Rankine Cycle for Waste Heat Recovery of IC Engines

Downloaded from SAE International by Brought to you by the University of Texas Libraries, Wednesday, April 16, 2014 04:06:18 AM

2009-01-0174

Rankine Cycle for Waste Heat Recovery of IC Engines J. Ringler, M. Seifert, V. Guyotot and W. Hübner BMW Group Research and Technology

Copyright © 2009 SAE International

ABSTRACT This paper deals with an analysis of the potential of the Rankine cycle as an additional power generation process, which uses the waste heat of a car engine. A review of different heat recovery technologies leads to the identification of the steam process as a favorable solution for automotive applications. Two basic configurations (exhaust gas only; exhaust gas plus coolant) are selected from numerous illustrated Rankine cycle layouts for a detailed evaluation of heat recovery based on a four-cylinder internal combustion engine (ICE). Furthermore, the impact of different working fluids is studied. A parametric analysis of the efficiency of two selected systems is accomplished with the Dymola simulation tool, followed by a direct comparison between both configurations for different engine technologies and vehicle speeds. Finally, test bench measurements of the applied engine coupled with a heat recovery device are carried out. It can be demonstrated that waste heat recovery can produce an additional power output of about 10% at typical highway cruising speeds.

INTRODUCTION In recent years the automobile industry has made great progress in improving engine efficiency. So far, the corresponding R&D activities have mainly focused on the optimization of fuel/air mixture preparation and the combustion process [1; 2]. Nevertheless, today’s engine

SAE Int. J. Engines | Volume 2 | Issue 1

efficiency is reaching its technical limit and will not be sufficient enough to meet future fuel economy targets without additional measures. Electrical hybrid concepts are one well-known approach [3]. A relatively new perspective is to take a closer look at the "periphery" around the engine with all its units for converting energy. The application of an optimized energy management system can also contribute to the reduction of fuel consumption [4]. But what about the biggest and, as yet, untapped source of energy in the car: engine waste heat? Even modern combustion engines achieve a maximum efficiency of approximately 40%. Hence, at least 60% of the energy in the fuel is lost via exhaust gas and engine coolant. In this study the issue of heat recovery of combustion engines is addressed for automotive applications.

REVIEW ON HEAT RECOVERY SYSTEMS Several physical effects can be used to recover energy from the waste heat of a combustion engine (Figure 1). The utilization of the waste energy is the first determining key factor for the overall efficiency of each technology illustrated. While turbo-machines (e.g. turbochargers) can only use pressure gradients and/or kinetic energy fractions of the exhaust gas, other technologies are tapping the much bigger amount: waste heat (Figure 1).

67


Gear

. y h g c r e e n M E

Auxillaries Driving Resistances

y g r e n E

Kinetic + Pressure Exhaust Exhaust Energy

Coolant Energy

n o i t r a c o i f l Acoustic y p t p i Stirling x a e e l v p i t Joule m o 1-LoopCatalytic o m Rankine (A) C o t Combined u a Turbines

2-LoopRankine 1-LoopRankine (B)

Seebeck

POSSIBLE SYSTEM CONFIGURATIONS OF THE RANKINE STEAM CYCLE As shown in Figure 2 the underlying ideal Rankine steam cycle can be divided into four processes, each associated with a change of state of the working fluid:

• Process 1-2: The isentropic compression of the liquid working fluid.

Turbine Turbo-machines

In addition to the heat utilization, the efficiency of the process itself plays an important role. Turbo-machines and thermo-chemical technologies are not considered, because their efficiency benefits are strongly linked to the combustion engine and, hence, hard to assess. Furthermore, turbo-machines, and in particular turbochargers, can be applied in combination with the thermal processes illustrated in Figure 1. Current efficiencies of thermo-electrical devices are lower than those of thermo-dynamic processes, but future material development could offer the chance to efficiently generate electrical power directly on-board [5]. The two remaining thermo-dynamic processes (Joule, Rankine) are well-established in gas and steam cogeneration plants. Due to its better efficiency the steam cycle is used for lower temperature levels (300-900°C) as found in the exhaust gas of an IC engine. To summarize the Rankine steam cycle can be identified as a favorable approach for the recuperation of waste heat.

Heat utilization

Thermo-dynamic

Thermo-chemical

Thermo-electric

Thermo-acoustic

Figure 1 Energy utilization vs. complexity of different heat recovery systems

The differences in the thermal processes in heat utilization are mainly caused by the ability to use the temperature difference. Chemical reactions have to run within a required temperature range (>400–500°C) and can therefore only utilize heat above this temperature level. Due to the process management, the realization of a high heat input is much more critical for devices based on the thermo-acoustic effect or the Stirling cycle. The heat transfer by gas-to-gas of the Joule process and by gas-to-material of a thermoelectric device makes the heat utilization for a given heat exchanger surface more difficult in comparison to the evaporation procedure of a Rankine cycle. Moreover, the latter approach offers the possibility to recover the released heat of both the exhaust gas and the coolant system. In general, system complexity (e.g. size, weight, system integration) is increasing with growing heat utilization. But the complexity of some technologies (Stirling, thermoacoustic) is disproportionally high in comparison to the heat recovery achievable for automotive applications. For this reason these systems are excluded in the following evaluation.

• Process 2-3: The isobaric heating, evaporation and superheating of the medium (2 →3) Qmed,in by supplying the thermal loss energy. • Process 3-4: The isentropic expansion of the vaporous medium (3 →4) in an expansion machine, generating mechanical power. • Process 4-1: The isobaric condensation of the medium back into the liquid phase (4 →1) along with dissipation of heat Q med,out into the environment.

T [K]

3 Tsuperheating h3

Qmed,in

Wt,pump

2

h2

  

1 h1

 Tcond.

~ Tmed,in

4

Tevap.

Qmed,out

4s

~ h4 Tmed,out

s [kJ/(kg K)]

Figure 2 Fundamentals of the Rankine cycle

68


Wt,exp


Assuming an ideal process, the thermal efficiency of the Rankine process is as follows:

E.

System A E.

C. E.  Rankine, ideal

=

W t, exp − W t, pump Q med, in

=

(h3 − h4 ) − ( h2 − h1 ) (h3 − h2 )

1

(1)

System B C.

 Rankine , ideal

= 1−

Qmed, out Qmed,in

~ T med, out = 1− ~ T med, in

7

E.

C.

Based on this fundamental cycle, various designs can be identified to recover engine waste heat. These configurations differ in the utilization of the heat source (coolant system and/or exhaust gas) and the number of system components (e.g. heat exchangers, expanders), as well as in the number of separate loops (Figure 3). In a dual or multiple loop system the working fluid of each sub-system is not exchanged. For example, layout 1 illustrates a dual-loop apparatus which combines a high temperature (HT) loop and a low temperature (LT) loop. The HT loop only recovers the heat of the exhaust gas. The LT loop recovers the residual HT heat, the coolant heat and the remaining exhaust gas heat. The two separate loops are coupled with a heat exchanger. Information on this configuration can be found in [6] where, besides the additional power, other automotive aspects such as the geometrical integration in a passenger car are also discussed. A complete system evaluation of all the configurations displayed, including the package requirements as well as cost effectiveness would exceed the framework of this paper. Hence, in order to access the potential of heat recovery for automotive application we will only focus on the additional power. Furthermore, the evaluation is restricted to two basic single loop systems which are studied in detail in the following section:

• System A: Utilization only of exhaust gas • System B: Utilization of exhaust gas and coolant


E.

3

System Configurations

(2)

The thermal efficiency as given in equation (1) gauges the extent to which energy input to the working fluid passing through the heat exchanger is converted to the net work output. The values for technical work of the pump (W t,pump) and the expander (W t,exp) can be calculated from the differences in the enthalpy of the working fluid (Figure 2), whereas Q med,in gives the heat transferred from the loss energy of the engine to the process (2→3). The thermal efficiency can be also expressed in terms of heat transfers as shown in equation (2) where Q med,out gives tthe heat released during the condensation process (4 →1). As illustrated in Figure 2 the mean lower and upper process temperature ~ ~ correspond to Tmed,out and Tmed,in , respectively.

2

6

5 E.

C. E.

4 C.

C. E.: Exhaust C.: Coolant

Figure 3 Conceivable designs for the heat recovery of the engine waste heat with steam cycles

Using the example of the two selected system configurations the role of the heat source as well as the impact of the operating parameters on the heat recovery potential were analyzed. The presented simulations, calculations and experimental investigations were performed for the most common engine design, a fourcylinder engine with a stoichiometric combustion [2].

WASTE HEAT FLOWS For every Rankine cycle knowledge of the temperature level of the heat sources is essential for the selection of a working fluid and the optimization of the operating conditions. Hence, the temperature of the coolant and exhaust is needed. Furthermore, the ratio of the corresponding heat flows is required for the comparison of the net output of system A and B. This ratio is given by: q=

Q coolant Q exhaust

(3)

Test bench measurements of heat flows were carried out for the studied engine. First, the temperature level of the coolant was raised to ~115°C and kept constant in order to optimize the utilization of this heat source. It is important to point out that no negative effects on the engine efficiency could be observed. The exhaust gas temperature sensors were positioned behind the catalyst, where the inlet of the exhaust gas heat exchanger was placed. Figure 4 shows the measured data together with the road resistance curve for the top gear. Changing the vehicle speed from 45 to 80 mph causes the exhaust gas temperature to increase from ~500 to ~800°C. As indicated by the isolines, engine load has also a significant influence on the temperature level of the exhaust gas. Nevertheless, the road resistance curve for the top gear will be the focus in the following evaluation, bearing in mind that heat flows might be very different for highly dynamic driving profiles.

69


SCREENING AND SELECTION OF THE WORKING FLUID

                     

  

  

     

  

  

              



  

  

 



 

  

  

   

 

















  

Figure 4 Exhaust gas temperatures in the gasoline engine map

Figure 5 shows the heat flows for various vehicle speeds (45, 55, 70 and 80 mph). For velocities between 45 and 80 mph, the exhaust gas heat flow ranges between ~5 to ~30 kW. At the same time, the coolant heat flow rises from ~9 to ~20 kW. It is obvious that with increased driving speed more power is needed, which in turn will result in higher waste heat flows. However, it is interesting to note that the ratio of the heat flows changes significantly with the driving speed. The ratio lines in Figure 5 indicate that at ~55 mph the waste heats have a ratio close to 1:1, whereas at lower speeds the coolant, and at higher speeds the exhaust gas, dominates as a heat source. 50

5 . 1 = q

1 = q

40

] 30 W k [ t n a l o o c

0 . 5 q =

80 mph

Q20

70 mph 55 mph

10

45 m ph

0 0

10

20 30 Qexhaust [kW]

40

50

  Figure 5 Heat flows Q und Q at different vehicle exhaust coolant cruising speeds

70


In a Rankine steam cycle, the working fluid, which is repeatedly vaporized, expanded and re-condensed, plays a key role in determining the potential as well as the cost effectiveness of a heat recovery system. The work output of a steam process for a given process temperature gradient differs significantly for various working fluids. The technical work of the expansion is given by the difference in the enthalpies of the working fluid entering and leaving the expansion machine, as shown in equation (1). In order to optimize the work output for a given temperature gradient, the evaporation enthalpy should be as high as possible. Water exhibits the highest evaporation enthalpy (~2250 kJ/kg), followed by alcohols (methanol ~1100 kJ/kg, ethanol ~820 kJ/kg). Methanol was dismissed a priori, due to its known health risks. Other organic substances such as toluene have much lower evaporation heat levels (~350 kJ/kg). Based on the evaporation enthalpy, one would expect water to be the preferred working fluid for any heat recovery system based on the Rankine steam process. However, in most real world applications the utilization of waste heat is limited by technical restrictions (maximum and/or minimum pressure). Having a future economically feasible heat recovery system in mind, the operating conditions were constrained as follows. The minimum condensation level is set at 70°C. Note that in any Rankine heat recovery application the largest fraction of the utilized heat has to be relieved into the ambient air. In order to limit the effort for safety measures, to reduce material costs and to avoid contamination through leakages the maximum temperature was limited to 300°C and the minimum and maximum pressures were restricted to 0.7 and 10 bar, respectively. Furthermore, the maximum evaporation temperature for system B is limited by the temperature of the coolant (T~115°C). Since a positive temperature difference is needed to transfer the heat and to ensure the cooling of the engine, the corresponding evaporation temperature (T evap,B) was set to be ~110°C. Figure 6 shows how the limits discussed restrict the temperature gradient which can be used for heat utilization. Considering the minimum pressure of 0.7 bar the condensation temperature had to be raised for water (Tcond,H2O~90°C) and toluene (T cond,PhCH3 ~100°C) for both systems, as indicated by the diamond symbols. The triangle symbols indicate the constant upper temperature level for all fluids in case of system B. Restricting the upper pressure for system A leads to the highest evaporation temperature for toluene (T evap,A,PhCH3~220°C), followed by water (T evap,A,H2O~180°C) and ethanol (Tevap,A,EtOH~150°C). Using these constraints the potential of the selected working fluids can be validated by applying equation (1).


For system A water delivers the highest thermal efficiency whereas for system B ethanol is the preferable working fluid. Other organic substances such as toluene provided significantly lower outputs for both systems and are therefore not considered further. Note that a detailed quantitative analysis is given in the next section. 240 Condensation (both systems)

220

Evaporation (system A)

200

Evaporation (system B)

180 ] 160 C ° [ e r 140 u t a r e p120 m e T 100

Limitation by T coolant for systemB

80

PARAMETRIC ANALYSIS

Limitation by T cond. for bothsystems

60 Minimum pressure for both systems

40

Maximum pressure for system A

20 0,1

1 10 Evaporation pressure [bar] Ethanol (EtOH)

Water (H2O)

rates and requires larger heat exchanger surfaces. This has negative effects on the size and net output of the system. Furthermore, most organic compounds such as alcohol are not very stable at temperatures above 200300°C. In addition, alcohol is critical in terms of safety and environmental aspects. Ethanol is a flammable liquid and at ambient temperature it can form explosive vapors. Hence, special safety measures have to be applied. W ith respect to material compatibility, alcohol is also a problematic substance. Water, on the other hand, has other disadvantages. It acts highly corrosively as a superheated vapor and its freezing point is not as low as desired for an automotive application. To summarize, from the evaluated working fluids (water, alcohols and further organic substances) water is the preferable working fluid for system A whereas for system B an alcohol (e.g. ethanol) would be the right choice.

100

A simulation model was developed with the tool Dymola in order to compare alternative heat recovery systems based on the Rankine cycle for different engine types. The modeling work was carried out in the Modelica ([7]) programming language. The interface of the simulation tool for the system configuration chosen in this study is shown in Figure 7.

Toluene (PhCH3)

Figure 6 Process temperature and pressure for system A and B for different working fluids

So far the ability to perform work by a working fluid has been the focus of this paper. However, with respect to a future automotive application, further aspects have to be considered.The preferred working fluids, water and ethanol, are discussed in the following section using the criteria listed in Table 1. Requirement High expansion work output High enthalpy of vaporization Thermal stability

Impact on

Water

Ideal system efficiency

+

System size Required pumping power System durability Range of application Compliance with car

Low freezing point application: antifreeze request

2250kJ/kg

Ethanol 820kJ/kg

+

-

0°C

-114°C

Environment aspects

Waste disposal

+

0

Material compatibility

Material choice Cost and weight

0

-

Safety

Explosion protection due to critical flashpoint

N.A

Price (€/l)

Costs

+

16.6°C -

Table 1 Comparison of working fluids

System size and weight are always important issues in automotive applications. Ethanol, with its lower evaporation heat level, has significant higher circulation


Figure 7 Simulation interface configured for system A and B

As shown in Figure 2 the process consists of the following key components: the expander, the pump and the condenser, as well as the two heat exchangers. Since the coolant heat exchanger can be virtually deactivated, system A and B can be simulated using the same simulation program. Each component was physically treated as a control volume with four inlet and outlet variables. The three in- and output variables, temperature, pressure and mass flow define the thermodynamic state of the working fluid up- and downstream of the control volume.

71


Furthermore, the fourth variable refers to the energy transfer (input/output: heat transfer; heat losses; technical work) for each component. All these parameters are linked via the thermodynamic properties of the working fluid. Note that chemical data of any working fluid can be integrated. This approach allowed a significant reduction of the complexity in the steam process model. Using the simulation tool a parametric analysis was conducted to investigate the influence of operating conditions and the properties of the heat source on the system performance of an ideal system. All relevant process variables are illustrated in Figure 2. As shown in equation (2), the thermal efficiency is given by the ratio of the provided and dissipated heat. In other words, the best efficiency is achieved when heat input is kept to the highest possible and heat output to the lowest possible temperature level. This relationship is illustrated in Figure 8 where the thermal efficiency for both working fluids is plotted in the relevant evaporation pressure range. The upper plot of Figure 8 shows the increase of the ηRankine,ideal with evaporation temperature (pressure) for system A. Constraining the operating evaporation pressure of water to 0.7-10 bar as illustrated in Figure 6 ηRankine,ideal will lead to ~19% as indicated by the thick line in Figure 8. This value increases by ~4% (number in brackets) if no lower pressure limit is applied. Note that the values in the brackets give the absolute efficiency increase when the condensation temperature was set to 70°C (corresponding condensation pressure 0.3 bar). At a temperature of ~110°C (indicated by dotted lines), which corresponds to fixed evaporation level of system B, the efficiency is only about 6%, and hence considerably lower than for ethanol (~10.5%), which is displayed in the lower plot. This finding is consistent with the results concerning the potential of the working fluids. As pointed out in the working fluid section, system B can use only a small temperature gradient; thus reducing the heat utilization by increasing the condensation temperature has great impact on the efficiency as shown in Figure 8 lower plot. Enhancing the lower temperature from 70°C to 90°C decreases the efficiency by 5% (number in brackets). The thermal efficiency of both systems can be slightly increased when the working fluid is superheated. Although not very obvious in Figure 8 the the isolines are steeper for water than for ethanol. Hence superheating has a greater impact on the efficiency of water. This is due to the stronger intermolecular attraction forces caused by hydrogen bonds. Increasing the working fluid superheating temperature from 200 to 300°C will raise the absolute efficiency value by ~1% for water and ~0.3% for ethanol. The effect is more significant at lower than at higher evaporation pressures. Nevertheless the effect is rather small.

72


250

40 26 (+3.25)

240

30

230

25

24 (+3.5)

220

20

210

] 200 r C ° e190 [ t a e r W180 u 170 t a r 160 e p 150 m 140 e t 130 n o i 120 t a r 110 o 100 p a v 120 E l 115 o n a 110 h t E

22 (+3.65)

15

20 (+3.75)* 18 (+4)*

10 8

16 (+4.25)*

6

14 (+4.35)*

5 4

12 (+4.5)*

3

10 (+4.75)* 8 (+5)* 6 (+5.25)*

2 1 4

12 (-5)*

] r a b [ e r u s s e r p n o i t a r o p a v E

11 (-5)* 3

10 (-5)*

105

9 (-5)* 2.2

100 200

210

220

230

240

250

260

270

280

290

300

Superheating temperature [°C]

Figure 8 Thermal efficiency (ηRankine, ideal) of a water Rankine process as a function of evaporation and superheating temperature. - Upper plot: water for Tcond.=90°C (*number in brackets give the absolute increment for Tcond.=70°C). - Lower plot: ethanol for Tcond.=70°C (*number in brackets give the absolute increment for Tcond.=90°C)

To summarize, the thermal efficiency is mainly determined by the evaporation and condensation temperature. For system B in particular, the realizable condensation temperature level plays a significant role. So far the focus has been on the evaporation temperature and its effect on the thermal efficiency of a Rankine process. Hence the next question to be addressed is how the selected evaporation temperature influences the ideal power output of the system. The ideal power output for system A and B is given by the following expressions: Pideal , A = Q exhaust ⋅  Rankine, ideal, A ⋅ (ξ A )

(4)

Pideal , B = Q exhaust ⋅  Rankine , ideal, B ⋅ (ξ B + q)

(5)

where ξA and ξB is the heat recovery efficiency of the exhaust heat exchanger for system A and B. No heat recovery coefficient is given in equation (5) for the energy released from the coolant because engine coolant mass flow is arranged as a closed loop. Therefore, assuming no heat losses, the heat recovery efficiency is 1.


In contrast, the exhaust gas energy refers to ambient conditions, but the exhaust leaves the heat exchanger at a much higher temperature level. The transferred heat for the exhaust gas is determined by the difference of the cooled exhaust gas temperature (T exhaust,in) and the temperature of the exhaust gas leaving the heat exchanger (T exhaust,out). Texhaust,out in turn depends on the evaporation and superheating temperature. Hence ξ is a function of T evap, Tsuperheating and Texhaust,in in equation (6).

ξ=

Qmed, in Qexhaust

= f (T evap., T superheating , T exhaust,in )

(6)

As regards superheating, it was set to a maximum of 300°C if not limited by the exhaust gas temperature. For system B, with its fixed evaporation temperature of 110°C, the heat recovery efficiency ξB increases with the exhaust gas temperature. The higher the difference between evaporation temperature and exhaust gas temperature the more heat can be recovered and hence the power output also rises. As regards system A, the relationship is somewhat more complicated. On the one hand, an increasing evaporation temperature raises the thermal efficiency; on the other hand it decreases the transferred heat ( ξA) into the system. Depending on the exhaust gas temperature, an optimum evaporation temperature can be found. Figure 9 shows the ideal system efficiency as a function of evaporation and exhaust gas temperatures for the working fluid water. The data is plotted for a lower pressure level of 0.7 bar. First of all, the focus is on how the exhaust gas temperature influences the system efficiency when the evaporation temperature is kept constant. Increasing the exhaust gas temperature from 200 to 800°C causes the efficiency to increase from ~5 to ~18% at an evaporation pressure level of 10 bar. This trend is due to the increased heat recovery efficiency, as already pointed out for system B. Following the system efficiency for a given exhaust gas temperature the following trend can be observed in Figure 9. For a given exhaust gas temperature, the system efficiency increases with increasing evaporation temperature up to a temperature which corresponds to the maximum efficiency. A further increase in the evaporation temperature causes the product ηRankine,ideal ⋅ ξA to become lower. Hence, an ideal evaporation temperature can be found for a given exhaust gas temperature which is indicated by the black line (optimum line) in Figure 9. It is interesting to note that the ideal evaporation temperature rises with the exhaust gas temperature. For exhaust gas temperatures above 400°C, the optimum evaporation is found outside the data range, as indicated by the dashed line. In this context it is important to point out that a restriction of the operating pressure due to technical limitations will decrease the system efficiency, in particular at higher exhaust gas temperatures.


] C ° [ e r u t a r e p m e t n o i t a r o p a v E

290 280 270 260 250 240 230 220 210 200 190 180 170 160 150 140 130 120 110 100

50 E

40 ] v r 

30 25







20

 

15



e L i n m u t i m O p





10 8 6 5 4 E 3

    

200

250

300

350

400

450

500

550

a a p b [ o r e a r t u i o s n s e p r r p e s n s o i u t r a e r o [ b p a a r v ]



600



650



700

2 



750

1 800

Exhaust gas temp eratur e [°C]

Figure 9 System efficiency (ηRankine,ideal•ξA) of a water Rankine process (system A) as a function of evaporation and exhaust gas temperature at Tcond.=90°C

SYSTEM COMPARISON The analysis in the above section demonstrates how the process parameters of a Rankine cycle can be optimized. It also reveals how the technical constraints of the operating conditions can restrict the system performance. For system A, the upper and lower pressure limit (in other words the pressure ratio of the expander) is the key parameter. With respect to system B, the realization of a low condensation temperature and the corresponding low pressure is the main challenge. Hence, the setting of the operating range has a significant impact on the simulation results for the following evaluation of both systems. Based on the process limitations already presented, the minimum pressure was set to 0.7 bar, corresponding to a condensation temperature of 90°C for system A and 70°C for system B (Figure 6). As regards the upper pressure of system A, the optimum value was chosen as long as it did not exceed the maximum pressure limit of 10 bar. For system B, the evaporation was kept constant at 3 bar (corresponding to T evap.=110°C). Superheating was limited to 300°C for both setups. Using these parameter settings the ratio of the ideal power output for system A and B (P ideal,B /Pideal,A) can be calculated based on equation (4) and (5). The ratio (Pideal,B /Pideal,A) is a function of q, ηRankine and ξ which, in turn, depends on the exhaust gas temperature. Pideal , B Pideal , A

=

 Rankine, ideal , B

⋅ (ξ B + q)

 Rankine, ideal , A ⋅ ξ A

(7)

Hence, in Figure 10 the power ratio is plotted as a function of exhaust gas temperature (T IN) and q. The plotted data can be interpreted as follows: The fat isoline which separates the plot in two sections corresponds to a ratio of 1. Above this line system B yields higher outputs, whereas system A shows the higher potential below this line. An increasing deviation from the 1:1 ratio line will result in a steadily growing difference in power

73


output between the two systems. The isolines also reveal a second trend. For a given waste heat ratio q, the relative potential of system A increases with the exhaust gas temperature, which is due to the enhanced heat recovery efficiency. 1,50 45 m ph

B m e t s y s

1,25 55 mph 1,00

] [ q

A m e t s y s

70 mph

0,75

80 mph

0,50

Hence the simulation results presented cannot be interpreted as a general recommendation regarding a preference for either system at a definite vehicle speed. The results rather demonstrate that the simulation tool can be used in order to compare different heat recovery setups based on the Rankine cycle for any given engine type.

RESULTS OF TEST BENCH MEASUREMENTS The presented experiments were mainly carried out on an engine test bench. Figure 11 illustrates a four-cylinder engine, which is equipped with a dual-loop system (see configuration 1 in Figure 3) in such a test environment.

0,25 300

350

400

450

500

550

600

650

700

750

800

850

Exhaust gas temperature [ °C]

Figure 10

Power output ratios

The next question is how Figure 10 can be interpreted in terms of vehicle speed. Generally, one would expect system A to have a higher potential at higher velocities (associated with a larger waste heat fraction in the exhaust at high temperature level), while the opposite should be the case at lower velocities (associated with larger waste heat fraction in the coolant and lower exhaust gas temperatures). Figure 11

This is consistent with the simulation results for our engine application. Plotting the heat flow data discussed above reveals that system B would be the preferred heat recovery system in the speed range of 45 to 55 mph, but at higher speeds of 70 and 80 mph system A yields a higher output. At around 65 mph, where the road resistance curve intersects the 1:1 power ratio, both systems have comparable potential. In this context it is important to point out that other engine types may have very different waste heat properties for the same vehicle speed. Hence the position and shape of the road resistance curve in Figure 10 would be changed. At the same vehicle speed a turbo charged engine would have a reduced exhaust gas temperature associated with a lower fraction of waste heat in the exhaust gas. Hence the road resistance curve is relocated towards lower T IN and q, as illustrated in area I. Combustion processes, such as diesel or stratified gasoline, can enhance this trend even further, as indicated by the also displayed area II. Things get even more complicated when the temperature of the coolant is altered. Note that the progression of the 1:1 ratio critically depends on the realizable engine coolant temperature. A higher temperature level would shift this line to lower q ratios, whereas the opposite would be the case for lower temperatures.

74


Dual-Loop configuration on test bench

The underlying two loops can be separated with slight modifications to this assembly. The two subsystems correspond to the setup of systems A and B. The previous analysis indicated that system B is more favorable than system A for typical highway cruising speeds (45–70 mph) in combination with the engine type used. For this reason the experimental data focuses on this configuration. Besides the limiting factors under ideal conditions, as discussed in the previous section, further losses occur in the real process:

• Heat losses convection)

by

radiation

(conduction

and

• Real heat transfer from coolant/exhaust gas to working fluid • Pressure losses • Auxiliary losses (real pumping, sensors, etc.) • Real expansion (friction of expander + nonisentropic expansion)


To minimize the heat losses, the apparatus is insulated in critical areas (e.g., the exhaust pipe downstream of catalysts, the heat exchanger, expander, tubes upstream of the pump and downstream of the expander). Moreover, components which are relevant for the drop in system pressure (e.g. tubing, inlet and outlet of expander) are optimized. As depicted in Figure 12 operation of the engine in combination with system B produces an additional power net output of between 0.7-2 kW at relevant stationary loads for highway cruising (45-70 mph). In a wide operating range these values correspond to 65-70% of the ideal power output for the chosen operating conditions (T evap.=110°C, T superheating∼300°C and Tcond.=70°C). However, a visible decrease to 50% occurs at relatively high loads. The losses are mainly caused by the non-ideal expansion and the heat transfer performance, which starts to drop especially at higher mass flows in the exhaust.

5

x % · Pideal

4,5

90%

] 4 W k [ 3,5 r e 3 w o p2,5 r e d 2 n a p1,5 x E 1

80% 70% 60% 50% 40% 30% 20%

0,5

10%

0

Figure 12

45

50

55

60 65 70 Velocity [mph]

4 0 k W

                     

2 0 k W

3 0 k W

5 0 k W

Padd [%]



 1 0 k W

  

 



 

 

5 k W

  

    

 

 

 





   

 

 













  

Figure 13

Relative additional power net output for system B in the engine map. - Grey area Padd [%] - Isolines refer to the engine power - Different driving profile indicated by thick lines (solid line: road resistance curve, dotted line: acceleration, dashed line: deceleration)

100%

Test bench measurements

40



75

80

85

Real power net output of system B in comparison to the ideal system performance

Additional power outputs amounting to 10% of the engine power could be provided close to the road resistance curve. At higher engine loads this value decreases somewhat to about 8%, whereas lower engine loads are associated with significant higher potential (up to 15%). This is mainly due to the fact that the engine becomes more efficient at higher load. Another interesting result is the fact that the additional power measured is relatively constant along the road resistance curve. This trend can be explained by the reduced system efficiency at higher velocities (80 mph) as mentioned above. Hence, plotting the same data for an ideal system would show an increased system performance in this operating range. Nevertheless, the operation of the present hardware setup of system B at typical highway velocities (4580 mph) leads to a considerable increase in efficiency.

CONCLUSIONS The measurements performed provide the data basis to implement efficiency maps of each system component in the simulation tool. After this calibration the model can be used to evaluate different heat recovery setups depending on engine type, system layout and working fluid, as well as the operating parameters. So far the focus has been on the real power output. In the following, the additional relative power gain provided by the recuperation of waste heat will be discussed. Figure 13 shows the measured relative power net output of system B in the engine map (grey field). The plotted data covers most conditions relevant for highway cruising: constant-speed driving (indicated by the road resistance curve) and dynamic engine operation (acceleration: dotted line; deceleration: dashed line).


Different heat recovery technologies were reviewed and compared from an automotive perspective. The Rankine steam cycle is identified as a favorable approach for the recuperation of waste heat and is studied in detail in this paper. Feasible designs for the recovery of engine waste heat using the steam cycle were illustrated. Two basic single-loop systems (system A: exhaust gas only and system B: exhaust gas plus coolant) were discussed in detail to provide a fundamental analysis of heat recovery technology based on a Rankine cycle. A thermodynamic investigation revealed that water would be a preferable working fluid for system A, which uses exhaust gas as a high temperature level heat source (T>300°C). However, for system B, which uses a lowtemperature heat source in addition (T~100°C), such as the coolant, an alcohol (e.g. Ethanol) would be the right choice.

75


A simulation model was developed in order to evaluate the two alternative systems for different engine types. A parametric analysis was conducted with this tool in order to investigate the effects of the determining thermodynamic parameters on the system performance. System B shows a higher potential at typical highway speeds (45-70 mph) for the engine type chosen (4-cylinder, stoichiometric combustion) and operating conditions. Nevertheless, this cannot be interpreted as a general recommendation. Heat source parameters, which are deeply influenced by engine type and load profile, as well as operating parameters, which are limited by technical constraints (pressure level, ambient temperature), have significant effects on the net power output. Based on the simulation results, test bench results for system B are discussed in detail. Under relevant stationary conditions, additional power outputs between 0.7-2 kW could be demonstrated. This corresponds to an increase in engine performance in the range of 10% close to the road resistance curve for the top gear. Hence the operation of the Rankine cycle system presented leads to a remarkable increase in fuel efficiency. A further important step has been taken in the introduction of waste heat recovery systems in automotive applications.

REFERENCES 1. Langen, P., Missy, S., Schwarz, C., Schuenemann, E., New BMW 6- and 4-cylinder petrol engines with high precision injection and stratified combustion, th presented at the 28 Internationales Wiener Motorensymposium, 2007. 2. Liebl, J., Klüting, M., Poggel, J., Missy, S. (2001). Der neue BMW Vierzylinder-Ottomotor mit Valvetronic, MTZ Journal, 62(6). 3. Göschel, B., Einsatzmöglichkeiten für Hybridantriebeth in BMW Fahrzeugen, presented at the 25 Internationales Wiener Motorensymposium, 2004. 4. Liebl, J., Eder, A., Wärmemanagement: ein Beitragth zu BMW Efficient Dynamics, presented at the 6 Tagung Wärmemanagement des Kraftfahrzeugs, Haus der Technik Berlin, 2008. 5. Treffinger, P., Häfele, Ch., Weiler, T., Eder, A., Richter, R., Mazar, B., Energierückgewinnung durch Wandlung von Abwärme in Nutzenergie (Recovery of energy through conversion of waste heat), presented at VDI Tagung, 2008. 6. Freymann, R., Strobl, W., Obieglo, A. (2008). The Turbosteamer: A system introducing the principle of cogeneration in automotive applications, MTZ Journal, 69(5). 7. Introduction to physical modeling with Modelica, Kluwer Academic Publishers, Tiller, M., 2001.

76


CONTACT Juergen Ringler, BMW Forschung und Technik GmbH, email: [email protected] Marco Seifert, BMW Forschung und Technik GmbH, email: [email protected] Vianney Guyotot, BMW Forschung und Technik GmbH, email: [email protected] Walter Huebner, BMW Forschung und Technik GmbH, email: [email protected]

DEFINITIONS, ACRONYMS, ABBREVIATIONS Symbol

Designation

Usual Unit

 Q

heat flow

kW

η h T P p ξ s Wt Q

efficiency specific enthalpy temperature power pressure heat recovery efficiency entropy technical work heat

% kJ/kg °C kW bar % kJ/kg.K kWh kJ

Abbreviation cond. evap. exp in out q system A system B H2O EtOH PhCH3 add IC med

Designation condensation evaporation expander input output waste heat ratio heat recovery system with exhaust gas only heat recovery system exhaust gas plus coolant water ethanol toluene additional internal combustion working fluid

Rankine Cycle for Waste Heat Recovery of IC Engines

Recommend Documents