2009-01-0967

B-COOL Project - Ford Ka and Fiat Panda R-744 MAC Systems
C. Malvicinoa and R. Seccardini a
Centro Ricerche Fiat S.C.p.A.
M. Markowitzb, K. Schuermannsb, A. Bergamic, C. Arnaudd, R. Hallere, C. Petitjeane, C. Struppf, N. Lemkef, D. Clodicg, C. Posth and A. Hafneri
Ford GmbH, Maflow S.p.A., Delphi S.A., Valeo Thermal System S.A., Braunschweig Technical University, h i g Ecoles des Mines - CEP, Hydro, Sintef
Copyright 2009 SAE International

ABSTRACT

The B-COOL project, funded by the European Union, is devoted to the development of low cost and efficient R-744 system for small cars. In the framework of this initiative a Fiat Panda and a Ford Ka prototype have been realized adopting two different R-744 systems and a testing procedure has been identified and adopted to qualify the vehicles in terms of fuel consumption and thermal comfort performance. The Project started in March 2005 and ended in November 2008, this paper presents the major project outcomes on R-744 mobile air conditioning systems (efficiency and related fuel consumption and LCCP, costs, architecture) synthesizing the remaining technology open issues
Methods to assess performance, fuel annual consumption and environmental impact were identified, within the Project, and constituted a first step towards EU new standards.

INTRODUCTION

The B-COOL Project was fully devoted to the development of a low cost and high efficiency airconditioning system based on a vapour compression cycle using CO2 - identified with the acronym R-744 when used as refrigerant fluid. Figure 1 - The B-COOL consortium
The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAEs peer review process under the supervision of the session organizer. This process requires a minimum of three (3) reviews by industry experts. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. ISSN 0148-7191 Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. SAE Customer Service: Tel: 877-606-7323 (inside USA and Canada) Tel: 724-776-4970 (outside USA) Fax: 724-776-0790 Email: CustomerService@sae.org *9-2009-01-0967* SAE Web Address: http://www.sae.org Printed in USA
The EU, as Greenhouse Gas emission reduction measure, proposed the ban for Mobile Air Conditioning systems of fluids having a Global Warming Potential higher than 150 (i.e. R134a) with possible future complementary measures - e.g. measurement of the MAC fuel consumption and this initiative represents a challenge and an opportunity for OEMs and Mobile A/C Suppliers to increase their competitiveness. R-744 is one promising candidate to replace the present used fluid, named R134a. Besides safety, reliability and efficiency, the additional cost, estimated in the range of 70 - 150 Euro with reference to the low priced car systems, represents a serious challenge to its diffusion. This is even more relevant considering the lower priced cars that constitute up to 80% of the present EU car market considering the recent EU enlargement.
Two versions of the R-744 system have been designed and realized for each car to evaluate the R-744 technology soundly.
R-744 AS A REFRIGERANT FOR MOBILE AIR CONDITIONING
Compared to conventional refrigerants, the most remarkable property of R-744 is the low critical temperature of 31.1 so a vapour compression cycle C, operating at normal ambient temperatures works close to and even above the critical pressure of 7.38 MPa. This leads to three distinct features of R-744 systems: Heat is rejected at supercritical pressure in many situations. The system will then use a transcritical cycle that operates partly below and partly above the critical pressure. High-side pressure in a transcritical system is determined by refrigerant charge or the expansion device and not by saturation pressure. The system design thus has to consider the need for controlling high-side pressure to ensure optimized COP and sufficient cooling capacity. The pressure level in the system is high (around 3-10 MPa). Components therefore have to be redesigned to fit the properties of R-744. Due to smaller volumes of piping and components, the stored explosion energy in a R-744 system is equal to a conventional system. A benefit of high pressure is that the required compressor displacement is reduced by 80-90% for a given cooling capacity. Compressor pressure ratios are low, thus giving favourable conditions for high compressor efficiency. Large temperature variation glide occurs during heat rejection. At supercritical or near-critical pressure, all or most of the heat transfer from the refrigerant takes place by cooling the compressed gas without phase change. The heat exchanger is therefore called gas cooler instead of condenser. Gliding temperature can be useful in heat pumps for heating water or air. With proper heat exchanger design the refrigerant can be cooled to a few degrees above the entering coolant (air, water) temperature, and this contributes to high COP of the system. Due to the difference of R-744 thermo-physical properties and cycle characteristics compared to HFC refrigerants, typical efficiency curves (COP, Coefficient Of Performance: cooling capacity divided by power input) show different trends with different ambient temperatures. R-744 tends to be more efficient at lower ambient temperatures, while HFC systems may be slightly more efficient at higher ambient temperatures. This tendency has been verified for a variety of applications such as mobile air-conditioning and supermarket refrigeration. The intersection of the two depicted curves varies depending on various factors such as cycle layout and

4=5 2=3

Condenser

Gas Cooler

Compressor 1
Expansion Valve Compressor Evaporator

Expansion Valve

6 7=8=1

Evaporator

R-134a loop

R-744 loop

Figure 2 - R-744 and R-134a systems schemes. The Internal Heat Exchanger is adopted to assure appropriate efficiency to the R-744 loop The Project has been carried out by a consortium constituted by 2 major OEMs, 4 suppliers and three acknowledged excellence centres gathering skilled European scientists and engineers in this specific field. The project has been focused, at first, to the identification of the most appropriate testing procedures so to be able to qualify in realistic way a mobile air conditioning in terms of fuel consumption and performances (thermal comfort). A specific activity has also been launched to verify the safety-related issues. The major effort has been devoted to the development of the A/C systems for a Fiat Panda with automatic air conditioning and a Ford KA with manual control. The cars have been fully characterized following the identified procedures before and after the R-744 A/C system installation.
component efficiency. It shall be emphasized that in this situation, it would be misleading to base the comparison indicated in Figure 2 on design-point conditions, which typically are at an extreme ambient temperature. A more sensible basis for comparison is to use mean/average conditions, or to apply a seasonal analysis based on climatic variation, as applied in LCCP (Life Cycle Climate Performance) calculations.

TESTING PROCEDURE

Two procedures have been identified to characterize of a passenger car air conditioning system in terms of fuel consumption and cooling performance
25 Temperature (C) Time (min) 1500
Outlet Temperature Cabin Temperature Speed

2000 Speed (km/h)

Figure 2 - COP of R-744 and R-134a systems at varying ambient temperature R-744 are widely diffused for low temperature refrigeration and start to be applied as heat pump and air conditioner [1]. a) Heat pumps and Air Conditioning: Heat pump water heaters, heat pumps for tap water heating, were commercialized in Japan in 2001 for both residential and commercial applications. Systems adapted to European conditions are under development. One of the major advantages is that these transcritical systems are able to provide water at high temperatures (90 without a C) substantial drop in COP, compared to systems using HFC as a refrigerant. b) Commercial refrigeration: R-744 is an important refrigerant alternative to HFCs in commercial refrigeration systems. Some of the major companies have introduced direct systems using solely R-744 as a refrigerant with sub/transcritical cycles, depending on ambient temperature. Also in the light commercial sector, i.e. stand-alone equipment such as bottle coolers and vending machines, some of the major companies have introduced R-744 technology. c) Mobile Air Conditioning: Mobile Air conditioning is the application with the largest HFC emissions and the second largest GHG emissions in R-744-equivalent resulting from refrigerant emissions. Hafner et al (2004) compared R-744 mobile air-conditioning systems to R134a and R-152a systems based on experimental and climate data from different cities around the world. Compared to HFC-134a R-744 showed an LCCP reduced by %.

Figure 3 The testing cycle based on the New European Driving Cycle (NEDC) that has been adopted to assess the fuel consumption of mobile air conditioning systems.
FUEL CONSUMPTION. The procedure has been conceived to be feasible in the existing testing benches and to be representative of a real use and has been derived from a study carried out by Armines and CRF [2]. The procedure is based on a modified NEDC cycle (Normal European Driving Cycle) where four elementary Urban Cycles have been added to evaluate the effect of the cool-down transient as well as of steady state conditions during the urban cycle (Figure 3). The test can be performed in a climatic chamber equipped with rolling bench and does not require major changes to the existing testing facilities and standard testing procedures. To assure an acceptable accuracy level each test has to be repeated at least three times. Equivalent thermal conditions without solar irradiation have been identified under the hypothesis that the solar cabin soaking can be represented by an air temperature increase. This hypothesis introduces an approximation but simplifies in a crucial way the testing procedure requirements making it applicable in almost all the existing facilities (climatic chamber with rolling benches and emission and consumption measurement systems). The ambient testing conditions are as follows: 28 and 50% R.H. - European Summer: these C conditions can be considered representative to classify the air conditioning system with regards to the fuel consumption and thermal comfort. The A/C system set point: 20 C.
35 and 60% R.H. - Severe Summer: representative of C very high thermal load (non-European). The A/C system set point: 23 C. 15 and 70% R.H. - Dehumidification: to consider the C use of the A/C as a dehumidifier. The A/C system set point: 20 C. All tests are performed with the A/C in fresh air mode
COMPRESSOR: 60 cc scroll, transmission ratio = 1.32 CONDENSER: 574 x 315 x16 Serpentine Parallel Flow with integrated dryer EVAPORATOR: 185 x 188 x 58, plates and Fins EXPANSION DEVICE: thermostatic expansion valve LINES: 3 lines
A specific procedure has been defined to represent in a realistic way the use of manual A/C systems and thermal comfort [3].
FORD KA: 1.3 Gasoline with manual A/C (2005 my) COLOUR: Black COMPRESSOR: 90 cc scroll, transmission ratio = 1.40

Tem perature (C)

CONDENSER: 400 x 382 x 20

NP K NP K NP K ,'/(

EVAPORATOR: 210 x 240 x 81 EXPANSION DEVICE: orifice & accumulator LINES: 4 lines

2XWOHWV

R-744 SYSTEM ARCHITECTURE

Time (min)

*$6 &22/(5
+3 6(1625 ,17(51$/ +($7 (;&+$1*(5 &+$5*( 3257
Figure 4 - Cool Down test (43 - 35% R.H., 900 C 2 W/m solar irradiation). The test start when the air temperature at head level reached 60 A/C is set C. at maximum power and in recirculation mode COOL DOWN. The test is devoted to qualify the air conditioning system in terms of cooling performance under severe thermal load (see figure 4) and should be performed in a climatic wind tunnel with solar irradiation simulation.

/3 6(1625 &2035(6625

),/7(5

(9$325$725

[ 25,),&(

$&&808/$725

Figure 6 - B-COOL R-744 A/C system scheme
BASELINE VEHICLE CHARACTERISTICS
Two low segment cars have been selected as baseline vehicles: Both the developed R-744 systems have a similar architecture (Figure 5) based on variable displacement piston compressor, internal heat exchanger and orifice expansion device and an accumulator. The compressors, of different type, are of piston type with have 29 cc displacement modified to have a maximum displacement of 20 cc.
THE FIAT PANDA B-COOL SYSTEM

),$7 3$1'$

Figure 5 The Ford Ka and Fiat Panda used to realize the B-COOL vehicle demonstrators FIAT PANDA: 1.2 l Gasoline with automatic A/C COLOUR: black
Two versions of R-744 system have been conceived, one with a separate internal heat exchanger (figure 7a) and one with the heat exchanger integrated in the accumulator (Figure 7b). The components have been designed and realized by Dephi, the line, the accumulator and the internal heat exchanger by Maflow. Sensata supplied the temperature
and pressure sensors. The expansion device is a fixed orifice (0.55 mm) with a by pass at 12 MPa (Egheloff) The evaporator fits in the HVAC module not requiring major changes and the gas cooler has the same face area of the baseline condenser. The gas cooler fan, that is located on the left hand in the baseline system, has been moved in a more central position for a more uniform air flow.
(9$325$725 ,17(51$/ +($7 (;&+$1*(5 ),/7(5 $&&808/$725 25,),&(
The components have been designed and realized by Valeo, the A/C lines and the Internal heat exchanger by Maflow, who also provided the accumulator; Sensata supplied the temperature and pressure sensors. The expansion device is a fixed orifice (0.50 mm) with a by pass at 12.5 MPa. The heat exchangers have the same face area of the baseline components. The accumulator replaces the R134a accumulator of the baseline system. The co-axial tube internal heat exchanger is designed as a separate component. The gas cooler size is severely limited by the front-end package.

SYSTEM PERFORMANCE

The demonstrator vehicles have been characterized following the procedure above described before and after the installation of the R-744 systems.

Testing Conditions

Temperature Humidity Air Enthalpy 15 C 28 C 35 C 70% R.H. 50% R.H. 60% R.H. 34 J/kg 59 J/kg 91 J/kg Fuel Overcunsumption [% of baseline @ 28 C]
9$5,$%/( ',63/$&(0(17 FF &2035(6625 ZLWK UHGXFHG GLVSODFHPHQW WR

*$6 &22/(5 FF

Figure 7a - First version of B-COOL Fiat Panda R744 system. The compressor position, between the engine and the firewall led to a characteristic design of the Internal Heat Exchanger
25,),&( (9$325$725 ),/7(5
Panda - R-134a Panda - R-744 1st Panda - R-744 2nd Ka - R-134a Ka - R-744
Figure 7b - Second version of B-COOL Fiat Panda R-744 system: the Internal Heat Exchanger has been integrated in the accumulator.
Table 1: measured fuel overconsumption on a NEDC based testing cycle - % of the baseline system (R-134a) fuel overconsumption at 28 50% C R.H. Note that the baseline Ford Ka fuel consumption is rather higher than the Fiat Panda baseline fuel consumption.
Temperature Humidity Air Enthalpy Panda - R134a Panda - R-744 1st Panda - R-744 2nd Ka R134a Ka - RC 28 C 35 C 70% R.H. 50% R.H. 60% R.H. 34 J/kg 59 J/kg 91 J/kg Themal Comfort [1-10 scale]
THE FORD KA B-COOL SYSTEM
The Ford KA R-744 system has a more conventional lay out due to the fact that the compressor is located in front of the engine.

n.a. n.a. n.a. n.a. n.a.

8.1 8.1 8.1 8.2 8.5

7.3 7.3 7.7 7.3 7.8

25,),&( ,17(51$/ +($7 (;&+$1*(5
Table 2: Thermal comfort in arbitrary units measured on a NEDC based testing cycle FUEL CONSUMPTION. The measured data are reported in the tables 1 and 2. The results of the two versions of the Panda system have been included to show the effect of the system changes (compressor displacement increase from 15 cc to 20 cc and adoption of the IHX).
Figure 8 - Ford KA B-COOL R-744 A/C system
Both the R-744 systems shows a slightly higher fuel consumption at higher thermal load (35 C). The Fiat Panda R-744 1st version system shows also a slightly decrease of the thermal comfort at 35 that is C fully recovered with the second version of the system. The Ford Ka has a better performance at 35 due to C the evaporator temperature control, the baseline produced a bit too low outlet temperature. The lower increase, in percent, of the fuel consumption of the Ford Ka with respect to the reference test point (baseline @ 28 to is due to the quite high baseline A/C system C) fuel consumption value,

3.C - 60% R.H. 2.5 Fuel Add. Cons. [l/100 km]
bench data: where the analysis is based on theoretical vehicle models, with typical engine efficiencies. The defined thermal load (f{tambient}) of the vehicle(s) and the corresponding cooling demand is the basis for performance tests carried out in test benches as shown in figure 10.
2500 Required compressor shaft power [W] 2000

HFC R744

Ambient temperature [C] 50
1.C - 70% R.H. 1.- 50% R.U C 0.5 Panda - R-134a 0.100 Ambient Conditions (H - kJ/kg) Panda - R-744 1st Panda - R-744 2nd
Figure 10 - Bench test data (@ equal cooling capacities) used as input for LCCP calculation for Ford Ka
vehicle data: the analysis is based on measured fuel consumption as function of ambient temperatures (figure 9), which can be applied for various climates.
Fuel consumption measured on-board Fiat Panda R134a Fiat Panda R-744 Fiat Panda R- 744 2nd* Life Time Emissions Athens Paris Trondheim kg CO2/ 12Years 377
Figure.9 - Fiat Panda fuel consumption increase due to the air conditioning vs. ambient air enthalpy. The data have been used as input for the LCCP. COOL DOWN. The Table 3 synthesizes the results of the cool down test comparing the baseline vehicles with the R-744 A/C system equipped demonstrators. The results shows that the R-744 system are able to guarantee adequate cooling performance event at high thermal load and the second version of the Panda system with increased compressor displacement allows to achieve better performance at the end of the cycle (idling). It should be highlighted that this increase of cooling power is paid with a fuel consumption increase (see table 1 and figure 9).
*@ higher thermal comfort
Table 4:a LCCP (Life Cycle Climate Performance) estimation for R-134a and R-744 B-COOL systems tested on board (see figure 9).
Fuel consumption measured on bench Ford Ka R-134a Ford Ka R-744 Life Time Emissions Athens Paris Trondheim kg CO2/ 12Years 552 275
Table 4:b LCCP estimation for R-134a and R-744 BCOOL systems tested at bench (see figure 10). The LCCP estimation in table 4 has been performed considering that: the materials are from the same region (e.g. Al from North Europe) and assembled in the same plant (e.g. France), this implies that 1 kg of CO2 is emitted for each kg of and the installed A/C system Entire life time HFC-loss are estimated as equal to 0.4 - 0.65 kg, including 15-25% loss during service (1x in central & northern EU; 2x in southern EU) and 50% recovery at End of Life (EoL). Service is assumed to be requested after 150g HFC loss in S-EU and 180g in C&N-EU

T Outlet Mean [ C]

Panda - R134a Panda - R744 1st Panda - R744 2nd Ka R134a Ka - R744

9.9 10.4 8.0 14.0 12.0

9.3 5.4 7.0 9.0 5.0

7.4 4.9 7.0 7.0 5.0

10.6 7.1 5.0 7.0 5.0

16.9 14.3 9.0 17.5 19.0

Table 3: cool down test cycle at 43 30% R.H. and C, W/m. see figure 4
LCCP - LIFE CYCLE CLIMATE PERFORMANCE
Two approaches [4] have been applied within B-COOL to estimate the LCCP of a mobile air conditioning system:
Both the calculation methods have been applied: using the vehicle data (Fig. 9) for the Fiat Panda using the test bench data for the Ford KA (Fig. 10) The LCCP calculation has been performed considering three different climate regions: Athens, Paris and Trondheim. The results of the analysis are reported in table 4. There is evidence that the R-744 system has a lower LCCP in all the three evaluated conditions. Even at higher fuel consumptions of the R-744 system at an improved thermal comfort, the reasonable HFC-leakage rates results in higher LCCP values for the R-134a systems. In addition to that the data in the table also show that both the adopted LCCP calculation approaches confirm the ranking between the two systems. It should be underlined that, when bench test data are used as input, the procedure risks to underestimate the LCCP value, as the tables show: the Fiat Panda MAC system has a significant lower fuel consumption than the Ford Ka system when measured on board. To estimate accurately the effect of vehicle fuel consumption of the MAC during a driving cycle when bench data are use as input a sophisticated and well tuned vehicle model is required. The system bench test can not take into account the effect of on board installation. The use of on board measurements allows to obtain a more reliable value of the LCCP value.
To evaluate more accurately the risk an mean C Feet C head C experimental 7 activity has been carried out considering the most severe 6 conditions: a vessel of 1-liter volume, containing 400 g of
5 C (% vol.) t (s) 3000 4000
Figure 9 - R-744 concentration at driver place with 4 passengers, maximum ventilation, re-circulation. Sudden leak: all the charge is released in 60 s. maximum leak
C Feet 14 C (% vol.) 3000 t (s) C head C mean
Figure 11 - R-744 concentration at driver place with 4 passengers, no ventilation, re-circulation. Sudden leak: all the charge is released in 60 s. R-744 and temperature controlled has been used to simulate the A/C system leak in a Fiat panda cabin. The R-744 concentration is measured by means of sensors (accuracy 20 ppm in the range of 0-10000 ppm) placed at the driver and back passenger places at breath and at feet level.

ON BOARD SAFETY

R-744 is a non-toxic refrigerant (as classified in EN 378), however at concentration equal or higher than 3% vol. it causes stimulating effect on the respiratory centre and could be lethal at concentration higher than 9% vol. In the framework of the B-COOL project, tests and theoretical analysis have been performed to assess the risk associated to the R-744 leak in the cabin so to evaluate if safety devices are required. The R-744 A/C system has an internal volume of around 1.2 l to 1.4 l with a charge in the range of 350 g. The empty cabin volume of a B-segment car is around 2.1 m. The highest peak R-744 concentration is theoretically reached when the entire refrigerant of the A/C system is discharged in the cabin and 4 passengers are on board (air volume reduction of 200 l approx. and the R-744 emission due to the respiration). These unrealistic conditions lead to a maximum peak concentration of 12% vol.

Figure 12 - Two mini-sheds used to evaluate the system tightness in the ARMINES-CEP laboratory
A test matrix has been followed considering different leak rate, ventilation level, recirculation and passenger presence. The air outlet has been oriented to the drivers head, and the vents at the front passenger seat are closed so to increase the R-744 concentration at the drivers head. On the basis of literature data and tests it has been found that the R-744 concentration increase due to the passenger presence can be estimated in: 0.5 %vol./passenger: recirculation and no ventilation 0.1 %vol./passenger: recirculation and max ventilation The tests represent worst-case scenarios with the highest pressure in the evaporator and the front panel outlets oriented to the face of the occupant. The tests results show that the most critical cases are: leak in recirculation mode and no ventilation sudden leak in recirculation mode and full ventilation
optimized compressor non-corrugated flexible lines. sound insulation and vibration damping material. mufflers.
LEAK RATE AND SYSTEM RELIABILITY
This issue has been one of the most common in the history of R-744 MAC systems, and also is important within the context of the B-COOL Project. The leakage rate needs to be kept under control so that the system only needs to be serviced within the specified timeframe, while offering good performance. If a component leaks out of the specified rate, the system will loose its charge and will stop working, requiring a refill. The most critical issue is the leakage through the compressor shaft seal. Other sources of leaks are the seals and fittings, but metallic seals and good tightening of the fittings are to be used to keep the leakage within specifications. The B-COOL system leak rate has been measured adopting the concept first developed for the measurement of leak rate of A/C systems running with R-134a [5] and based on the measure of the R-744 concentration in an accumulation volume named minished. The measured leak rate was not acceptable, indicating that further improvements are required to reach an annual leak rate of about 50g/year that can be considered a reasonable target maintenance/recharge.

and indicate that the risks due to the R-744 release in the cabin can be prevented safely by: managing properly the recirculation flap detecting critical leak by means of conventional diagnostic tools (e.g. pressure and temperature sensor monitoring), so no additional sensors are strictly required
In addition to that, it is important to remark that: If the leak happens when the engine and/or the electrical systems are OFF (e.g. parking) sensors or other active devices are useless because not active and R-744 concentration drops rapidly to non critical values just when the door is open If the leak happens when the engine and/or the electrical systems are ON a critical leak can be detected elaborating properly the signal of pressure and temperature sensors and the information available on the vehicle network. The activation of fresh air mode and ventilation can prevent any risk.

COST ANALYSIS

The B-COOL project included the prediction of system cost with a very ambitious target of an additional cost of 30 Euro/system. In order to insure the coherency of cost estimates and to have a common baseline the cost estimation has been performed with the following main assumptions the given costs are the ones paid by the car manufacturers which means prices for the suppliers as reference an average R-134a system cost has been identified on the basis of the Fiat Panda and Ford KA A/C loops and components. The present cost has been reduced according to the market trend to estimate the cost in 2011 the electronic control has been excluded the reference year is 2011 for R-744 production
The noise and vibration may represent an issue for the R-744 systems. The compressor is the main source, while the lines are another risky element. The B-COOL system has shown limited problems related to NVH. The NVH level is aligned with the baseline vehicle characteristics. In general several options can be considered and are still under study to decrease the negative NVH characteristics:
production volume assumption is 300.000 units/year for all the components except gas cooler and evaporator where the selected volumes considered are the full production capacity (i.e. 1.4 million units/year).
diffusion, but smaller displacement compressors (15 cc) guarantee better efficiency. As it has been previously mentioned, on the short term the cost will be significantly higher than present R-134a. If all the OEMs were to switch to R-744 technology and production volumes were increase, the cost would likely decrease but would hardly reach the same level of R134a system. The technical developments within the B-COOL Project have led to specific solutions for the use of this technology in small cars. The B-COOL project demonstrated that the R-744 technology for A-B segment cars seems technologically affordable even if reliability and system additional cost are still open issues that need to be further investigated.

As synthesized in the table 6, the cost of a 2011 R-744 A/C system ranges from 1.5 to 2 times the cost of a 2011 R-134a loop (e.g. +100 Euro +150 Euro). This estimation is far from the original B-COOL target of an additional cost +30 Euro or in other words a target of 1.2 times the cost of a 2011 R-134a A/C loop.
Compressor Evaporator Condenser/Gas Cooler Lines, Accumulator, IHX Refrigerant Sensors Expansion Device Total
R-134a R-744 (reference) Min (%) Max (%) 1 1.3 1.1.5 1.0.8 1.2.8 3.0.3 0.1.0 1.1.3 1.6 1.5 1.8

REFERENCES

[1] www.R744.com [2] Mthode de mesure et mesures des surconsommations de climatisations automobiles Convention ADEME 067 - RAPPORT FINAL Rfrence ARMINES 20152 - Jugurtha BENOUALI, Denis CLODIC (ARMINES), C. MALVICINO, S. MOLA (CRF) [3] Mobile air conditioning fuel consumption & thermal comfort assessment procedure, C. Malvicino (a), S. Mola (a), D. Clodic(b) - (a) Centro Ricerche Fiat, (b) Ecole des Mines de Paris, Center for Energy and Processes. IIR Gustav Lorentzen Conference on Natural Working Fluids, Trondheim, Norway, May 28-31, 2006 [4] Global environmental &economic benefits of introducing R-744 mobile air conditioning, Armin Hafner & Petter Neks, SINTEF Energy Research Trondheim,, Norway, 2nd International Workshop on Mobile Air Conditioning and Auxiliary Systems. Turin, Italy, November 2007 [5] Measurement of Leak Flow Rates of Mobile Air Conditioning (MAC) Components - How to Reach a Generic Approach, SAE 2007-01-1186, SAE 2007 World Congress, Yingzhong Yu, Denis Clodic, Ecole des Mines de Paris, Center for Energy and Processes
Table 5: cost range of a R-744 A/C system, relative to a 2011 R-134a A/C system (unitary cost).
Unfortunately, for first applications in 2011 on small cars, it seems very hard to reduce the todays given figures in a significant way.

CONCLUSIONS

Within the B-COOL EU-funded project a R-744 air conditioning system has been conceived, developed and installed on two vehicle demonstrators representative of the European A-B segment: a Fiat Panda and a Ford Ka. The R-744 air conditioning systems have been fully characterized on bench and on board and compared with the baseline R-134a systems. The results demonstrate that the performance issues have been solved and R-744 A/C system can achieve the same efficiency level of present R-134a systems, even if further developments and testing are required to reach the same reliability levels as with R134a systems. The system efficiency will increase when right-sized compressors will be available. The 28cc externally controlled variable capacity compressors, designed for C-segment cars, when used on small cars, as in this study, in partial load, have a lower efficiency. Those compressors can be used to validate the rest of the components, and in a first phase of R-744 system

CONTACT

Carloandrea Malvicino Centro Ricerche Fiat, Head of Thermal Systems Depts. +390119083260, e-mail: carloandrea.malvicino@crf.itf

From Back of the Envelope to Large-scale Simulation: Public Policy Evaluation and Support in Complex Domains
Paul R. Kleindorfer paul.kleindorfer@insead.edu
Paul Dubrule Professor of Sustainable Development & Distinguished Research Professor, INSEAD Anheuser-Busch Professor of Management Science, Emeritus, The Wharton School, University of Pennsylvania Winter Simulation Conference--Baltimore December 7, 2010

Topics for Discussion

Complexity and multi-party stakeholder values drive the use of large-scale simulation modeling Illustrative Examples from Recent Research and Policy Initiatives
Cat Risk Insurance Implementing the European Postal Directive Commercial Fleet Operations
Salience of Legitimation/Validation Problems
Complexity and Interdependence are Increasing
Globalisation, including Opportunities/Risks in Emerging Economies, Outsourcing and Off-Shoring and Extended Supply Chains Energy and Agriculture (and other traditionally independent sectors) now deeply intertwined because of sustainability concerns and policies Customer empowerment and forward and temporal extensions of Supply Chains to encompass customers Growth in importance of electronic markets of global scope Information, logistics infrastructure & technology are all moving to provide virtual integration of the resulting economic structures
Growth in International Trade*
Total Exports (M & S) 2000 = $8 Trillion Total Exports (M & S) 2009 = $20 Trillion

Globalization

increasing cross-border trade flows increasing demand for cross-border logistic & other services Outsourcing

Technology Drivers

Decision aids Communications e-Commerce Open Innovation

Market Liberalization

INTERDEPENDENCIES = SYSTEMIC RISKS

Business Transformation

Integrated Service Offerings integration with business processes
Increasing deregulation and liberalization/WTO Markets & Politics/Sustainability
*Figures are in current $s. Source: WTO--http://stat.wto.org/Home/WSDBHome.aspx; for a
Dicussion, see Kleindorfer and Wind, The Network Challenge, Wharton Publ. 2009.
Growth of Supporting Infrastructure for Logistics and Contracting
Reformulating & Extending Risk Management
Epistemic risk (aka knowledge risk): Risk arising from our lack of knowledge (episteme = Greek for knowledge) Aleatory risk: Risk arising from underlying randomness in nature (alea = Latin word for dice) Behavioral risk: Risk arising from human factors Important consequences of the differences Epistemic Risk & Behavioral Risk can be reduced Aleatory Risk remains Black Swans (Nassim Taleb) are unexpected events that arise because of epistemic risk. These are exacerbated by behavioral biases of myopia and availability. Reducing epistemic risk has become a central issue in modern risk management.
Nassim Talebs Assault on Normality
Daily Movements of DJIA Between 1916 and 2003 (The Economist, Jan 24, 2009)
Daily Movement Number of Days Predicted by the Normal Distribution One day in 300,000 years Actual Number of Days 48
> 3-4 % > 4-5% >7%
It is not so much that we are beset by Black Swans, but by mental models that do not prepare us for the appearance of many events, which we then classify as Black Swans!
Navigating the Knowledge Spectrum

Newer Approaches

Scenario planning Influence diagrams Peripheral vision Real options analysis Multi-agent simulation Certainty Risk Uncertainty Ambiguity Systems Thinking Idealized Design Strategic Gaming Stakeholder Mapping Complexity theory Chaos/Ignorance

Known Traditional Tools

Cost Benefit Analysis Net Present Value Linear Programming Point forecasting Optimization theory Utility theory

Unknown

Unknowable
Decision trees Bayesian updating Single agent simulation Portfolio theory Stochastic modeling Insurance & Hedging
Adapted from Schoemaker (2002)
This framework goes back to Frank Knight and Friedrich von Hayek. For details see P.R Kleindorfer (2010). Reflections on Decision Making under Uncertainty. URL: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1310239##

http://www.weforum.org/documents/riskbrowser2010/risks/#

Economics

Geopolitics

Environment

Society

Technology

Summary of the Argument
Globalization, market liberalization and the evolving economic order have given rise to an increasing need to understand complex interactions between company strategies, markets and public policy Large-scale simulation models are becoming the vehicle of choice for understanding these interactions in the public policy arena Some Examples
Catastrophe Modeling for Natural Hazards & Climate Change Regional and global liberalization of Markets Operational risks for banks and financial institutions Vaccination and global health problems Many, many others we take only 3 to illustrate the emerging process
Example: Catastrophe Models
Various modeling firms have put into place large-scale simulation models to evaluate the consequences of catastrophes BEFORE THE FACT, including the impact of various mitigation strategies
AIR Worldwide EQECAT Risk Management Solutions (RMS)
There are also publicly available models such as HAZUS (developed for FEMA), Los Alamos National Lab Model for Pandemics and Regional Models for SA, Europe, Middle East, Africa and Asia, underlying current OECD policy valuation of cat risks Uses of Catastrophe Models (from property damage to public health consequences)
Originally for insurance and reinsurance companies to assess exposures following Hurricane Andrew in 1992, reinforced by many events since then Climate change issues now a central area of application Extended to evaluate terrorism risks Extensions to evaluate risks and strategies for coping with pandemics
Simulation as an Integrative Framework connecting Science, Private and Public Policies
Scientists Estimating Probabilities of Disasters (e.g. USGS, NOAA) Engineers Estimating Damage from Disasters
Organizations Modeling Distribution of Losses from Disasters (e.g. AIR, EQE, RMS)
Reinsurance and Insurance Industry

Financial Instruments

Financial Institutions
Complementary Roles of Mitigation and Insurance
Reinsurance and Insurance Industry Financial Instruments

Insurance Protection Against Losses Encourages Mitigation
Inspections Enforcing Provisions Certification Construction Industry and Real Estate Sector
Property and Lives at Risk ---------------------Decision Processes of Firms & Homeowners
Insurance Requirements Cost of Capital
Joint Mitigation Strategies Joint Risk Bearing Strategies

Public Sector Agencies

Capturing Epistemic & Aleatory Risk
Nature of Hazard Sources Recurrences & Attenuation Assets and Lives at Risk Location Type
Aleatory Risks Hazard Related Behavioral Asset Related
Cat Simulation Historical Loss Data Engineering Data Fragility Estimates Losses (Economic & Lives) Direct Indirect Public & Private
Epistemic Risks Data Related Model Related Theory Related
Example: Charleston, SC Region
Mapping of Total Building Count
Charleston, South Carolina Region
Occupancy Mapping Example: Residential, Low Code & Low Inferior
OCCUPANCY W1 S1L S2L S5L C1L C2L C3L PC1 PC2L RM1L RM2L URML MH RES1 92% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 8% 0% RES2 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 100% RES3 62% 0% 3% 0% 2% 2% 0% 0% 0% 5% 4% 22% 0% RES4 48% 5% 4% 4% 8% 4% 0% 3% 3% 3% 3% 15% 0% RES5 7% 7% 6% 6% 17% 6% 3% 8% 6% 5% 5% 24% 0% RES6 22% 11% 8% 8% 8% 3% 2% 4% 3% 5% 4% 22% 0% W1 S2L C1L PC1 URML C3L RM1L RM2L Wood light frame S1L Steel moment frame low-rise (LR) Steel moment frame mid-rise (MR) S5L Steel frame w/unreinforced masonry, LR Concrete moment frame, low-rise C2L Concrete moment frame, MR Precast concrete tilt-up walls PC2L Precast concrete frame w/conc. shear, LR Unreinforced masonry bearing walls, LR MH Mobile Homes Concrete frame w/unreinforced masonry walls, LR Reinforced masonry bearing walls w/wood or metal deck diaphragms, LR Reinforced masonry bearing walls w/wood or metal deck diaphragms, MR
Assumptions underlying Epistemic Risk

Example: Charleston, SC

Sources: A few hundred background & characteristic events provide the representative set Magnitude Conversion: Boore and Atkinson (1987) Attenuation: Project 97 relationship (i.e. 1/2Frankel et al., 1996 + 1/2Toro et al., 1997) Soils Mapping: Soil class D (stiff soils), NEHRP amplification factors (FEMA, 1997), and no liquefaction potential Inventory: HAZUS default database Vulnerability: HAZUS default fragility curves

Results: Mean EP curve

Epistemic Uncertainty & Loss Exceedance Probability Curves
Uncertainty in Probability
Probability p(L) that losses will exceed L

Uncertainty in Loss

Loss, L (in dollars)

Using Modeling for Policy Moving: P(100) to.004 A. Original Situation B. Effects of Mitigation C. Underwriting Choices D. Risk Transfer Strategies
1.2 1.0 0.8 0.6 0.4 0.2 0.300

Probability (%)

Loss ($M)
Probability (%) Probability (%) Probability (%) Probability (%)
1.2 1.0 0.8 0.6 0.4 0.2 0.100
Probability (%) Probability (%) Probability (%)

$1,000

$1,200

$1,400

$1,600

$1,800

Miami Beach E. Dade Co. Key West E. Palm Beach Co. E./S.C. Co's Monroe Co. W. Pinellas Co. Martin Co. Mid-N.E. Co's Hialeah Ft. Lauderdale Brow ard Co. W. Palm Beach Co. W. Dade Co. St. Peters burg West Co's Miami E. Duval Co. Rem. Brevar d Co. Rem. Pinellas Co. Tampa Rem. Escambia Co. Osceola Orlando Polk Co. Jacksonville Rem. State W. Duval Co.
Using Cat Modeling for Regulation Florida Advisory Loss Costs

Territory

Indicated

Implemented

Key Challenges for Research
Understanding and dealing with epistemic risk and uncertainty
Categorizing epistemic risk and linking it to choice Determining key clusters of scenario or decision variables that determine really good or really bad outcomes
Validation of models and conveying their strengths and weaknesses to multiple stakeholders with varying levels of sophistication
Situation 1: Insurers and Reinsurers -- well advanced Situation 2: Regulators, the Public and Modeling Firms remains a tough uphill climb
Climate Change is the BIG event on the horizon, requiring massive simulation models which will clearly meet many validation challenges, including wide differences in values & beliefs.
Selected References on Cat Models
Grace, Martin F., Robert W. Klein, Paul R. Kleindorfer, Michael Murray, Catastrophe Insurance: Supply, Demand and Regulation, Kluwer Academic Publishers, Boston, 2003. Mulvey, John et al., Dynamic financial analysis for multinational insurance companies. In S. Wallace and W. Ziemba, North Holland Handbook on Finance, 2005. Grossi, Patricia and Howard Kunreuther (eds), Catastrophe Modeling: A New Approach for Modeling Risk, Springer Verlag, New York, 2005. Kleindorfer, Paul R., Climate Change and Insurance, in Beyond PML: Frequency and Severity, AonBenfield, Sydney. September, 2009.
Liberalization and Restructuring of Network Industries Example: The Postal Sector
Liberalization and Privatization: A Global Phenomenon in the Postal Arena
EU Starting in early 1990s Privatization: Netherlands, Germany, Austria. USA PAEA (2006) Latin America and EE In search of service quality and economies of scale Asia Let a 100 flowers bloom! Australia, New Zealand, Canada Social contract/charter emphasizing corporatization and commercialization

Drivers & Levers of Postal Reform
Cost of and Demand for the USO Changing market structure in communications markets and related electronic substitution Hidden costs (health & pension benefits) Not-so-hidden costs (X-inefficiencies, including those related to technological progress) Outdated regulatory practices Research and lessons from other sectors Hope

Path of Postal Reform

Market Structure Competition (FMO)
Commercial Public Enterprise Privatized Investor-Owned Enterprise

Public

PO/USP Governance Private
Planned Path of Postal Reform

Monopoly

Economic Analysis of Liberalization

Themes

Market(s) (Complements or Substitutes?) Efficiency and Cost Structure of the USP
Drivers that influence the Impacts
Postal scale and density,. lettermail, parcels, direct mail Electronic competition Labor cost, efficiency & Technological change USP Financing needs Benefits of Workable competition achieved Scope & location of entry Employment impacts Price and product changes in response to competition Variety of impacts across countries
Impact of scope and Counter density, frequency characteristics of of delivery, USO Regulatory policies Licensing, pricing, access,
Diversity of driver values across countries
Structure of the Economic Model
Following dAlcantara & Amerlynck (2007)

Model Computation

Incumbent
USO / Non USO products Innovation Scenarios Price Caps & Tariffs Uniformity Restrictions Tariffs per destination Network Access Tariffs
Regulatory Restrictions Cost basis for Access

Model Foundations

Micro-segmentation

Model Outputs

Impact on Incumbent / Competitor Volume
Customers Geog. Households Licences Business Products Electronic substitution MS% Switching function

Demand Model

Incumbent profile
Cost structure Fixed Tour / variable costs
Optimal Decisions Market Equilibrium Tariffs

Impact on Operators

Market shares FTEs Cash flows P&L s 2008 2012

Competitors profile

Cost structure Access or bypass investment rules

Competitors

Product definitions Innovation Quality & Features Bypass strategy Tariffs per destination

USO Definition Access rules Regulatory cond.
Price caps, Quality Info., Sanctions
Results of various scenarios were analyzed by simulating the impact on representative countries
Low impact able to sustain USO under FMO with changes that can readily be expected to occur under market conditions; Average impact modest potential loss of market share to entrants, with some rebalancing measures necessary, amongst which restructuring of USP operations, in order to sustain USO under FMO; High impact significant potential loss of market share to entrants, so that rebalancing measures are imperative, amongst which restructuring of USP operations and possible USO subsidies, in order to sustain USO under FMO.
Data from across EU were used to determine impacts (with predictions from existing studies)
Name Scale Urbanisation rate (one element of Postal Density) Labour cost ratio Demand elasticity Automation % in sorting Automation % in sequencing Flexible USP workforce in collection Flexible USP workforce in sorting Flexible USP workforce in transport Flexible USP workforce in delivery Franchised counters Definition Mail items per person per year Percentage of population in urban area (UN standard) Ratio of freelancer vs employee FTE cost (Eurostat) Variation in % of volume for a 1% variation of price Percentage of mail item automatically processed Percentage of mail item automatically processed Percentage of freelancer FTE hired by USP in activity Percentage of freelancer FTE hired by USP in activity Percentage of freelancer FTE hired by USP in activity Percentage of freelancer FTE hired by USP in activity Percentage of counters franchised Lo-Imp 500 45% 90%.2 90% 75% 0% 26% 57% 29% 46% Av-Imp 200 70% 75%.4 50% 25% 0% 6% 10% 8% 28% Hi-Imp 10 95% 60%.6 10% 10% 0% 0% 0% 0% 0%
Example of Results: Urbanisation rate and Postal scale
Postal scale (items per inhabitant) 200 100
GR, HU, IS, IE, LV, BG, CY, CZ, EE, IT, LT, LI, PL, PT, RO, SK, MT, ES SI AT, FI, NL, CH BE, DK, FR, DE, LU, NO, SE, UK
High urbanisation puts USPs less at risk as they tend to have less expensive rural routes to support. HoweverThere are complex interactions under competitive entry.

50% 60% 70% 80% 90% 100% Urbanization rate (UN standards)
Low postal scale typically implies greater risk and greater need for re-balancing measures
7% 19% 37% Scale (items per inhabitants per year) 14-100 100-200 200-300 300-400 >400 Balance to finance

11% 26%

A countervailing factor is upstream market opening through access/worksharing, which can contribute to increasing scale by generating more bulk mail volumes, particularly in direct mail.

Important

Limited

Significant

Low impact country Average impact country High impact country Postal scale
Labour cost differential has a MAJOR impact on the market share of new entrants
Labour cost ratio directly impacts overall cost-to-serve, hence the competitive positioning of the incumbent.
Low impact country Average impact country High impact country New entrant market share Significant Limited Important 40% 50% 60% 70% 80% 90% 100% Labour cost ratio
Labour Cost Ratio = Estimated Ratio of Entrants Hourly Wage to Incumbents
Conclusion of EU Studies (and reflected in the 3rd Postal Directive.2011): Country differences are crucial!

Ready to Go!

121-145 101-120 80-100 65-80
Significantly above average Above average Below average Significantly below average

IE DK NO SE LT

BE LU FR

DE CZ AT IT SI

SK HU RO

BG PT ES GR

Selected References on Postal Reform & Modeling
Crew, Michael A., Paul R. Kleindorfer and James I. Campbell, Jr. (eds.), Handbook of Worldwide Postal Reform, Edward Elgar, Cheltenham, UK: 2008. Crew, Michael A. and Paul R. Kleindorfer (eds.), Heightening Competition in the Postal and Delivery Sector, Edward Elgar, Cheltenham, UK: 2010. Crew, Michael A. and Paul R. Kleindorfer (eds.), Reinventing the Postal Sector in an Electronic Age, forthcoming, Edward Elgar, Cheltenham, UK: 2011.
Sustainable Mobility and Commercial Fleet Operations: Electric Vehicles
Vanessa Chocteau (La Poste)
David Drake (INSEAD) Robert Hain (WHU) Paul R. Kleindorfer (INSEAD) Andrei Neboian (WHU) Renato Orsato (INSEAD) Alain Roset (La Poste) Stefan Spinler (WHU)

Sustainable Fleet Operations in the Postal Sector, INSEAD Working Paper 2010/30/ISIC/TOM
5 Poles of the EV Problem for a Postal Operator-PO
Operator (PO) Real-options framework to evaluate alternatives Multiple objectives (financial, carbon, labour, ) Electricity supplier Improved sales and load factor Externalities (V2G synergies with wind) Automotive Manufacturer Sustainable mobility strategies and related technology scenarios Interacting economies of scale with overall EV market penetration and EV pricing/design/infrastructure Carbon accounting, financing (JI) and markets Impact on total PO value chain Transport-specific activities Government Taxes on oil (revenue) Subsidies and regulations
Emissions from various energy & EV power systems
Energy Source Oil Oil Oil Motive Energy for Automobile Gasoline Diesel Diesel Power System Internal Combustion Internal Combustion Internal Combustion EV with LI Ion Battery EV with LiIon Battery Liters of Gas Gr of CO2 Eqv./100 km Eqv./km * * Gr of CO2 Eqv/kmAuto Alone

6.7 6.4 5.2 6.0 2.2

EU Elect Mix Electricity ** Wind Electricity
This calculation is for a VW Gulf-sized auto * Includes energy/CO2 for all upstream activities ** Electric Power Mix based on EU-15 use of nuclear, coal, gas, oil & renewables 2008

German Data 2008.

Sustainable Fleet Initiatives are aligned with Broader Sustainability Agenda

Social Impact

Beneficial reputation for both local and global community Increased employee motivation Increased legitimacy in dicussions with regulators and the public

Economic/Profit Impact

Reduced CO2 emissions as potential revenue generator via JI credits Energy savings Potential attractiveness for sustainable portfolios
The impact of Low-carbon Fleet Operations

Environmental Impact

Reduced CO2 emissions Reduced particulate matter in urban areas Noise reduction Reduced level of ozone precursors
Sustainable Fleet Initiatives & POs vehicle expenses 2010 Figures show EV Costs >> ICE Costs
PO Transportation Expenditures
La Poste Total Number of Vehicles Number of Collection & Delivery Vehicles that are candidates for switch to EVs* Current Vehicle Expenses Capital and Depreciation Maintenance and Operating Costs Fuel Costs Total Vehicle Expense 45,000 30,000 USPS 200,000 142,000

Illustrative Comparison of (EV) and (ICE) Cost and Environmental Performance
EV Purchase Price () Resale Value at End of Year 5 () Vehicle Expenses () Annual Capital Cost Annual Operating Cost Annual Maintenance Cost Total Annualized Vehicle Expense CO2e Emissions (tons/year) Annual Value of CO2e Emission Reductions () Other Environmental Benefits TOTAL ANNUAL COST/vehicle 25,000 5,000 5,225 1,029 1,110 7,364 0.0 54.00 --7,418 ICE 10,000 2,000 2,090 1,139 2,221 5,450 1.8 0.0 --5,450
Millions of Euros 93.6 132.9 68.8 295.3
Millions of Dollars 169.8 777.7 148.3 1095.8
Motivation for Electricity Supplier
Generally part of smart grid initiatives to improve energy efficiency at distribution and consumer level Improved load factor from off-peak sales (true for most commercial fleet operators) Vehicle to Grid (V2G) and Battery to Grid (B2G) operations allow a potentially valuable storage source, given growth in wind power
During vehicle life timebidirectional charges After vehicle life of battery ended, use banks of batteries for centralized storage
Relatively inexpensive hook-up for unidirectional flows. More complicated for bidirectional flows and battery parcs.
Vehicle to Grid (V2G) and Commercial Fleets*
*For an introduction to V2G issues from a GRID perspective., see http://www.magicconsortium.org/_Media/test-v2g-in-pjm-jan09.pdf
Basic alternatives/options in vehicle replacement with electric vehicles (EVs)

option value

Continue using ICVs
Find a practical solution to transfer investment situation into option terms; Identify relevant stochastic processes; Scientific and numerical valuation for practical decision making;

Combination

Refurbish with EVs
Additional economic value for La Poste through replacement with EVs using a dynamic scenario;
Allow for modification possibilities to introduce more sophisticated and detailed assumptions after first results; Identify the sustainability impact of the vehicle replacement strategy (Saved total CO2 emissions);

Managerial Implication s

Best replacement path and resulting design of the replacement strategy for La Poste (including timing and amount to be replaced); Improved measurable carbon footprint of La Poste;
The ultimate goal is to identify the flexibility value of renewal strategies and their sustainability impact
Input parameters and decisions have influence over different time horizons
Cost impact Decision horizon

2015 2020

End of project decision

Maintenance costs

Irrelevant time horizon all vehicles reached end of lease
Demand EV supply Leasing rate (ICV and EV) Electricity Price Fuel Price Battery Price and Salvage Value

Leasing rate (ICV)

Complete model minimizes total fleet costs by choosing appropriate vehicle portfolios

Breach portfolio

ICV(t, age) Limitation on breach Vehicle stock (acq.) EV(t) ICV(t) Breached ICVs Total stock Demand (Mail/Parcel Market)

EV supply

ICV(t)

M L B M L

BEV(t) BIV(t)

EV(t) ICV(t)

Electr. Fuel
Saved Running costs Legend:

Saved impact costs

Impact costs Optimize

Running costs

Fixed Variable
Total costs of the fleet discounted to present time

Generated

Upper bound

Acquisition portfolio

15-year horizon Real Options Solution Framework
Every year of operation, the fleet operator can decide on either replacing disused vehicles with ICVs or EVs i.e binomial decision; It would seem natural to replace the entire cohort (of same age vehicles) whenever a replenishment decision is made. Each state corresponds to a year of operation and fleet composition (ICVs and EVs); Movement to the right on the option map corresponds to time progression, movement down to increasing proportion of EVs in the fleet, represented by the degree of electrification (a); Initial movement down is afflicted with a one-time outlay of I(x, t); Base case (BAU scenario) could exclude the possibility to introduce EVs i.e. every year, the fleet operator replaces retired ICVs with new ICVs; Each state transition is affected by acquisition costs of new vehicles (qI and qE respectively for vehicle type I and E); Near-optimal solutions can be computed and simulated as to their financial and operational implications
One output is the flexibility (option) value of the setting - i.e. the option premium over the BAU case
Some Lessons/Reflections on EV Project
Collaborative innovation! Multiple-stakeholder risk sharing: Structuring of the interorganizational risk and benefit profiles and determination of total benefits and instruments for risk sharing (e.g. in pricing of initial leased vehicles, tariff setting with EdF, CO2 credits, ) Note: A significant further contribution to the payoff could arise from V2G operations, integrated with renewables/wind--Making good on this will require significant investments in the grid to allow load-following battery reserves to function properly. Given the complexity of this matter, large-scale simulation modeling will be required.
Commonalities & Differences
Project Cat Modeling for Extreme Weather Events Liberalization of the Postal Sector Stakeholders (Re-)Insurers Regulators Property Owners Municipalities Postal Operators & Employees Regulators Public (USO) Entrants Fleet Operator Electric Utility Auto Manufcter Government Performance Objectives Profitability Solvency Risk-based Premiums Key Modeling Challenges Legitimation Challenges Modeling Heterogeneous assumptions & stakeholders related epistemic Complex risks science/engine ering models

Heterogeneous Validation of Economic demand model stakeholders Efficiency Complex Public service Validation of economic objectives PO cost data models & data Country (USO) questions differences Competition NPV/Financial Risk sharing Carbon/GHG Innovation Optimization and Bounding Pricing models Battery prices Risk Sharing Media and misinformation Complexity of optimization
EVs and Commercial Fleet Operations
Validation: From G. B. Kleindorfer et al. Mgmt Sci (1998)
Approach Rationalism & Positivism Focus Logical Justification Truth Guarantor Logic and empirical verification Predictive success theory-free observations to test theories Survival thru paradigms shifts & answer key ?s Knowledge growth with participation Represenative Philosophers Descartes, Russell & Whitehead Pierce Friedman Popper Lakatos Hegel Kuhn Polanyi Habermas Bernstein Gadamer Validation Approach Rational Foundation & Empirical Verification Shown by predictive accuracy Continued testing to eliminate faulty models Accordance with expert opinion, wide acceptance Participation and consensus in the outcome

Instrumentalism

Theory as framework for prediction Theory as framework for prediction Progressive historical growth of knowledge Interpretation through dialog & practice

Falsificationism

Kuhnianism
Hermeneutics Legitimation

Conclusions

Globalization and Collaboration!
Increased networking and its interdependencies have made knowledge-based risk management a center piece for companies, societies and governments. Structuring and reducing epistemic risks for problems with multiple stakeholders has placed increased emphasis on collaboration, validation & legitimation. Large-scale simulation models are playing a fundamental role in connecting ex ante and ex post outcomes and choices and in understanding and reducing epistemic risks for major public policy initiatives.