Evaluation of a small-scale waste heat recovery organic Rankine cycle

Applied Energy 192 (2017) 146–158

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Evaluation of a small-scale waste heat recovery organic Rankine cycle Antti Uusitalo ⇑, Juha Honkatukia, Teemu Turunen-Saaresti Lappeenranta University of Technology, School of Energy Systems, Laboratory of Fluid Dynamics, P.O. Box 20, 53851 Lappeenranta, Finland

h i g h l i g h t s
Utilization of exhaust gas heat by means of small-scale ORC was studied.
High molecular weight siloxane MDM was used as the working fluid.
Experimental study was carried out at different operational conditions.
The potential for designing small-scale ORC systems with high efficiency was confirmed.

a r t i c l e

i n f o

Article history: Received 6 October 2016 Received in revised form 24 January 2017 Accepted 27 January 2017

Keywords: Waste heat recovery Organic Rankine cycle Organic fluid Siloxane Heat transfer

a b s t r a c t In recent years, the use of small-scale organic Rankine cycles (ORC) in exhaust gas heat recovery of reciprocating engines has been intensively studied. In this paper, the working fluid selection and experimental results of a small-scale ORC unit utilizing exhaust heat of a diesel engine are presented and discussed. Based on the working fluid selection study, siloxane MDM was evaluated as the most suitable fluid for the experimental system. The experiments were conducted with the aim of studying and analyzing the capability of the ORC process of recovering heat from the diesel engine exhaust. The high pressure MDM vapor was expanded through an expansion valve; thus, no power was extracted from the experimental setup and the main focus was on studying the performance of the process heat exchangers. The system under study was identified to be capable of efficiently recovering the waste heat of the exhaust gases, and the potential of using high molecular weight and high critical temperature fluids as the working fluids in high-temperature, small-scale ORC applications was confirmed. It was concluded that when using siloxane MDM as the working fluid, the requirements for the process sealing to withstand low vacuum conditions as well as the effective removal of non-condensable gases during the operation can be identified as one of the major challenges in achieving the targeted power output from this type of ORC systems. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction Organic Rankine Cycles (ORC) are similar to conventional Rankine cycles with the exception that an organic fluid is used as the working fluid. The use of ORCs enables to design small-scale power systems capable of operating at relatively low temperature levels [1]. ORC technology has been proven to be commercially viable especially in geothermal and biomass power plants [1], while the use of ORC in heat recovery applications has attracted increasing interest in recent years [1,2]. Previous studies e.g. [3–6] have investigated the use of ORCs for recovering waste heat from reciprocating engines showing the potential of using ORCs for recovering waste heat from both small- and large-scale engines, and discussed on the critical design aspects of this type of heat recovery systems. ⇑ Corresponding author. E-mail address: [email protected] (A. Uusitalo). http://dx.doi.org/10.1016/j.apenergy.2017.01.088 0306-2619/Ó 2017 Elsevier Ltd. All rights reserved.

Another potential application has been identified as exhaust gas recovery from gas turbines [7–9]. In addition, interest in the use of small-scale ORCs, ranging from 1 kW to 20 kW, in other applications such as the use of ORCs as domestic power units [10,11] has been on the rise in recent times. The selection of a suitable working fluid is one of the most important steps in the design of ORC systems, as it significantly impacts the performance of the cycle as well as the optimal cycle configuration and dimensions of the process components. Based on the shape of the vapor saturation curve on a temperatureentropy diagram, working fluids can be classified into three categories: dry fluids, isentropic fluids, and wet fluids [12,13]. In ORC systems the fluids having a dry expansion are often preferred [1]. When one selects a working fluid for an ORC, issues related to thermodynamic performance, expander type and the size of the heat exchangers, as well as condensing pressure and evaporation pressure have to be taken into account [14–16]. There are also several

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

147

Nomenclature Latin alphabet h specific enthalpy, kJ/kg P power, kW p pressure, bar qm mass flow rate, kg/s T temperature, °C U overall heat transfer coefficient, W/m2/K M molecular weight, kmol/kg m mass, kg n rotational speed, rpm v specific volume, m3/kg

eg in LMTD l out p rec s t v wf

Greek alphabet g efficiency, – / heat rate, kW e recuperator effectiveness, – x angular speed, rad/s

Abbreviations EG exhaust gas LUT Lappeenranta University of Technology MDM Octamethyltrisiloxane OP operational point ORC organic Rankine cycle PS plate and shell WF working fluid

Subscripts avg average c cycle eV evaporator

other important considerations and criteria, such as the thermodynamic and physical properties of the fluid, fluid stability at high temperatures, fluid compatibility with materials, environmental impacts, safety, fluid availability, and price [17]. This study concentrates on small-scale and high-temperature ORCs, meaning that the temperature level of the heat source is above 300 °C which is typical for exhaust gas temperature levels in reciprocating engines. For high-temperature ORC applications, fluids with high critical temperature and high molecular weight, such as heavy hydrocarbons and siloxanes have been identified as potential working fluid candidates in some previous studies [5,7,18]. This type of fluids have suitable thermodynamic properties for regenerative high-temperature ORC applications, including high conversion efficiency and dry expansion in the system expander [5,19,20]. When using a working fluid with a high critical temperature, the temperature and enthalpy drop over the process expander is typically relatively low leading to high expander outlet temperatures, which then favors the use of a recuperator in the system [5,19]. In general, the use of high critical temperature fluids results in a large expansion ratio over the expander, low condensing pressure, and the system can reach a high thermodynamic efficiency in high-temperature applications [21]. These kinds of fluids can be considered as suitable candidates for stationary small-scale ORCs, such as for waste heat recovery from small diesel generators, gas turbines, or externally fired domestic ORCs. However, the use of fluids having a high critical temperature and high molecular weight is not necessarily an optimal choice in moving applications, such as in heavy duty trucks, due to the restricted available space for the process heat exchangers and other components [2]. The thermal and chemical stability of the fluid can be identified as an important fluid selection criteria, particularly in the case of hightemperature ORCs. Erhart et al. [22] carried out an experimental study on working fluid decomposition in 7 large-scale hightemperature ORC plants. Their results indicated that decomposition of the fluid occurred over a long time period, and that the degradation of the fluid was caused by the use of petroleumbased lubricants in the system alongside the high working fluid temperatures in the process.

exhaust gas inlet logarithmic mean temperature difference liquid outlet pump recuperator isentropic turbine vapor working fluid

Different types of expanders have been proposed or considered for small-scale ORCs including different types of turbines [23–25], screw expanders [26], piston expanders [25], or vane expanders [10]. The majority of the small-scale ORC systems adopt a screw or scroll expander instead of a turbine and fluids with relatively low critical temperatures are often favored. However, the maximum pressure ratios over volumetric expanders are limited below 5–15 [27,28] which restricts the achievable cycle efficiency, especially in high-temperature ORC applications [21]. The turbine types considered in ORC systems are typically axial [29,30] or radial [31] turbines. Single-stage turbines have been reported to be capable of reaching pressure ratios over 100 which is typical of high-temperature ORCs using fluids of high molecular complexity and high critical temperature[32,33]. In this project, the temperature difference between the heat sink and exhaust gas is high which results in the use of high pressure ratio in the process and allows to reach high thermodynamic efficiency for the cycle. Based on a literature review of small-scale ORCs, the most of the research efforts have been concentrating on the use of fluids having a relatively low critical temperature and low pressure ratio in the process. Thus, there exists only a limited amount of available experimental data concerning the use of high molecular weight and high critical temperature fluids in small-scale ORC applications. In this paper, the working fluid selection and experimental results of a small-scale high-temperature and high pressure ratio ORC using octamethyltrisiloxane (MDM) as the working fluid are presented. The targeted power level for the ORC in this study is about 10–15 kW that would be suitable especially for waste heat recovery in about 100–200 kW scale engine systems. The aim of this paper is to study especially the performance of the heat exchangers and the capability of the experimental setup to recover the exhaust gas heat. In addition, the potential for producing mechanical power with the studied ORC system was studied and evaluated numerically. Based on the numerical and experimental results, the technical feasibility of this kind of ORC systems and the most critical design aspects are discussed and highlighted.

148

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

2. Methods The thermodynamic analysis of the ORC was carried out by using a commercial thermodynamic library Refprop [34] for solving the fluid state at each process node. The well-known calculation principles of Rankine cycles were applied. In the thermodynamic analysis, the exhaust gas inlet temperature of 400 °C and mass flow rate of 0.31 kg/s were selected according to the exhaust gas values obtained from the engine manufacturer. The exhaust gas outlet temperature of 200 °C was selected in order to avoid the exhaust gas temperature values below acid dew point [5]. Condensing temperature of 60 °C was used according to the available cooling capacity. Only a subcritical process was considered in this study and thus, the evaporation pressure was limited by using a maximum evaporator pressure of 90% of the critical pressure. The maximum working fluid temperature of 300 °C was selected to ensure the thermal stability of the fluid [35], and a small degree of superheating of 10 °C was used which is required to ensure the fluid exiting the once-through type of evaporator without any liquid droplets. The expander efficiency of 75% was used in this analysis which was evaluated to be a suitable value for small-scale ORC turbines according to the previous studies [5,23,31,32]. It should be noted that the obtainable turbine efficiency is highly dependent on the turbine type, working fluid, and operating conditions and a more detailed analysis on the obtainable turbine efficiency with different fluids is beyond the aims of this paper. The pressure losses in the process components were excluded in this preliminary analysis. In addition, a preliminary sizing of a radial turbine was included in the thermodynamic evaluation by using the non-dimensional design parameters, namely the specific speed and specific diameter. According to [36], the optimum values for radial turbines of about 0.5 for the specific speed and 3 for the specific diameter were used in this preliminary analysis for evaluating the turbine rotational speed and the diameter with different fluids and process design parameters. The specific speed can be defined as

Ns ¼

xq0:5 v

0:75

Dhs

ð1Þ

where x is the angular speed. The angular speed is defined as

x ¼ 2pn:

ð2Þ

As the value for specific speed was selected to reach a high efficiency for a radial turbine and the volumetric flow rate and isentropic enthalpy drop were calculated in the thermodynamic analysis, a suitable rotational speed was solved for each fluid and operational condition from the definition of specific speed. In the experimental section of this paper, the measured temperatures, pressures, and mass flow rates were used. The temperatures and pressures were measured at the inlet and outlet of each heat exchanger with the exception that there were no measurements between the recuperator outlet and condenser inlet. The fluid enthalpy and entropy were calculated at each process node by using the measured temperature and pressure and by assuming the fluid in the system being pure MDM. The evaporator heat rate was defined from the working fluid side by using the measured working fluid mass flow rate and the measured working fluid pressure and temperature at the evaporator inlet and outlet

/ev ¼ qm;wf ðhev;out  hev;in Þ:

ð3Þ

The recuperator heat rate was defined by using the measured working fluid mass flow rate as well as temperature and pressure at the recuperator inlet and outlet on the liquid side.

/rec ¼ qm;wf ðhrec;l;out  hrec;l;in Þ:

ð4Þ

There were no temperature measurement at the recuperator outlet on the vapor side. Thus, the vapor enthalpy hrec;v;out and temperature T rec;v;out were evaluated by using the measured vapor inlet temperature, recuperator heat rate, and working fluid mass flow rate. The recuperator effectiveness was defined as

e¼

T v;in  T v;out : T v;in  T l;in

ð5Þ

The power consumption of the feed pump was defined as

Pp ¼ qm;wf

v Dp gp

:

ð6Þ

An efficiency of 50% was used for the feed pump in analyzing the feed pump power consumption during the experiments. The potential for producing mechanical power from the vapor expansion was estimated by using the measured pressure and temperature at the control valve inlet as well as the measured pressure at the control valve outlet. Deriving from the definition of the isentropic efficiency, the enthalpy at the expander outlet was solved as follows

ht;out ¼ ht;in  gs ðht;in  ht;out;s Þ:

ð7Þ

The mechanical power was defined as

Pt ¼ qm;wf ðht;in  ht;out Þ:

ð8Þ

The cycle efficiency was defined as

gc ¼

Pt  Pp : /ev

ð9Þ

The measured operational temperatures of the evaporator and recuperator were compared to the analytical results of heat exchanger models that were used for predicting the heat exchanger performance at different operational conditions. The heat exchanger calculation model was based on the design values of the heat exchanger, namely the flow rates, temperatures, total heat transfer area, and flow arrangement. The flow rate and inlet temperature of the exhaust gas as well as the flow rate and inlet temperature of the working fluid were used for calculating the outlet temperature on both sides of the evaporator. Based on this approach the heat transfer coefficient in the experiments was evaluated. A similar approach was adopted for the liquid and vapor working fluid in the recuperator analysis. 3. Thermodynamic analysis of different working fluids The suitability of different dry working fluids were evaluated in the preliminary phase of the project according to the thermodynamic performance, pressure and temperature levels, and preliminary design of the components [21]. Here, four working fluids that were evaluated as the most suitable fluid candidates in the preliminary analysis were selected for in-depth analysis. The selected fluids are siloxanes MDM and MM, hydrocarbon toluene, and hydrofluorocarbon R245fa. The main specifications of the studied fluids are presented in Table 1. All of the selected fluids have been previously used in commercial ORCs or considered as suitable candidates for waste heat recovery ORC applications [1,33,37]. The thermodynamic process design for the four selected fluids is illustrated in temperature-entropy diagram for the case of highest evaporation pressure and efficiency in Fig. 1a–d. The simulated expander power, cycle efficiency, working fluid mass flow rate, and turbine optimal rotational speed as a function of evaporation pressure are presented in Fig. 2a–d. The process operating with MDM and toluene can operate close to the maximum working fluid temperature limit of 300 °C and the

149

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158 Table 1 Specifications of the studied fluids. Fluid

Chemical formula

Classification

T crit [°C]

pcrit [bar]

M [kg/kmol]

pcond ð@60  CÞ [bar]

C3H3F5 C7H8 C8H24O2Si3 C6H18OSi2

Hydrofluorocarbon Aromatic hydrocarbon Linear siloxane Linear siloxane

154 319 291 245.5

36.5 41.3 14.2 19.4

134.0 92.1 236.5 162.4

4.63 0.19 0.04 0.26

R245fa Toluene MDM MM

300

sat. liquid sat. vapor toluene

300 250

o

o

Temperature, [ C]

250

Temperature, [ C]

sat. liquid sat. vapor R245fa

200 150 100

150 100 50

50 0

200

−0.5

0

0.5

1

0 1

1.5

1.2

1.4

(a)

300

sat. liquid sat. vapor MDM

300

2

sat. liquid sat. vapor MM

o Temperature, [ C]

250

o

Temperature, [ C]

1.8

(b)

250 200 150 100 50 0

1.6

Specific entropy, [kJ/kgK]

Specific entropy, [kJ/kgK]

200 150 100 50

−0.5

0

0.5

1

1.5

0

−0.5

0

0.5

Specific entropy, [kJ/kgK]

Specific entropy, [kJ/kgK]

(c)

(d)

1

1.5

Fig. 1. Thermodynamic process on temperature-entropy diagrams with different working fluids. (a) Toluene, (b) R245fa, (c) MDM, and (d) MM.

highest thermodynamic performances were simulated by using these fluids. In general, the simulated power output increases as the evaporation pressure is increased. The fluids with lower critical temperatures, namely MM and R245fa, have the maximum expander inlet temperature well below 300 °C. This can be explained by the low critical temperature of these fluids which restricts the maximum evaporation temperature and maximum expander inlet temperature in subcritical cycles. To reach higher temperatures with these fluids in a subcritical process, a high degree of superheating would be required. However, it has been reported that the increase in superheating does not increase the ORC performance when a dry fluid is used [38]. In addition, supercritical fluid conditions could be considered in order to increase the power output. The highest expander power outputs were calculated for toluene and the lowest for R245fa. The preliminary radial turbine design indicated that a turbine operated with MDM would have

a larger turbine wheel and lower rotational speed when compared to the other studied fluids by using the turbomachinery design rules in [36]. According to the thermodynamic analysis, R245fa was not evaluated to be a suitable fluid candidate because the other three fluids showed significantly higher thermodynamic performances. This is mainly because of the relatively low critical temperature of the fluid which restricts the cycle from reaching high efficiency when operating at subcritical pressure levels. In addition, the expander would be fast rotating, having a rotational speed of about 70,000–100,000 rpm, with impractically small dimensions, having a turbine diameter of about 30 mm. The use of R245fa would have allowed to have a condensing pressure well above the atmospheric pressure that would have removed the need for using high degrees of vacuum in the low pressure side of the process. The use of toluene was estimated to be a promising option purely from the thermodynamic point of view, but the use of toluene would also require a fast rotating and small expander

150

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

18

26 24

Cycle efficiency, [%]

Expander power, [kW]

16 14 12 10 R245fa toluene MDM MM

8 6 0

5

22 20 18 16 14 R245fa toluene MDM MM

12 10 10

15

20

25

30

8 0

35

5

Evaporation pressure, [bar]

10

(a) x 10 R245fa toluene MDM MM

Turbine rotational speed, [rpm]

Working fluid mass flow rate, [kg/s]

0.3 0.25 0.2 0.15 0.1 0.05 0 0

5

10

15

20

25

30

35

30

35

(b) 4

10 0.35

15

Evaporation pressure, [bar]

20

25

30

35

8

R245fa toluene MDM MM

6

4

2

0 0

5

10

15

20

25

Evaporation pressure, [bar]

Evaporation pressure, [bar]

(c)

(d)

Fig. 2. Results of (a) simulated expander power, (b) cycle efficiency, (c) working fluid mass flow rate, and (d) the optimal rotational speed of the turbine as a function of evaporation pressure.

wheel. Thus, toluene was evaluated as an inappropriate working fluid for the targeted power level of about 10–15 kW. The final choice was done between siloxanes MDM and MM. After a careful evaluation, MDM was selected as the most suitable working fluid for the experimental setup. The selection of the working fluid was a compromise taking into account the simulated power output, heat exchanger design, and preliminary turbine design. The disadvantage of using MDM as the working fluid is the low pressure level in the condenser resulting in high pressure ratio in the process that is required for achieving a high cycle performance. In addition with MDM, the volumetric flow rate on the low pressure side of the process is high when compared to the use of many other fluids which leads to the use of large-sized heat exchangers and process piping on the low pressure side of the process. More detailed information on the working fluid selection and turbine design for the experimental setup are presented in [21].

4. Description of the experimental setup The final design parameters of the experimental setup are presented in Table 2 and the process on a temperature-entropy diagram is presented in Fig. 3. After the working fluid was selected the final design values were selected iteratively according to the

Table 2 Setup design values. qm;wf [kg/s]

0.2

/ev [kW] pev;out [bar] T ev; out [°C] /rec [kW] e [–] T cond [°C] pcond [bar] pvalve;out [bar] P mech [kW]

67 7.9 265 40 0.68 57 0.03 0.07 12

results obtained in the thermodynamic analysis taking into account the predicted component efficiencies and pressure losses as well as the more detailed design and operational requirements of each process component. The working fluid temperature below 300 °C at the evaporator outlet was selected in order to reduce the risk of thermal decomposition of MDM in the evaporator [35]. In addition, the availability of adding a turbine to the system in a later phase of the project was taken into account more thoroughly in this design phase.

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

300

o Temperature, [ C]

250

sat. liquid sat. vapor process

200 150 100 50 0 −0.8

−0.6

−0.4

−0.2

0

0.2

0.4

0.6

0.8

Specific entropy, [kJ/kgK] Fig. 3. Process design values on temperature-entropy diagram.

Fig. 4. Overview of the experimental setup in the LUT laboratory.

An overview of the experimental setup is presented in Fig. 4. The experimental rig included a diesel generator for producing waste heat for the ORC. The engine has 6 cylinders, the maximal power output is approximately 200 kW, the rotational speed is 1500 rpm, and the measured exhaust gas temperature level in the engine exhaust pipe was slightly below 400 °C in the experiments. A simplified process diagram showing the main components of the experimental setup are presented in Fig. 5. The process included the heat exchangers, namely the evaporator, recuperator, and condenser; as well as the other main components including the control valve, storage tank, main feed pump, pre-feed pump, and a vacuum pump. The main design values of the process heat exchangers are presented in Table 3.

151

A control valve in the vapor line was used for expanding the high pressure vapor and for controlling the evaporation pressure and working fluid mass flow rate in the cycle. Liquid MDM is fed to the evaporator where it is preheated, vaporized, and superheated as a result of the heat extracted from the hot exhaust gas. The evaporator is a once-through type and it produces superheated vapor which expands in the control valve. The evaporator is connected directly to the exhaust pipe; thus, no thermal oil circuit is included in the system. After the control valve, vapor at superheated state is introduced to the recuperator in where the hot vapor preheats the liquid working fluid. From the recuperator, vapor enters the condenser where it is further desuperheated and finally condensed back into liquid. Cold water is used as a cooling medium in the condenser. After the condenser the liquid working fluid is then collected in the storage tank. From the storage tank, liquid MDM is pumped to the recuperator by the pre-feed pump and the main feed pump. The pre-feed pump is used to prevent cavitation in the main feed pump. In the recuperator, heat is transferred from the superheated vapor MDM to the liquid MDM which is preheated before it is fed to the evaporator. The recuperator is a counterflow heat exchanger. The recuperator effectiveness of slightly below 0.7 was selected in the design since it allows moderate pressure drop on the vapor side and on the other hand reduced the recuperator size when compared to higher degrees of recuperation. In addition, the use of higher degree of recuperation would result in increased exhaust gas temperatures at the evaporator outlet which would reduce the amount of heat that can be transferred from the exhaust gas to the working fluid. It was estimated that there is a pressure drop on the low pressure vapor side of about 0.04 bar in the recuperator and condenser. A vacuum pump was used in the system for maintaining a subatmospheric pressure on the low pressure side of the process. The vacuum pump was operated before each test run to remove noncondensable gases, mainly air, from the system. In addition, the vacuum pump could be operated during the experiments to remove air and other gases from the system in order to lower the pressure on the low pressure side of the process. The system pressure, vapor temperature at the evaporator outlet and fluid flow rate were controlled by changing both the rotational speed of the feed pump and by using the control valve. In addition, the diesel generator power output was controlled in order to attain the desired exhaust gas temperature level and mass flow rate. The temperature of the vapor working fluid was kept below 300 °C in the experiments in order to avoid overheating and thermal decomposition of the fluid [22,35]. The temperature measurements were carried out by using commercial thermocouples and the pressure measurements were carried out by using commercial pressure sensors. The working fluid mass flow rate was measured by using a positive displacement meter having an accuracy of ±0.5%. 5. Experimental results In this section, the experimental results are presented. First, the studied operational conditions are presented. Second, the performances of the evaporator, recuperator and condenser are analyzed in more detail. Finally, the vapor expansion is analyzed and potential for extracting power from the expansion is evaluated based on the experimental results. 5.1. Operational conditions The experiments were carried out by using different evaporation pressures, working fluid mass flow rates, evaporator heat

152

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

Fig. 5. Simplified process diagram of the experimental setup.

Table 3 Heat exchanger design values. Heat exchanger Evaporator Recuperator Condenser

Type

Atot [m2]

U avg [W/m2K]

m [kg]

DT LMTD [K]

PS PS PS

30.4 16.0 7.0

31 43 201

515 218 168

72 59 40

rates, and condensing conditions. The diesel engine power output varied from 110 kW to 160 kW in the experiments. The results of four different operational conditions (OP) are presented in Table 4. The measured operational conditions are illustrated on a temperature-entropy diagram in Fig. 6a and b. In OP1 a lower condensing temperature of 33 °C was used as compared to OP2-OP4 having a condensing temperature close to

Table 4 Studied operational conditions.

qm;wf [kg/s] pvalve;in [bar] T valve;in [°C] /ev [kW] pvalve;out [bar] pvalve;in /pvalve;out [–] P p [kW] T cond [°C] pcond [bar]

OP 1

OP 2

OP 3

OP 4

0.17 8.1 261 59.0 0.15 52.5 0.33 33 0.09

0.14 6.9 277 48.0 0.22 32.1 0.24 55 0.14

0.15 4.2 263 51.0 0.22 19.2 0.16 55 0.15

0.13 7.1 256 42.1 0.21 34.6 0.26 52 0.15

the design value of 57 °C. OP1 has the highest working fluid mass flow rate, evaporator heat rate, and evaporation pressure of the studied operational points. The working fluid temperature at the evaporator outlet was set above 250 °C to ensure the superheated vapor conditions at the evaporator outlet. In addition, a small degree of superheating was used in order to better control the system. The working fluid mass flow rate varied from 0.13 kg/s to 0.17 kg/s in the measurements. Thus, the working fluid flow rate was slightly lower in the experiments than the designed mass flow rate of 0.2 kg/s. The evaporator heat rate varied between 42 kW and 59 kW in the measurements while the evaporator was designed for 67 kW heat rate. This was mainly because the diesel generator produced smaller amount of exhaust gas heat in the studied conditions than was used in the ORC design. The vapor temperature at the evaporator outlet was close to the design value in all cases OP1-OP4. The evaporation pressure and vapor temperature at the evaporator outlet were relatively close to the design value of 7.9 bar in OP1 and slightly lower evaporation pressure was adopted in OP2 and OP4. OP3 had an evaporator outlet pressure of 4.2 bar in order to study the performance of the cycle at low evaporation pressure conditions. The experimental system was designed to reach a pressure ratio of over 100. In the experiments, a maximum pressure ratio over the control valve of 52.5 was attained in OP1 representing the highest evaporation pressure and the lowest condensing pressure of the operational conditions studied. The pressure ratio was lower than the one used in the system design mainly because the measured condensing pressure was significantly higher as compared

153

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

Temperature, [ oC]

250

sat. liquid sat. vapor OP1 OP2

300 250

Temperature, [ oC]

300

200 150 100 50

sat. liquid sat. vapor OP3 OP4

200 150 100 50

0 −0.8 −0.6 −0.4 −0.2

0

0.2

0.4

0.6

0.8

0 −0.8 −0.6 −0.4 −0.2

0

0.2

0.4

Specific entropy, [kJ/kgK]

Specific entropy, [kJ/kgK]

(a)

(b)

0.6

0.8

Fig. 6. Thermodynamic process on temperature-entropy diagram in OP1-OP4.

to the saturated pressure of MDM at the measured condenser temperature and to the design control valve outlet pressure. 5.1.1. Evaporator The results of the evaporator analysis are presented in Table 5 and the evaporator temperature diagrams showing the temperature profile of the working fluid and the exhaust gas as a function of transferred heat rate are presented for OP1-OP4 in Fig. 7a–d, respectively. The relative heat rate in Fig. 7a-d represents the portion of the transferred heat power of the total heat power of the heat exchanger at different locations in the heat exchanger. The evaporation temperature of the working fluid was calculated with Refprop [34] by using the saturated temperature corresponding to the measured evaporator outlet pressure. The evaporator inlet and outlet temperatures on the working fluid side and exhaust gas side were measured. The amount of heat that is needed for preheating, evaporating, and superheating the fluid was calculated by solving the specific enthalpy at the measured temperature and pressure using Refprop [34]. In Table 5 and in Fig. 7a–d, the exhaust gas temperature decreases in the evaporator in all four cases by about 200 °C while the temperature increase of the working fluid is in the order of magnitude of 100 °C in the studied operational conditions. The highest temperature changes in the working fluid and exhaust are in OP1 which has the highest evaporator heat rate of the studied operational points. The degree of superheating varies between the measured conditions. The smallest degrees of superheat of 5.7 °C and 9.0 °C were measured in OP1 and OP4 having the highest evaporation pressures, while OP2 and OP3 represented significantly higher degrees of superheating of 32.1 °C and 46.5 °C. The working fluid temperature was kept above 250 °C in all the studied cases; thus, the degree of superheating increased as the evaporation pressure was decreased. In the evaporator, the smallest temperature difference (pinch point) between the exhaust gas and the working fluid was mea-

sured at the cold end of the evaporator in all the studied operational conditions. Similar results for the location of the pinch point in the evaporator have been obtained in a previous thermodynamic study on using MDM for recovering approximately 400 °C exhaust gas heat [5]. Typically the smallest temperature difference between the exhaust gas and working fluid is located at the point where the evaporation of the working fluid starts [6]. The smallest pinch point temperature difference of less than 3 °C was measured in OP2 and the highest pinch point temperature difference of 13.5 °C was measured in OP1. According to the measured temperatures, the evaporator performance was evaluated also numerically. In OP1, OP3, and OP4 the overall heat transfer coefficient of 42.9–46.1 W/m2/K was calculated for reaching the measured exhaust gas and working fluid temperatures at the evaporator inlet and outlet. These values are at maximum well over 40% higher than the design heat transfer coefficient of 31 W/m2/K obtained from the heat exchanger manufacturer. This deviation can be partly explained by the fouling resistance that was used in the dimensioning of the heat exchanger. It was analyzed, that if the heat transfer coefficient were 31 W/m2/K as used in the design of the evaporator the working fluid would have been in wet vapor conditions at the evaporator outlet. In OP2 having the significantly small measured temperature difference at the cold end of the evaporator, the heat exchanger calculation model estimated a need for significantly high overall heat transfer coefficient of 70.4 W/m2/K in order to reach the measured evaporator temperatures. It should be noted that the calculated overall heat transfer coefficient is sensitive to the measured temperatures and especially to the temperature difference at the cold end of the evaporator which explains the high simulated heat transfer coefficient in OP2. Fig. 8a presents the measured exhaust gas and working fluid temperatures at the evaporator inlet and outlet and Fig. 8b presents the measured evaporator outlet pressure during the first 6000 s of a test run. In general, the temperature on the working

Table 5 The results of the evaporator analysis.

Evaporator heat rate [kW] Exhaust gas temperature drop [°C] Working fluid temperature increase [–] Pinch point temperature difference [°C] Degree of superheating [–]

OP 1

OP 2

OP 3

OP 4

59.0 233 117 13.5 5.7

48.0 222 114 2.7 32.1

51.0 221 100 7.9 46.5

42.1 196 99 10.6 9.0

154

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

400

400 MDM EG

350

300 o

Temperature, [ C]

o Temperature, [ C]

300 250 200 150

200 150 100

50

50 0

20

40

60

80

0

100

20

40

60

Relative heat rate, [%]

(a)

(b)

80

100

80

100

400 MDM EG

350

MDM EG

350 300 o

Temperature, [ C]

300 250 200 150

250 200 150

100

100

50

50

0

0

Relative heat rate, [%]

400

o Temperature, [ C]

250

100

0

MDM EG

350

0

20

40

60

80

100

0

0

20

40

60

Relative heat rate, [%]

Relative heat rate, [%]

(c)

(d)

Fig. 7. Temperature change in the evaporator on working fluid and exhaust gas side as a function of transferred thermal power. (a) The temperature diagram of OP1, (b) OP2, (c) OP3, and (d) OP4.

fluid side can be remained relatively constant during the operation and the working fluid temperature follows well the exhaust gas temperature. The temperature peaks and pressure changes visible in Fig. 8a and b are mainly due to the changes in control valve position and changes in main feed pump rotational speed. Heating up the evaporator takes several minutes due to the large mass of the evaporator and the transient response for the changing conditions is rather slow during the operation. The working fluid flow through the evaporator was started when the exhaust temperature was above 300 °C and this explains the relatively long delay in the increase of evaporation pressure and temperature on the working fluid side during the system start. 5.1.2. Recuperator and condenser The results of the recuperator analysis are presented in Table 6. The recuperator temperature diagrams showing the temperature profile of the liquid and vapor working fluid as a function of transferred heat rate are presented in Fig. 9a–d for OP1-OP4, respectively. The working fluid temperature on the liquid side was measured at the recuperator inlet and outlet and on the vapor side at the recuperator inlet. The vapor state at the recuperator outlet was evaluated by solving the energy equation and by using Refprop [34]. Recuperator effectiveness varied from 0.58 to 0.65 according to the measurements which are close to the recuperator effectiveness value of 0.68 used in the recuperator design. In general, the tem-

perature increase on the liquid side and the temperature decrease on the vapor side were relatively close to the recuperator design values. The recuperator heat rate values varied between 27.5 kW and 35.9 kW in the experiments while the design heat rate was 40 kW. This was mainly because the flow rate of the working fluid was lower when compared to the design flow rate. The minimum temperature difference between the vapor and liquid was measured at the cold end of the recuperator in all the studied operational conditions which is well in line with the temperature profile of the recuperator estimated in the thermal design of the recuperator. In the cases of OP1 and OP4, a higher values of recuperator effectiveness and smaller temperature differences between the vapor and liquid working fluid were found when compared to OP2 and OP3. This can be mainly explained that OP1 and OP4 had the lowest vapor temperatures at the control valve outlet. The heat exchanger calculation model estimated the overall heat transfer coefficient in the recuperator to be from 20.8 W/m2/K to 31.4 W/ m2/K in OP1-OP4. These overall heat transfer coefficients that were evaluated for the recuperator are well below the value of 43 W/m2/ K obtained from the heat exchanger manufacturer. The deviation can be mainly explained by the lower flow rates in OP1-OP4 when compared to the design values of the recuperator which negatively affects on the heat transfer in the recuperator. The measured temperatures on the liquid side and vapor side inlet as well as the estimated temperature at the vapor side outlet are presented in Fig. 10 presenting data for the first 6000 s of a test

155

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

10

350

o

Temperature, [ C]

300 250 200 150 100

EG in EG out WF in WF out

50

Evaporator outlet pressure, [bar]

400

8

6

4

2

0

0 0

1000

2000

3000

4000

5000

6000

0

1000

2000

3000

Time, [s]

Time, [s]

(a)

(b)

4000

5000

6000

Fig. 8. Measured temperatures at the evaporator inlet and outlet on working fluid and exhaust gas side (a) and measured evaporator outlet pressure (b).

Table 6 The results of the recuperator analysis.

Recuperator effectiveness [–] Recuperator heat rate [kW] Minimum temperature difference [°C]

OP 1

OP 2

OP 3

OP 4

0.65 35.9 66

0.58 29.5 85

0.59 33.1 82

0.64 27.5 64

run. The results show that temperature levels in the recuperator have no significant fluctuations during the operation. The temperature levels in the recuperator vapor inlet and liquid outlet slightly increases during the operation following the trend of temperature increase of the exhaust gas. In general, the temperature levels on the recuperator vapor side inlet were slightly higher than the temperature levels used in the recuperator design. This can be mainly explained by the small temperature drop in the vapor expansion through the control valve while the extraction of power from the expansion in a later phase of the project was taken into account in the recuperator design. As presented in the results, the minimum temperature differences between the vapor and liquid working fluid are relatively large when compared to the minimum temperature differences in the evaporator. The heat transfer in the recuperator could have been improved by increasing the heat transfer area in the recuperator. However, as presented in the evaporator analysis, the smallest temperature difference between exhaust gas and working fluid occurred at the cold end of the evaporator. Thus, increasing the heat transfer in the recuperator would result in higher exhaust gas temperatures at the evaporator outlet and might affect negatively on the performance of the cycle due to the reduced amount of heat transferred in the evaporator. In addition, the increase in the recuperator size would increase the pressure loss on the low pressure side of the process which would affect negatively on the potential of extracting power from the vapor expansion due to the reduced pressure ratio over the expander. The condenser temperature diagrams showing the temperature profile of the liquid and vapor working fluid as a function of transferred heat rate are presented in Fig. 11a for OP1 having the lowest condensing temperature and in Fig. 11b for OP2 having the highest condensing temperature. In the estimation of the condenser heat transfer rates and performance, the measured temperature at the condenser outlet was used as a condensing temperature. Thus in this assumption, the possible subcooling of the liquid working fluid was neglected because the small amount of subcooling in the condenser was estimated to have only a minor impact on the overall

performance of the condenser. The amount of cooling required to remove the superheating was calculated from the energy balance and by using Refprop [34] for determining the fluid enthalpies. The condenser temperature profiles presented in Fig. 11a and in b show that relatively large portion of the cooling power is required for removing the superheating of the vapor. Thus, it can be concluded that there might be further potential for increasing the system efficiency by increasing the heat transfer in the recuperator. However, as mentioned earlier in this paper, the effects of increasing the heat transfer in the recuperator on the cycle performance would require more in-depth analysis, taking into account especially the effects on the evaporator performance and pressure losses on the low-pressure side of the process. 5.1.3. Potential for extracting power from the expansion Based on the experimental results, the capacity for producing mechanical power from the expansion was also analyzed numerically using the scenario that an expander had been added to the system. The analysis was carried out for OP1 representing the highest pressure ratio and working fluid mass flow rate. The measured pressure and temperature at the expansion valve inlet, the measured pressure at the expansion valve outlet, and the measured working fluid mass flow were used in the analysis. Additionally, a similar analysis was conducted using the valve design outlet pressure of 0.07 bar instead of the measured pressure at the expansion valve outlet. This was done in order to evaluate the potential of producing mechanical power from the expansion in a case the sealing of the system was improved and higher degree of vacuum were to be obtained. Turbine isentropic efficiency of 75% was used in the analysis which was evaluated to be an achievable value for a small-scale high expansion ratio radial turbine according to previous studies [5,23,32,33]. However, it should be noted that there are high uncertainties in the obtainable turbine isentropic efficiency for this type of system because in the literature there are few experimental studies pertaining to the operation of small-scale high-pressure supersonic turbines operating with non-conventional working fluids.

156

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

250

250

liquid side vapor side

200 o

Temperature, [ C]

Temperature, [ oC]

200

150

100

50

0 0

250

150

100

50

20

40

60

80

0 0

100

20

40

60

Relative heat rate, [%]

Relative heat rate, [%]

(a)

(b) 250

liquid side vapor side

80

100

80

100

liquid side vapor side

200 o

Temperature, [ C]

200

Temperature, [ oC]

liquid side vapor side

150

100

50

0 0

150

100

50

20

40

60

80

100

0 0

20

40

60

Relative heat rate, [%]

Relative heat rate, [%]

(c)

(d)

Fig. 9. Temperature change in the working fluid vapor and liquid side as a function of transferred thermal power. (a) The temperature diagram of OP1, (b) OP2, (c) OP3, and (d) OP4.

300

Temperature, [ oC]

250

liquid in liquid out vapor in vapor out (calc)

imately 9.8 kW which corresponds to cycle efficiency of approximately 16.1%. 6. Discussion

200

By using the measured control valve inlet and outlet pressure and temperature and by calculating the corresponding specific enthalpies for MDM, it was observed that a enthalpy drop ranging from about 10–50 kJ/kg was calculated over the control valve by using Refprop database. However, expansion in a valve without any power extraction can be generally treated as nearly isenthalpic. This deviation between the valve inlet and outlet specific enthalpy was estimated to be caused by four possible factors.

150 100 50 0 0

1000

2000

3000

4000

5000

6000

Time, [s] Fig. 10. Temperatures on the recuperator vapor and liquid side.

The results show that if a turbine isentropic efficiency of 75% was reached the potential for producing mechanical power from the expansion is about 8.1 kW in OP1 if the measured valve outlet pressure was adopted. This corresponds to cycle efficiency of about 13.2%. If the turbine design outlet pressure of 0.07 bar were reached the maximum mechanical power output would be approx-

1. The working fluid might have contained small portions of other substances than MDM. The high condenser pressure at the measured temperature would also support that the working fluid contained small portions of other substances having a higher condensing pressure when compared to pure MDM. One explanation could be either a thermal or chemical decomposition of the working fluid or presence of other impurities in the process. The possible thermal decomposition might have been caused by the ‘‘hot spots” in the evaporator, since the exhaust gas temperature level in the experiments was well above the thermal stability threshold of linear siloxanes [35].

157

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

150 working fluid cooling water

working fluid cooling water

o

Temperature, [ C]

80

o

Temperature, [ C]

100

60

40

100

50

20

0 0

20

40

60

80

100

Relative heat rate, [%]

(a)

0 0

20

40

60

80

100

Relative heat rate, [%]

(b)

Fig. 11. Temperature profile of the working fluid and cooling water as a function of transferred thermal power. (a) The temperature diagram of OP1 and (b) OP2.

2. Heat losses to the environment between the measurement points. 3. Inaccuracies in the measured temperature and pressure data. 4. Inaccuracies in the thermodynamic model of MDM [39]. The high pressure in the condenser that was observed during the system operation was estimated to have mainly been caused by the presence of small amounts of non-condensable gases, such as air, in the system despite the efforts to remove all the gases by using a vacuum pump before each test run. A small leakage of air to the system was also detected after the experimental campaign despite the efforts to make the experimental setup fully closed. It can be concluded that if an expander were added to the system the power output would be very sensitive to the pressure level in the condenser. Therefore, it would be important to improve the removal of non-condensable gases and sealing of the experimental setup to reach lower condensing pressure. The estimated power outputs of 8.1 kW and 9.8 kW are lower than the mechanical power output of approximately 12 kW which was evaluated in the design phase of the experimental setup. This can be mainly explained by the lower working fluid mass flow rate obtained in the experiments when compared to the design mass flow rate used in the system design. The working fluid mass flow rate could be increased by using higher diesel engine power outputs in the future tests. In addition, the measured pressure loss between the control valve outlet and the condensate tank was about 0.05–0.08 bar while this pressure loss was estimated to be slightly lower of about 0.04 bar in the process design phase. The pressure loss between the valve outlet and the condenser reduces the pressure ratio in the expansion and thus, it negatively affects on the amount of power that can be extracted from the expansion. The occurrence of possible decomposition of the fluid should be analyzed in more detail in the future. It should also be noted that the most significant effects of the possible decomposition of the working fluid will be revealed after a long operational time [22,35]. 7. Conclusions The objective was to study the suitability of high-pressure ratio and high-temperature ORC systems for recovering thermal energy from exhaust gases of about 100–200 kW scale engines. After a careful evaluation, siloxane MDM was evaluated as the most suitable fluid for the experimental setup. The thermodynamic aspects that were used as the criteria for the working fluid selection and the experimental results were presented. The experiments were

carried out in order to study and analyze the performance of the heat exchangers and other process components and to study the technical feasibility of high molecular weight and high critical temperature fluids in this type of waste heat recovery system. In general, the experiments were successful and showed the potential of using high molecular weight fluids in hightemperature and small-scale ORC applications. The results showed that the thermal energy in the exhaust gas was efficiently recovered by the ORC system and the design temperature and pressure of the working fluid at the evaporator outlet were reached. In addition, the experiments were carried out using lower evaporation pressures and different condensing temperatures in order to study the performance of the process at different pressure and temperature levels. The four operational points under study adopted slightly lower evaporator heat rates, recuperator heat rates, and working fluid mass flow rates as compared to the respective values used in the design of the experimental setup. A maximum evaporator heat rate of 59 kW was obtained in the test runs and the results indicated that the evaporator reached a pinch point temperature of less than 15 °C in all the four cases. The pinch point temperature difference was measured at the cold end of the evaporator in all the studied operational conditions. Based on the measured results the overall heat transfer coefficient was estimated to be higher in the evaporator and lower in the recuperator when compared to the heat exchanger design values. Based on the experimental campaign, the potential for producing mechanical power with the proposed system was evaluated. Maximums of 8.1 kW and 9.8 kW of mechanical power were estimated according to if the measured valve outlet pressure were used and if the design valve outlet pressure were used, respectively. These values for the power outputs were obtained by using expander efficiency of 75%. Indeed, there are high uncertainties in the obtainable expander efficiency for this type of systems because in the literature there are only few experimental studies concerning the operation of small-scale, high-pressure, and supersonic turbines operating with non-conventional working fluids. A turbogenerator featuring a radial turbine will be added to the experimental setup in the future in order to study the power extraction from the expansion more in detail. The experimental results and the estimation for the potential of producing power from the expansion highlighted the importance of maintaining a high degree of vacuum on the low pressure side of the process during operation. Thus, it is recommended to pay additional attention on the process sealing and on the removal of non-condensable gases from the system. In addition, it was con-

158

A. Uusitalo et al. / Applied Energy 192 (2017) 146–158

cluded that the presence of small portions of other substances than the pure working fluid in this type of ORC systems would cause notable deviations in the thermodynamic behavior of the process, especially on the low pressure side of the process, which sets further requirements on the fluid stability and purity. Acknowledgements Funding for the project was received from Lappeenranta University of Technology. M.Sc. Timo Koponen is greatly acknowledged for participating in the experimental work. The authors would also like to thank Professor Emeritus Jaakko Larjola for all of his help and guidance during the project. In addition, the assistance from those participating in the design and construction phase of the process is highly appreciated. References [1] Colonna P, Casati E, Trapp C, Mathijssen T, Larjola J, Turunen-Saaresti T, et al. Organic Rankine cycle power systems: from the concept to current technology, applications and an outlook to the future. J Eng Gas Turb Power 2015;137. http://dx.doi.org/10.1115/1.4029884. [2] Sprouse III C, Depcik C. Review of organic Rankine cycles for internal combustion engine exhaust waste heat recovery. Appl Therm Eng 2013;51 (1–2):711–22. [3] Bombarda P, Invernizzi CM, Pietra C. Heat recovery from Diesel engines: a thermodynamic comparison between Kalina and ORC cycles. Appl Therm Eng 2010;30(2–3):212–9. [4] Di Battista D, Mauriello M, Cipollone R. Waste heat recovery of an ORC-based power unit in a turbocharged diesel engine propelling a light duty vehicle. Appl Energy 2015;152:109–20. [5] Uusitalo A, Turunen-Saaresti T, Honkatukia J, Colonna P, Larjola J. Siloxanes as working fluids for mini-ORC systems based on high-speed turbogenerator technology. J Eng Gas Turb Power 2013;135(4):042305. 1–9. [6] Uusitalo A, Honkatukia J, Turunen-Saaresti T, Larjola J. A thermodynamic analysis of waste heat recovery from reciprocating engine power plants by means of organic Rankine cycles. Appl Therm Eng 2014;70(1):33–41. [7] Invernizzi C, Iora P, Silva P. Bottoming micro-Rankine cycles for micro-gas turbines. Appl Therm Eng 2007;27(1):100–10. [8] Chacartegui R, Sanchez D, Munoz JM, Sanchez T. Alternative ORC bottoming cycles for combined cycle power plants. Appl Energy 2009;86(10):2162–70. [9] Carcasci C, Ferraro R, Miliotti E. Thermodynamic analysis of an organic Rankine cycle for waste heat recovery from gas turbines. Energy 2014;65:91–100. [10] Qiu G, Shao Y, Li J, Liu H, Hao L, Riffat SB. Experimental investigation of a biomass-fired ORC-based micro-CHP for domestic applications. Fuel 2012;96:374–82. [11] Bracco R, Clemente S, Micheli D, Reini M. Experimental tests and modelization of a domestic-scale ORC (Organic Rankine Cycle). Energy 2013;58:107–16. [12] Hung TC, Shai TY, Wang SK. A review of organic Rankine cycles (ORCs) for the recovery of low-grade waste heat. Energy 1997;22(7):661–7. [13] Liu BT, Chien KH, Wang CC. Effect of working fluids on organic Rankine cycle for waste heat recovery. Energy 2004;29(8):1207–17. [14] Quoilin S, Declaye S, Tchanche BF, Lemort V. Thermo-economic optimization of waste heat recovery Organic Rankine Cycles. Appl Therm Eng 2011;31 (14):2885–93. [15] Maraver D, Royo J, Lemort V, Quoilin S. Systematic optimization of subcritical and transcritical organic Rankine cycles (ORCs) constrained by technical parameters in multiple applications. Appl Energy 2014;117:11–29. [16] Vivian J, Manente G, Lazzaretto A. A general framework to select working fluid and configuration of ORCs for low-to-medium temperature heat sources. Appl Energy 2015;156:727–46. [17] Chen H, Goswami DY, Stefanakos EK. A review of thermodynamic cycles and working fluids for the conversion of low-grade heat. Renew Sust Energy Rev 2010;14(9):3059–67.

[18] Lang W, Colonna P, Almbauer R. Assessment of waste heat recovery from a heavy-duty truck engine by means of an ORC turbogenerator. J Eng Gas Turb Power 2013;135(4):042313. [19] Fernandez FJ, Prieto MM, Suarez I. Thermodynamic analysis of hightemperature regenerative organic Rankine cycles using siloxanes as working fluids. Energy 2011;36(8):5239–49. [20] Lai NA, Wendland M, Fischer J. Working fluids for high-temperature organic Rankine cycles. Energy 2011;36(1):199–211. [21] Uusitalo A. Working fluid selection and design of small-scale waste heat recovery systems based on organic Rankine cycles. Doctoral thesis, Lappeenranta University of Technology, Finland; 2014 [22] Erhart T, Gölz J, Eicker U, van den Broek M. Fluid stability in large scale ORCs using siloxanes: long-term experiences and fluid recycling. In: Proceedings of 3rd international seminar on ORC power systems, October 12–14, 2015, Brussels, Belgium; 2015. [23] Kang SH. Design and experimental study of ORC (organic Rankine cycle) and radial turbine using R245fa working fluid. Energy 2012;41(1):514–24. [24] Borsukiewicz-Gozdur A. Experimental investigation of R227ea applied as working fluid in the ORC power plant with hermetic turbogenerator. Appl Therm Eng 2013;56:126–33. [25] Seher D, Lengenfelder T, Gerhardt J, Eisenmenger N, Hackner M, Krinn I. Waste heat recovery for commercial vehicles with a rankine process. In: 21st Aachen colloquium automobile and engine technology, Aachen, Germany, 7–9 October 2012. [26] Wang EH, Zhang HG, Fan BY, Ouyang MG, Zhao Y, Mu QH. Study of working fluid selection of organic Rankine cycle (ORC) for engine waste heat recovery. Energy 2011;36(5):3406–18. [27] Lemort V, Guillaume L, Legros A, Declaye S, Quoilin S. A comparison of piston, screw and scroll expanders for small scale Rankine cycle systems. In: Proceedings of the 3rd international conference on microgeneration and related technologies, Naples, Italy, 2013 April 15–17. [28] Song P, Wei M, Shi L, Danish SN, Ma C. A review of scroll expanders for organic Rankine cycle systems. Appl Therm Eng 2015;75:54–64. [29] Klonowicz P, Borsukiewicz-Gozdur A, Hanausek P, Kryllowicz W, Brüggemann D. Design and performance measurements of an organic vapour turbine. Appl Therm Eng 2014;63(1):297–303. [30] Da Lio L, Manente G, Lazzaretto A. Predicting the optimum design of single stage axial expanders in ORC systems: is there a single efficiency map for different working fluids? Appl Energy 2016;167:44–58. [31] Costall AW, Hernandez AG, Newton PJ, Martinez-Botas RF. Design methodology for radial turbo expanders in mobile organic Rankine cycle applications. Appl Energy 2015;157:729–43. [32] Verneau A. Supersonic turbines for organic fluid rankine cycles from 3 to 1300 kW: small high pressure ratio turbines. Von Karman Institute for Fluid Dynamics Lecture series, 1987, 7, June15–18. [33] van Buijtenen JP, Larjola J, Turunen-Saaresti T, Honkatukia J, Esa H, Backman J, Reunanen A. Design and validation of a new high expansion ratio radial turbine for ORC application. In: 5th European conference on turbomachinery. Prague, 2003, March 17–22. [34] Lemmon EW, Huber ML, McLinden MO. NIST reference fluid thermodynamic and transport properties (REFPROP), Version 9.0, National Institute of Standards and Technology; 2010. [35] Invernizzi C, Bonalumi D. Thermal stability of organic fluids for Organic Rankine Cycle systems: chapter 5 in Organic Rankine Cycle (ORC) power systems: technologies and applications. In: Macchi E, Astolfi M, editors. Woodhead Publishing Series in Energy, Number 107; 2016. [36] Balje OE. Turbomachines: a guide to design, selection and theory. NewYork: John Wiley; 1981. [37] Head AJ, De Servi CM, Casati EIM, Colonna P. Design specifications and requirements of the orchid set-up: version 1.1. Delft University of Technology; 2015. [38] Saleh B, Koglbauer G, Wendland M, Fischer J. Working fluids for lowtemperature organic Rankine cycles. Energy 2007;32(7):1210–21. [39] Colonna P, Nannan NR, Guardone A. Multiparameter equations of state for siloxanes: [(CH3)3-Si-O1/2]2-[O-Si-(CH3)2]i = 1, 3, and [O-Si-(CH3)2]6. Fluid Phase Equilib 2008;263(2):115–30.