- Email: [email protected]
isobutane mixtures in heat pumps: An experimental investigation
Accepted Manuscript Title: Propene/isobutane mixtures in heat pumps: an experimental investigation Author: Valerius Venzik, Dennis Roskosch, Burak Atakan PII: DOI: Reference:
S0140-7007(17)30047-6 http://dx.doi.org/doi: 10.1016/j.ijrefrig.2017.01.027 JIJR 3539
To appear in:
International Journal of Refrigeration
Received date: Revised date: Accepted date:
15-9-2016 19-1-2017 25-1-2017
Please cite this article as: Valerius Venzik, Dennis Roskosch, Burak Atakan, Propene/isobutane mixtures in heat pumps: an experimental investigation, International Journal of Refrigeration (2017), http://dx.doi.org/doi: 10.1016/j.ijrefrig.2017.01.027. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Propene/isobutane mixtures in heat pumps: an experimental investigation
Valerius Venzik, Dennis Roskosch, Burak Atakan Thermodynamics, IVG, University of Duisburg-Essen, 47057 Duisburg, Germany [email protected] Highlights
Experimental investigations of the influence of fluid mixtures on the overall process
Pressure drop: negative influence on the temperature glide of a binary mixture
Temperature glide: positive influence on the exergy losses within the heat exchangers
Comparison of theoretical predicted and measured COPs
Abstract Zeotropic mixtures in heat pumps, based on thermodynamic analysis, should lead to higher coefficients of performance (COP) due to the temperature glide which decreases exergy losses in the heat exchangers. However, fluid mixtures influence every component of a plant and the total system performance. In addition to the various theoretical studies in this field, a laboratory scale vapor compression heat pump test rig was designed and set up. In the present experimental investigations, the operating performance for the pure fluids isobutane and propene, and their mixtures are systematically investigated. COPs and exergetic efficiencies as a function of evaporation temperature, compressor speed and composition of the mixture are presented and compared with a theoretical approach. Contrary to theoretical expectations, the experimental results show only a slight increase of the COP for the mixture, compared to the better pure fluid, because heat exchanger pressure drops reduce the temperature glide.
Keywords: vapor compression system, heat pumps, refrigeration mixture, hydrocarbon, temperature glide 1.
Introduction
Today, fluids for thermodynamic cycles like vapour compression heat pumps or ORCs (organic rankine cycles) are commonly HFCs (hydrofluorocarbons) and in some cases the earlier fluids of the substance groups CFCs (chlorofluorocarbons) or HCFCs (hydrochlorofluorocarbons) are still used. Regarding ecological
1 Page 1 of 28
aspects (ozone depletion, greenhouse effect) and recent political restrictions those fluids have to be replaced in many countries. Various aspects have to be considered in fluid selection; besides those ecological aspects, non-toxicity, nonflammability, good thermodynamic properties, also the miscibility with the commonly applied lubricants and good material compatibility are important. Today, scientific efforts are mainly focused on thermodynamic criteria like vapor pressures, thermal efficiency or transport properties; if halogenated hydrocarbons shall be avoided, so-called natural refrigerants like alkanes or their mixtures remain as a relatively flexible choice. They could be used as pure compounds but also as mixtures. Unlike for pure fluids, for zeotropic mixtures the saturation temperature is a function of composition and is not constant for an isobaric evaporation or condensation. This physical characteristic is called temperature glide (TG). The main idea of using mixtures in a thermodynamic cycle, is the targeted exploitation of the temperature glide to reduce the mean temperature difference between the refrigerant and the secondary fluid, as well as the reduction of the difference between the thermodynamic mean temperatures of the heat source and the heat sink. This could lead to a better performance of the process, mainly if the heat source or heat sink changes its temperature considerably within the heat transfer process, the matching can be improved. Also, using fluid mixtures allows to find more choices as working fluids fulfilling all or at least the most important requirements. Several investigations on pure HCs (hydrocarbons) refrigerants and their mixtures were published in recent years, mainly aiming to substitute CFCs and HFCs; some of them shall be summarized briefly. Chang et. al. [1] published experimental results of a heat pump cycle reporting the performance and heat transfer characteristics of four pure HCs (isobutane, butane, propane and propene) as well as of two mixtures composed of propane with isobutane and butane. Their aim was finding alternative fluids for R22. In case of the binary mixtures, they showed that with increasing mass fraction of propane, the cooling and heating flow rates increase. In cooling mode, they found that both mixtures have higher COPs than R22. However, the COP for the best mixture if compared to the better pure fluid (propane) increases only slightly. Besides, in heating mode, the best COPs were reached with pure propane. Richardson and Butterworth [2] investigated the performance of propane and propane/isobutane mixtures in a vapor compression system with the aim of substituting R12. The mixture with 50 % propane reached very similar saturation characteristics and an enhanced COP compared to R12. Based on these findings, they assumed that the mentioned mixture and the pure fluid propane may be used in unmodified R12 systems. Mani and Selladurai [3] investigated a HC mixture composed of 68 % propane and 32 % isobutane (by mass) in a vapor compression refrigeration cycle. Here, the mixture reached on average an increased cooling flow rate by 20 – 30 % with respect to R12 and R134a. The thermodynamic performance of ethane and propane mixtures in a heat pump was considered by Park and Jung [4]. Regarding their investigation, the cooling and heating flow rates increase with increasing the fraction of ethane while the COP decreases. For a fraction of 4 % ethane similar COPs and cooling/heating 2 Page 2 of 28
flow rates were reached compared to R22; however, the pure fluid propane showed the highest COP. Similar approaches for domestic refrigerators regarding propane and isobutane mixtures were followed by [5],[6]. Furthermore, several investigations on ternary mixtures were carried out. Akash and Said [7] studied the use of LPG (30 % propane, 55 % n-butane and 15 % isobutane by mass fraction) in a domestic refrigerator. It was reported that the cooling flow rate and the COP are always higher compared to R12. Other ternary mixtures like butane/propane/R134a were also considered by other authors, mainly with the aim of finding a ‘Drop-In’ fluid for R12 (e.g.: [8]). Lee et. al. [9] investigated the thermodynamic performance of a water source heat pump driven with the fluorinated hydrocarbons R32/R152. Mixtures in the range between 20-50 % of R32 with the aim of finding an alternative working fluid for R22. They found that the COP of the mixture is up to 15.8 % higher than the one of R22, however, the filling capacity of the mixture is actually 27 % less than R22. Further practical as well as theoretical approaches can be found in the review of Mohanraj et. al. [10]. Apart from using mixtures in heat pumps, binary mixtures are widely discussed and investigated in the field of ORC working fluids (e.g. [11]). Most investigations are focused on finding a ‘Drop-In’ fluid for the direct replacement of the commonly used (CFCs or HFCs) with similar cooling or heating flow rates as observed with the CFCs or HFCs. Summarizing the literature, the usual approach of finding new fluids includes a trial of several pure fluids or mixtures and analysing important fundamental values like heat flow rates and the COPs, but without obtaining a deeper understanding of the reason for their changes. In general, the different parts like heat exchangers and compressors were not analysed separately and systematically as a function of process parameters. The influence of mixtures on the single components and on the entire cycle, e.g. induced by the non-isothermal isobaric evaporation, was not systematically analysed, so far. In the present paper, these aspects are investigated systematically for a binary fluid mixture in order to learn more about the influence of using mixtures in thermodynamic cycles, also for supporting theoretical fluid searching approaches with detailed data. Regarding the common theoretical fluid screenings [12],[13] as well as the recent inverse engineering approaches [14],[15], the implemented process models are relatively simple and do not reproduce the influence of a fluid on the interaction of various parts as an overall system. For example pressure drops and the fluid dependent change of the isentropic efficiency of the compressor are generally not considered in detail. Especially, fluid mixtures change the performance of a cycle: heat transfer coefficients are decreased [1],[16] and based on different flow rates of vapor and liquid, composition shifts can occur [17]. Those aspects are not considered in actual theoretical approaches, so far; however they can strongly change the fluid ranking. Thus, experimental investigations and characterizations of the influence of fluid mixtures on the whole process constitute an essential basis for improving theoretical approaches. Hydrocarbons like isobutane, propane and propene are environmentally friendly (based on the low global warming potential, GWP) refrigerants and widely discussed in our society and used in household refrigerators. 3 Page 3 of 28
In this paper, two pure fluids R1270 (propene) and R600a (isobutane) and their mixtures are investigated experimentally. This mixture system has, compared to mentioned fluids, the highest temperature glide (dependent on our boundary conditions). Sometimes the flammability of these hydrocarbons is used as an argument against their usage, but the non-flammable alternatives in this temperature regime are halogenated hydrocarbons with a high GWP and, thus, not a better choice. In addition, it should be kept in mind that pure hydrocarbons like R600a or R290 are already used in many European refrigerators and allowed by the legislations, proving that the safety risks are controllable. The influence of main parameters like mixture composition and compressor rotation speed on the heating flows, compressor power input, temperature glide, COP and exergy losses of individual components are investigated here. For three different compressor rotation speeds and a constant condensing temperature (TH = 30°C) the evaporation temperature is varied between 0 °C and 6 °C; those temperature levels were selected, because they are typical for ground source heat pumps for building heating applications. The results are discussed in detail and some of them are compared to theoretical predictions.
4 Page 4 of 28
2. 2.1
Experiments Experimental setup and measurement equipment
The experimental setup with the main components of the heat pump is schematically shown in Fig. 1; it includes a compressor, an expansion valve, a condenser, an evaporator, liquid tank reservoir and further auxiliary valves for controlling the mass flow of the heat source and sink. The states are numbered in Fig. 1, because we refer to them in a later figure. The hermetic reciprocating compressor (GEA Bock HGHC12P), especially developed for hydrocarbons (isobutane, propane and propene), has two cylinders and a maximum power input of 2.2 kW. Using a frequency inverter, the rotation speed can continuously be varied and controlled between 1050 min-1 and 2100 min-1 while the electrical power consumption is monitored. Both heat exchangers, evaporator and condenser, are designed as counter flow double-pipe heat exchangers, because their heat transfer characteristics are well known. Each heat exchanger consists of 10 straight sections with a length of 1.4 m connected by return bends. In Fig. 1 only 3 straight sections with 2 bends are shown for both heat exchangers, for the sake of brevity. The total heat transfer areas of the evaporator and condenser are 0.5 m² and 0.75 m², respectively. The refrigerant flows through the inner tube (evaporator: inner diameter = 10 mm; wall thickness = 1 mm, condenser: inner diameter = 13 mm; wall thickness = 1.5 mm), while the secondary heat transfer fluid (water) flows through the annulus (inner diameter: 16 mm evaporator, 19 mm condenser and a wall thickness of 1.5 mm for both). At the inlets and outlets, as well as at every return bend, of the heat exchangers the temperature of the refrigerant and the secondary heat transfer fluid are measured by resistance thermometers. Furthermore, the pressure is recorded at the inlets and outlets. In order to inspect the thermodynamic state of the refrigerant, visually, four sight glasses are installed. Next to the sight glasses, sampling points are mounted in order to take specimens of the refrigerant. A gas chromatograph with a relative error of 2 % (Agilent Technologies 7890B GC System) is used to determine the exact composition of the mixture; potential composition shifts of the mixture at different positions within the plant can, thus, be observed. The refrigerant mass flow is quantified accurately by a Coriolis flow meter (KROHNE MFS3081 K 1.5 E). Initially, within this study, temperature and pressure transducers at the inlet of the flowmeter (Fig. 1, grey background) were not installed yet, because much smaller pressure drops through the flowmeter were expected.
However, the use of Coriolis flow meters lead to both, heat losses and pressure drops (dependent on the fluid and the compressor rotation speed between 2 and 9 bar); and thus, temperature and pressure transducers were subsequently installed, in order to characterize the flowmeter as well as the compressor in detail. Coriolis flow meters are generally not suitable to measure two-phase flows. Therefore, despite the high pressure losses within the flow meter, it has to be ensured phase changes are avoided. Two positions within the cycle are 5 Page 5 of 28
suitable accordingly, behind the condenser in the liquid phase or behind the compressor in the gas phase. Commonly, the subcooling of the refrigerant at the outlet of the condenser is small (subcooling temperature of 1 K) which, in combination with the pressure drop of the Coriolis flow meter, leads to a high risk of a partial evaporation inside the Coriolis flow meter and to incorrect measurements. This risk does not exist at the outlet of the compressor and therefore it appeared to be the best position for the flow meter in order to measure the refrigerant mass flowrates accurately. The mass flow of the secondary heat transfer fluid in the evaporator and condenser are recorded by turbine flow sensors, in Fig. 1 named rotameter (GEMS Sensors&Controls FT110). A fine metering needle valve is used as expansion valve, and coupled with a servo-motor which allows controlling the evaporation pressure via PC. The liquid tank reservoir (7.5 liter) is necessary on the one hand to perform experiments for different compressor rotation speeds (the refrigeration charge depends on the compressor rotation speed) and on the other hand to compensate fluctuations during the experiments while changing the conditions. The charge amount of the cycle strongly depends on the fluid and its mean density; for example, in case of pure isobutane the charge amount is about 152 g and in case of propene about 730 g. The charge amount for the mixtures is, depending on the composition, between those two values. Based on our experience, a too high charge amount does not influence the steady state performance of the cycle as long as the installed fluid reservoir is not fully filled with liquid. In contrast, a lack of refrigerant in the cycle leads to process changes like incomplete condensation and may distort the results. In order to prevent this, the heat pump is filled with the necessary charge amount related to the maximal possible compressor rotation speed and the maximal possible evaporator temperature. In case of operation points with lower mass flow rates (e.g. at a compressor rotation speed of 1050 min-1) the excess refrigerant is buffered in the refrigerant reservoir. In order to minimize the heat loss of the whole heat pump, the heat exchangers, every measurement sensor as well as every connecting pipe were well insulated with polyurethane foam of 20 mm wall thickness. Regarding temperature and pressure measurements, resistance thermometers (PT 1000) of type AA and calibrated relative pressure transducer (piezo stainless steel sensor ICS IMP331) are used. Overall 40 resistance thermometers, 6 pressure sensors, 2 rotameters, and one Coriolis flow meter are installed. Those components as well as the servo-motor of the expansion valve and the frequency inverter are connected to a PC by an USB multifunction DAQ device (Labjack U12). The software tool LabVIEW [18] was used for monitoring, recording and controlling the experiments. The corresponding values are recorded every second within a measuring cycle. Based on given tolerances from the manufacturers of the used measuring equipment the statistical error was estimated (error propagation) with respect to the COPH; the maximal relative error regarding the COPH is 4.63 %. To validate this value, selected experiments for different compositions, compressor rotation speeds, evaporation and condensation temperatures were reproduced. This resulted in a maximum relative deviation of 2.5 %.
6 Page 6 of 28
2.2
Investigated parameters
For a systematic analysis of the interaction between the working fluid (-mixture) and the process, combinations of different evaporation and condensation temperatures are investigated for the pure fluids isobutane and propene and for 5 compositions of them. Table 1 shows the investigated fluid compositions, while the investigated combinations of evaporator and condensation inlet temperatures are listed in Table 2. Furthermore, all these experimental series were carried out for three different compressor rotation speeds; 1050 min-1, 1500 min-1 and 2100 min-1. Water is used as secondary heat transfer fluid for both, the evaporator (heat source) and the condenser (heat sink), respectively. The water mass flow rate through the evaporator is kept constant at 8 kg min-1. The water mass flow through the condenser, as well as the opening position of the expansion valve are both control parameters for adjusting the evaporation and condensation pressures, and thus, for the selected temperatures. By varying the water mass flow rate in the heat sink without changing the opening position of the expansion valve, the whole cycle can be shifted to higher or lower temperature levels without changing the spread of both temperature levels (evaporation and condensation temperature). In contrast, changing the opening positon of the expansion valve, allows to vary the spread between the low and high temperature levels. In summary, the evaporator inlet temperature can be adjusted independent of the flow rate of the heat source (as long as no pinch point problem occurs in the evaporator). Based on this decoupling a systematic investigation of the influence of the evaporator inlet temperature is possible. The water mass flow rate in the condenser was varied between 1.5 and 7.5 kg min-1. The water inlet temperature of the evaporator and condenser are held constant at 17 °C and 14 °C, respectively. Depending on the investigated condensation temperature, the heat sink water outlet temperature varied between 26 and 47 °C. For a condensation temperature of 30 °C the outlet temperature of the heat sink is 34 °C ±2 °C. This temperatures are sufficiently high for heating floor applications. After the process reached steady state conditions, the data was recorded for about 20 minutes; the data analysis is based on averaged values. 2.3
Data reduction
For all thermodynamic analyses, the enthalpies and entropies for the different states are needed. Those were taken from the NIST REFPROP [19] database for the working fluids and for water at the measured temperatures and pressures. In the saturated regime, the enthalpy of the working fluid was calculated from the energy balance in the heat exchanger (which is assumed to be adiabatic towards the surrounding) and enthalpy and pressure were used as independent variables, together with the known composition (see below). The heat flow rates for cooling
and heating
were easily calculated by means of the first law of thermodynamics,
written for water as an example in Eq. 1. The European sign convention for heat and work is used: both are positive, when they are transferred to the system. The enthalpies are based on the measured inlet and outlet temperatures.
7 Page 7 of 28
Eq. 1
In combination with the measured power input to the compressor, the coefficient of performance for the heat pump is given by:
Eq. 2
The COP of a reversible (Carnot) cycle was calculated in order to compare the experimental values with the theoretical limit:
Eq. 3
Commonly, the temperatures
H
and
L
are the thermodynamic mean temperatures of the heat source and heat
sink (condenser and evaporator). If only the cycle shall be analysed, the Carnot-COP calculated with the thermodynamic mean temperatures of the working fluid can also be very helpful. Most of the thermodynamic states are fully defined by the recorded temperatures and pressures. The evaporator inlet state is saturated, thus, for pure fluids temperature and pressure are not sufficient to define the state. Basically, there are two different ways of calculating the evaporator inlet state; one is assuming an adiabatic expansion (hco,out = hev,in) and the other is assuming an adiabatic evaporator (Δ
Ref
=Δ
w,L).
Certainly, neither of the two processes are strictly adiabatic, but the temperature differences between the fluid and the environment is in our case higher for the expansion valve, thus, assuming an adiabatic evaporator seems to be more accurate and was used here. From the energy balance of the evaporator, the enthalpy at the evaporator inlet is derived:
Eq. 4
The heat losses of the evaporator were estimated and are very small compared to the transferred heat, justifying the assumption of adiabatic conditions. For fluid mixtures, the combination of temperature and pressure is sufficient to define a state within the two phase region.
8 Page 8 of 28
Besides the COPH, the analysis of exergies [20],[21]. The exergy losses
is a suitable tool to evaluate the performance of a process
of each component indicate the potential for improvements, this is especially,
important for fluid mixtures. The reduction of exergy losses in the heat exchangers by means of the targeted use of the non-isothermal isobaric phase change is one of the main ideas for using mixtures as working fluids. The exergy loss for a component i is calculated from the difference between entering and exiting exergy flow rates:
Eq. 5
Here,
and
are inlet and outlet exergy flow rates, respectively; Pel is the compressor power input and
the exergy flow rate induced by a heat flow. The exergy for each state is given by:
Eq. 6
The non-isothermal isobaric phase change is one basic physical property of zeotropic mixtures and known as temperature glide (TG). In general, regarding a fixed pressure and composition the TG is defined as the temperature difference between the dew and the bubble point:
Eq. 7
Behind the throttle, where a saturated fluid is entering the evaporator, the theoretical (isobaric) TG can be calculated by:
Eq. 8
Due to pressure losses, the real TGs of the working fluid in the heat exchangers differ from the theoretical ones. The determination of the real TG requires the knowledge of the starting and ending temperature of the phase change. Regarding exemplarily the evaporator, the inlet temperature is known but due to the fact that the temperatures of the working fluid within the evaporator are measured only every 1.4 m, the exact state at the end of the evaporation is unknown. Therefore, an alternative procedure was used: From the determined temperature gradients in each heat exchanger section of 1.4 m length it is easily recognized whether
9 Page 9 of 28
evaporation is completed in a section or not. From the temperature change in the last section with two-phase fluid, the evaporation endpoint temperature can be determined within ± 1 K.
Eq. 9
Temperatures and pressures were initially not measured in front of the Coriolis flow meter, yet; however, those flow meters commonly lead to pressure drops and to heat losses. In order to determine the exergy loss in the compressor, the state at the exit of the compressor or the entrance state of the flow meter is inevitably needed. Based on 17 independent measurements for the pure fluids isobutane and propene and for different operating points, the characteristic pressure drop of the flow meter was fitted as a function of outlet state and mass flow rate; this fit was used here for data evaluation. From these independent measurements, the mean absolute error of the compressor exergy losses were evaluated to be 12 W (min: 0.14 W; max: 61.26 W), leading to a mean relative error of ± 3 % here.
10 Page 10 of 28
3.
Results and discussion
Experiments for two pure fluids and five mixtures at several compressor rotation speeds and different combinations of evaporation and condensation temperatures were carried out. Here, the detailed analysis focusses on the variation of the evaporation temperature. Fig. 2 shows the COPH as a function of the mole fraction of propene (xR1270), for the evaporator inlet temperatures: 0 °C, 3 °C, 6 °C and for two different compressor rotation speeds (n): 1050 min-1 (Fig. 2a), 2100 min-1 (Fig. 2b). Throughout, the condensation temperature is kept constant at 30 °C. Based on the limited evaporator heat transfer area, it was not possible to carry out the experiments for the whole composition range for n = 2100 min-1 and evaporator inlet temperatures of 3 °C and 6 °C. First of all, it can be seen that the COPH increases with the evaporator inlet temperature TL. Furthermore, a direct comparison of Fig. 2a and Fig. 2b shows a remarkable reduction of the COPH with the compressor rotation speed. With respect to the COPH change with the mole fraction of propene the COPH always increases first, passes through a maximum and finally slightly decreases for pure propene (R1270).
The maxima are always found at a propene mole fraction of 72 % and are only slightly better (5 % - 7 %) than the one of propene, which is the better performing pure fluid here. The maximum COPH (5.49) is found for TL = 6 °C and n = 1050 min-1; in this case the COPH of propene is about 5.11. Regarding the pure fluids, propene (R1270) always performs better than isobutane (R600a); e.g. in case of an evaporation temperature of 6 °C and n = 1050 min-1, the COPH of propene is about 20 % higher compared to isobutane. The power consumptions per mass flow rate of the compressor (abbreviated as specific work, wel) as a function of the mixture composition for two different compressor rotation speeds 1050 min -1 and 2010 min-1 are exemplary shown for TL = 0 °C in Fig. 3. It can be observed that the specific work decreases with increasing mole fraction of propene, goes through a minimum and increases again for higher propene fractions. The specific work for pure isobutane is always higher than the one of propene. Compared to the composition of the maximum COPH (Fig. 2) the minimum of the specific compressor work can be found at the same composition, leading to the assumption that the compressor work has the major impact on the COP of the entire process.
Furthermore Fig. 4 shows the COPH change with compressor rotation speed for TL = 6 °C and xR1270 = 43.47 %. Also, the specific work of the compressor, and the absolute value of heat flow rates per mass flow rate of fluid (abbreviated as specific heat, qh) of the condenser are plotted for 3 different rotation speeds. The specific work of the compressor also increases with the rotation speed (at the same pressure ratio), thus the compressor becomes less efficient. Within the compressor, the higher losses are induced by higher 11 Page 11 of 28
piston velocities and accelerations, increased friction (piston – cylinder wall) and losses through the compressor valves, caused by the higher mass flow rates. The higher irreversibility leads to increased fluid outlet temperatures which also lead to larger transferred heat per mass qh. But the compressor work increases more than qh thus, the COPH decreases. Assuming sufficient heat exchanger areas and adiabatic conditions for the condenser and the connection tubes, the gradients of qH and wel would be equal; however, even then the COPH would decrease. The main reason of the compressor work increase with compressor rotation speed can be found by inspecting the pressure drops of the whole cycle. Fig. 5 shows the pressure drops in the evaporator and condenser together with the mass flow rates of the working fluid as a function of the compressor rotation speed. First of all, it is clear that the mass flow rate of the working fluid increases with higher rotation speeds; furthermore, the pressure drops also increase. From common approaches for tube friction it is known that the pressure drop is a strong function of the flow velocity, which again is proportional to the mass flow rate. The gradient as well as the value of Δp are larger for the evaporator than for the condenser; this is due to the lower flow cross section of the evaporator which also influences the pressure drops grossly.
A further reason for the larger pressure drop in the evaporator compared to the condenser are the differences in the thermodynamic states. Both heat exchangers are in large parts perfused by saturated fluid, however, the average vapor quality (over the length) is larger in the evaporator than in the condenser; commonly, pressure drops for vapors are higher than for liquids at comparable conditions. Finally, for a constant condenser inlet temperature, the pressure drop in the evaporator leads to an increased pressure ratio of the compressor, raising the power need. The increase of the COPH (see Fig. 2) with increasing evaporator inlet temperature TL at a constant condensation temperature is what is expected from the Carnot-cycle. With increasing TL the quotient between the thermodynamic mean temperatures Strictly speaking,
L
L
and
H
decreases and according to Eq. 3 the COPH increases, too.
can only be increased until the pinch point is reached. The reduction of the condensation
temperature has similar effects, the COP rises; however the outlet temperature of the secondary fluid is also decreased and the use of the heat pump vanishes. The most interesting aspect of the measurements illustrated in Fig. 2 is the composition dependence of the COPH. In general, it is assumed that the use of fluid mixtures has a great potential for improving thermodynamic cycles, but our measurements show only a minor increase of the COPH compared to the better performing pure fluid propene. Similar results for a mixture of isobutane and propene were also found by Chang et. al. [1]. In their cooling cycle the COP of the best mixture was 6 % higher than the one of the better performing pure fluid; for heating, the highest COP was reached by pure propene. Taking into account the need of larger heat exchangers due to smaller convection coefficients for mixtures, leading to such small 12 Page 12 of 28
advantages in the performance, the use of a mixture would probably not be justified. Theoretical approaches often show a greater benefit for using mixtures; in the following the reasons for this mismatch between experiment and theory is discussed.
Fig. 6 shows the results of different calculations, all for an evaporation temperature of 3 °C and a condensation temperature of 30 °C. The simulations are based on simple thermodynamic balances of the main components (evaporator, compressor, condenser, expansion valve) as they have been carried out by many authors (e.g. [12],[13]) with the aim of evaluating fluids in cycles and in order to create fluid rankings. Four different cases are considered here: (a) an isentropic compressor ηc,s = 1.0, (b) an irreversible compressor with constant isentropic efficiency ηc,s = const. = 0.585, and two cases with fluid composition dependent isentropic efficiencies ηc,s = f(x1270) (c) without and (d) with x1270 dependant pressure losses in the heat exchangers. Additionally, the measured COPs for these conditions (e) which are also included in Fig. 2a are shown. For cases (c) and (d), the isentropic efficiencies as well as the pressure losses as a function of the mole fraction of propene (summarised in Table 3) are taken from measurements at the given conditions. The constant isentropic efficiency (case b) is the average value of the xR1270 dependant isentropic efficiencies. For the isentropic compressor case, the COPH reaches a maximum at a propene mole fraction of 70.71 %. The mixture COP is significantly better by 26.71 % and 30.02 % compared to the pure fluids isobutane and propene, respectively. For a reduced constant isentropic efficiency of the compressor (Fig. 6b), there is a significant reduction of the COPH which is independent on the mole fraction. Furthermore, the advantage of the mixture compared to the pure fluids is strongly reduced. While isentropic efficiencies decrease, the compressor outlet temperature as well as the thermodynamic mean temperature increase. Thus, the COP H is decreased and the impact of the TG on the thermodynamic mean temperatures becomes smaller. Interestingly, now the order of the COPs of the pure fluids for case b is different than the experimentally determined ones, isobutane performs better than propene; only by the implementation of mole fraction dependant isentropic efficiencies the simulations lead to the experimentally found ranking of the pure fluids. However, in this case only a minor advantage of propene compared to isobutane remains but in comparison to case (b) the order of the pure fluids changes. Otherwise, the curve is not influenced strongly. Starting from pure isobutane (ηc,s = 0.55) the measured isentropic efficiency of the compressor increases, then passes a maximum (ηc,s = 0.66) at a propene mole fraction of 72 % and decreases again towards pure propene (ηc,s = 0.60). Regarding the fluid dependant compressor performance, approaches to simulate such cycles with a constant isentropic efficiency seem to be not suitable for correct assessments and thus, should be combined with a realistic compressor model as long as a single compressor is used with different mixtures. Probably, the highest (constant) isentropic efficiency could be reached for each mixture with fluid optimized compressors, in part justifying previous approaches. With additional implementation of the pressure losses in the evaporator and condenser (Fig. 6d), the values of 13 Page 13 of 28
the COPH come closer to the experimental values, although the experimental values remain much lower. The curve 5d is shifted to lower values and the advantage of the mixture against pure propene becomes lower, again.
One of the main reasons for the low experimental COPs is the additional pressure drop through the Coriolis flowmeter. Fig. 7 shows the process in a p-h diagram for the same operating conditions as the simulations are done for two different compressor rotation speeds; the state numbering is the same as in Fig. 1. Due to the Coriolis flow meter, the increase in pressure pcompressor achieved by the compressor (state 1-state 2) in both cases is higher compared to the pressure increase which would actually be needed to reach the condensation temperature of 30 °C without a flowmeter. The pressure drop in the flow meter (in Fig. 7 from point 2 to 3) depends strongly on the refrigerant mass flow rate and the mass flow rate depends on the compressor rotation speed. In the two cases shown in Fig. 7 the pressure drops are 2.98 bar for a compressor speed of 1050 min-1 (a) and 5.39 bar for 2100 min-1 (b). Depending on the fluid and the compressor rotation speed, typical pressure drops through the flowmeter are found to be between 2 and 9 bar, in general. This additional pressure drop has to be compensated by the compressor, and thus, the power consumption of the compressor increases. To investigate the influence of the flow meter on the COPH, some additional measurements were carried out for nearly the same mixture compositions and operating conditions without the Coriolis flow meter, using a bypass; they are shown in Fig. 8.
Fig. 8 shows experimental COPH values with / without the use of the Coriolis flow meter and simulated COPH as a function of the propene mole fraction, all for TL = 3 °C, TH = 33 ± 2 °C and a compressor rotation speed of 1050 min-1. It is seen that the COPH increases considerably if the Coriolis flow meter is omitted. The usage of the Coriolis flow meter leads to an increased compressor power consumption by a factor of up to 1.2 while H
increases only by a factor of 1.02. Nevertheless, comparing the simulation results of Fig. 6 with the
measured COPH in Fig. 8 (without the use of the flow meter), some deviations remain. It is assumed that these originate from the used isentropic efficiencies of the compressor which were fitted to measurements with the flow meter, while the isentropic efficiencies decrease with lower pressure drops, as is the case without flow meter. Thus, the simulations lead to overestimated COPH values compared to the measurements without flow meter. As can be seen in Fig. 8, by using process parameters, as summarised in Table 4, which are measured without a flow meter, the deviations between the simulations and experiments decrease. Apart from the negative influence on the value of the COPH, the investigations show that the use of the Coriolis flowmeter does not influence the composition at maximum COPH and thus, the comparability of the fluid ranking is conserved. But since it is not clear if this conclusion is generally valid; detailed investigations
14 Page 14 of 28
on this topic will follow in near future. Despite all disadvantages of using flow meters, the mass flow rate of the refrigerant is of great importance and Coriolis flow meters are the tools of choice. Apart from the flow meter, there are further sources of losses which decrease the efficiency of the cycle and probably lead to deviations between simulations and measurements. The thermal losses of the heat exchangers, the sight glasses and all connection pipes are between 26 Watt and 214 Watt, in total, depending on the fluid and the operation point. This corresponds to 3 % - 15 % of the
value and thus leads to a
decrease of the COPH of 15% at maximum. Maximal thermal losses are detected for isobutane, explaining the highest deviation between the simulation and the measurement in this case.
The pressure drops in the evaporator and the condenser as well as the mass flow rate as a function of the mole fraction of propene are exemplarily shown in Fig. 9 for a compressor rotation speed of 1050 min-1 and TL = 0°C, TH = 30°C. The density of propene (for the same conditions) is approximately two times higher than the one of isobutane, therefore the mass flow rate of the working fluid increases with the propene mole fraction and the pressure drop increases with the mass flow rate. Due to the smaller tube diameters, the pressure drop again changes stronger with the propene content in the evaporator than in the condenser. This increase of the pressure drop in the evaporator (see Fig. 7) has to be overcome by the compressor and requires a larger power per mass flow rate. Fig. 10 shows T-h-plots of the pure fluid isobutane (a, left) and for the best isobutane/propene mixture (b, right); for both, the heat exchangers are assumed to be non-isobaric. Additionally, the temperature changes of the secondary fluids as well as the thermodynamic mean temperatures are depicted. As shown in Fig. 10a, the pressure drops affect the process also in another way: for a pure fluid, a pressure drop leads to non-isothermal conditions in the saturation regime. As a consequence, the thermodynamic mean temperatures in the condenser and in the evaporator are decreased and the ratio of L
and
H
becomes smaller leading to smaller COPs (Eq. 3). The main thermodynamic advantage of using
zeotropic mixtures which is the temperature glide (TG) resulting in reduced exergy losses and in an increased ratio between
L
and
H
can be seen by comparing the both T-h-plots. The temperature glide in the condenser
becomes larger for the mixture while it is inverted in the evaporator. This leads on the one hand to a decreased difference between
L
and
H
and on the other hand to a better match between the temperature gradient of the
secondary fluid and the working fluid in the heat exchangers. However, these advantages are counter-acted here by pressure drops which reduce the real temperature glides and lead to the only minor advantage of the mixture.
This leads to the question which operating conditions may be increasing or reducing pressure drops and thus influencing the COP. Fig. 11 shows a comparison between measured and calculated evaporator TGs
15 Page 15 of 28
(evaluated for isobaric conditions) as a function of the propene mole fraction, determined for two different compressor rotation speeds 1050 min-1 and 2100 min-1, respectively. The ideal TG starts with 0 K for pure isobutane, increases till reaching a maximum of 6.87 K at a composition of xR1270 = 72 % and decreases to 0 K for pure propene, again. The measured TG starts and ends at negative values for the pure fluids which is due to the pressure drops. At a propene mole fraction of about 11.9 % the TG is large enough to over-compensate the pressure drop and the temperature gradient of the working fluid during evaporation gets positive. The experimental curves are nearly parallel to the theoretical ones, but shifted to smaller values. At higher compressor rotation speed, the differences between the calculated and measured TG increase due to the higher pressure drops (see Fig. 5 ) at higher Re numbers, reducing the small advantage of a zeotropic mixture (compare Fig. 2). Although the targeted improvement of a thermodynamic cycle with zeotropic fluids is reduced strongly if pressure drops get important, this could be improved using mixtures with higher temperature glides. But mixtures with a high temperature glide commonly are composed of fluids with strongly differing evaporation pressures which usually differ strongly in their densities, too. Although this makes things more complicated, this field seems worth further investigations. Apart from the COPH, the exergy losses in different parts of the cycle are also composition dependent; this shall be analysed finally. In Fig. 12 the specific exergy losses of the main components (including the flowmeter) as a function of xR1270 are exemplary shown for a compressor rotation speed of 1050 min-1 and TL = 0 °C, TH = 30 °C. Except for the throttle and the flowmeter, all exergy loss curves pass a minimum. The exergy losses in the heat exchangers are only slightly dependent on composition. This is easily explained by the boundary conditions: the heat capacity flow rates of the secondary fluids are high and consequently the temperature gradients are small; thus, even in case of pure fluids the exergy losses are already small. Here, the positive influence of the temperature glide on the COPH is mainly due to the decreased gap between
H
and
L.
The
highest exergy losses are found in the compressor which is also partially determined by the pressure drops in the heat exchangers; therefore this component is most important. The minimum exergy loss occurs at a mixture composition of xR1270 = 72 %; this corresponds directly to the composition of the maximum COPH. It is observed that the specific exergy loss the flowmeter increases slightly with increasing mole fraction of propene, because of the increasing pressure drop with increasing propene mole fraction.
4.
Conclusion
In the present work, the influence of two pure fluids R1270 (propene) and R600a (isobutane) and their zeotropic mixtures on the COP of a heat pump for different evaporator inlet temperatures, condensation temperatures and compressor rotation speeds are investigated experimentally. The COPH is only slightly changing with composition; a COP improvement of 5 - 7 % relative to propene was found. This improvement 16 Page 16 of 28
results mainly from the decrease of the difference between the two thermodynamic mean temperature levels for heat transfer to and from the cycle and is based on the temperature glide (TG). The exergy losses in the heat exchangers are only lowered slightly by the use of these mixtures, with relatively small differences in saturation temperatures at the investigated pressures. The evaluation of the experimental data as well as a theoretical analysis show that pressure drops as well as the non-ideality of the compressor are mainly responsible for the findings. Also, the pressure drops within the heat exchangers reduce the temperature glide and thus, reduce the main advantage of the mixture usage in thermodynamic cycles. Non-ideal compressors lead to higher outlet temperatures and thus, to increased thermodynamic mean temperatures in the condenser. In consequence, the temperature glide has a minor influence on the reduction of the mean temperature differences. The presented results seem to indicate that the use of fluid mixtures is not justified regarding the small improvement of the COPH. However, in case of lower heat capacity flows of the secondary fluids the use of fluid mixtures could be more advantageous. Also, one has to design the tube diameters carefully in order to reduce pressure drops in the heat exchangers and further parts if zeotropic mixtures shall retain their thermodynamic advantage compared to pure compounds. Besides, the pure fluid propene always performs about 20 % better than isobutane, which would contradict simple theory. The present investigation shows that this behavior can only be reproduced correctly by implementing validated mixture dependent isentropic efficiencies, as long as a single compressor is used for all fluid compositions. Neglecting this, may lead to wrong (mixed) working fluid rankings for thermodynamic cycles. Increasing the compressor rotation speed increases the irreversibility within the process as well. Thus, on the one hand the values of the COPH are decreased and on the other hand the advantage of the mixture usage compared to the pure fluid becomes smaller. Increasing the evaporation temperature or decreasing the condensation temperature results in a dropped gap between the both thermodynamic mean temperatures and thus, the COPH rises. However, the degrees of freedom regarding those temperatures are commonly very small. In future, fluid mixtures with a greater temperature glide as well as secondary fluids with lower heat capacity flows should be investigated.
5.
Acknowledgment
We thank KROHNE Germany for provision of technical measuring equipment. Technical assistance by Stephan Steinbrink and Andreas Görnt as well as laboratory support by Andreas Lennert is gratefully acknowledged.
17 Page 17 of 28
6.
6.1
Appendix
Nomenclature coefficient of performance of a heat pump
[-]
exergy flow rate of each state
[kW]
specific exergy loss of each component
[kJ kg-1]
exergy loss flow rate
[kW]
specific enthalpy
[kJ kg-1]
mass flow rate
[kg s-1]
rotation speed of the compressor
[min-1]
electrical power consumption of the compressor
[kW]
pressure
[bar]
specific heat flow
[kJ kg-1]
quality
[-]
heat flow
[kW]
specific entropy
[kJ kg-1 K-1]
temperature
[K]
temperature glide
[K]
specific work of the compressor
[kJ kg-1]
mole fraction
[-]
Greak symbols enthalpy flow rate change p
pressure drop
[mbar]
isentropic efficiency of the compressor
[-]
Subscripts bubble
bubble point
c
compressor
co
condenser
dew
dew point 18 Page 18 of 28
ev
evaporator
H
at high level (e.g. temperature, enthalpy, mass flow rate, heat flow rate in condenser)/ heat pump
i
component
in/ out
inlet/outlet
L
at low level (e.g. temperature, enthalpy, mass flow rate, heat flow rate in evaporator)
m
middle (e.g. middle temperature at low/high level)
Ref
working fluid (refrigerant)
s
isentropic
vap
vapour
w
secondary heat transfer fluid (water)
0
fluid properties/conditions at 20 °C and 1,01 bar
Acronyms CFCs
chlorofluorocarbons
COP
coefficient of performance
DAQ
data Acquisition
HC
hydrocarbon
HCFCs
hydrochlorofluorocarbons
HFCs
hydrofluorocarbons
ORC
organic rankine cycle
TG
temperature glide
19 Page 19 of 28
6.2
References
[1] Chang Y, Kim M, Ro S. Performance and heat transfer characteristics of hydrocarbon refrigerants in a heat pump system. International Journal of Refrigeration 2000;23(3):232–42. [2] Richardson RN, Butterworth JS. The performance of propane/isobutane mixtures in a vapourcompression refrigeration system. International Journal of Refrigeration 1995;18(1):58–62. [3] Mani K, Selladurai V. Experimental analysis of a new refrigerant mixture as drop-in replacement for CFC12 and HFC134a. International Journal of Thermal Sciences 2008;47(11):1490–5. [4] Park K, Jung D. Performance of heat pumps charged with R170/R290 mixture. Applied Energy 2009;86(12):2598–603. [5] Jung D, Kim C, Song K, Park B. Testing of propane/isobutane mixture in domestic refrigerators. International Journal of Refrigeration 2000;23(7):517–27. [6] Kuijpers L, Wit J de, Janssen M. Possibilities for the replacement of CFC 12 in domestic equipment. International Journal of Refrigeration 1988;11(4):284–91. [7] Akash BA, Said SA. Assessment of LPG as a possible alternative to R-12 in domestic refrigerators. Energy Conversion and Management 2003;44(3):381–8. [8] Tashtoush B, Tahat M, Shudeifat MA. Experimental study of new refrigerant mixtures to replace R12 in domestic refrigerators. Applied Thermal Engineering 2002;22(5):495–506. [9] Lee H, Kim H, Kang D, Jung D. Thermodynamic performance of R32/R152a mixture for water source heat pumps. Energy 2012;40(1):100–6. [10] Mohanraj M, Muraleedharan C, Jayaraj S. A review on recent developments in new refrigerant mixtures for vapour compression-based refrigeration, air-conditioning and heat pump units. Int. J. Energy Res. 2011;35(8):647–69. [11] Braimakis K, Preißinger M, Brüggemann D, Karellas S, Panopoulos K. Low grade waste heat recovery with subcritical and supercritical Organic Rankine Cycle based on natural refrigerants and their binary mixtures. Energy 2015;88:80–92. [12] Brown JS. Predicting performance of refrigerants using the Peng–Robinson Equation of State. International Journal of Refrigeration 2007;30(8):1319–28. [13] Saleh B, Wendland M. Screening of pure fluids as alternative refrigerants. International Journal of Refrigeration 2006;29(2):260–9. [14] Roskosch D, Atakan B. Reverse engineering of fluid selection for thermodynamic cycles with cubic equations of state, using a compression heat pump as example. Energy 2015;81:202–12. [15] Lampe M, Stavrou M, Bücker HM, Gross J, Bardow A. Simultaneous Optimization of Working Fluid and Process for Organic Rankine Cycles Using PC-SAFT. Ind. Eng. Chem. Res. 2014;53(21):8821–30. 20 Page 20 of 28
[16] Wen M, Ho C, Hsieh J. Condensation heat transfer and pressure drop characteristics of R-290 (propane), R-600 (butane), and a mixture of R-290/R-600 in the serpentine small-tube bank. Applied Thermal Engineering 2006;26(16):2045–53. [17] Rajapaksha L. Influence of special attributes of zeotropic refrigerant mixtures on design and operation of vapour compression refrigeration and heat pump systems. Energy Conversion and Management 2007;48(2):539–45. [18] LabVIEW Professional Development System: National Instruments; 2013. [19] Lemmon, E.W., Huber, M.L., McLinden, M.O. NIST Standard Reference Database 23: Reference Fluid Thermodynamic and Transport Properties-REFPROP. 9th ed. Gaithersburg: National Institute of Standards and Technology; 2013. [20] Ahamed JU, Saidur R, Masjuki HH. A review on exergy analysis of vapor compression refrigeration system. Renewable and Sustainable Energy Reviews 2011;15(3):1593–600. [21] Ganjehsarabi H, Güngor A, Dincer I. Exergetic performance analysis of Dora II geothermal power plant in Turkey. Energy 2012;46(1):101–8.
21 Page 21 of 28
heat source out p
PC controlled
T
M
p
evaporator
FI
5
p
1 M
sight glass
p
T
expansion valve
p
T
T
compressor
rotameter
motor
heat source in heat sink out
2
p p
CF
liquid tank reservoir
p
condenser
3
T
T
p T p
T
4 sample point
sampling device
T
heat sink in suction line discharge gas line discharge liquid line
T p CF FI
temperature pressure Coriolis flow meter frequency inverter
Fig. 1 Schematic diagram of the experimental setup
a
b 6
TL= 0°C
TL= 3°C
TL= 6°C
6
1050 min-1
4
3
2
TL= 3°C
TL= 6°C
2100 min-1
5
COPH [-]
COPH [-]
5
TL= 0°C
4
3
0.0
0.2
0.4
0.6
0.8
2
1.0
mole fraction propene
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene
Fig. 2: COPH as a function of mixture composition; determined for the indicated evaporator inlet temperatures and for compressor rotation speeds of a: 1050 min-1and b: 2100 min-1. Condenser inlet temperature TH = const. = 30 °C
22 Page 22 of 28
1050 min-1
2100 min-1
specific work [kJ kg-1]
160
140
120
100
80
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 3: Specific work of the compressor as a function of mixture composition for compressor rotation speed of 1050 min -1 and 2100 min-1, TL = 0 °C and TH = 30 °C
qH
wel
COPH
6
370
5
290 4 210 3
130 50
COPH [-]
specific heat, work [kJ kg-1]
450
1050
1500
2100
2
-1
rotation speed [min ] Fig. 4: COPH, specific compressor work and condenser heat as a function of the compressor rotation speed for x R1270 = 43.47 %, TL = 6 °C and TH = 30 °C
23 Page 23 of 28
pressure drop [mbar]
pcondenser
mRef
0.011
480
0.010
360
0.009
240
0.008
120
0.007
0
1050
1500
2100
mass flow rate [kg s-1]
pevaporator
600
0.006
-1
rotation speed [min ] Fig. 5 Pressure drops and mass flow rates of the working fluid as a function of the compressor speed for x R1270 = 43.47 % and TL = 6 °C, TH = 30 °C
a
15
b
c
d
e
COPH [-]
12
9
6
3
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 6: COPH as a function of the mole fraction of propene (R1270) for TL = 3 °C and TH = 30 °C; a: ηc,s = 1.0, b: ηc,s = const. = 0.585, c: ηc,s = measured values depending on x R1270 , d: ηc,s and dpcond,evap= measured values depending on x R1270 , e: measured COPs for 1050 min-1
24 Page 24 of 28
a
b 100
1050 min-1
4
10
pcompressor without flow meter
3 5
1
2
200
1
400
2100 min-1
pcompressor with flow meter
pressure [bar]
pcompressor with flow meter
4
10
pcompressor without flow meter
2 3
5
pevaporator
600
1
800
200
specific enthalpy [kJ kg-1]
1
400
pevaporator
600
800
specific enthalpy [kJ kg-1]
Fig. 7: p-h diagram for xR1270 = 72 %, TL = 3 °C , TH = 30 °C and a: compressor rotation speed of n = 1050 min-1 and b: compressor rotation speed of n = 2100 min-1
Coriolis flowmeter 7
bypass
simulation
1050 min-1
6
COPH [-]
pressure [bar]
100
5 4 3 2
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 8: Comparison of the COPH with / without use of a Coriolis flow meter and simulation depending on x R1270 for n = 1050 min-1, TL = 3 °C and TH = 35 ± 2 °C
25 Page 25 of 28
pevaporator
pcondenser
mRef
0.011
1050 min-1
480
0.009
360
0.008
240
0.006
120
0.005
0
0.0
0.2
0.4
0.6
0.8
1.0
mass flow rate [kg s-1]
pressure drop [mbar]
600
0.003
mole fraction propene Fig. 9: Pressure drops as a function of x 1270 for a compressor rotation speed of 1050 min-1and TL = 0 °C, TH = 30 °C
a
b heat source
150
heat sink
1050 min-1
temperature [°C]
temperature [°C]
120
90 2
60
0 -30
heat sink
1050 min-1
120
30
heat source
150
3
TH
4 1
TL
2
60 30 0
5
200
90
400
600
-30
800
-1
3
TH TL
4
1 5
200
400
600
800
-1
specific enthalpy [kJ kg ]
specific enthalpy [kJ kg ]
Fig. 10: T-h-Diagrams of a: pure isobutane and b: isobutane/propene mixture (x R1270=72 %) at TL = 0 °C, TH = 30 °C and n = 1050 min-1
26 Page 26 of 28
8
1050 min-1
TG [K]
4 0 -4 -7 TGlide,measured 8
TGlide,ideal
2100 min-1
TG [K]
4 0 -4 -7
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 11: TG in the evaporator as a function of x R1270 for two compressor rotation speed 1050 min-1, 2100 min-1 and TL = 0°C, TH = 30 °C
specific exergy loss [kJ kg-1]
70.0
1050 min-1
eloss, evaporator eloss, compressor
52.5
eloss, condenser eloss, throttle eloss, flowmeter
35.0
17.5
0.0
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 12: Specific exergy losses of each component as a function of the mole fraction of propene, for a compressor rotation speed of 1050 min-1 and TL = 0 °C, TH = 30 °C
27 Page 27 of 28
Table 1: Investigated mixture composition and according temperature glide as derived by Eq. 7
R1270 mole fraction [%] 0.00 11.90 43.47 55.31 72.00 83.12 100.00
R600a mole fraction [%] 100.00 88.10 56.53 44.69 28.00 16.88 0.00
temperature glide [K] 0.00 5.59 7.82 7,95 8.15 7.56 0.00
Table 2: Experimental matrix for steady-state conditions
↓
L
[°C] TH → 0 3 6
25 x
30 x x x
35
40
x
x
Table 3: Messured isentropic efficiency of the compressor, evaporator/ condenser pressure drop depending on the mole fraction of propene for TL = 3 °C, TH = 30 °C and a compressor rotation speed of 1050 min-1
R1270 mole fraction [%] 0.00 11.90 43.47 55.31 72.00 83.12 100.00
isentropic efficiency ηc,s [%] 55.05 49.27 59.40 60.91 65.95 64.65 60.17
pressure drop evaporator [mbar] 79.18 101.5 151.58 171.10 202.41 232.30 255.30
pressure drop condenser [mbar] 70.12 85.34 97.21 108.05 114.51 124.34 143.80
Table 4: Messured isentropic efficiency of the compressor, evaporator/ condenser pressure drop depending on the mole fraction of propene for TL = 3 °C, TH = 35 ± 2°C °C and a compressor rotation speed of 1050 min-1
R1270 mole fraction [%] 0.00 18.83 59.24 83.46 100
isentropic efficiency ηc,s [%] 55.75 55.78 65.64 65.43 67.21
pressure drop evaporator [mbar] 80.38 120.08 203.80 258.96 295.33
pressure drop condenser [mbar] 36.23 64.52 81.16 88.06 124.52
28 Page 28 of 28
S0140-7007(17)30047-6 http://dx.doi.org/doi: 10.1016/j.ijrefrig.2017.01.027 JIJR 3539
To appear in:
International Journal of Refrigeration
Received date: Revised date: Accepted date:
15-9-2016 19-1-2017 25-1-2017
Please cite this article as: Valerius Venzik, Dennis Roskosch, Burak Atakan, Propene/isobutane mixtures in heat pumps: an experimental investigation, International Journal of Refrigeration (2017), http://dx.doi.org/doi: 10.1016/j.ijrefrig.2017.01.027. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Propene/isobutane mixtures in heat pumps: an experimental investigation
Valerius Venzik, Dennis Roskosch, Burak Atakan Thermodynamics, IVG, University of Duisburg-Essen, 47057 Duisburg, Germany [email protected] Highlights
Experimental investigations of the influence of fluid mixtures on the overall process
Pressure drop: negative influence on the temperature glide of a binary mixture
Temperature glide: positive influence on the exergy losses within the heat exchangers
Comparison of theoretical predicted and measured COPs
Abstract Zeotropic mixtures in heat pumps, based on thermodynamic analysis, should lead to higher coefficients of performance (COP) due to the temperature glide which decreases exergy losses in the heat exchangers. However, fluid mixtures influence every component of a plant and the total system performance. In addition to the various theoretical studies in this field, a laboratory scale vapor compression heat pump test rig was designed and set up. In the present experimental investigations, the operating performance for the pure fluids isobutane and propene, and their mixtures are systematically investigated. COPs and exergetic efficiencies as a function of evaporation temperature, compressor speed and composition of the mixture are presented and compared with a theoretical approach. Contrary to theoretical expectations, the experimental results show only a slight increase of the COP for the mixture, compared to the better pure fluid, because heat exchanger pressure drops reduce the temperature glide.
Keywords: vapor compression system, heat pumps, refrigeration mixture, hydrocarbon, temperature glide 1.
Introduction
Today, fluids for thermodynamic cycles like vapour compression heat pumps or ORCs (organic rankine cycles) are commonly HFCs (hydrofluorocarbons) and in some cases the earlier fluids of the substance groups CFCs (chlorofluorocarbons) or HCFCs (hydrochlorofluorocarbons) are still used. Regarding ecological
1 Page 1 of 28
aspects (ozone depletion, greenhouse effect) and recent political restrictions those fluids have to be replaced in many countries. Various aspects have to be considered in fluid selection; besides those ecological aspects, non-toxicity, nonflammability, good thermodynamic properties, also the miscibility with the commonly applied lubricants and good material compatibility are important. Today, scientific efforts are mainly focused on thermodynamic criteria like vapor pressures, thermal efficiency or transport properties; if halogenated hydrocarbons shall be avoided, so-called natural refrigerants like alkanes or their mixtures remain as a relatively flexible choice. They could be used as pure compounds but also as mixtures. Unlike for pure fluids, for zeotropic mixtures the saturation temperature is a function of composition and is not constant for an isobaric evaporation or condensation. This physical characteristic is called temperature glide (TG). The main idea of using mixtures in a thermodynamic cycle, is the targeted exploitation of the temperature glide to reduce the mean temperature difference between the refrigerant and the secondary fluid, as well as the reduction of the difference between the thermodynamic mean temperatures of the heat source and the heat sink. This could lead to a better performance of the process, mainly if the heat source or heat sink changes its temperature considerably within the heat transfer process, the matching can be improved. Also, using fluid mixtures allows to find more choices as working fluids fulfilling all or at least the most important requirements. Several investigations on pure HCs (hydrocarbons) refrigerants and their mixtures were published in recent years, mainly aiming to substitute CFCs and HFCs; some of them shall be summarized briefly. Chang et. al. [1] published experimental results of a heat pump cycle reporting the performance and heat transfer characteristics of four pure HCs (isobutane, butane, propane and propene) as well as of two mixtures composed of propane with isobutane and butane. Their aim was finding alternative fluids for R22. In case of the binary mixtures, they showed that with increasing mass fraction of propane, the cooling and heating flow rates increase. In cooling mode, they found that both mixtures have higher COPs than R22. However, the COP for the best mixture if compared to the better pure fluid (propane) increases only slightly. Besides, in heating mode, the best COPs were reached with pure propane. Richardson and Butterworth [2] investigated the performance of propane and propane/isobutane mixtures in a vapor compression system with the aim of substituting R12. The mixture with 50 % propane reached very similar saturation characteristics and an enhanced COP compared to R12. Based on these findings, they assumed that the mentioned mixture and the pure fluid propane may be used in unmodified R12 systems. Mani and Selladurai [3] investigated a HC mixture composed of 68 % propane and 32 % isobutane (by mass) in a vapor compression refrigeration cycle. Here, the mixture reached on average an increased cooling flow rate by 20 – 30 % with respect to R12 and R134a. The thermodynamic performance of ethane and propane mixtures in a heat pump was considered by Park and Jung [4]. Regarding their investigation, the cooling and heating flow rates increase with increasing the fraction of ethane while the COP decreases. For a fraction of 4 % ethane similar COPs and cooling/heating 2 Page 2 of 28
flow rates were reached compared to R22; however, the pure fluid propane showed the highest COP. Similar approaches for domestic refrigerators regarding propane and isobutane mixtures were followed by [5],[6]. Furthermore, several investigations on ternary mixtures were carried out. Akash and Said [7] studied the use of LPG (30 % propane, 55 % n-butane and 15 % isobutane by mass fraction) in a domestic refrigerator. It was reported that the cooling flow rate and the COP are always higher compared to R12. Other ternary mixtures like butane/propane/R134a were also considered by other authors, mainly with the aim of finding a ‘Drop-In’ fluid for R12 (e.g.: [8]). Lee et. al. [9] investigated the thermodynamic performance of a water source heat pump driven with the fluorinated hydrocarbons R32/R152. Mixtures in the range between 20-50 % of R32 with the aim of finding an alternative working fluid for R22. They found that the COP of the mixture is up to 15.8 % higher than the one of R22, however, the filling capacity of the mixture is actually 27 % less than R22. Further practical as well as theoretical approaches can be found in the review of Mohanraj et. al. [10]. Apart from using mixtures in heat pumps, binary mixtures are widely discussed and investigated in the field of ORC working fluids (e.g. [11]). Most investigations are focused on finding a ‘Drop-In’ fluid for the direct replacement of the commonly used (CFCs or HFCs) with similar cooling or heating flow rates as observed with the CFCs or HFCs. Summarizing the literature, the usual approach of finding new fluids includes a trial of several pure fluids or mixtures and analysing important fundamental values like heat flow rates and the COPs, but without obtaining a deeper understanding of the reason for their changes. In general, the different parts like heat exchangers and compressors were not analysed separately and systematically as a function of process parameters. The influence of mixtures on the single components and on the entire cycle, e.g. induced by the non-isothermal isobaric evaporation, was not systematically analysed, so far. In the present paper, these aspects are investigated systematically for a binary fluid mixture in order to learn more about the influence of using mixtures in thermodynamic cycles, also for supporting theoretical fluid searching approaches with detailed data. Regarding the common theoretical fluid screenings [12],[13] as well as the recent inverse engineering approaches [14],[15], the implemented process models are relatively simple and do not reproduce the influence of a fluid on the interaction of various parts as an overall system. For example pressure drops and the fluid dependent change of the isentropic efficiency of the compressor are generally not considered in detail. Especially, fluid mixtures change the performance of a cycle: heat transfer coefficients are decreased [1],[16] and based on different flow rates of vapor and liquid, composition shifts can occur [17]. Those aspects are not considered in actual theoretical approaches, so far; however they can strongly change the fluid ranking. Thus, experimental investigations and characterizations of the influence of fluid mixtures on the whole process constitute an essential basis for improving theoretical approaches. Hydrocarbons like isobutane, propane and propene are environmentally friendly (based on the low global warming potential, GWP) refrigerants and widely discussed in our society and used in household refrigerators. 3 Page 3 of 28
In this paper, two pure fluids R1270 (propene) and R600a (isobutane) and their mixtures are investigated experimentally. This mixture system has, compared to mentioned fluids, the highest temperature glide (dependent on our boundary conditions). Sometimes the flammability of these hydrocarbons is used as an argument against their usage, but the non-flammable alternatives in this temperature regime are halogenated hydrocarbons with a high GWP and, thus, not a better choice. In addition, it should be kept in mind that pure hydrocarbons like R600a or R290 are already used in many European refrigerators and allowed by the legislations, proving that the safety risks are controllable. The influence of main parameters like mixture composition and compressor rotation speed on the heating flows, compressor power input, temperature glide, COP and exergy losses of individual components are investigated here. For three different compressor rotation speeds and a constant condensing temperature (TH = 30°C) the evaporation temperature is varied between 0 °C and 6 °C; those temperature levels were selected, because they are typical for ground source heat pumps for building heating applications. The results are discussed in detail and some of them are compared to theoretical predictions.
4 Page 4 of 28
2. 2.1
Experiments Experimental setup and measurement equipment
The experimental setup with the main components of the heat pump is schematically shown in Fig. 1; it includes a compressor, an expansion valve, a condenser, an evaporator, liquid tank reservoir and further auxiliary valves for controlling the mass flow of the heat source and sink. The states are numbered in Fig. 1, because we refer to them in a later figure. The hermetic reciprocating compressor (GEA Bock HGHC12P), especially developed for hydrocarbons (isobutane, propane and propene), has two cylinders and a maximum power input of 2.2 kW. Using a frequency inverter, the rotation speed can continuously be varied and controlled between 1050 min-1 and 2100 min-1 while the electrical power consumption is monitored. Both heat exchangers, evaporator and condenser, are designed as counter flow double-pipe heat exchangers, because their heat transfer characteristics are well known. Each heat exchanger consists of 10 straight sections with a length of 1.4 m connected by return bends. In Fig. 1 only 3 straight sections with 2 bends are shown for both heat exchangers, for the sake of brevity. The total heat transfer areas of the evaporator and condenser are 0.5 m² and 0.75 m², respectively. The refrigerant flows through the inner tube (evaporator: inner diameter = 10 mm; wall thickness = 1 mm, condenser: inner diameter = 13 mm; wall thickness = 1.5 mm), while the secondary heat transfer fluid (water) flows through the annulus (inner diameter: 16 mm evaporator, 19 mm condenser and a wall thickness of 1.5 mm for both). At the inlets and outlets, as well as at every return bend, of the heat exchangers the temperature of the refrigerant and the secondary heat transfer fluid are measured by resistance thermometers. Furthermore, the pressure is recorded at the inlets and outlets. In order to inspect the thermodynamic state of the refrigerant, visually, four sight glasses are installed. Next to the sight glasses, sampling points are mounted in order to take specimens of the refrigerant. A gas chromatograph with a relative error of 2 % (Agilent Technologies 7890B GC System) is used to determine the exact composition of the mixture; potential composition shifts of the mixture at different positions within the plant can, thus, be observed. The refrigerant mass flow is quantified accurately by a Coriolis flow meter (KROHNE MFS3081 K 1.5 E). Initially, within this study, temperature and pressure transducers at the inlet of the flowmeter (Fig. 1, grey background) were not installed yet, because much smaller pressure drops through the flowmeter were expected.
However, the use of Coriolis flow meters lead to both, heat losses and pressure drops (dependent on the fluid and the compressor rotation speed between 2 and 9 bar); and thus, temperature and pressure transducers were subsequently installed, in order to characterize the flowmeter as well as the compressor in detail. Coriolis flow meters are generally not suitable to measure two-phase flows. Therefore, despite the high pressure losses within the flow meter, it has to be ensured phase changes are avoided. Two positions within the cycle are 5 Page 5 of 28
suitable accordingly, behind the condenser in the liquid phase or behind the compressor in the gas phase. Commonly, the subcooling of the refrigerant at the outlet of the condenser is small (subcooling temperature of 1 K) which, in combination with the pressure drop of the Coriolis flow meter, leads to a high risk of a partial evaporation inside the Coriolis flow meter and to incorrect measurements. This risk does not exist at the outlet of the compressor and therefore it appeared to be the best position for the flow meter in order to measure the refrigerant mass flowrates accurately. The mass flow of the secondary heat transfer fluid in the evaporator and condenser are recorded by turbine flow sensors, in Fig. 1 named rotameter (GEMS Sensors&Controls FT110). A fine metering needle valve is used as expansion valve, and coupled with a servo-motor which allows controlling the evaporation pressure via PC. The liquid tank reservoir (7.5 liter) is necessary on the one hand to perform experiments for different compressor rotation speeds (the refrigeration charge depends on the compressor rotation speed) and on the other hand to compensate fluctuations during the experiments while changing the conditions. The charge amount of the cycle strongly depends on the fluid and its mean density; for example, in case of pure isobutane the charge amount is about 152 g and in case of propene about 730 g. The charge amount for the mixtures is, depending on the composition, between those two values. Based on our experience, a too high charge amount does not influence the steady state performance of the cycle as long as the installed fluid reservoir is not fully filled with liquid. In contrast, a lack of refrigerant in the cycle leads to process changes like incomplete condensation and may distort the results. In order to prevent this, the heat pump is filled with the necessary charge amount related to the maximal possible compressor rotation speed and the maximal possible evaporator temperature. In case of operation points with lower mass flow rates (e.g. at a compressor rotation speed of 1050 min-1) the excess refrigerant is buffered in the refrigerant reservoir. In order to minimize the heat loss of the whole heat pump, the heat exchangers, every measurement sensor as well as every connecting pipe were well insulated with polyurethane foam of 20 mm wall thickness. Regarding temperature and pressure measurements, resistance thermometers (PT 1000) of type AA and calibrated relative pressure transducer (piezo stainless steel sensor ICS IMP331) are used. Overall 40 resistance thermometers, 6 pressure sensors, 2 rotameters, and one Coriolis flow meter are installed. Those components as well as the servo-motor of the expansion valve and the frequency inverter are connected to a PC by an USB multifunction DAQ device (Labjack U12). The software tool LabVIEW [18] was used for monitoring, recording and controlling the experiments. The corresponding values are recorded every second within a measuring cycle. Based on given tolerances from the manufacturers of the used measuring equipment the statistical error was estimated (error propagation) with respect to the COPH; the maximal relative error regarding the COPH is 4.63 %. To validate this value, selected experiments for different compositions, compressor rotation speeds, evaporation and condensation temperatures were reproduced. This resulted in a maximum relative deviation of 2.5 %.
6 Page 6 of 28
2.2
Investigated parameters
For a systematic analysis of the interaction between the working fluid (-mixture) and the process, combinations of different evaporation and condensation temperatures are investigated for the pure fluids isobutane and propene and for 5 compositions of them. Table 1 shows the investigated fluid compositions, while the investigated combinations of evaporator and condensation inlet temperatures are listed in Table 2. Furthermore, all these experimental series were carried out for three different compressor rotation speeds; 1050 min-1, 1500 min-1 and 2100 min-1. Water is used as secondary heat transfer fluid for both, the evaporator (heat source) and the condenser (heat sink), respectively. The water mass flow rate through the evaporator is kept constant at 8 kg min-1. The water mass flow through the condenser, as well as the opening position of the expansion valve are both control parameters for adjusting the evaporation and condensation pressures, and thus, for the selected temperatures. By varying the water mass flow rate in the heat sink without changing the opening position of the expansion valve, the whole cycle can be shifted to higher or lower temperature levels without changing the spread of both temperature levels (evaporation and condensation temperature). In contrast, changing the opening positon of the expansion valve, allows to vary the spread between the low and high temperature levels. In summary, the evaporator inlet temperature can be adjusted independent of the flow rate of the heat source (as long as no pinch point problem occurs in the evaporator). Based on this decoupling a systematic investigation of the influence of the evaporator inlet temperature is possible. The water mass flow rate in the condenser was varied between 1.5 and 7.5 kg min-1. The water inlet temperature of the evaporator and condenser are held constant at 17 °C and 14 °C, respectively. Depending on the investigated condensation temperature, the heat sink water outlet temperature varied between 26 and 47 °C. For a condensation temperature of 30 °C the outlet temperature of the heat sink is 34 °C ±2 °C. This temperatures are sufficiently high for heating floor applications. After the process reached steady state conditions, the data was recorded for about 20 minutes; the data analysis is based on averaged values. 2.3
Data reduction
For all thermodynamic analyses, the enthalpies and entropies for the different states are needed. Those were taken from the NIST REFPROP [19] database for the working fluids and for water at the measured temperatures and pressures. In the saturated regime, the enthalpy of the working fluid was calculated from the energy balance in the heat exchanger (which is assumed to be adiabatic towards the surrounding) and enthalpy and pressure were used as independent variables, together with the known composition (see below). The heat flow rates for cooling
and heating
were easily calculated by means of the first law of thermodynamics,
written for water as an example in Eq. 1. The European sign convention for heat and work is used: both are positive, when they are transferred to the system. The enthalpies are based on the measured inlet and outlet temperatures.
7 Page 7 of 28
Eq. 1
In combination with the measured power input to the compressor, the coefficient of performance for the heat pump is given by:
Eq. 2
The COP of a reversible (Carnot) cycle was calculated in order to compare the experimental values with the theoretical limit:
Eq. 3
Commonly, the temperatures
H
and
L
are the thermodynamic mean temperatures of the heat source and heat
sink (condenser and evaporator). If only the cycle shall be analysed, the Carnot-COP calculated with the thermodynamic mean temperatures of the working fluid can also be very helpful. Most of the thermodynamic states are fully defined by the recorded temperatures and pressures. The evaporator inlet state is saturated, thus, for pure fluids temperature and pressure are not sufficient to define the state. Basically, there are two different ways of calculating the evaporator inlet state; one is assuming an adiabatic expansion (hco,out = hev,in) and the other is assuming an adiabatic evaporator (Δ
Ref
=Δ
w,L).
Certainly, neither of the two processes are strictly adiabatic, but the temperature differences between the fluid and the environment is in our case higher for the expansion valve, thus, assuming an adiabatic evaporator seems to be more accurate and was used here. From the energy balance of the evaporator, the enthalpy at the evaporator inlet is derived:
Eq. 4
The heat losses of the evaporator were estimated and are very small compared to the transferred heat, justifying the assumption of adiabatic conditions. For fluid mixtures, the combination of temperature and pressure is sufficient to define a state within the two phase region.
8 Page 8 of 28
Besides the COPH, the analysis of exergies [20],[21]. The exergy losses
is a suitable tool to evaluate the performance of a process
of each component indicate the potential for improvements, this is especially,
important for fluid mixtures. The reduction of exergy losses in the heat exchangers by means of the targeted use of the non-isothermal isobaric phase change is one of the main ideas for using mixtures as working fluids. The exergy loss for a component i is calculated from the difference between entering and exiting exergy flow rates:
Eq. 5
Here,
and
are inlet and outlet exergy flow rates, respectively; Pel is the compressor power input and
the exergy flow rate induced by a heat flow. The exergy for each state is given by:
Eq. 6
The non-isothermal isobaric phase change is one basic physical property of zeotropic mixtures and known as temperature glide (TG). In general, regarding a fixed pressure and composition the TG is defined as the temperature difference between the dew and the bubble point:
Eq. 7
Behind the throttle, where a saturated fluid is entering the evaporator, the theoretical (isobaric) TG can be calculated by:
Eq. 8
Due to pressure losses, the real TGs of the working fluid in the heat exchangers differ from the theoretical ones. The determination of the real TG requires the knowledge of the starting and ending temperature of the phase change. Regarding exemplarily the evaporator, the inlet temperature is known but due to the fact that the temperatures of the working fluid within the evaporator are measured only every 1.4 m, the exact state at the end of the evaporation is unknown. Therefore, an alternative procedure was used: From the determined temperature gradients in each heat exchanger section of 1.4 m length it is easily recognized whether
9 Page 9 of 28
evaporation is completed in a section or not. From the temperature change in the last section with two-phase fluid, the evaporation endpoint temperature can be determined within ± 1 K.
Eq. 9
Temperatures and pressures were initially not measured in front of the Coriolis flow meter, yet; however, those flow meters commonly lead to pressure drops and to heat losses. In order to determine the exergy loss in the compressor, the state at the exit of the compressor or the entrance state of the flow meter is inevitably needed. Based on 17 independent measurements for the pure fluids isobutane and propene and for different operating points, the characteristic pressure drop of the flow meter was fitted as a function of outlet state and mass flow rate; this fit was used here for data evaluation. From these independent measurements, the mean absolute error of the compressor exergy losses were evaluated to be 12 W (min: 0.14 W; max: 61.26 W), leading to a mean relative error of ± 3 % here.
10 Page 10 of 28
3.
Results and discussion
Experiments for two pure fluids and five mixtures at several compressor rotation speeds and different combinations of evaporation and condensation temperatures were carried out. Here, the detailed analysis focusses on the variation of the evaporation temperature. Fig. 2 shows the COPH as a function of the mole fraction of propene (xR1270), for the evaporator inlet temperatures: 0 °C, 3 °C, 6 °C and for two different compressor rotation speeds (n): 1050 min-1 (Fig. 2a), 2100 min-1 (Fig. 2b). Throughout, the condensation temperature is kept constant at 30 °C. Based on the limited evaporator heat transfer area, it was not possible to carry out the experiments for the whole composition range for n = 2100 min-1 and evaporator inlet temperatures of 3 °C and 6 °C. First of all, it can be seen that the COPH increases with the evaporator inlet temperature TL. Furthermore, a direct comparison of Fig. 2a and Fig. 2b shows a remarkable reduction of the COPH with the compressor rotation speed. With respect to the COPH change with the mole fraction of propene the COPH always increases first, passes through a maximum and finally slightly decreases for pure propene (R1270).
The maxima are always found at a propene mole fraction of 72 % and are only slightly better (5 % - 7 %) than the one of propene, which is the better performing pure fluid here. The maximum COPH (5.49) is found for TL = 6 °C and n = 1050 min-1; in this case the COPH of propene is about 5.11. Regarding the pure fluids, propene (R1270) always performs better than isobutane (R600a); e.g. in case of an evaporation temperature of 6 °C and n = 1050 min-1, the COPH of propene is about 20 % higher compared to isobutane. The power consumptions per mass flow rate of the compressor (abbreviated as specific work, wel) as a function of the mixture composition for two different compressor rotation speeds 1050 min -1 and 2010 min-1 are exemplary shown for TL = 0 °C in Fig. 3. It can be observed that the specific work decreases with increasing mole fraction of propene, goes through a minimum and increases again for higher propene fractions. The specific work for pure isobutane is always higher than the one of propene. Compared to the composition of the maximum COPH (Fig. 2) the minimum of the specific compressor work can be found at the same composition, leading to the assumption that the compressor work has the major impact on the COP of the entire process.
Furthermore Fig. 4 shows the COPH change with compressor rotation speed for TL = 6 °C and xR1270 = 43.47 %. Also, the specific work of the compressor, and the absolute value of heat flow rates per mass flow rate of fluid (abbreviated as specific heat, qh) of the condenser are plotted for 3 different rotation speeds. The specific work of the compressor also increases with the rotation speed (at the same pressure ratio), thus the compressor becomes less efficient. Within the compressor, the higher losses are induced by higher 11 Page 11 of 28
piston velocities and accelerations, increased friction (piston – cylinder wall) and losses through the compressor valves, caused by the higher mass flow rates. The higher irreversibility leads to increased fluid outlet temperatures which also lead to larger transferred heat per mass qh. But the compressor work increases more than qh thus, the COPH decreases. Assuming sufficient heat exchanger areas and adiabatic conditions for the condenser and the connection tubes, the gradients of qH and wel would be equal; however, even then the COPH would decrease. The main reason of the compressor work increase with compressor rotation speed can be found by inspecting the pressure drops of the whole cycle. Fig. 5 shows the pressure drops in the evaporator and condenser together with the mass flow rates of the working fluid as a function of the compressor rotation speed. First of all, it is clear that the mass flow rate of the working fluid increases with higher rotation speeds; furthermore, the pressure drops also increase. From common approaches for tube friction it is known that the pressure drop is a strong function of the flow velocity, which again is proportional to the mass flow rate. The gradient as well as the value of Δp are larger for the evaporator than for the condenser; this is due to the lower flow cross section of the evaporator which also influences the pressure drops grossly.
A further reason for the larger pressure drop in the evaporator compared to the condenser are the differences in the thermodynamic states. Both heat exchangers are in large parts perfused by saturated fluid, however, the average vapor quality (over the length) is larger in the evaporator than in the condenser; commonly, pressure drops for vapors are higher than for liquids at comparable conditions. Finally, for a constant condenser inlet temperature, the pressure drop in the evaporator leads to an increased pressure ratio of the compressor, raising the power need. The increase of the COPH (see Fig. 2) with increasing evaporator inlet temperature TL at a constant condensation temperature is what is expected from the Carnot-cycle. With increasing TL the quotient between the thermodynamic mean temperatures Strictly speaking,
L
L
and
H
decreases and according to Eq. 3 the COPH increases, too.
can only be increased until the pinch point is reached. The reduction of the condensation
temperature has similar effects, the COP rises; however the outlet temperature of the secondary fluid is also decreased and the use of the heat pump vanishes. The most interesting aspect of the measurements illustrated in Fig. 2 is the composition dependence of the COPH. In general, it is assumed that the use of fluid mixtures has a great potential for improving thermodynamic cycles, but our measurements show only a minor increase of the COPH compared to the better performing pure fluid propene. Similar results for a mixture of isobutane and propene were also found by Chang et. al. [1]. In their cooling cycle the COP of the best mixture was 6 % higher than the one of the better performing pure fluid; for heating, the highest COP was reached by pure propene. Taking into account the need of larger heat exchangers due to smaller convection coefficients for mixtures, leading to such small 12 Page 12 of 28
advantages in the performance, the use of a mixture would probably not be justified. Theoretical approaches often show a greater benefit for using mixtures; in the following the reasons for this mismatch between experiment and theory is discussed.
Fig. 6 shows the results of different calculations, all for an evaporation temperature of 3 °C and a condensation temperature of 30 °C. The simulations are based on simple thermodynamic balances of the main components (evaporator, compressor, condenser, expansion valve) as they have been carried out by many authors (e.g. [12],[13]) with the aim of evaluating fluids in cycles and in order to create fluid rankings. Four different cases are considered here: (a) an isentropic compressor ηc,s = 1.0, (b) an irreversible compressor with constant isentropic efficiency ηc,s = const. = 0.585, and two cases with fluid composition dependent isentropic efficiencies ηc,s = f(x1270) (c) without and (d) with x1270 dependant pressure losses in the heat exchangers. Additionally, the measured COPs for these conditions (e) which are also included in Fig. 2a are shown. For cases (c) and (d), the isentropic efficiencies as well as the pressure losses as a function of the mole fraction of propene (summarised in Table 3) are taken from measurements at the given conditions. The constant isentropic efficiency (case b) is the average value of the xR1270 dependant isentropic efficiencies. For the isentropic compressor case, the COPH reaches a maximum at a propene mole fraction of 70.71 %. The mixture COP is significantly better by 26.71 % and 30.02 % compared to the pure fluids isobutane and propene, respectively. For a reduced constant isentropic efficiency of the compressor (Fig. 6b), there is a significant reduction of the COPH which is independent on the mole fraction. Furthermore, the advantage of the mixture compared to the pure fluids is strongly reduced. While isentropic efficiencies decrease, the compressor outlet temperature as well as the thermodynamic mean temperature increase. Thus, the COP H is decreased and the impact of the TG on the thermodynamic mean temperatures becomes smaller. Interestingly, now the order of the COPs of the pure fluids for case b is different than the experimentally determined ones, isobutane performs better than propene; only by the implementation of mole fraction dependant isentropic efficiencies the simulations lead to the experimentally found ranking of the pure fluids. However, in this case only a minor advantage of propene compared to isobutane remains but in comparison to case (b) the order of the pure fluids changes. Otherwise, the curve is not influenced strongly. Starting from pure isobutane (ηc,s = 0.55) the measured isentropic efficiency of the compressor increases, then passes a maximum (ηc,s = 0.66) at a propene mole fraction of 72 % and decreases again towards pure propene (ηc,s = 0.60). Regarding the fluid dependant compressor performance, approaches to simulate such cycles with a constant isentropic efficiency seem to be not suitable for correct assessments and thus, should be combined with a realistic compressor model as long as a single compressor is used with different mixtures. Probably, the highest (constant) isentropic efficiency could be reached for each mixture with fluid optimized compressors, in part justifying previous approaches. With additional implementation of the pressure losses in the evaporator and condenser (Fig. 6d), the values of 13 Page 13 of 28
the COPH come closer to the experimental values, although the experimental values remain much lower. The curve 5d is shifted to lower values and the advantage of the mixture against pure propene becomes lower, again.
One of the main reasons for the low experimental COPs is the additional pressure drop through the Coriolis flowmeter. Fig. 7 shows the process in a p-h diagram for the same operating conditions as the simulations are done for two different compressor rotation speeds; the state numbering is the same as in Fig. 1. Due to the Coriolis flow meter, the increase in pressure pcompressor achieved by the compressor (state 1-state 2) in both cases is higher compared to the pressure increase which would actually be needed to reach the condensation temperature of 30 °C without a flowmeter. The pressure drop in the flow meter (in Fig. 7 from point 2 to 3) depends strongly on the refrigerant mass flow rate and the mass flow rate depends on the compressor rotation speed. In the two cases shown in Fig. 7 the pressure drops are 2.98 bar for a compressor speed of 1050 min-1 (a) and 5.39 bar for 2100 min-1 (b). Depending on the fluid and the compressor rotation speed, typical pressure drops through the flowmeter are found to be between 2 and 9 bar, in general. This additional pressure drop has to be compensated by the compressor, and thus, the power consumption of the compressor increases. To investigate the influence of the flow meter on the COPH, some additional measurements were carried out for nearly the same mixture compositions and operating conditions without the Coriolis flow meter, using a bypass; they are shown in Fig. 8.
Fig. 8 shows experimental COPH values with / without the use of the Coriolis flow meter and simulated COPH as a function of the propene mole fraction, all for TL = 3 °C, TH = 33 ± 2 °C and a compressor rotation speed of 1050 min-1. It is seen that the COPH increases considerably if the Coriolis flow meter is omitted. The usage of the Coriolis flow meter leads to an increased compressor power consumption by a factor of up to 1.2 while H
increases only by a factor of 1.02. Nevertheless, comparing the simulation results of Fig. 6 with the
measured COPH in Fig. 8 (without the use of the flow meter), some deviations remain. It is assumed that these originate from the used isentropic efficiencies of the compressor which were fitted to measurements with the flow meter, while the isentropic efficiencies decrease with lower pressure drops, as is the case without flow meter. Thus, the simulations lead to overestimated COPH values compared to the measurements without flow meter. As can be seen in Fig. 8, by using process parameters, as summarised in Table 4, which are measured without a flow meter, the deviations between the simulations and experiments decrease. Apart from the negative influence on the value of the COPH, the investigations show that the use of the Coriolis flowmeter does not influence the composition at maximum COPH and thus, the comparability of the fluid ranking is conserved. But since it is not clear if this conclusion is generally valid; detailed investigations
14 Page 14 of 28
on this topic will follow in near future. Despite all disadvantages of using flow meters, the mass flow rate of the refrigerant is of great importance and Coriolis flow meters are the tools of choice. Apart from the flow meter, there are further sources of losses which decrease the efficiency of the cycle and probably lead to deviations between simulations and measurements. The thermal losses of the heat exchangers, the sight glasses and all connection pipes are between 26 Watt and 214 Watt, in total, depending on the fluid and the operation point. This corresponds to 3 % - 15 % of the
value and thus leads to a
decrease of the COPH of 15% at maximum. Maximal thermal losses are detected for isobutane, explaining the highest deviation between the simulation and the measurement in this case.
The pressure drops in the evaporator and the condenser as well as the mass flow rate as a function of the mole fraction of propene are exemplarily shown in Fig. 9 for a compressor rotation speed of 1050 min-1 and TL = 0°C, TH = 30°C. The density of propene (for the same conditions) is approximately two times higher than the one of isobutane, therefore the mass flow rate of the working fluid increases with the propene mole fraction and the pressure drop increases with the mass flow rate. Due to the smaller tube diameters, the pressure drop again changes stronger with the propene content in the evaporator than in the condenser. This increase of the pressure drop in the evaporator (see Fig. 7) has to be overcome by the compressor and requires a larger power per mass flow rate. Fig. 10 shows T-h-plots of the pure fluid isobutane (a, left) and for the best isobutane/propene mixture (b, right); for both, the heat exchangers are assumed to be non-isobaric. Additionally, the temperature changes of the secondary fluids as well as the thermodynamic mean temperatures are depicted. As shown in Fig. 10a, the pressure drops affect the process also in another way: for a pure fluid, a pressure drop leads to non-isothermal conditions in the saturation regime. As a consequence, the thermodynamic mean temperatures in the condenser and in the evaporator are decreased and the ratio of L
and
H
becomes smaller leading to smaller COPs (Eq. 3). The main thermodynamic advantage of using
zeotropic mixtures which is the temperature glide (TG) resulting in reduced exergy losses and in an increased ratio between
L
and
H
can be seen by comparing the both T-h-plots. The temperature glide in the condenser
becomes larger for the mixture while it is inverted in the evaporator. This leads on the one hand to a decreased difference between
L
and
H
and on the other hand to a better match between the temperature gradient of the
secondary fluid and the working fluid in the heat exchangers. However, these advantages are counter-acted here by pressure drops which reduce the real temperature glides and lead to the only minor advantage of the mixture.
This leads to the question which operating conditions may be increasing or reducing pressure drops and thus influencing the COP. Fig. 11 shows a comparison between measured and calculated evaporator TGs
15 Page 15 of 28
(evaluated for isobaric conditions) as a function of the propene mole fraction, determined for two different compressor rotation speeds 1050 min-1 and 2100 min-1, respectively. The ideal TG starts with 0 K for pure isobutane, increases till reaching a maximum of 6.87 K at a composition of xR1270 = 72 % and decreases to 0 K for pure propene, again. The measured TG starts and ends at negative values for the pure fluids which is due to the pressure drops. At a propene mole fraction of about 11.9 % the TG is large enough to over-compensate the pressure drop and the temperature gradient of the working fluid during evaporation gets positive. The experimental curves are nearly parallel to the theoretical ones, but shifted to smaller values. At higher compressor rotation speed, the differences between the calculated and measured TG increase due to the higher pressure drops (see Fig. 5 ) at higher Re numbers, reducing the small advantage of a zeotropic mixture (compare Fig. 2). Although the targeted improvement of a thermodynamic cycle with zeotropic fluids is reduced strongly if pressure drops get important, this could be improved using mixtures with higher temperature glides. But mixtures with a high temperature glide commonly are composed of fluids with strongly differing evaporation pressures which usually differ strongly in their densities, too. Although this makes things more complicated, this field seems worth further investigations. Apart from the COPH, the exergy losses in different parts of the cycle are also composition dependent; this shall be analysed finally. In Fig. 12 the specific exergy losses of the main components (including the flowmeter) as a function of xR1270 are exemplary shown for a compressor rotation speed of 1050 min-1 and TL = 0 °C, TH = 30 °C. Except for the throttle and the flowmeter, all exergy loss curves pass a minimum. The exergy losses in the heat exchangers are only slightly dependent on composition. This is easily explained by the boundary conditions: the heat capacity flow rates of the secondary fluids are high and consequently the temperature gradients are small; thus, even in case of pure fluids the exergy losses are already small. Here, the positive influence of the temperature glide on the COPH is mainly due to the decreased gap between
H
and
L.
The
highest exergy losses are found in the compressor which is also partially determined by the pressure drops in the heat exchangers; therefore this component is most important. The minimum exergy loss occurs at a mixture composition of xR1270 = 72 %; this corresponds directly to the composition of the maximum COPH. It is observed that the specific exergy loss the flowmeter increases slightly with increasing mole fraction of propene, because of the increasing pressure drop with increasing propene mole fraction.
4.
Conclusion
In the present work, the influence of two pure fluids R1270 (propene) and R600a (isobutane) and their zeotropic mixtures on the COP of a heat pump for different evaporator inlet temperatures, condensation temperatures and compressor rotation speeds are investigated experimentally. The COPH is only slightly changing with composition; a COP improvement of 5 - 7 % relative to propene was found. This improvement 16 Page 16 of 28
results mainly from the decrease of the difference between the two thermodynamic mean temperature levels for heat transfer to and from the cycle and is based on the temperature glide (TG). The exergy losses in the heat exchangers are only lowered slightly by the use of these mixtures, with relatively small differences in saturation temperatures at the investigated pressures. The evaluation of the experimental data as well as a theoretical analysis show that pressure drops as well as the non-ideality of the compressor are mainly responsible for the findings. Also, the pressure drops within the heat exchangers reduce the temperature glide and thus, reduce the main advantage of the mixture usage in thermodynamic cycles. Non-ideal compressors lead to higher outlet temperatures and thus, to increased thermodynamic mean temperatures in the condenser. In consequence, the temperature glide has a minor influence on the reduction of the mean temperature differences. The presented results seem to indicate that the use of fluid mixtures is not justified regarding the small improvement of the COPH. However, in case of lower heat capacity flows of the secondary fluids the use of fluid mixtures could be more advantageous. Also, one has to design the tube diameters carefully in order to reduce pressure drops in the heat exchangers and further parts if zeotropic mixtures shall retain their thermodynamic advantage compared to pure compounds. Besides, the pure fluid propene always performs about 20 % better than isobutane, which would contradict simple theory. The present investigation shows that this behavior can only be reproduced correctly by implementing validated mixture dependent isentropic efficiencies, as long as a single compressor is used for all fluid compositions. Neglecting this, may lead to wrong (mixed) working fluid rankings for thermodynamic cycles. Increasing the compressor rotation speed increases the irreversibility within the process as well. Thus, on the one hand the values of the COPH are decreased and on the other hand the advantage of the mixture usage compared to the pure fluid becomes smaller. Increasing the evaporation temperature or decreasing the condensation temperature results in a dropped gap between the both thermodynamic mean temperatures and thus, the COPH rises. However, the degrees of freedom regarding those temperatures are commonly very small. In future, fluid mixtures with a greater temperature glide as well as secondary fluids with lower heat capacity flows should be investigated.
5.
Acknowledgment
We thank KROHNE Germany for provision of technical measuring equipment. Technical assistance by Stephan Steinbrink and Andreas Görnt as well as laboratory support by Andreas Lennert is gratefully acknowledged.
17 Page 17 of 28
6.
6.1
Appendix
Nomenclature coefficient of performance of a heat pump
[-]
exergy flow rate of each state
[kW]
specific exergy loss of each component
[kJ kg-1]
exergy loss flow rate
[kW]
specific enthalpy
[kJ kg-1]
mass flow rate
[kg s-1]
rotation speed of the compressor
[min-1]
electrical power consumption of the compressor
[kW]
pressure
[bar]
specific heat flow
[kJ kg-1]
quality
[-]
heat flow
[kW]
specific entropy
[kJ kg-1 K-1]
temperature
[K]
temperature glide
[K]
specific work of the compressor
[kJ kg-1]
mole fraction
[-]
Greak symbols enthalpy flow rate change p
pressure drop
[mbar]
isentropic efficiency of the compressor
[-]
Subscripts bubble
bubble point
c
compressor
co
condenser
dew
dew point 18 Page 18 of 28
ev
evaporator
H
at high level (e.g. temperature, enthalpy, mass flow rate, heat flow rate in condenser)/ heat pump
i
component
in/ out
inlet/outlet
L
at low level (e.g. temperature, enthalpy, mass flow rate, heat flow rate in evaporator)
m
middle (e.g. middle temperature at low/high level)
Ref
working fluid (refrigerant)
s
isentropic
vap
vapour
w
secondary heat transfer fluid (water)
0
fluid properties/conditions at 20 °C and 1,01 bar
Acronyms CFCs
chlorofluorocarbons
COP
coefficient of performance
DAQ
data Acquisition
HC
hydrocarbon
HCFCs
hydrochlorofluorocarbons
HFCs
hydrofluorocarbons
ORC
organic rankine cycle
TG
temperature glide
19 Page 19 of 28
6.2
References
[1] Chang Y, Kim M, Ro S. Performance and heat transfer characteristics of hydrocarbon refrigerants in a heat pump system. International Journal of Refrigeration 2000;23(3):232–42. [2] Richardson RN, Butterworth JS. The performance of propane/isobutane mixtures in a vapourcompression refrigeration system. International Journal of Refrigeration 1995;18(1):58–62. [3] Mani K, Selladurai V. Experimental analysis of a new refrigerant mixture as drop-in replacement for CFC12 and HFC134a. International Journal of Thermal Sciences 2008;47(11):1490–5. [4] Park K, Jung D. Performance of heat pumps charged with R170/R290 mixture. Applied Energy 2009;86(12):2598–603. [5] Jung D, Kim C, Song K, Park B. Testing of propane/isobutane mixture in domestic refrigerators. International Journal of Refrigeration 2000;23(7):517–27. [6] Kuijpers L, Wit J de, Janssen M. Possibilities for the replacement of CFC 12 in domestic equipment. International Journal of Refrigeration 1988;11(4):284–91. [7] Akash BA, Said SA. Assessment of LPG as a possible alternative to R-12 in domestic refrigerators. Energy Conversion and Management 2003;44(3):381–8. [8] Tashtoush B, Tahat M, Shudeifat MA. Experimental study of new refrigerant mixtures to replace R12 in domestic refrigerators. Applied Thermal Engineering 2002;22(5):495–506. [9] Lee H, Kim H, Kang D, Jung D. Thermodynamic performance of R32/R152a mixture for water source heat pumps. Energy 2012;40(1):100–6. [10] Mohanraj M, Muraleedharan C, Jayaraj S. A review on recent developments in new refrigerant mixtures for vapour compression-based refrigeration, air-conditioning and heat pump units. Int. J. Energy Res. 2011;35(8):647–69. [11] Braimakis K, Preißinger M, Brüggemann D, Karellas S, Panopoulos K. Low grade waste heat recovery with subcritical and supercritical Organic Rankine Cycle based on natural refrigerants and their binary mixtures. Energy 2015;88:80–92. [12] Brown JS. Predicting performance of refrigerants using the Peng–Robinson Equation of State. International Journal of Refrigeration 2007;30(8):1319–28. [13] Saleh B, Wendland M. Screening of pure fluids as alternative refrigerants. International Journal of Refrigeration 2006;29(2):260–9. [14] Roskosch D, Atakan B. Reverse engineering of fluid selection for thermodynamic cycles with cubic equations of state, using a compression heat pump as example. Energy 2015;81:202–12. [15] Lampe M, Stavrou M, Bücker HM, Gross J, Bardow A. Simultaneous Optimization of Working Fluid and Process for Organic Rankine Cycles Using PC-SAFT. Ind. Eng. Chem. Res. 2014;53(21):8821–30. 20 Page 20 of 28
[16] Wen M, Ho C, Hsieh J. Condensation heat transfer and pressure drop characteristics of R-290 (propane), R-600 (butane), and a mixture of R-290/R-600 in the serpentine small-tube bank. Applied Thermal Engineering 2006;26(16):2045–53. [17] Rajapaksha L. Influence of special attributes of zeotropic refrigerant mixtures on design and operation of vapour compression refrigeration and heat pump systems. Energy Conversion and Management 2007;48(2):539–45. [18] LabVIEW Professional Development System: National Instruments; 2013. [19] Lemmon, E.W., Huber, M.L., McLinden, M.O. NIST Standard Reference Database 23: Reference Fluid Thermodynamic and Transport Properties-REFPROP. 9th ed. Gaithersburg: National Institute of Standards and Technology; 2013. [20] Ahamed JU, Saidur R, Masjuki HH. A review on exergy analysis of vapor compression refrigeration system. Renewable and Sustainable Energy Reviews 2011;15(3):1593–600. [21] Ganjehsarabi H, Güngor A, Dincer I. Exergetic performance analysis of Dora II geothermal power plant in Turkey. Energy 2012;46(1):101–8.
21 Page 21 of 28
heat source out p
PC controlled
T
M
p
evaporator
FI
5
p
1 M
sight glass
p
T
expansion valve
p
T
T
compressor
rotameter
motor
heat source in heat sink out
2
p p
CF
liquid tank reservoir
p
condenser
3
T
T
p T p
T
4 sample point
sampling device
T
heat sink in suction line discharge gas line discharge liquid line
T p CF FI
temperature pressure Coriolis flow meter frequency inverter
Fig. 1 Schematic diagram of the experimental setup
a
b 6
TL= 0°C
TL= 3°C
TL= 6°C
6
1050 min-1
4
3
2
TL= 3°C
TL= 6°C
2100 min-1
5
COPH [-]
COPH [-]
5
TL= 0°C
4
3
0.0
0.2
0.4
0.6
0.8
2
1.0
mole fraction propene
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene
Fig. 2: COPH as a function of mixture composition; determined for the indicated evaporator inlet temperatures and for compressor rotation speeds of a: 1050 min-1and b: 2100 min-1. Condenser inlet temperature TH = const. = 30 °C
22 Page 22 of 28
1050 min-1
2100 min-1
specific work [kJ kg-1]
160
140
120
100
80
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 3: Specific work of the compressor as a function of mixture composition for compressor rotation speed of 1050 min -1 and 2100 min-1, TL = 0 °C and TH = 30 °C
qH
wel
COPH
6
370
5
290 4 210 3
130 50
COPH [-]
specific heat, work [kJ kg-1]
450
1050
1500
2100
2
-1
rotation speed [min ] Fig. 4: COPH, specific compressor work and condenser heat as a function of the compressor rotation speed for x R1270 = 43.47 %, TL = 6 °C and TH = 30 °C
23 Page 23 of 28
pressure drop [mbar]
pcondenser
mRef
0.011
480
0.010
360
0.009
240
0.008
120
0.007
0
1050
1500
2100
mass flow rate [kg s-1]
pevaporator
600
0.006
-1
rotation speed [min ] Fig. 5 Pressure drops and mass flow rates of the working fluid as a function of the compressor speed for x R1270 = 43.47 % and TL = 6 °C, TH = 30 °C
a
15
b
c
d
e
COPH [-]
12
9
6
3
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 6: COPH as a function of the mole fraction of propene (R1270) for TL = 3 °C and TH = 30 °C; a: ηc,s = 1.0, b: ηc,s = const. = 0.585, c: ηc,s = measured values depending on x R1270 , d: ηc,s and dpcond,evap= measured values depending on x R1270 , e: measured COPs for 1050 min-1
24 Page 24 of 28
a
b 100
1050 min-1
4
10
pcompressor without flow meter
3 5
1
2
200
1
400
2100 min-1
pcompressor with flow meter
pressure [bar]
pcompressor with flow meter
4
10
pcompressor without flow meter
2 3
5
pevaporator
600
1
800
200
specific enthalpy [kJ kg-1]
1
400
pevaporator
600
800
specific enthalpy [kJ kg-1]
Fig. 7: p-h diagram for xR1270 = 72 %, TL = 3 °C , TH = 30 °C and a: compressor rotation speed of n = 1050 min-1 and b: compressor rotation speed of n = 2100 min-1
Coriolis flowmeter 7
bypass
simulation
1050 min-1
6
COPH [-]
pressure [bar]
100
5 4 3 2
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 8: Comparison of the COPH with / without use of a Coriolis flow meter and simulation depending on x R1270 for n = 1050 min-1, TL = 3 °C and TH = 35 ± 2 °C
25 Page 25 of 28
pevaporator
pcondenser
mRef
0.011
1050 min-1
480
0.009
360
0.008
240
0.006
120
0.005
0
0.0
0.2
0.4
0.6
0.8
1.0
mass flow rate [kg s-1]
pressure drop [mbar]
600
0.003
mole fraction propene Fig. 9: Pressure drops as a function of x 1270 for a compressor rotation speed of 1050 min-1and TL = 0 °C, TH = 30 °C
a
b heat source
150
heat sink
1050 min-1
temperature [°C]
temperature [°C]
120
90 2
60
0 -30
heat sink
1050 min-1
120
30
heat source
150
3
TH
4 1
TL
2
60 30 0
5
200
90
400
600
-30
800
-1
3
TH TL
4
1 5
200
400
600
800
-1
specific enthalpy [kJ kg ]
specific enthalpy [kJ kg ]
Fig. 10: T-h-Diagrams of a: pure isobutane and b: isobutane/propene mixture (x R1270=72 %) at TL = 0 °C, TH = 30 °C and n = 1050 min-1
26 Page 26 of 28
8
1050 min-1
TG [K]
4 0 -4 -7 TGlide,measured 8
TGlide,ideal
2100 min-1
TG [K]
4 0 -4 -7
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 11: TG in the evaporator as a function of x R1270 for two compressor rotation speed 1050 min-1, 2100 min-1 and TL = 0°C, TH = 30 °C
specific exergy loss [kJ kg-1]
70.0
1050 min-1
eloss, evaporator eloss, compressor
52.5
eloss, condenser eloss, throttle eloss, flowmeter
35.0
17.5
0.0
0.0
0.2
0.4
0.6
0.8
1.0
mole fraction propene Fig. 12: Specific exergy losses of each component as a function of the mole fraction of propene, for a compressor rotation speed of 1050 min-1 and TL = 0 °C, TH = 30 °C
27 Page 27 of 28
Table 1: Investigated mixture composition and according temperature glide as derived by Eq. 7
R1270 mole fraction [%] 0.00 11.90 43.47 55.31 72.00 83.12 100.00
R600a mole fraction [%] 100.00 88.10 56.53 44.69 28.00 16.88 0.00
temperature glide [K] 0.00 5.59 7.82 7,95 8.15 7.56 0.00
Table 2: Experimental matrix for steady-state conditions
↓
L
[°C] TH → 0 3 6
25 x
30 x x x
35
40
x
x
Table 3: Messured isentropic efficiency of the compressor, evaporator/ condenser pressure drop depending on the mole fraction of propene for TL = 3 °C, TH = 30 °C and a compressor rotation speed of 1050 min-1
R1270 mole fraction [%] 0.00 11.90 43.47 55.31 72.00 83.12 100.00
isentropic efficiency ηc,s [%] 55.05 49.27 59.40 60.91 65.95 64.65 60.17
pressure drop evaporator [mbar] 79.18 101.5 151.58 171.10 202.41 232.30 255.30
pressure drop condenser [mbar] 70.12 85.34 97.21 108.05 114.51 124.34 143.80
Table 4: Messured isentropic efficiency of the compressor, evaporator/ condenser pressure drop depending on the mole fraction of propene for TL = 3 °C, TH = 35 ± 2°C °C and a compressor rotation speed of 1050 min-1
R1270 mole fraction [%] 0.00 18.83 59.24 83.46 100
isentropic efficiency ηc,s [%] 55.75 55.78 65.64 65.43 67.21
pressure drop evaporator [mbar] 80.38 120.08 203.80 258.96 295.33
pressure drop condenser [mbar] 36.23 64.52 81.16 88.06 124.52
28 Page 28 of 28