as heating/cooling loads will aect the system performance and most likely the conclusions drawn from the results in this thesis.

The main objective is to suggest methods for improving the use of solar collectors in the SAGSHP installation. The system overall design is previously determined by the responsible designers and the design parameters investigated in this thesis is the sizes of the two storage tanks in the solar collector loop and the control strategies for utilizing the collected heat. The objective is to investigate strategies for the solar collector loop which gives the most use of the solar collector yield. The solar collector area or the mass ow of brine in the solar collector loop are not subject to investigation.

(13)

1.2.1 Research questions

The objective of the thesis has been formulated into three specic research questions presented in the list below. The two temperature levels referred to as low and high have a oating temperature range but have dierent levels relative to each other. The high temperature tank is used for DHW production thus producing hot water at a maximum temperature level of around 65^◦C while the low temperature tank is used for increasing brine temperature to the heat pumps and thus limits at 30^◦C. The dierent temperature levels and their limits and control strategies are discussed further in the report.

1. What is the optimum control strategy for prioritizing either the low or the high temperature storage tank?

2. How does the low temperature storage tank size aect system performance?

3. How does the high temperature storage tank size aect system performance?

1.3 Previous work

Over the past few years a comprehensive work focusing on solar and heat pump systems has been carried out within the International Energy Agency solar heating and cooling programme. The specic project has been carried out under the name SHC Task 44 and includes a wide range of publications which have been helpful when conducting this thesis project. Several of the referenced papers in this thesis have been produced within the frame of IEA SHC Task 44.

Starting with the solar collectors the selection of publications is overwhelm- ing and a comprehensive range of applications is covered in the literature.

Bunea et al. (2012) investigates the performance of solar collectors under low temperature conditions which is likely to occur if the solar collectors are used as source for heat pumps. They found that unglazed collectors are more ef-

cient than glazed at plate collectors under low temperature operation due to the many secondary heat gain eects that occur if the solar collectors are not screened from the environment by a glazing. This raised some discussion in this thesis since the solar collectors used in the studied system are glazed

at plate collectors even though they can work as heat pump source.

The use of solar collectors for increasing the heat pump Coecient Of Performance(COP) was investigated by (Girard et al., 2015). Some potential

(14)

is shown but for locations with low solar irradiation the benet over a year is close to negligible. Pärisch et al. (2014) concludes that even if solar heat is used directly in a system it aects the evaporation temperatures for the heat pumps as it relives load from the heat pumps and less heat is extracted from the ground. Regarding control strategies, both (Pärisch et al., 2014), (Banister and Collins, 2015) and (Rui and Wang, 2012) show that heat used directly in a SAGSHP system is more valuable(leads to lower electricity use of the system) than heat used indirectly as source side of the heat pump. The same conclusion is drawn by (Kjellsson, 2009) and (Kjellsson et al., 2010) but with the added condition that the ground heat exchanger is correctly sized. If the ground collector is too small(too short boreholes) the benet of recharging the ground with solar heat is higher. The same conclusions is drawn by (Bertram, 2014) through TRNSYS simulation studies. Also (Haller and Frank, 2011) concludes that if temperatures are benecial for direct use of the solar heat it is more ecient than indirect use. Many of these results have been the basis of developing the control strategies in this thesis project.

Control strategies seem to be a more investigated area than the eect of storage sizes. There exists many papers that studies the eect of storage on the hot heat pump side (condenser side) but for storage on the cold side the papers are few. Gomri and Boulkamah (2011) investigates dierent solar collector areas and storage tank sizes in an all solar heat pump system. The storage tank is placed on the cold side(evaporator side). The conclusions are that the collected heat is not proportional to the increase in solar collector area since it is dependent on the storage tank size. For a certain collector area there is an optimum storage tank size and the temperature level of the storage increases with smaller storage tanks. It was shown that smaller storage tank sizes led to higher heat pump COP since the temperature fed to the evaporator is higher.

(15)

2 Method

The research questions have been investigated by modelling the system in TRNSYS. A literature study has been conducted to put the work into a larger context and to nd support for the conclusions drawn. The literature study has also focused on the modelling of certain components to be able to make correct choices of the models used in TRNSYS and also to be able to evaluate the constraints that certain mathematical simplications oers.

2.1 System description and control strategies

The full system ow chart is drawn in Figure 2.1. The gure provides an overview of the system and the dierent parts are described in detail in the coming sections. The abbreviations used are set by the designers and follows Swedish standard. For example the cold storage tank in the solar collector loop is referred to as VS51-AK01. The rst four letters are the system name: VS stands for Värme Sekundär(Heating secondary), the number 5 meaning that it is a solar collector system and number 1 meaning that it is the rst system of this type. The second four letters are the component name: AK stands for AcKumulator(Accumulator i.e. storage tank) and 01 is the tank number in the system. These abbreviations will be used to refer to components throughout this thesis but their Swedish translations will be omitted. A reference of the dierent ow chart symbols used is presented in Appendix A.

The system provides space heating, cooling and DHW to 7553 m² hotel, residential and conference complex also including a restaurant. Building loads and key gures are presented in table 2.1.

Table 2.1: Building loads and key gures

Heated area 7553 m²

Building latitude 57.75^◦N

Occupancy 219 persons/day

DHW demand 50 l/person and day

DHW yearly energy load 229 MWh (30 kWh/m²) Space heating yearly energy load 241 MWh (32 kWh/m²) Space cooling yearly energy load 70 MWh (9 kWh/m²)

Boreholes are acting as main source for the heat pumps and also as source for free cooling in summer time. There are two sets of heat pumps. The VS01 VMP01-04 provide space heating in winter and work as chillers in summer

(16)

Figure 2.1: System ow chart

(17)

time whenever the free cooling is unable to cover the cooling demand. VS02- VMP01 provide DHW whenever the solar collectors are unable to meet the demand. The four space heating and one DHW heat pump shown in the schematic are a result of a pre design and during the modelling it turned out that the heat pumps were to few to be able to cover peak demands. Therefore the simulation model was built with more heat pumps than shown in the ow chart. This is described in detail in section 2.1.5.

The solar collectors can either provide direct production of DHW at tank VS51-AK02 or act as source for the heat pumps at tank VS51-AK01. Even though the solar collector loop and the strategies for controlling it is the main focus in this thesis the whole system has been modelled in detail to ensure that the correct temperature levels are achieved in the dierent parts of the system as this will aect the use of the solar collectors and how the loop should be controlled. The dierent parts of the system are described in detail in the following sections.

2.1.1 Solar collector loop

The solar collector loop (which is shown in gure 2.2) is using an antifreeze mixture of Water/Ethanol as working media with an ethanol mass concentra- tion of 35% giving a freezing point of -25^◦C.

The tank sizes and control strategies are the subject of investigation in this thesis and the dierent investigated strategies are described in section 2.4.1. However there are some basic control strategies that are applied in all investigated cases. The tanks can be by-passed by the three-way valves VS51-SV31 and SV32 respectively. The basic criteria for by-passing a storage tank is if the temperature of the uid coming from the solar collectors is lower than either the tank temperature or the uid entering the heat exchanger coil in that storage tank. This strategy is applied to not have the solar collector

uid cool down the tanks. Another strategy also applied in all cases is the control of the circulation pump VS51-PV01. The nominal ow of this pump is set to give a temperature increase through the solar collectors at maximum yield to around 8^◦C. The temperature dierence of the uid running through the solar collectors is monitored and if the solar collector is loosing energy the ow is set to a minimum of 1 kg/h just to still be able to monitor the temperature dierence. A third basic strategy is that the low grade storage tank, VS51-AK01, is always by-passed in summer time when there is a need for free cooling and the heat pumps acts as chillers.

(18)

Figure 2.2: Solar collector loop

2.1.2 Kitchen refrigeration

The restaurant kitchen has a need for refrigeration and the condenser heat rejected has to be taken care of in the system. This is done in two stages where the uid rst passes the tank VS02-AK01 for preheating of DHW and then passes a heat exchanger VS01-VX01 where the rest of the heat is rejected to the source side of the space heating and DHW heat pumps. This is shown in gure 2.3.

(19)

Figure 2.3: Kitchen refrigeration

The kitchen refrigeration design was however not nalized at the time of writing this thesis. The operation times and loads were also considered to be very uncertain so the kitchen refrigeration was omitted in the TRNSYS simulations. The components are included in the model but due to the large uncertainty concerning their operation they were switched o in the simulation. Therefore no heat is transferred in VS02-AK01 or VS01-VX01 and all connected circulation pumps are switched o.

2.1.3 DHW

The system for heating DHW is shown in gure 2.4. DHW is heated in several stages. First cold water passes VS02-AK01 but since the Kitchen refrigeration is not in operation it is not heated. Secondly it enters VS51-AK02 and is heated by the solar collectors. After the tank the incoming DHW is mixed with the hot water circulation. If the temperature at VV01-GT44 is above the set point the DHW is directed out to the system via VV01-SV31. If the temperature is too low the DHW is directed via the storage tanks VV01- AK01-03 which are heated by the DHW heat pump. If no ow is run via the storage and temperature of the DHW still is above the set point(which can be the case if solar irradiation is very high) there is a possibility of mixing in cold water at VV01-SV32 as a precaution. Outgoing DHW temperature set point

(20)

Figure 2.4: DHW system

is 55^◦C and the storage is kept at 65^◦C. A 5^◦C temperature drop is assumed over the hot water circulation system.

2.1.4 Heat pumps source side

The heat pump source side is shown in gure 2.5. The main source for the heat pumps is the borehole storage. Heat can also be introduced from the solar collectors via storage tank VS51-AK01 or from the condenser heat of the kitchen refrigeration. The circulation pump KB01-PK01 pumps brine through the heat pump evaporators and KB01-PK02 pumps brine to the space cooling systems. As the kitchen refrigeration is switched o the strategies for

(21)

Figure 2.5: Heat pump source system

controlling it is of minor importance in this thesis. However in winter time when there is a need for heating, the kitchen refrigeration condenser heat is passed through KB01-SV51 thus going directly to heat pump evaporators to increase their eciency. In summer time when there is a need for space cooling the kitchen refrigeration condenser heat is distributed directly into the borehole storage to dissipate as much as possible of the heat in the ground. In summer the heat pumps are controlled to keep a certain forward temperature at KB01-GT43(the forward temperature to the space cooling systems) and only started if the free cooling from the boreholes isn't able to cover the demand.

2.1.5 Space heating/cooling load side

In heating case the heat pumps produce heating water to be stored in VS01- AK01. Water is taken from the tank and mixed to a proper forward temperature in the three way valve VS01-SV31. In the TRNSYS model this valve was

(22)

however replaced by three way valves at each secondary heating system. The system has four air handling units(AHU) and two oor heating systems serv- ing dierent buildings. In the modelling these were combined to two loads, one for the AHU's and one for the oor heating systems. Cooling can be provided both to the AHU's and to the oors via the oor heating pipes. The cooling load is modelled as one load with the same temperature requirements regardless of ambient temperature(which is how the system will be run in real operation). When the heat pumps are started in summer time to provide cooling the condenser heat can be rejected via the dry-cooler in the KM01 system. The heat pumps are provided with dierent temperature levels at the condenser depending on whether they are run in heating or cooling case. In the heating case the heat pumps are controlled to keep a specic temperature in the lower part of the storage tank at VS01-GT45 which is set by the tem-

(23)

perature requirements on the heating system. In cooling case the dry-cooler is operated to give a return temperature to the heat pump condenser of 30^◦C.

In heating case the number of heat pumps running is determined by the temperature in the storage tank VS01-AK01. The temperature is controlled at the bottom outlet of the tank and the set point in winter case is 40^◦C which gives a heating control signal for the heat pumps. The cooling control signal comes from the brine forward temperature to cooling coils(at KB01-GT43).

The set point is 15^◦C which is the required forward temperature for the cooling coils to provide the necessary amount of cooling. Start signal is given to the heat pumps depending on these two control signals. The controller listens to the cooling control signal in summer and the heating control signal in winter.

If the controlled temperature falls below the set point(or above in cooling case) another heat pump i started. No frequency controlled heat pumps are implemented and the load is met solely by stepping in and out heat pumps depending on the demand.

Worth noting is that in the pre-design of the system which has been the base for this thesis project the number of heat pumps to be installed was chosen to four. This turned out to be to few when the simulations got started and the heat pumps were unable to keep the desired heating set point. Therefore in the TRNSYS simulation the system has been given the possibility to run up to a maximum of seven heat pumps for space heating and choose whatever is needed, even though the schematic still only shows four space heating/cooling heat pumps. It has turned out in the simulations that six heat pumps are enough for keeping the desired set points.

2.2 TRNSYS modelling

TRNSYS, which is an acronym for TRaNsient SYstem Simulation Tool, is a simulation software designed for studying transient behaviour mainly in energy systems. The software has an extensive library of predened components that are commonly used in energy systems. The possibility of programming own components makes TRNSYS suitable of studying any process that includes a transient behaviour and not only energy systems(TESS, 2015). The TRNSYS version used is 17.02.0004.

TRNSYS components are referred to as Types and each Type is followed by a number dening that specic component. This terminology will be used from now on in this thesis. With TRNSYS a library of standard Types is included.

When studying more complex system the standard library is not enough and the major part of the Types used in this thesis project are Types developed by TESS inc.(Thermal Energy Systems Specialists). These Types are commonly

(24)

used in research all over the world and are very mature component models.

The TESS libraries used are of version 17.2.01.

The procedure of modelling and simulating in TRNSYS begins in a drag and drop graphical interface where Types are organized and the relations set by connections. Once the model is dened a so called deck le is created which basically is a text le containing all necessary information about which types that are used, their connections, parameters and initial values, external

les etc. The deck le is then read by the TRNSYS kernel which executes the simulation and writes the user dened outputs to text les or in a graphical interface.

For the system considered in this thesis project TRNSYS is an ideal simulation tool. The system contains solar thermal collectors, ground storage, storage tanks etc. which all can be represented by Types from the TESS libraries. The use of controllers makes it possible to simulate the system as it is intended to be run in reality where systems set points are weather and load dependent.

2.3 Component models

In this section the major Types used in the simulation are described. If there exists several Types representing the same component in the libraries the choice of the specic Type is motivated by a short mathematical reference and literature references. For full mathematical references of the types used the reader is referred to the TRNSYS and TESS documentations referenced in each section. Types of minor importance for the simulation(such as T- junctions, le readers and printers) are omitted in this report. The complete TRNSYS deck le is attached in Appendix B for reference. A summary of the major Types used are given in table 2.2.

Table 2.2: Types used in TRNSYS simulation

Component TRNSYS/TESS Type

Solar Collectors Type 539 Heat pumps Type 927 Ground storage Type 557a Storage tanks Type 534 Circulation pumps Type 742

Controllers Type 1669 and 23

(25)

2.3.1 Solar collector

The Solar collector model chosen for this system was the TESS libraries Type 539. Type 539 models a at plate collector with a correction for the collector thermal capacitance. The governing equation is a general rst order dier- ential equation developed by Due and Beckman. A complete mathematical reference of the model can be found in (TESS, 2012d).

The solar collector eciency is calculated using the Hottel-Whillier equation (equation 2.1)¹which is widely used in collector testing. The coecients needed for the Hottel-Whillier equation can easily be obtained from manufacturers. However care should be taken when using the coecients as they can be dened on dierent collector areas depending on where the test was performed. Typically test from Europe uses aperture area while coecients produced in USA are dened by gross collector area (TESS, 2012d).

η = η₀− a₁ti− ta

I_T − a₂(ti− ta)²

I_T (2.1)

where

ti is the uid inlet temperature to the collector [^◦C]

ta is the temperature of the ambient air [^◦C]

IT is the total incident radiation on the collector surface [W]

The specic collector used in this project is manufactured by Aquasol in Swe- den.

Table 2.3: Solar collector data and coecients for use with equation 2.1 Collector absorber area 100 m²

Optical eciency, η0 0.821

a₁ 3.557 W/m²K

a2 0.0073 W/(mK)²

The type of collector used is in this case is chosen by the client based on mainly aesthetic reasons and is not subject to investigation. There is however a need of commenting on the choice of using glazed at plate collectors since the thermal performance varies among types and other collector types could be

1In some cases referred to as the General solar collector eciency equation

(26)

more thermally benecial in this specic application. A general energy balance for a solar collector is given in equation 2.2 (Bertram et al., 2012). Besides heat absorption from short and long wave solar radiation( ˙qrad,S and ˙qrad,L) a solar collector gains heat from other secondary eects such as convective heat transfer(sensible ˙qair,sens and latent ˙qair,lat), heat conduction( ˙qc) and from rain falling on the collector surface( ˙qrain). The latent heat gains can also be divided into several sub-eects such as condensation, evaporation, freezing or melting on absorber surface.

Q˙gain

A_coll = ˙qgain= ˙qrad,S+ ˙qrad,L+ ˙qair,sens+ ˙qair,lat+ ˙qc+ ˙qrain (2.2)

In a glazed at plate collector several of these eects are negligible since the absorber area is protected from the surrounding air by the glazed surface. Therefore the mathematical description of the Type used in this thesis project does not include eects besides radiation and thermal losses obtained from the collector eciency correlation. However in the studied application the solar collectors will be used in a wide range of operating temperatures since it will both supply heat for DHW and work as assistance for the heat pumps on the evaporator side. Therefore it's not unreasonable to expect the collector working temperature to be below the ambient temperature or the dew point of the ambient air at certain running conditions. In these cases some of the secondary eects can have a major impact on the collector thermal performance and the use of an unglazed collector could be more benecial.

Bunea et al. (2012) concluded that unglazed collectors are more suitable in heat pump applications since they are more ecient under low temperature or night operation than at plate collectors.

Commenting on the thermal capacitance of the solar collectors Type 539 was chosen to give the system some inertia and not respond unreasonably fast when solar irradiation changes quickly. Deng et al. (2015) developed a series expansion model based on the same governing equations as Type 539 with a correction for the thermal capacitance of the solar collector in order to predict collector performance in cases with sharp solar irradiation changes. The thermal capacitance however is not easily determined and is rarely provided by the collector manufacturer. Deng et al. (2015) measured the response time constant for a solar thermal air collector and found the thermal capacitance to be 23.5 kJ/K·m². The collector is comparable to the ones used in this project in terms of materials and total weight.

D'Antoni and Saro (2012) performed a literature review on massive solar collectors which have a massive material behind the absorber area for the

(27)

purpose of heat storage. The thermal capacitances of the reviewed collectors were all above 100 kJ/K·m². These gures denes an order of magnitude for the thermal capacitance. The solar collectors utilized in this system is however not massive and the capacitance was chosen to 30 kJ/K·m²to give the system some thermal inertia². The thermal capacitance of the solar collectors is beside the focus of study in this thesis why no further investigations were carried out in this topic.

2.3.2 Ground collector

For the borehole collector model the Tess libraries Type 557a is used. The Type models a vertical borehole eld using U-tube heat exchangers. There are not many Types to choose from in TRNSYS when it comes to ground storage. For modelling U-pipe heat exchangers there is only Type 557a and 558b where the only dierence is the expression for the pipe to ground thermal resistance. In Type 557b the thermal resistance is used as an input when in Type 557a the dierent material properties for the pipes, borehole lling and ground are used as input. The choice of Type 557a was made due to the fact that the thermal resistance is unknown since no thermal response test is yet performed at the site. Instead the pipe and ground properties were estimated to be used as inputs to Type 557a. Ground properties was extracted from a bedrock map of the area and the corresponding thermal properties where taken from (Banks, 2009).

Type 557a is based on the work done by (Hellström, 1989) and models the borehole eld as a cylindrical ground storage. The boreholes are uniformly distributed within the storage volume and the borehole conguration can not be modied. Therefore Type 557a is not suitable if the aim is to study the thermal interaction between boreholes. For this thesis the ground storage is not in main focus and this simplication can be justied. However the borehole conguration and ground properties will aect the heat transfer from the ground and consequently the temperature of the uid leaving the boreholes.

2In the early stages of modelling dierent capacitances where tested and the eect on the system performance in the range of 25-45 kJ/K·m² was found to be very small

(28)

2.3.3 Heat pump

There are several ways of modelling heat pumps with dierent levels of detail.

Dott et al. (2013) divides heat pump modelling into three dierent categories depending on the application and need for detailed results. The three categories are Calculation methods, Dynamic system behaviour and Heat pump design. In the rst the aim is to provide simple and fast heat pump system performance comparison often presented in a seasonal coecient of performance(SCOP). Simple performance map models are often the basis for such methods. The second level of detail, which is suitable for this thesis project, are models aiming on analysing a dynamic system behaviour over time. Also in these models the basis are mainly performance maps but with a higher detail level of data required. The third category, where the heat pump design is in focus, requires the highest level of detail. These models are often based on refrigerant cycle calculations with component performance maps or detailed physical models.

TRNSYS oers the possibility of using several levels of detail in heat pump modelling. Refrigerant cycle models can be implemented by connecting TRN- SYS to an external software such as Engineering Equation Solver(EES). How- ever such a software connection, which has to be made for each simulation time step, signicantly increases the total simulation time. As such level of detail isn't required in this project the focus was instead to nd a suitable performance map model. An easy way to develop a heat pump performance map model is to use a biquadratic curve-t equation(developed from measured data) which can easily be implemented via an equation box in TRNSYS. The method is commonly used and is described by several authors. One example is (Carbonell et al., 2012).

For this thesis project the choice was made to use a performance map model and implementing it via the TESS libraries Type 927 Water to water heat pump. Type 927 takes measured and normalized performance data and interpolates within the range of the supplied data. The mathematical reference for Type 927 can be found in (TESS, 2012b).

The heat pumps utilized in this project is manufactured by Rydell & Lem- bke in Sweden and are of types EcMulti903p with R407C as refrigerant for the space heating system and R134a for the heating of DHW. Rated char- acteristics of the heat pumps are presented in table 2.4. Type 927 requires data les for a variety of tested boundary conditions. The amount of data required were however not available from the manufacturer so a part of this thesis project was spent testing the heat pumps to develop performance data in the manufacturers own test facility

(29)

Table 2.4: Rated heat pump conditions for the Rydell & Lembke EcMulti903p.

Rated conditions accordning to EN14511(i.e. 30/35^◦C at condenser side and 0/-3^◦C at evaporator side)

Variable EcMulti903p-R407C EcMulti903p-R134a

Evaporator ow[kg/s] 1.94 1.37

Condenser ow [kg/s] 1.66 1.07

Heating capacity [kW] 36.2 22.0

Cooling capacity [kW] 27.7 17.7

Compressor power draw [kW] 8.2 5.0

Test data was developed for a wide range of inlet temperatures on both condenser and evaporator side to prevent Type 927 from having to extrapolate performance data which can be very uncertain. The heat pump utilizing R407C was tested for two dierent evaporator as well as condenser mass ows to also capture the ow dependency of the performance. The R134a heat pump, which will heat domestic hot water, was only tested for one mass ow on each side. Type 927 however requires a data le with at least two mass

ows for each side. This problem was worked around by putting in the same performance data for several made up mass ows thus telling TRNSYS that the performance is ow independent. The cases tested is shown in table 2.5.

Table 2.5: Heat pump test cases for development of input to Type 927 data

les

Variable EcMulti903p-R407C EcMulti903p-R134a

Evaporator ow[kg/s] 1.57/1.95 1.38

Evaporator inlet temperatures [^◦C] 0/7/14 0/7/14

Condenser ow [kg/s] 1.50/0.98 1.38

Condenser inlet temperatures [^◦C] 30/38/46 52/60/68

Heating and cooling capacity as well as compressor power were measured for all cases and passed on to the Type 927 input le. The experimental results and normalized performance data of the heat pumps are presented in Appendix C.

(30)

Type 927 requires data les for both heating and cooling modes(Type 927 is modelled as a reversible heat pump). However only heating mode (i.e.

condenser and evaporator always on same side) was used and the cooling performance data le was left as default in the model. Even in summer case when the heat pumps work as chillers they are in heating mode but the heat rejected in the condenser is not utilized.

2.3.4 Storage tank model

For the storage tank model TESS libraries Type 534 was chosen. Type 534 can be used both with and without a coiled heat exchanger immersed. Therefore the Type ts this project well as it can be used both for the DHW-storage and space heating storage as well as the coiled tanks in the solar collector loop. The tank model is not very mathematically complicated. The tank is divided into discrete sections called nodes. In each node the uid is considered perfectly mixed thus each node is isothermal. Heat transfer occurs between adjacent nodes(above and below a node) through uid movements and conduction. At each time step the uniform temperature of each node is passed as input to the next time step. The tank also interacts thermally with the environment. The immersed heat exchanger is also divided into a user specied number of nodes in the same manner as for the storage tank. Heat is transferred between the heat exchanger and the surrounding uid via conduction and convection. The user species the dimensions of the coiled tube and where it is placed in the heat exchanger. The complete mathematical description of Type 534 can be found in (TESS, 2014).

The coiled tanks used in this thesis project has the heat exchanger coil evenly distributed between all thermal nodes in the storage tank and are counter ow coupled. The dimensions of the tank, coils and number of immersed coils are modelled after the Swedish manufacturer Götlinds Svets AB.

However when performing the parametric study on storage tank sizes the tank height and coiled tube dimensions were xed and only the volume was changed thus giving tanks with larger diameter for larger volumes.

2.3.5 Circulation pumps

The circulation pump models is used is from TESS Libraries Type 742 which is a variable speed pump with the owrate as input. The power draw of the pump is calculated from a given pressure drop and and pump eciency. The pumps are assumed to have an overall eciency of 60% and the pressure drops are estimated individually for each circulation pump. Pressure drops are however

(31)

given as constant, independently of the ow rate, so estimated values are given as mean pressure drops(at the most expected ow rate). Estimated pressure drops for the circulation pumps within the system boundaries(as dened in section 2.5.1) are presented in 2.6. The pumps also exchanges heat with the pumped uid. Mathematical reference of Type 742 can be found in (TESS, 2012c).

Table 2.6: Estimated pressure drops Pump Pressure drop [kPa]

VS01-PV01 70 VS03-PV01 60 VS51-PV01 90 KB01-PK01 140

2.3.6 Controllers

Two types of controllers are used in the model. TESS libraries Type 1669

Proportional controller is used for controlling the forward temperature on the secondary heating systems. The Type was however found to be unsatis- factory when it came to controlling the parts of the system that has faster response. Controlling of the forward temperature is done based on the outdoor temperature and the response is quite slow. The controller is described in (TESS, 2012a).

For the other controllers implemented the standard libraries Type 23 PID controller was implemented. Type 23 models a traditional PID controller with PID-parameters and a gain constant. The mathematical description of Type 23 can be found in (SEL et al., 2014).

There is need for a short comment on the PID-parameters used in Type 23. A PID controller works by combining three components into a correction for the control signal. A proportional deviation from the set point, an integrated term which is the sum of the deviation over a certain time range and a derivative term which is predicted from the slope of the deviation. Each term is multiplied by a gain constant(the respective PID-parameter) and combined into a single error value. Type 23 also uses a gain constant which is a multi- plier that acts on all three parts of the error. By decreasing the gain constant the controller can be slowed down and vice versa. By putting a negative gain constant the controller gives a higher control signal if the controlled value is

(32)

above the set point and thus works as a cooling control signal.

Even though the PID-parameters have their equivalents in real controllers the parameters cannot be transferred from a real case into a TRNSYS model which singles this Type out from others used in TRNSYS where the values have the same order of magnitude in the model as they would have in a real system. The reason for this is that the controllers are not responding in real time but rather in discrete time steps equal to the simulation time step. Therefore the PID-parameters are dependent on the simulation time step. Trimming the parameters in the model was quite time consuming. A large time step made it necessary to slow the controllers down since the controllers else would overcompensate and start to oscillate between time steps ultimately leading to convergence problems in the simulation. To eliminate the convergence problems the time step was reduced to as short as 10 s at some point. A further trimming of the PID parameters made it possible to increase the simulation time step to 20 s thus almost cutting simulation time in half. A key step was to have a very small gain constant(in the order of magnitude of 0.001) which took some time to realize since the Type 23 default value is 1.

2.4 Control strategies

The control strategies for the solar collector loop is subject to investigation in this thesis and are described in section 2.4.1. Besides the strategies coupled with the utilization of the solar collectors there are a couple of strategies that all aect the system behaviour and the temperature levels in dierent loops.

The most important control strategies in play in the system are described in the coming sections.

2.4.1 Solar collectors

The control strategies for the solar collector loop is the main focus in this study. Besides the basic control strategies already explained in section 2.1.1 which apply to all cases four dierent strategies have been tested.

Strategy A: Prioritize VS51-AK01

This strategy will be referred to as Prio AK01. It includes the condition that as long as the temperature coming from the solar collectors is below 35^◦C the hot storage tank VS51-AK02 will be by-passed and all the ow given to VS51-AK01 thus increasing evaporation temperatures of the heat pumps.

(33)

Strategy B: Prioritize VS51-AK02

This strategy will be referred to as Prio AK02. The condition implemented is that whenever the VS51-AK02 controller calls for heat, storage tank VS51- AK01 is by-passed. The control signal of VS51-AK01 is multiplied by 1 minus the control signal of VS51-AK02. This way tank VS51-AK02 is always prioritized and VS51-AK01 doesn't get any ow as long as VS51-AK02 does.

Strategy C: VS51-AK01 and VS52-AK02 in series

This strategy will be referred to as AK01-AK02 series. It includes no further conditions besides the basic control strategies. This means that each tank can call for ow regardless of what the other tank control signal is. This is the strategy that was rst planned to be used by the designing company. If the heat coming from the solar collectors is high grade enough it is used for DHW production. If the temperature after VS51-AK02 still is higher than the temperature in VS51-AK01 it is led through the tank and more energy is delivered to the system. This way solar collector energy can be delivered to the system at both temperature levels at the same time.

Strategy D: Only VS51-AK02

This strategy will be referred to as Only AK02. In this strategy the VS51- AK01 control signal is always 0 and no heat is transferred in the tank thus eliminating the possibility of the system to utilize solar collector as heat source for the heat pumps. The strategy was not included in the rst plan of testing.

Instead it emerged from the results of the other simulations. As the system will be built with booth tanks this strategy is unlikely to be implemented but it still proved to be an interesting strategy to investigate.

2.4.2 Summer/Winter denition

There is a need for the system to know whether it's summer or winter since dierent set points are used depending on the season. For example the space heating storage tank (VS01-AK01) has to keep a certain temperature in winter to be able to provide the correct forward temperature to the dierent space heating systems. In summer time this tank does not have to have a certain temperature but the heat pumps are still rejecting heat since they are used as chillers. Therefore the set point can be decreased and controlled by the dry-coolers to a lower tank temperature than during winter.

(34)

0 1

-20 -15 -10 -5 0 5 10 15

Control signal

24-hour mean ambient temperature Winter Summer

Figure 2.6: Control strategy for summer/winter case

Another example, which is connected to the solar collector controls, is that during summer when there is a need for cooling the solar collectors are not allowed to transfer any heat to the evaporators of the heat pumps. This would put the free cooling out of play and increase the run time of the heat pumps.

Therefore storage tank VS51-AK01 is only fed heat from the solar collectors when it is winter.

The break point for deciding the system is to be run in summer or winter operation is when the ambient temperature passes 12 ^◦C. To not get oscil- lations in the system due to constant changing between summer and winter(which can happen during spring and autumn months) a 24 hour mean value of the ambient temperature is used as input to the determining func- tion.

2.4.3 System temperatures for space heating and cooling

To ensure that the heat pumps are given the correct operating conditions the heating and cooling loads were divided by their respective required forward temperatures. The space cooling was set to require a constant forward temperature of 15^◦C regardless of outdoor temperature, which is the requirement of the cooling coils in the AHU's. The heating forward temperatures were controlled by outdoor temperature dependent functions. The forward temperature of the respective heating system for dierent outdoor temperatures

(35)

0 10 20 30 40 50 60 70

-20 -15 -10 -5 0 5 10 15

Forward temperature [°C]

Ambient temperature [°C]

Floor heating AHU heating coil

Figure 2.7: Forward temperatures for space heating systems

are shown in gure 2.7. By separating these systems a more realistic return as well as forward temperature for the heat pumps can be achieved thus giving the space heating heat pumps realistic operating conditions.

2.5 Result analysis

When analysing combined solar and heat pump systems there are numerous ways of assessing performance. Depending on the system conguration and the focus of study dierent performance gures can be suitable and as the results from this thesis will show there is a need for presenting several dierent performance indicators to fully be able to assess the performance of the system.

There is also a big dierence whether the economic performance of the system is in focus or if the focus is at some other aspect of the system, e.g. renewable fraction.

2.5.1 System boundaries

The system is modelled in it's entirety in order to simulate running conditions that are as close the real case as possible. Therefore some parts of the system is modelled but outside the scope of study in this thesis. Even though modelling the complete system has major advantages to simpler models it also imposes uncertainties to the results since many quantities implemented might

(36)

be unknown(e.g. pressure drop in secondary heating systems).

The study object in this thesis is the control and conguration of the solar collector loop and how it aects the system performance. Therefore only the components directly aected by the solar collector performance are included within the system boundaries. The performance is partly assessed by measuring the electricity use of the aected components for the dierent control strategies and storage tank sizes.

The system boundaries for performance assessment are set to include the solar collector loop and its circulation pump(VS51-PV01), the heat pumps for space heating and DHW production and their respective circulation pumps on condenser(VS01-PV01 and VS03-PV01) and evaporator side(KB01-PK01).

All other circulation pumps are omitted since their working conditions are not aected by the control of the solar collector loop.

In some performance gures there is a need of comparing the electricity use of the system to a reference system. The reference system in this thesis is dened as the same modelled system but with the solar collector circulation pump switched o. This way the system energy performance can be assessed without the inuence of the solar collector loop and the energy performance for dierent control strategies can be related to this reference case.

2.5.2 Performance indicators

As previously mentioned the performance of the system can be assessed using dierent indicators which all enlighten dierent aspects of the system performance. In order to present a complete picture as possible of the system performance a number of dierent performance indicators are used in this thesis project. The performance indicators are complemented by a general economic analysis. The indicators are presented in this section and a discussion of the benets and aws is presented in section 4.3.

Solar Fraction

Solar Fraction, SF , is a well known and commonly used performance gure used when analysing systems utilizing solar heat. The indicator describes the fraction of the total heating load in the system that is covered by solar collectors. Though well accepted, the denition of SF varies in the literature and there is a need for dening the indicator to not confuse it with other denitions. Malenkovic et al. (2013) makes a comprehensive study of dierent standards and their denition of SF . Common for all denitions presented is that only heat that is used directly is accounted for a solar heat and heat used

(37)

indirectly via a heat pump is omitted. The dierence between the dierent denitions is how they treat heat losses from storage tanks. A denition not accounting for heat storage losses can possibly have a solar fraction higher than 1. Hawlader et al. (2001) investigates a direct expansion solar heat pumps system with solar collectors as only source. Even though all solar heat is used indirectly via the heat pump the solar fraction is used and dened as the fraction of the collected solar heat to the total heating load.

SF = QSC

QSH+ QDHW (2.3)

where

QSC is the heat collected at the solar collectors [kWh]

QSH is the space heating load [kWh]

Q_DHW is the domestic hot water load [kWh]

The denition used in this thesis is shown in equation 2.3. No distinction is made between heat that is used directly or indirectly and the storage thermal losses are neglected.

Fractional energy savings

Fractional Energy Savings, fsav, is dened by Malenkovic et al. (2013) and expresses the eect a system conguration or control strategy has on the electricity use compared to the reference system. Electricity use is especially interesting since it is the only external energy input to the system. All other energy inputs in terms of heat from solar collectors or boreholes are free energy which is not charged in money. The denition used in this thesis is shown in equation 2.4.

fsav= Qel,savings

Qel,no solar

(2.4) where

Qel,savings is the dierence in electricity use between the reference sys-

tem and investigated control strategy/system conguration [kWh]

Qel,no solar is the electricity use of the reference system [kWh]

(38)

Specic electricity savings

Even though the Fractional Energy Savings quanties the saved electricity in comparison with a reference system it is also interesting to look at absolute numbers to clearly show the order of magnitude of the savings. Instead of presenting total savings the specic electricity savings per heated building area was chosen to be able to compare it to other installations if the reader wishes to do so. The denition is shown in equation 2.5.

qel,savings=Q_el,savings

Atemp (2.5)

where

Atemp is the total heated building area [m²] Solar savings eciency

The Solar savings eciency, ηSS, is dened in this thesis project and no support in literature or previous research has been found. The reason for dening this performance indicator is an attempt to dene the value the solar heat brought into the system has. The indicator is dened as the fraction of the electricity savings of a certain strategy or system conguration(as dened above) to the total heat brought into the systems via the solar collectors.

η_SS= Q_el,savings

QSC

(2.6) Although failing to present a good picture of system performance in some cases the performance indicator still introduces some interesting aspects of the eciency of systems that utilize renewable energy to reduce electricity consumption. The indicator is further discussed in section 4.3.

(39)

Economic analysis

As the project includes investigating dierent storage tank sizes the question of economic feasibility naturally occurs. An electricity price of 1.1 SEK/kWh excluding VAT is assumed and the price of storage tanks are taken as an av- erage from the price list of Götlinds Svets AB which is roughly 7000 SEK/m³ excluding VAT. Analysis is made by comparing simple pay-o values of dierent tank congurations and calculating a dierence in pay-o compared to the reference system. This is dened in equation 2.7. As the results will show the dierence in performance between dierent storage tank congurations turned out to be very small and a deeper economic analysis(e.g. discounted payback or another analysis including interest rate and lifetime) is not motivated.

P ayOf f = 7000 · ∆Vstorage

1.1 · ∆Q_el,savings (2.7)

where

∆V_storageis the combined storage tank volume increase for the two stor-

age tanks in the solar collector loop [m³]

∆Q_el,savings is the change in electricity savings [kWh]

Important to notice is that equation 2.7 does not show the pay o time for the solar collector installation but instead shows the pay-o time of adding extra storage to the solar collector loop. The equation also implies that adding extra storage always leads to electricity savings which for some control strategies have turned out to not be the case. This is discussed in section 4.4.

(40)

3 Result

The result from the simulations are presented in the following sections. The strategies are compared to each other as well as the variation of performance within a certain strategy due to dierent storage tank congurations. The main results, which will be shown in the description of the results, is that the choice of control strategy is dominating over the tank size choices in terms of system performance. Small dierences can be seen for dierent sizes of the storage tanks but the variation between control strategies is much higher.

3.1 Solar fraction

Using the denition presented in equation 2.3 the solar fraction of the system is presented in gure 3.1. The bars represent the mean value for the control strategy and the error bars shows the range within the control strategy for dierent storage tank sizes. Simulations were run for tank sizes between 2 and 9.5 m³ each in steps of 1.5m³ (and combining all possible combinations).

The solar fraction reaches from 0.13 for Prio AK01 and AK01-AK02 series to around 0.10 for Only AK02.

0 0,02 0,04 0,06 0,08 0,1 0,12 0,14 0,16

Solar Fraction [-]

Prio AK01 Prio AK02 Only AK02 AK01-AK02 series

Figure 3.1: Solar Fraction

(41)

3.2 Fractional energy savings

The fractional energy savings are shown in gure 3.2. fsav is dened in equation 2.4 and both heat pump and circulation pump electricity is accounted for(as dened by the system boundaries). This is one of the most interesting results since there seem to be a negative correlation between fractional energy savings and the solar fraction in this case. The strategies with the lowest solar fraction are the ones with the highest fractional energy savings meaning that more electricity is saved at the same time as the solar collectors harvest less heat.

As previously seen the eect of control strategies dominate the eect of dierent storage tank congurations. Variation between control strategies ranges from 0.066 for Prio AK01 and AK01-AK02 series to 0.093 for Only AK02 making the latter 40 % higher than the lowest fractional energy savings simulated.

0 0,02 0,04 0,06 0,08 0,1 0,12

Fractional energy savings [-]

Prio AK01 Prio AK02 Only AK02 AK01-AK02 series

Figure 3.2: Fractional energy savings [-]

3.3 Solar savings eciency

The solar savings eciency is shown in gure 3.3. As the fractional energy savings is increasing with lower solar fraction for the system the solar savings eciency follows the same pattern as the fractional energy savings but the variation between strategies is even higher than seen for the fractional energy savings. The highest eciency is seen for Only AK01 which has a maximum of 0.46 meaning that each kWh of solar heat harvested by the solar collectors

(42)

0 0.1 0.2 0.3 0.4 0.5

ηss

Prio AK01 PrioAK02 Only AK02 AK01-AK02 series

Figure 3.3: Solar savings eciency

leads to 0.46 kWh electricity savings for the system. The lowest eciency is seen for Prio AK01 of 0.23.

3.4 Solar collector behaviour

There is a major dierence in the performance of the solar collectors depending on the control strategy. Figure 3.4 shows the total collected heat at the solar collectors for the four investigated control strategies. The inuence of the storage tank congurations seems to be dependent on the control strategy.

Prio AK02 stands out in the way that storage tank sizes seems to have a big inuence where for the other strategies the inuence seem to be negligible.

The collected heat has a range of 32 % from the highest yield(Prio AK01 and AK01-AK02 series) to the lowest Only AK02. The previously shown Solar Fraction is directly correlated to the harvested solar heat which explains why the two gures (3.1 and 3.4) show the same pattern.

The main explanation for the large dierence in collected solar heat between strategies is that the control strategy highly aects the temperature level in the solar collector loop. Figure 3.5 shows hourly mean values of the outlet temperature from the solar collectors at two of the control strategies over one year. The temperatures are taken from simulations with the storage tanks in the solar collector loop being 2 m³each. During summer there is no dierence in outlet temperature since the two strategies have the same conditions during the summer and no heat is introduced into the low grade storage

Solar assisted ground source heat pump system - modelling and simulation