Master Level Thesis

(1)

Master Level Thesis

European Solar Engineering School

No.197, August 2015

Replacing Finned Copper with

Corrugated Stainless Steel, for

the Heat Exchangers of a Solar

Combisystem Store -

Performance and Economic

Evaluation

Master thesis 30 hp, 2015 Solar Energy Engineering Author:

(2)

(3)

I Abstract

The importance of investigating cost reduction in materials and components for solar thermal systems is crucial at the present time. This work focuses on the influence of two different heat exchangers on the performance of a solar thermal system. Both heat exchangers studied are immersed helically coiled, one made with corrugated stainless steel tube, and the other made with finned copper tube with smooth inner surface.

A test apparatus has been designed and a simple test procedure applied in order to study heat transfer characteristics and pressure drop of both coils. Thereafter, the resulting experimental data was used to perform a parameter identification of the heat exchangers, in order to obtain a TRNSYS model with its corresponding numerical expression. Also a representative small-scale combisystem model was designed in TRNSYS, in order to study the influence of both heat exchangers on the solar fraction of the system, when working at different flow rates. It has been found that the highest solar fraction is given by the corrugated stainless steel coil, when it works at the lowest flow rate (100 l/hr). For any higher flow rate, the studied copper coil presents a higher solar fraction. The advantageous low flow performance of stainless steel heat exchanger turns out to be beneficial for the particular case of solar thermal systems, where it is well known that low flow collector loops lead to enhanced store stratification, and consequently higher solar fractions.

(4)

II

List of Figures

Figure 1 Growth and installed capacity of solar thermal, PV and wind. Based on (Mauthner, et al., 2015) 1 Figure 2 Solar thermal and PV prices evolution. Based on AEE INTEC (2014) and Feldman, et al.

(2012) ... 2

Figure 3 Thesis outlook ... 4

Figure 4 Test apparatus used for testing of different heat exchanger configurations (Dahm, et al., 1998) (Published with permission of the authors) ... 12

Figure 5 Test apparatus used by Colorado et al. (2011) to test immersed HCHX´s... 14

Figure 6 Cold water temperature and daily DHW consumption from Persson (2004) and CEN (2000) .... 18

Figure 7 corrugated stainless steel and finned coper heat exchangers ... 20

Figure 8 Profile of the corrugated stainless steel pipe ... 21

Figure 9 Profile of the finned copper pipe ... 21

Figure 10 Test apparatus ... 23

Figure 11 T4 and outlet temperature. Start and end of test ... 26

Figure 12 UA value vs ΔTlm for copper, at the tested flow rates... 27

Figure 13 UA value vs ΔTlm for steel, at the tested flow rates ... 28

Figure 14 Effectiveness of both copper and steel, for all tests ... 29

Figure 15 Store stratification, when Ts,avg is 37 °C, for copper ... 30

Figure 16 Store stratification, when Ts,avg is 37 °C, for steel ... 30

Figure 17 Pressure drop ... 31

Figure 18 TRNEdit Parameter identification procedure ... 32

Figure 19 Parameter identification. Real and simulated outlet temperature of test S-300 ... 34

Figure 20 Parameter identification. Real and simulated heat to store of test S-300 ... 35

Figure 21 UA value vs flow rate vs (Ti + TNX) for the copper HX ... 36

Figure 22 UA value vs flow rate vs (Ti + TNX) for the steel HX ... 36

Figure 23 TRNSYS combisystem model diagram ... 39

Figure 24 TRNSYS solar loop diagram ... 39

Figure 25 TRNSYS DHW loop ... 40

Figure 26 TRNSYS SH loop ... 41

Figure 27 Simulation model energy balance ... 44

Figure 28 SF and Ti,AVG for both HX at all flow rates ... 45

(7)

V

Figure 30 EAUX and payback period for different HX lengths, assuming an electricity price of 1.5

SEK/kWh, and a heat exchanger price of 31.3 SEK/m ... 49

List of tables

Table 1 Profile description of the corrugated stainless steel pipe ... 21

Table 2 Profile description of the finned copper pipe ... 22

Table 3 Flow regime of the copper heat exchanger ... 22

Table 4 UA value increment due to higher flow rates. For both heat exangers ... 28

Table 5 Copper HX parameter identification results and accuracy ... 33

Table 6 Steel HX parameter identification results and accuracy ... 34

Table 7 Main settings and expected combisystem performance ... 38

Table 8 Type 12c set values ... 41

Table 9 ΔTON and ΔTOFF for the solar loop controller ... 43

Table 10 Final economic analysis ... 47

(8)

VI Nomenclature

γ Controller control function

δ Curvature ratio

ε Heat exchanger effectiveness

∈ Parameter identification fitting factor

η Efficiency

μ Dynamic viscosity [kg/m·s]

𝜌 Density [kg/m3_]

𝑣 Mean fluid velocity [m/s]

A Area [m2_]

bHX,1 Exponent for flow dependency

bHX,2 Exponent for temperature dependency

bHX,3 Exponent for temperature dependency

CP Isobaric specific heat capacity [J/kg·K]

d Diameter [m]

De Dean number

E Energy [Wh]

FHX Factor for time dependency

FR Collector heat removal factor

G Irradiance [W/m2_]

k Thermal conductivity [W/m·K]

l Length [m]

L Characteristic linear dimension [m]

𝑚̇ Mass flow [kg/s]

nHX N° of nodes occupied by the heat exchanger

NMAX Maximum number of nodes in Type 340

NTU Number of Transfer Units

P Power [W]

Ph Coil pitch [m]

Q Heat [Wh]

Q̇ Heat transfer rate [W]

Re Reynolds number

S Absorbed solar radiation per unit area [W/m2_]

SHX Start-up factor of the heat exchanger

t Time [s]

T Temperature [°C]

ΔTON Solar control on dead-band [K]

ΔTOFF Solar control off dead-band [K]

U Overall heat transfer coefficient [W/m2_·K]

V Volume [l ; m3_]

(9)

VII

UL Collector heat loss coefficient [W/m2·K]

UA UA value [W/K]

(UA)HX TRNSYS deck value of UA value [kJ/hr·K]

Acronyms

Cu Copper

CFD Computational Fluid Dynamics

CHS Collector Hardware Simulator

CW Cold Water

DN Diameter Nominal

DHW Domestic Hot Water

HX Heat Exchanger

HXHX Helically Coiled Heat Exchanger

PV Photovoltaic

R&D Research and Development

SDHW Solar Domestic Hot Water

SEK Swedish krona

SF Solar Fraction

SH Space Heating

Ton Tonne (1000 kg)

Under scripts

a Ambient

aCW Temperature amplitude of cold water

aCW Mean value of cold water

c Heat exchanger cold side

calc Calculated

col Collector

EL Electrical

h Heat exchanger hot side

(10)

(11)

1

1 Introduction

1.1 Background

Among the different solar energy technologies, solar water heating has attained a high level of technological development and can provide an advantageous correlation between cost and effectiveness from a life cycle perspective (World Energy Council, 2013, p. 8.7). By the beginning of 2014 the total installed capacity worldwide accounted for 374.7 GWTH, above

other renewable energy technologies such as PV and wind. The major part of the power installed, about 84%, corresponds to small scale solar water systems installed in single family houses (Mauthner, et al., 2015).

Concepts such as the future 2020 European regulation, “Nearly Zero Energy Buildings”, enhances the use of alternative energy sources, where solar thermal is expected to play an important role, thus the R&D efforts have been focused on cost reduction, increased solar fraction per building and improved reliability (European Technology Platform, 2012).

Despite the positive perspectives, the market development of the solar water heating sector haven’t been promising during the last years. Renewable electricity generation technologies, mainly photovoltaic (PV) systems have taken part of the market share that corresponded to solar thermal. Figure 1 shows the yearly growth of solar thermal, PV and wind sector.

Figure 1 Growth and installed capacity of solar thermal, PV and wind. Based on (Mauthner, et al., 2015)

The reasons behind the slowdown shown in the previous graph are diverse. Uncertainty in the future of heating technologies, policies and politics are part of the reasons, nevertheless the price development of solar thermal systems may be the most significant (AEE INTEC, 2014). Figure 2 show a stagnation in system prices during the last years, which contrasts with the drastic drop of PV systems price.

(12)

2

Figure 2 Solar thermal and PV prices evolution. Based on AEE INTEC (2014) and Feldman, et al. (2012) The previous figure confirms that cost reduction is one of the main research priorities in the solar thermal field, as stated in several reports. This reduction in the system cost would logically lead to a reduction of the initial investment required and payback period, thus attracting investors. Therefore the research in cost optimization is crucial for the future of the solar thermal technology, where the cost reduction must not compromise (if not improve) the performance of the system.

The maturity of the solar thermal technology makes the latter a challenging task. Nevertheless, several publications, such as European Technology Platform (2012, p. 27), whichsuggests that there is still gap of improvement in performance with a forecasted 10 % increment for 2020 and cost with a forecasted 50 % reduction for 2020. Such improvement is to be attained by means of using new materials, new design and manufacturing processes. Thus, the aforementioned reasons, combined with the increase of most of the commodities prices occurred in the last years, motivates the cost optimization of the solar thermal systems by means of using materials alternative to the traditional.

1.1.1 Solar combisystem and heat exchangers

Solar combisystems are solar thermal systems that supply both Domestic Hot Water (DHW) demand and Space Heating (SH) demand, which main components are the collector array and the thermal store. This type of systems are common in central Europe, where the SH demand is significant European Technology Platform (2012, p. 16).

When analyzing a solar combisystem, it turns out to be evident that apart of the solar collector, the components which ensure a good heat transfer and energy storage are key factors in the total performance of the system. Heat exchangers are devices used for any heat transfer process between two different fluids, and have an important impact in the overall system performance (Naphon & Somchai , 2004), affecting the number of on-off operations of the boiler and pumps (Figueiredo and Raimundo, 1996), as well as the heat transfer in the collector and tank stratification (Duffie & Beckman, 2006). The majority of the solar combisystems uses

(13)

3

heat exchangers for harnessing the solar energy, since the circulating is not only water, but a mixture of water and antifreeze (Drück, 2006, p. 2).

The most common type of heat exchanger (HX) used in small solar thermal systems are the Helically Coiled Heat Exchangers (HCHX), due to their simplicity, compactness and good thermal efficiency (Fernández-Seara, et al., 2013) and their large surface area per unit volume (Colorado, et al., 2011). Traditionally HCHX have been made with copper, due to its excellent thermal conductivity (k, 201 W/mK) that makes copper the best option available in the market, regarding heat conductivity. Nevertheless, as stated before the commodity prices increase occurred the last years have specially affected the copper price, which rose from 2000 $/Ton in 2000 to 7000 in 2014 $/Ton (Wieland, 2014), leading the manufacturers to find alternative materials for the heat exchangers in order to improve the cost-effectiveness without compromising the performance.

Among the possible alternatives, stainless steel seems to be the optimum candidate. It produces less fouling depositions (Kazi, et al., 2012), and despite its much lower thermal conductivity (k, 16 W/mK), several properties makes it a suitable option, and in fact it is already widely spread as heat exchanger material.

1.1.2 Stocksbro

Stocksbro is a Swedish manufacturer of hot water stores with immersed HCHX of copper, which can use several different energy sources (solar, pallet boiler, oil-electrical-gas heater). In this context, it is required to study the suitability of using stainless steel HX´s as replacement of copper HX´s.

1.2 Aim and objectives

The aim of this Master Thesis is to study the influence in the performance of a solar combisystem, caused by the change of the immersed HCHX material, from finned copper tubes with smooth inner surface to corrugated stainless steel tubes. Also economic figures are used to compare both heat exchangers. Furthermore, an optimized stainless steel HX length has to be found, in order to provide the manufacturer with the required information for replacement. Several objectives have to be accomplished in order to meet the desired aims. Development of a test-rig and test procedure for the studied HX´s, which allows to properly measure and identify their thermal performance and also study the pressure drop at different conditions of flow rate and temperature.

(14)

4

1.3 Method and tools

The work carried out in this project can be divided according to the objectives mentioned above. Three main parts can be distinguished, together with a literature review that encompasses all of them, the process is described in Figure 3 and below.

Literature review

As shown in the last figure, the literature review encompasses every step in this project. Providing guidance and argumentation to the decisions made. The main topics of literature review are described in Chapter 2.

Experimental: Setup design, procedure and HX tests

The first step in this study was the design of the test-rig, which allowed the HX testing under different conditions of temperature and flow rate. For which the test-rig required control and logging systems and a testing procedure that enabled reproducibility and repeatability of the tests, for the sake of a proper comparison between the two heat exchangers under study. This section of the project is explained in detail in Chapter 3.

TRNSYS

For the simulation part of the study, it was essential to use a software that enables a detailed modelling of the combisystem store and its HX´s. After analyzing the different options available, TRNSYS was the software selected due to its suitability for parameter identification, since it provides plenty of components models that can be configured with high degree of detail. Solar thermal systems are within the main applications available in TRNSYS.

Parameter identification

The behavior of a dynamic simulation component is described by numerical models. Thus, the result of such numerical models will depend on the model itself, the input values to the model, and the parameters used for the calculation (Drück, 2006, p. 12). Therefore, the parameter identification is the process by which the values of those parameters are found, so that the model represents properly the component to be modelled. In this study it is the link between the experimental and simulation procedures. The process was based on the identification of four of the parameters that rules the heat transfer coefficient (UA) equation of the solar loop

Test Apparatus Design

Test Procedure

Parameter

Identification Combisystem simulation

Final Results Literature Review

(15)

5

heat exchanger HX(Eq. 15) and was done by means of a TRNSYS parametric study. This is described in detail in Section 5.1.

Combisystem model design and simulation

Once the HX´s were modelled, the final step was to obtain results regarding the effect of each HX in a combisystem. For this, a TRNSYS combisystem model was designed, paying special attention to the collector loop design and hot water store. The process is described in Section 5.3. This last part of the methodology provided the final results, which complied with the aim of this thesis.

1.4 Boundary conditions

Time frame, and resources were a constraint in this work, and set the limitations which are described below.

1.4.1 Limitations in the experimental procedure

The main limitation of the experimental procedure is that only two heat exchangers were tested (finned copper and corrugated stainless steel), limiting the comparison options. Other limitation is the fact that only one of the heat exchangers of the testing store (described in Chapter 3) was tested, which corresponded to the solar loop. Since the focus of this thesis was to study the influence on a solar thermal system, the test of the DHW HX was neglected. Furthermore, the control of the test-rig allowed very basic control of the temperature inlet to the HX, as well as control of flow rate. It is assumed that a better solution is possible, which would allow to perform more realistic and accurate testing procedures. The same reason applies for the pressure drop study, which accuracy is difficult to determine, and the used method was a solution that accomplished with the time and resources available.

All tests were carried out with water, which implies a difference with real systems, where the fluid through the heat exchanger is a mixture of water and glycol.

Regarding the data logging, values of flow rate were not logged, but controlled throughout all the tests. This didn´t represent a major inconvenience to the development of this work, since each one of the testing sequences were performed at constant flow rate. Nevertheless, the logging of flow rate values, would represent a substantial improvement in the quality of the results, giving the chance to account for small variations in flow rate that occurred during the tests.

1.4.2 Limitations in the simulation procedure

In Section 5.2 it is mentioned the accuracy of the resulting parameter identification, which presents margin of improvement. Nevertheless, the final result was considered to be sufficient, considering the aim of this project. Also the validation range of the resulting HX models are unknown for some of the temperature values that may occur during the simulation.

(16)

6

Other limitation of the simulation procedure is the number of nodes considered in the store. 10 nodes were considered, which may affect the accuracy of temperature calculations. But since the main objective of the developed model was to compare performances, rather than provide accurate absolute values, 10 nodes was considered to be an acceptable number.

(17)

7

2 Literature Review

2.1 Immersed HCHX in solar thermal systems

There is a large amount of publications available regarding the performance of immersed HCHX´s, since they are extensively used in heat transfer applications, such as refrigeration, steam generation, chemical processes, nuclear reactions and domestic hot water. Among the different heat exchanger configurations that can be applied in small DHW stores, coiled heat exchangers are the most common, due to its compactness (great heat transfer area per unit of volume), simple manufacturing and maintenance, and good thermal efficiency (Fernández-Seara, et al., 2013).On the other hand HCHX enhances the inner convection heat transfer, when comparing with straight tubes HX´s (Prabhanjan, et al., 2002).

In HCHX´s, the curvature of the pipe causes a centrifugal force gives rise to a second flow pattern, which is perpendicular to the main axial flow. The fluid then moves from the inner wall of the tube, across the center, to the outer wall, and moves back to the inner wall through the pipe inner surface (as can be seen, two differentiated paths, from inner wall to outer, and from outer to inner wall). This secondary flow moves the fluid across the temperature gradient, and therefore enhances the heat transfer coefficient and the rate of heat transfer per unit length, which is considerably higher than the obtained with straight tubes, where this additional convective heat transfer mechanism doesn’t exist (Rennie & Raghavan, 2005) (Satapathy, 2009). The existence of this secondary flow was reported for the first time by Eustice (1911) by injecting ink into the flowing water as it is cited by Piazza & Ciofalo (2010, p. 1).

As stated before, the secondary motion is caused by the imbalance between the centrifugal and inertial forces in the cross section of the pipe. From the research carried out by Dean (1927) a new governing parameter emerged, the Dean Number (De, defined in the following equation),

which couples together inertial and centrifugal effects as cited by Piazza & Ciofalo (2010, p. 1).

𝐷_𝑒 = 𝑅_𝑒 √𝛿 (1)

Where δ is the curvature ratio, defined as:

𝛿 = 𝑑𝑖/𝑑𝑐𝑜𝑖𝑙 (2)

Where di is the internal tube diameter and dcoil is the coil diameter.

And Reynolds number is defined as:

𝑅𝑒 =

𝜌 𝑣 𝐿 𝜇

(3)

Where

𝜌 is the density of the fluid. 𝑣 is the mean velocity of the fluid

L is the characteristic linear dimension. For smooth tubes, L is equal to di.

(18)

8

Similar to the Reynolds number for straight pipes, the Dean number is used to characterize the flow in a helical coiled pipes (Jayakumar, 2012). Jayakumar (2012) stated that it has been found that one effect of the coil curvature is that it suppress the formation of turbulent fluctuations of the circulating fluid. Therefore, it has been seen that for helically coiled pipes the transition from laminar to turbulent flow (critical Re number) occurs at higher Re values than in straight

pipes. On the other hand, it has been proved that the aforementioned turbulence suppression effect is enhanced with higher values of δ (Jayakumar, 2012, p. 312).

Jayakumar (2012, p. 314) presents the results of several researches, regarding the laminar to turbulent transition. Based on those researches, Jayakumar (2012, p. 314) presents a correlation between Re number and δ for helically coiled pipes, based in De number. This correlation is

used in Section 3.1 to estimate the flow regime of the tested heat exchangers, which is crucial to the discussion of the experimental results.

Many other dimensionless quantities are used for describing the fluid dynamic processes of coiled pipes, and the studies dealing with them reaches high degree of specialization and complexity. Therefore, considering the scope of this study, only studies dealing basic concepts of heat transfer and fluid dynamics were searched in the literature, especially those dealing with HCHX and solar thermal systems.

2.1.1 Heat transfer in HCHX

The heat transfer rate through the coil is the result of three different parameters. The inner and outer convection processes, the conductivity of the wall and the fouling resistance of the coil. Which means that for a given coil the heat transfer rate is established only by the outer and inner heat transfer coefficients, apart from the fouling resistance (Fernández-Seara, et al., 2013).

Plenty of publications are focused on the description of the heat transfer processes occurring in an immersed HCHX´s. Such publications deals with high level of calculation complexity, where methods such as artificial neural networks (Colorado, et al., 2011) or numerical methods (Fernández-Seara, et al., 2007) are necessary in order to describe the heat transfer processes in HCHX´s. The main difficulty in these studies is that it is extremely complex to numerically characterize the natural convection process between the heat exchanger and the stored water, which stablishes the outer heat transfer coefficient. Other simplified approaches are present in the literature, such as simplified theoretical models (Figueiredo & Raimundo, 1995), where commonly it is considered a constant wall temperature or a constant heat flux as boundary conditions.

(19)

9

Section 2.2 and 2.3.1 of this chapter, offers more information regarding the UA estimation value of HCHX´s.

Effectiveness-NTU method

The calculation of the effectiveness (ε) and NTU is convenient for solar thermal systems, and these dimensionless parameters are used for several calculations, such as the estimation of temperature dead-band for the solar loop controller, as described in Section 2.3.3. The maximum heat transfer (Qmax) through a flat plate heat exchanger is given by the following

equation (Duffie & Beckman, 2006, p. 169).

𝑄_𝑚𝑎𝑥 = (𝑚̇ ∗ 𝐶_𝑝)_ℎ∗ (𝑇_ℎ𝑖𝑛− 𝑇_𝑐𝑖𝑛) (4) Where

(Cp)h is the heat capacity of the fluid in the hot side.

Thin is the input temperature of the hot side.

Tcin is the input temperature of the cold side.

Which for a HCHX under charging conditions can be written as:

𝑄𝑚𝑎𝑥 = (𝑚̇ ∗ 𝐶𝑝)ℎ∗ (𝑇𝑖𝑛− 𝑇𝑁𝑜𝑢𝑡) (5)

Where

Tin is the inlet temperature of the heat exchanger.

TNout is the tank temperature at the height of the outlet of the heat exchanger. Which in other

words is the temperature of the tank node at the height of the heat exchangers outlet.

Then the effectiveness (ε) can be defined as the ratio between the actual heat transferred (Q) and the maximum possible (QMAX) (Duffie & Beckman, 2006, p. 170).

𝜀 = 𝑄 𝑄𝑚𝑎𝑥 = (𝑚̇ ∗ 𝐶𝑝)𝑚𝑖𝑛∗ (𝑇𝑖𝑛− 𝑇𝑜𝑢𝑡) (𝑚̇ ∗ 𝐶𝑝)ℎ∗ (𝑇𝑖𝑛− 𝑇𝑁𝑜𝑢𝑡) (6) Where

(Cp)min is the lowest heat capacity of the fluids.

Tin is the inlet temperature of the heat exchanger.

Tout is the outlet temperature of the heat exchanger.

Giving a working equation for the heat exchanger as:

𝑄 = 𝜀 ∗ (𝑚̇ ∗ 𝐶_𝑝)_ℎ∗ (𝑇_𝑖 − 𝑇_𝑜) (7)

Then, the NTU will be given by the following equation (Duffie & Beckman, 2006, p. 170)

𝑁𝑇𝑈 = 𝑈𝐴

(𝑚̇ ∗ 𝐶_𝑝)

(8)

(20)

10

into the collector, as stated by Duffie & Beckman (2006, p. 433), and as it is implemented further down in this report, in Section 2.3.3.

2.1.2 Influence of geometrical parameters

Allocation in the store, dimension and shape of the heat exchangers are key in determining their thermal behavior. Ali (2004) studied the free convection heat transfer from the outer surface of HCHX´s, finding an enhancement of the overall heat transfer coefficient value by increasing δ or the number of coil turns, and keeping constant the rest of the geometrical parameters.

Fernández-Seara et al. (2013) developed and validated experimentally a numerical model, in order to study the influence of the main representative geometrical parameters in the heat transfer and pressure drop of a HCHX of a water storage tank. The representative geometrical parameters of the coil are tube outer diameter, coil diameter, pitch and length. Natural convection of the tube outer surface is considered as boundary condition. Being coil pitch the distance between the helix turns.

The results showed that an increment of the tube diameter causes a larger heat transfer rate to pressure drop ratio. On the other hand, it is asserted that the heat transfer is independent of other geometrical parameters, for a given value of inner area. Also it was found a strong influence of increasing diameter and decreasing pitch, in order to reduce the pressure drop. 2.1.3 Effect on Solar Thermal Systems

The effect of the HX performance in the overall efficiency of the solar thermal system is significant. The following equation represents the useful energy gain (Qu) from a flat plate

collector (Duffie & Beckman, 2006, p. 265).

𝑄𝑢 = 𝐴𝐶𝑜𝑙 𝐹𝑅 [𝑆 − 𝑈𝐿 (𝑇𝑖𝑛− 𝑇𝑎)] (9)

Where

Acol is the collector aperture area

S is the absorbed solar radiation per unit area UL is the collector overall heat loss coefficient

FR is the collector heat removal factor, which is defined in the following equation (Duffie &

Beckman, 2006, p. 264).

𝐹𝑅 =

𝑚 ̇ 𝐶_𝑝(𝑇_𝑖𝑛− 𝑇_𝑜𝑢𝑡) 𝐴𝑐𝑜𝑙 [𝑆 − 𝑈𝐿 (𝑇𝑖𝑛− 𝑇𝑎]

(10)

The heat removal factor FR relates the useful energy gain of the collector to the useful energy

gain if the whole collector was cooled down to the same temperature as the inlet fluid (maximum possible heat gain). The latter equations are very important for the performance of a solar thermal system based on flat plate collectors, and they determine the instantaneous efficiency, described as follows (Duffie & Beckman, 2006, p. 291).

𝜂 = 𝑄𝑢 𝐴𝑐 𝐺𝑇 =

𝐹_𝑅 [𝑆 − 𝑈_𝐿 (𝑇_𝑖𝑛− 𝑇_𝑎)] 𝐺𝑇

(11)

(21)

11

removal factor, consequently increasing the energy gain and thus the instantaneous efficiency. This is explained in detail by Duffie & Beckman (2006) in the Section 6.7 and gives a clear explanation of the importance of a good heat transfer in the store.

Apart of collector efficiency, the solar loop heat exchangers affect the store performance. Thermal stratification is an advantageous phenomena present in water tanks, which has positive influence in the performance of solar thermal system and storage efficiency. It consist in keeping hot water (top) and cold water (bottom) separated by gravitational stratification. Store geometry and type of charge-discharge process are among the main factors affecting stratification (Wahiba, et al., 2013). This was demonstrated by Hollands & Lightstone (1989) research, where it was shown that a store delivered 37 % more energy when was working under stratification, when compared with a fully mixed tank.

The temperature at the top of the tank is the result of many variables, among them, the heat exchanger type and flow rate of the solar loop. Furbo & Knudsen (2006) demonstrated that for low flow solar thermal systems, the resulting temperature at the top of the tank is close to the load temperature, when compared with systems of high flow, thus reaching a higher level of stratification, lower auxiliary consumption and higher solar fractions.

Plenty of studies highlight the advantage of low flow rates in solar thermal systems. Cristofari, et al. (2002) mentions three main advantages of low flow solar thermal systems.

 Higher stratification, which leads to higher SF

 Smaller piping diameter

 Less energy consumption for pumping

Logie, et al. (2010) studied the heat transfer performance and stratification of three immersed HCHX´s of different material and arrangements, for solar thermal systems. The main conclusion was that it was found a correlation between high convective heat transfer coefficient and low stratification. The reason behind this was that high flow rates enhanced the convective heat transfer due to turbulent flow, but on the other hand disrupted the stratification.

2.1.4 Finned copper tubes vs corrugated stainless steel tubes

The main topic of this thesis is focused in the comparison of a finned copper HCHX and a corrugated stainless steel HCHX. It was not found in the literature any study regarding this particular comparison. Nevertheless many publications study finned or corrugated separately. It is proven that corrugation in tubes enhances the convective heat transfer. Zachár (2010) carried out a CFD analysis on corrugated HCHX´s, in order to study the heat transfer enhancement of the corrugation patterns, and it was proved that the ratio of corrugation depth to tube diameter directly proportional to the heat transfer.

(22)

12

conductivity. Among the studied materials (copper, stainless steel, aluminum and brass) it was found that copper showed the highest scaling, while stainless steel the lowest.

2.2 Testing Methods

For the development of the testing apparatus required for this project, the literature review focused on publications that described the testing apparatus used, as well as the testing parameters such as flow rates, temperatures and other conditions. The literature available regarding this topic is extensive, therefore in order to narrow down the search, special emphasis was placed on investigations that involved immersed HCHX´s in solar thermal systems, although some investigations regarding mantle heat exchangers are also mentioned, since they provided useful information to design the test methodology. The following is a short compilation of the most relevant papers that have influenced this project, and which were the basis for the experimental methodology described in Chapter 3. Present performance test standards for solar thermal systems, such as BS EN 12977-3:2012 (British Standard Institution, 2012) are not considered in this literature research.

Dahm, et al. (1998) evaluated small solar thermal stores under different immersed HCHX configurations. The test sequence applied was a six-day indoor storage test, with representative conditions for the Swedish summer (no SH heating load). The aim of the testing apparatus developed was to test the stores and heat exchangers, as shown in Figure 4.

Figure 4 Test apparatus used for testing of different heat exchanger configurations (Dahm, et al., 1998) (Published with permission of the authors)

(23)

13

The size of the store tested was 750 l, and the size of the collector field emulated by the CHS 10 m2_{. The DHW load was set as 13 kWh daily load, which corresponded to the consumption}

of a Swedish household with 5 to 6 occupants. Regarding the test sequence, the initial temperature of the store was 20 °C, with a fully mixed tank. In order to obtain data regarding flow dependency of heat exchangers, three different flow rates were tested.

Knudsen and Furbo (2004) carried out experimental and numerical (CFD) analysis of mantle heat exchangers for SDHW systems, in order to know the influence in the thermal stratification, for two mantle inlet positions (top inlet and bottom inlet).

For the experiments in the indoor testing facility, water is used as mantle fluid with a constant flow rate of 0.4 l/min, starting from an initially stratified tank, with temperatures of 20 °C at the bottom and 69 °C at the top. The testing consists of a first heating period of 2 h with an inlet temperature of 70 °C, a second period of resting during 1 h and finally a 2 h period with 50 °C of inlet temperature (in total 3 h of testing).

Temperature sensors were placed in seven points inside the tank, as well as in the inlet and outlet positions of the mantle heat exchanger along with a flow meter, to measure the flow rate through the mantle. In this way, the tank is divided in 8 volumes (also called nodes), in a similar way as it was done by Dahm, et al. (1998), in the study described above. The vast majority of papers that study stratification in solar stores use store internal temperature sensors, being the measurement of each one of them representative for the temperature of the corresponding control volume (node).

The energy supply to the storage during the testing time (t) is given by the following equation (Knudsen & Furbo, 2004).

𝑄𝑠𝑡𝑜𝑟𝑎𝑔𝑒 = ∫ 𝑚̇ 𝑐𝑝(𝑇𝑖𝑛− 𝑇𝑜𝑢𝑡) 𝑑𝑡 𝑡𝑒𝑛𝑑

𝑡𝑠𝑡𝑎𝑟𝑡

(12)

Where ṁ is the flow rate through the mantle.

The performance of SDHW systems with mantle heat exchanger was also studied Soo Too et al. (2008) outdoor and controlled indoor conditions. The installation of sensors was similar to the one used by Knudsen & Furbo (2004), with measurement of inlet and outlet temperatures, flow rate and temperature of the nodes inside the store. The UA value of the mantle heat exchanger was calculated as follows (Soo Too, et al., 2008, p. 354).

𝑈𝐴_ℎ𝑥 = 𝑄̇

∆𝑇_𝑙𝑚 (13)

Where ΔTlm is the log mean difference between the hot and cold fluids of the heat exchanger,

and it is calculated as follows (Soo Too, et al., 2008, p. 354).

∆𝑇_𝑙𝑚 = (𝑇𝑖𝑛 − 𝑇𝑁𝑖𝑛) − (𝑇𝑜𝑢𝑡− 𝑇𝑁𝑜𝑢𝑡) 𝐿𝑛 ( 𝑇𝑖𝑛 − 𝑇𝑁𝑖𝑛

𝑇𝑜𝑢𝑡 − 𝑇𝑁𝑜𝑢𝑡)

(24)

14

The log mean difference is widely used for calculating the heat transfer coefficient in heat exchangers, and it is found in many publications. These last equations are used later on for the processing of the obtained experimental data, as it is done later on in Section 4.2.

Colorado et al. (2011) designed both physical and empirical models in order to describe the heat transfer of a HCHX. The experimental part of the study, which configuration is depicted in Figure 5, was carried out with a Copper-Zink alloy HCHX, with a thermal conductivity of 111 W/m ºC and different geometries. In all the tests the coils were heated up by water flowing at turbulent regime. Plexiglas was used for separating the coil turns with a specific distance to fix the pitch of the coils such that the ratio of Pb/do=2.5 for all the coils to be

tested. A schematic based on the test apparatus used by Colorado, et al. (2011) is depicted further down in Figure 5.

The hot water of the isothermal reservoir tank (left) was heated up to 80 ºC, thermocouples where placed in the coil, in order to measure the inlet and outlet temperature. The starting temperature of the coil tank (right) had an initial value of 5ºC and fully mixed (Colorado, et al., 2011). The design of this experiment had as objective to calculate the external overall average heat transfer coefficient, by using the internal (predicted) heat transfer coefficient, the measured temperatures and the known coil thermal conductivity. The flow rate applied was 0.833x10-4_m3_{/s (300 l/hr).}

Plenty of other researches follow similar methodologies, such as the studies carried out by Fernández-Seara et al. (2013) for the validation of a numerical model with experimental data, or Prabhanjan, et al. (2004), in order to study the outer heat transfer of copper helically coiled tubes, where the tests were performed in discharge process (circulating fluid being heated up in the HX), and three different flow rates were applied.

For many of the stages of this project, the value of density and heat capacity of water had to be adjusted according to the temperature of the fluid. Whenever it was possible, the following correlations were used for the calculation (DeWitt & Incropedia, 1996).

𝜌_𝑤 = 0.00001732224 𝑇3_{− 0.006092642 𝑇}2_{+ 0.01617533 𝑇 + 1000.201 (15)}

And for the heat capacity. Mixing HE Heater Pump Flow meter Test tank

(25)

15

𝐶_𝑝𝑤= 0.000003276 𝑇4_{− 0.0008014 𝑇}3_{+ 0.0772 𝑇}2_{− 2.964𝑇 + 4295.98} ₍₁₆₎

2.3 Simulation literature

Many different simulation approaches can be taken to study the performance of heat exchangers, hot water stores and solar thermal system, among them CFD are numerous. Nevertheless, in view of the aim of this project the search focused on publications of TRNSYS based simulations.

2.3.1 Type 340

Type 340 is a TRNSYS model of a stratified fluid storage tank with up to four heat exchangers, an internal electric heater and up to ten double ports for direct connection. Type 340 allows a flexible configuration, and it has been designed for simulation of solar thermal systems (Drück, 2006).

Type 340 works by simulating N fully mixed nodes of equal volume. The store can be charged and discharged indirectly, by means of the internal heat exchangers, or directly through the double ports. The heat loss of the store can be considered with individual values for the bottom, the top and four zones of the mid-section.

The instantaneous value of the head transfer capacity rate (UA)*HX can be specified for each

one of the up to four internal heat exchangers. The mathematical expression from which (UA)*HX is calculated is given by the following equation (Drück, 2006).

(𝑈𝐴)_ℎ𝑥,𝑠∗ 𝑛ℎ𝑥 = (𝑈𝐴)_ℎ𝑥,𝑠 𝑛ℎ𝑥 𝐹ℎ𝑥 𝑚̇ℎ𝑥 𝑏1_[𝑇 𝑖𝑛− 𝑇𝑁,𝑥]𝑏2 [ 𝑇_𝑖𝑛 + 𝑇_𝑁,𝑥 2 ] 𝑏3 (17) Where

nhx is the number of nodes occupied by the heat exchanger

(UA)HX is the heat transfer capacity rate from TRNSYS deck, in kJ/hrK.

TNX is the temperature of the corresponding tank node x, in °C.

b1 is the mass flow dependency exponent.

b2 is the exponent for dependency on the temperature difference between HX inlet and store.

b3 is the exponent for dependency on the mean temperature between the HX inlet and store.

Fhx is the time dependency factor.

The aforementioned time dependency factor accounts for the increase of natural convection with time caused by the inertia of the fluid in the tank, which results in an increment of (UA)HX. It is described by the following equation (Drück, 2006).

𝐹_ℎ𝑥,𝑡 = 1 𝑆_ℎ𝑥 ∫ 𝑚̇ 𝑡=𝑡−1 𝑡=0 (1 − 𝐹_ℎ𝑥) 𝑑𝑡_ℎ𝑥 (18) Where

(26)

16

Since the power transferred through the solar heat exchanger is relatively low when compared with the DHW heat exchanger, during the start-up the time dependency for the solar HX can be neglected. Nevertheless this doesn´t applies for the DHW loop, where the typical values are between 0.01 and 0.05 (Drück, 2006).

Type 340 allows modifying the number of nodes to consider NMAX, which ultimately defines

the degree of stratification in the store. There are different methods to calculate NMAX, but

Drück (2006) states that for well stratified stores, commonly used in solar thermal systems, the value of NMAX should be between 30 and 1000 to be realistic.

2.3.2 Parameter identification

“The results calculated by a numerical model during a dynamic simulation depend on the model itself, the input quantities fed into the model and the parameters used for the calculation. One way of determining certain parameters for the model is the numerical method of parameter identification” (Drück, 2006, p. 12), which requires measured data of the component.

The literature regarding TRNSYS parameter identification of solar thermal systems presents many examples where the components of a solar thermal system are modelled according to real test data.

Dahm, et al. (1998) developed several test sequences (test apparatus is described above in Figure 4) for individual components which served as basis for the posterior parameter identification in TRNSYS. The UA mass flow dependency parameter (b1 described in Eq. 17)

of the load side heat exchangers was identified from data obtained at three different flow rates. Nevertheless, the flow dependency of the UA value of the collector side heat exchanger was considered constant, since the mass flow of the collector loop in a real system has little variation during working conditions.

The validation process was carried out by a re-simulation of a six-day sequence test. The validation figures used were the solar fraction and the difference between the heat exchanger transferred energy in the real test and in the model, which was calculated as follows (Dahm, et al., 1998, p. 411).

∈ = 𝑄𝑐𝑎𝑐𝑙− 𝑄𝑚𝑒𝑎𝑠

𝑄𝑚𝑒𝑎𝑠 ∙ 100 % (19)

For the validation of the solar heat exchanger, it was found that the solar heat exchanger presented a v value of ± 3%, then the heat exchanger was not considered validated according to the standards valid at that time (ISO/DP 9459-4A, 1996). A similar approach is used in this work, in Section 5.2, to calculate how well the parameter identification fits to the measured values.

Dahm, et al. (1998) used the TRNSYS MULTIPORT Store Model, which is a former version of the Type 340. For the case of Type 340, which is the store model used in this study, the parameter identification focuses on four of the parameters that rules Eq. 15, b1, b2, b3, (UA)HX

and SHX.

(27)

17

140. It is shown that for heat exchangers the identified exponents of the UA expression are fairly constant despite of the change of system size. It is also mentioned that for heat exchangers that were not tested, the UA was assumed to be directly related to the area of the heat exchanger. This same assumption is taken in this thesis, as mentioned in Section 5.3, for the modelling of the DHW heat exchanger, which was not tested, and also it is applied in Section 6.1.5, for the heat exchanger length optimization process.

Regarding the tool utilized for the parameter identification process, the available literature shows that the main tool used is GENOPT-TRNOPT optimization tool.

2.3.3 Combisystem model design

As it is later on mention in Chapter 5, during the TRNSYS simulation modelling one of the requirements of the developed model was that it had to be representative of real combisystems, despite of its simplicity. The studies that served as basis for the model design are described in this section, where the main topics revised where sizing, expected output, load and control of combisystems.

Size and expected output

The relation between collector area (Acol) and store volume (Vs) was studied by Rodríguez-

Hidalgo, et al. (2012) in order to find an optimum store size for solar DHW systems located in Spain. The optimum value of Vs/Ac was found to be between 0.05 and 0.08. Nevertheless,

sizing of solar DHW systems and combisystems may vary significantly. Lund (2005, p. 61) mentioned that normal small scale domestic combisystems accounted an area Ac between 8 to

30 m2_{corresponding to a store V}

s of 300 to 1500 l respectively. While for solar DHW, Ac of 4

to 8 m2_{corresponded to a V}

s of 200 to 500 l.

Lundh, et al. (2010) studied the influence of store size on medium-size solar combisystems. In this research, it is mention that the solar fraction of combisystems normally covers between 20% and 25 % of the total energy consumption (Lundh, et al., 2010, p. 1095).

Control

The control features present on the store are crucial in the overall energy gain of the solar thermal system. Furbo, et al. (2005) has proven that the application of smart control measurements (smart tank) can improve the performance of small SDHW tanks between 5 % and 35 %.

In this project, the control strategy used in the solar loop of the developed TRNSYS model are based on the information presented in Duffie & Beckman (2006, p. 432). Section 5.3.3 of this report describes the on-off strategy applied, which is based on the measurement of two sensors, one located at the outlet of the collector and the other located inside the store. The difference between these two temperatures is used to turn on or off the pump of the solar system.

The overall performance of the system is affected by the temperature values that establish those on-off cycles. Duffie & Beckman (2006, p. 432) presents a calculation procedure for calculating the ratio between the dead-band for setting the pump on (ΔTON) or off (ΔTOFF).

(28)

18 ∆𝑇_𝑜𝑓𝑓 ≤ 𝐴𝑐𝑜𝑙 𝐹𝑅 𝑈𝐿

𝜀 (𝑚 ̇𝐶_𝑝)_𝑚𝑖𝑛 ∆𝑇𝑜𝑛 (20)

For the designed model, the resulting ratios from Eq. 20 are depicted in Table 9. Load

DHW consumption of households depends on many variables, such as season, geographical location, and mostly it depends on the occupants DHW consumption habits (Rodríguez-Hidalgo, et al., 2012). The values of DHW used for the simulation are taken from the study carried out by Persson (2004), where the DHW load of small Swedish single family house was used as input for TRNSYS simulations. The data was taken from the standard CEN (2000). The aforementioned data corresponded to DHW flow rate values, over one year, at a resolution of 6 min. The temperature of tap water has been modelled by using the following equation, which values are also based on the study carried out by Persson (2004, p. 62) and corresponds to data from Stockholm.

𝑇_𝐶𝑊 = 𝑇_𝑚𝐶𝑊+ 𝑇_𝑎𝐶𝑊 sin[360 (𝑡 + (273,25 − 𝐷_{𝑜𝑓𝑓𝐶𝑊}) 24) /8760] (21) Where

TmCW is the mean cold water temperature over the year, for Stockholm 8.5 °C

TaCW is the temperature amplitude of the cold water, for Stockholm 6.4 °C

t is the time in hours

DoffCW is the time offset in days, for Stockholm 80 days

Eq. 19 represents a sinusoidal function, which is plotted in Figure 6, together with the DHW consumption.

Figure 6 Cold water temperature and daily DHW consumption from Persson (2004) and CEN (2000)

0 100 200 300 400 500 600 700 0 2 4 6 8 10 12 14 16 0 1000 2000 3000 4000 5000 6000 7000 8000 DHW Co n su m p tio n (l) Te m p era th re ( °C )

Hour of the year

(29)

19

Regarding the SH load, as it is the case of DHW it depends on many variables, Nygren (2003) in his research mentions 22000 kWh/year as the average for single family houses with electric heating in Sweden, at the time of the study.

In this thesis, the developed TRNSYS model uses a Type 12, configured for mode 4, for modelling the SH load, as it is described in Section 0. The following are the main equations that rule Type 12 when it is configured for mode 4.

𝑄_𝑇 = 𝛾 𝜀 𝐶_𝑚𝑖𝑛 (𝑇_𝑖 − 𝑇_𝑅) (22)

Where

QT is the rate at which energy is transferred through the Type 12 heat exchanger

γ is a function with value 1 if there is flow through the Type 12, otherwise it is 0 TR is the room temperature.

𝑄_𝐿 = 𝑈𝐴 (𝑇_𝑅 − 𝑇_𝑎) − 𝑄𝑔𝑎𝑖𝑛 (23)

Where

QL is the instantaneous heat loss from the structure, minus the heat gains QGAIN

(30)

20

3 Experimental methodology

As stated in the literature review, in order to simulate the real behaviour of a solar thermal loop, under indoor conditions, it is required the use of a device capable of simulating the response of a collector. This was called Collector Hardware Simulator (CHS). Devices such a CHS were out of the limitations of this project. Based on that, it was decided the design of a simplified test apparatus, which design was based on the studies presented in Section 2.2. The test apparatus designed for this thesis is strongly influenced by the design used by Colorado et al. (2011) (presented in Section 2.2, Figure 5). The reason is its simplicity, and all the required equipment was available for the construction.

In Section 2.2, the installation of temperature sensors is described by many papers, Dahm et al. (1998), Knudsen & Furbo (2004) Soo Too, et al. (2008) and Colorado, et al. (2011). All of them have in common that temperature sensors are placed inside the store, at different heights in order to measure the temperature of the different water layers. On the other hand, the heat exchanger inlet and outlet temperatures are also recorded by temperature sensors, in the case of the study carried out by Colorado et al. (2011), these sensors are immersed in the circulating fluid. The sensor placement of the test apparatus of this thesis have been decided according to the aforementioned studies.

Regarding the testing conditions, Colorado, et al. (2011) describes a test with fully mixed store as initial testing condition. This have been implemented in the test procedure described below.

3.1 The heat exchangers

Two different heat exchangers were tested, a finned copper HX and a corrugated steel HX. The copper HX is an exemplary of the coils used in the current manufacturing process of the solar loop of small-medium size tanks. On the other hand, the steel coils were novel, and were manufactured for the purpose of this research. Both types of HX´s are shown in Figure 7.

(31)

21

The design of the new steel HX was done in order to obtain a fair comparison with the copper HX, thus the characteristics of the steel coil, such as pipe length, coil pitch, coil diameter and height and number of turns were set to match as close as possible the copper coil. For the tested heat exchanger coils the length, the coil diameter, the number of turns and the height had the same value. Nevertheless, as can be seen in Figure 7, for the steel HX it was required the use of a metallic support fixture in order that the coil could keep its shape, since the steel pipes used were considerably less rigid than the copper ones. This decision affects the comparison, nevertheless the metallic support fixture is an element used in most of the immersed corrugated steel HX, and it is expected to be part of the final product. Thus this assumption was considered as valid.

The copper tube presents a significantly high external area when compared with the steel one. This is due to the fins located in the outer part of the tube, which leads to a higher area per internal volume. The profile of each tube is depicted in Figure 8 and Figure 9, and the main features in Table 1 and Table 2. The nomenclature used is based on the figures.

Figure 8 Profile of the corrugated stainless steel pipe Table 1 Profile description of the corrugated stainless steel pipe

Material DN Φ D Φ d h n Thickness k A V

- - mm mm mm mm mm W/mK m2_/m _l/m

AISI 316 20 20.8 15.5 4.75 2 0.2 16 0.112 0.254

The profile of the finned copper pipe is described as follows.

(32)

22 Table 2 Profile description of the finned copper pipe

Material d1 d3 d4 d5 S2 h m k Ai Ao V

- mm mm mm mm mm mm mm W/mK m2_{/m m}2_{/m l/m}

Cu 99.9 % 18 12.4 14 21 0.8 3.5 2.3 201 0.038 0.2 0.12 Table 1 and Table 2 presents a significant difference in tube diameter and consequently tube internal volume. Initially a steel tube with the same diameter as the copper tube was tested, and it was noted that the pressure drop on the tube was unacceptably high for the purpose of the test, thus it was decided to use a steel tube with higher diameter than the copper one. As can be seen in the previous figures, the internal surface of the steel coil is corrugated, as it is its external surface. On the other hand, the copper tube presents the fins on the external surface, whilst its internal surface is smooth.

3.1.1 Re number: Flow regime of the tested coils

The flow regime of the tested heat exchangers is crucial when analyzing the results of the experimental process. For that, as explained in Section 2.1, it is required to calculate the Re

number for each coil. Nevertheless, the calculation of the Re number of the stainless steel coil

involves complex calculations, due to its corrugated inner surface. In this section, only the copper coil will be analyzed regarding the Re number.

As stated by Jayakumar (2012), for helically coiled pipes the critical Re number (transition from

laminar to turbulent flow) is dependent on the value of the curvature ratio δ (ratio of pipe diameter to coil diameter). δ is calculated according to Eq. 2. For the studied copper coil the resulting value of δ is 0.046. Jayakumar (2012) (as described in Section 2.1), presented a correlation between δ and the critical Re number, for helically coiled pipes. According to that

correlation, the corresponding critical Re number for the studied copper pipe is 8000.

The Re numbers are calculated by means of applying Eq. 3. Table 3 shows the Re number at

four different flow rates (which are the testing flow rates, described in Section 4.1). Table 3 Flow regime of the copper heat exchanger

Flow rate Re Regime

l/hr - -

100 5211 _{Re<8000 Laminar}

300 15604 _{Re>8000 Turbulent}

400 20805 _{Re>8000 Turbulent}

500 26007 _{Re>8000 Turbulent}

(33)

23

3.2 The test apparatus

Based on the discussion presented at the beginning of this chapter, the testing apparatus is shown in Figure 10, and it was used to perform tests of the HX in charging conditions (store being heated by the HX), with a constant inlet temperature of 50 °C and constant flow rate.

Figure 10 Test apparatus The reservoir vessel

The vessel presented at the left is the so called reservoir, consisting of a 900 l cylindrical steel vessel with a 6 kW internal heater located in the bottom section, just above the inlet from the testing heat exchanger. The reason of placing the heater at the bottom is that the water coming back from the heat exchanger has a temperature significantly lower than the required 50 °C. Thus it is required to heat it up again up to 50 °C, before it is being fed again into the HX, and in that way perform a test at constant inlet temperature, this is done by setting the thermostat control of the heater at the desired temperature value.

A 70 l open expansion vessel is installed at the top, in order to allow the expansion of the water volume content in the reservoir, when the water of the reservoir is at high temperature. The outlet port of the reservoir, which is connected to the inlet of the heat exchanger is located at the top.

(34)

24

know when to start the test (see Section 3.3), and they don´t provide useful data for the results analysis, thus these values were not logged.

A pump (Pump 2 in Figure 10) is connected to the reservoir through its bottom connection and the outlet T connection. This pump was used before starting the test to fully mix the reservoir and break down the stratification caused by the heater.

The store

The vessel located at the right is the so called store, consisting of a 500 l (1.8 m high) cylindrical steel vessel, covered with 90 mm of molded fire resistant polyurethane insulation. The filling port of the store is located at the top (in the lid) and in the bottom the draining port is located. The upper coil is used to cool down the store after the testing has concluded. As described in the following section, the inlet is connected to the cold water mains, and the outlet to the draining. The pump connected to the store (Pump 3 in Figure 10) has the same purpose as the pump connected to the reservoir, it is used to fully mix the store before the testing start. The thermocouples located inside the store, served as an indication to know the degree of mixing of the store, and therefore also indicate when the conditions were given for starting the test.

The coil at the bottom of the store is the heat exchanger to be tested. Pressure drop hoses

The pump located at the bottom interconnection (Pump 1 in Figure 10) of the two vessels is in charge of generating the required flow through the heat exchanger circuit. The pump model is Wilo-Star-RS 26/6, with single speed and three levels. The flow generated by the pump flows through the interconnections of the vessels, passing through the T connections where the transparent vertical hoses are connected, at the inlet and outlet of the testing heat exchanger. The difference in water level of these two hoses is used for the pressure drop calculation. For that purpose, both hoses are scaled. A description of the pressure drop estimation process is presented in Section 3.3.1.

Flow Control

The control of the experimental apparatus was performed manually. The flow rate control was performed through the use of the valve XX. The flow in the circuit was constrained depending on the position of this valve, thus that position was adjusted in order to obtain the desired flow in the circuit.

3.2.1 Sensor placement and logging system

The magnetic flow meter sensor model is Enermet MP-115, it is located in the upper interconnection. The sensor was equipped with a LCD display, and it was used to set the testing flow rate at the desired value. As it is mentioned in the limitations, Section 1.4.1, the flow sensor measurements were not logged. The sensor was located in the upper interconnection in order to avoid possible interference from the pump vibration. The measurements displayed by the flow meter were also constantly compared with the values of a spring restrained flow meter located at the inlet of the pump.

(35)

25

maintain them at a fixed position (not in direct contact with the bar, to avoid interferences). Thus, the store was divided into 7 nodes of the same volume, approximately 70 l each.

Another two thermocouples were placed at the inlet and outlet connections. For these sensors, a rubber hose section was added to the inlet and outlet connections. An orifice was perforated in each rubber piece and the thermocouple joint was installed inside the hose section, thereafter the orifice was properly sealed. Thus, these two sensors were placed immersed in the circulating fluid. Also extra insulation was placed over them, to improve the measurement quality.

A Pt-100 sensor, located in the interior of the logging device for temperature compensation of the thermocouples, was taking valid records of the thermocouples cold junction point.

EdgeBox V12 was used as data logger. 12 analog voltage inputs were used for the corresponding thermocouples, and as stated before, the measurements from the internal Pt-100 were also used as ambient temperature values. The data logger was connected to a computer, which was used to display the measurement values in real time, as well as performing and displaying calculations, such as heat transfer in the heat exchanger. The measurement interval for all the tests was set to 5 seconds, whilst the logging interval was set to 15 seconds, consequently each recorded value is an average of three measurements.

3.3 Test Sequence

The initial conditions for starting a test dealt mostly with the temperature of the vessels. For the start of all the tests, the reservoir vessel was fully mixed at 50 °C. The store vessel was also fully mixed at 30 °C. On the other hand, the thermostat of the electric heater was set to 50 °C. Once the system was set in the aforementioned state, the mixing pumps (Pump 2 and Pump 3 in Figure 10) were turned off, the valves adjusted to swap to the heat exchanger circuit, and then the pump of the HX circuit was turned on. Thereafter, by regulating the valve XX, the desired flow rate was established. That flow rate was kept constant during the whole test, and very little adjustments were required to keep the flow at the desired value.

(36)

26

Figure 11 T4 and outlet temperature. Start and end of test 3.3.1 Pressure drop test

The pressure drop tests were carried with two different methodologies, in order to validate the results. The results of these two processes are described in Section 4.4.

Method 1

This method is based on the level difference between the water content in the plastic hoses, located at the inlet and outlet of the HX. For this method, the pump was turned on whilst the valve XX was closed, which means that there was no flow through the heat exchanger. By means of a gradually opening of the valve XX the flow was increased step by step, for each step the level difference was recorded by using the scale on the hoses, until the valve was fully open, and the pump was set at its maximum level.

Method 2

This method is based on the pump curve, which correlates flow rate with the corresponding pump head. In this case the plastic hoses are not used and they are bypassed from the circuit. Valve XX is set in fully open position and the pump is turned on at level 1, then the measured flow rate is recorded. The process is repeated for all three levels of the pump, giving as a result three different flow rates. These values are then input in the pump curve, to find the corresponding pump head.

The corresponding pump head found, is equivalent to the pressure drop of the whole HX circuit. This means that that pressure drop value includes the pressure drop of all the circuit components, including the heat exchanger. Therefore, in order to know the pressure drop of the heat exchanger, the pressure drop of all the rest of the components in the circuit has to be calculated, and discounted from the found pump head value.

Master Level Thesis