Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure

Applied Energy xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

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

Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure Zlatina Dimitrova ⇑, François Maréchal EPFL, Industrial Process and Energy System Engineering Laboratory, SCI-STI-FM ME A2 402 (Bâtiment ME) Station 9, CH-1015 Lausanne, Switzerland

h i g h l i g h t s
The full hybrid electric vehicle suits for sustainable urban mobility and customer investment.
The full hybrid electric urban vehicle is efficient, with consumption less than 2 L/100 km.
The range extender vehicle is a technology for low CO2 emissions – less than 20 g/km CO2.
The total CO2 emissions for range extender and plug-in vehicles are sensitive to the use place.

a r t i c l e

i n f o

Article history: Received 9 January 2015 Received in revised form 11 September 2015 Accepted 15 September 2015 Available online xxxx Keywords: Multi-objective optimization Hybrid electric vehicles Energy conversion

a b s t r a c t The design criteria for modern sustainable development of vehicle powertrain are the high energy efficiency of the conversion system, the competitive cost and the lowest possible environmental impacts. In this article a multi-objective optimization methodology is applied on hybrid electric vehicles study in order to define the optimal powertrain configurations of the vehicle, estimate the cost of the powertrain equipment and show the environmental impact of the technical choices on the lifecycle perspective of the vehicle. The study illustrates optimal design solutions for low fuel consumption vehicles – between 2 L/100 km and 3 L/100 km. For that a simulation model of a hybrid electric vehicle is made. This model is coupled with a cost model for the vehicle. The techno–economic optimizations are performed for two case studies, illustrating the possibilities of the optimization superstructure. Firstly the life cycle inventory is written as a function of the parameters of the techno–economic model. In this way, the obtained environmental indicators from the life cycle assessment are calculated as a function of the decision variables for the vehicle design. In the second example the parameters of the energy distribution function are included as decision variables in the techno–economic optimization and are simultaneously optimized. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction In the traditional energy scenario by 2040, 90% of the global transportation will run on liquid petroleum based fuels ([1]). The proliferation of hybrid and other advanced vehicles, along with improvement of the conventional vehicle efficiency, will result in flattening the demand for petrol for personal transportation, even as the number of personal vehicles in the world doubles. The global competition for affordable energy and resources will lead to an increase of the diversification of energy sources, fuel types, and vehicles. This diversification will be greater in urban environments where the transport and distance requirements are ⇑ Corresponding author. E-mail address: [email protected] (Z. Dimitrova).

more compatible with diversified energy types and new energy distribution infrastructures ([2]). Alternative conversion technology and fuels are prospected for reducing the CO2 impact of the transport, but also the cost effectiveness of the proposed solution is researched. Thiel et al. [3] study the cost and environmental implications of the vehicle fleet CO2 emission regulation in the European Union. Bishop et al. [4] study the cost-effectiveness of alternative powertrains for reduced energy use and CO2 emissions in passenger vehicles. This work analyzed the cost-effectiveness of avoiding carbon dioxide (CO2) emissions using advanced internal combustion engines, hybrids, plug-in hybrids, fuel cell vehicles and electric vehicles across the nine UK passenger vehicles segments. The optimal design of from techno/economic and techno/ environmental point of view is researched.

http://dx.doi.org/10.1016/j.apenergy.2015.09.071 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

2

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

Nomenclature LCA life cycle assessment SoC state of charge v vehicle velocity (m/s) qbat battery capacity (kW h) el machine electric machine th engine thermal engine CML_01_short LCA method under OSMOSE Peugeot 3008 HY4 Peugeot vehicle with hybrid electric powertrain MOO multi objective optimization p_em power of electric motor (kW) p_th_engine power of thermal engine (kW) TTC ‘‘Toutes Taxes Comprise” – all taxes included REX Range Extender vehicle

Ribau et al. [5] research the efficiency, cost and life cycle CO2 optimization of fuel cell hybrid and plug-in hybrid urban buses. This study highlights the significance of the driving conditions and the conflict between the optimization of investment cost, efficiency and environmental impact in powertrain design optimization of these kinds of vehicles. The largest applied converters in passenger cars are the internal combustion engines – gasoline, diesel, adapted also for operating on biofuels. The state of the art today is considering the ‘‘tank-to-wheel” energy balance where basically for thermal powertrain, it is considered that 30% of the energy is used for the mobility, as mechanical power. The rest 70% are wasted – waste heat in the coolant 30% and waste heat in the exhaust gases 40% ([6]). The number of components that are necessary to realize modern future propulsion system is inexorably increasing. Hybrid electric vehicles are representative for the design complexity and the energy management strategy. Some studies research the energy efficiency analysis of a series plug-in hybrid electric bus with different energy management strategies and battery sizes. The efficiency improvement is based on the optimized energy management strategies and the benefit of that is the battery size reduction. Significant global efficiency improvements can be achieved once the system energy breakdown is identified, individuating the losses connected to each powertrain component. Finesso et al. [7] present a layout design and energetic analysis of a complex diesel parallel hybrid electric vehicle. The study focuses on the design, optimization and analysis of a complex parallel hybrid electric vehicle, and on the evaluation of its potential to reduce fuel consumption and NOx emissions. The optimal powertrain control strategy – in terms of the management of the power flows of the engine and electric machines, and of gear selection – is necessary in order to be able to fully exploit the potential of the hybrid architecture. Hybrid electric vehicles present advantages on the operating costs and the global economic model of investment. Operating costs should be considered as highlighted by Millo et al. [8]. Advanced powertrain vehicles reach lower consumption, low CO2 emission and the governments use environmental bonus for clean vehicles incitation. For personal transportation, the energy outlooks state that the hybrid electric vehicles will be an important part of the fleet in the future. Thus their optimization from efficiency, economic and environmental point of view is important. The cited works show that a lot of efforts are currently done in the optimization of the

NEDC new European drive cycle HEV hybrid electric vehicle PHEV Plug-in Hybrid Electric Vehicle Li- Ion lithium-ion CAFE corporate average fuel economy eq. equivalent PV photo voltaic GWP global warming potential ODP ozone deplation potential CO2 – eq. equivalent CO2 PSA Peugeot Société Anonyme R&D research and development

design and the energy management strategies of the advanced hybrid electric powertrains. They are all based on heuristic optimization approaches. Globally the optimization efforts concern the design of the powertrain and of the energy management strategies – based on iterations of design configurations and their operating and cost adaptation. A global optimization approach considering simultaneously the design and the operation optimal criteria for vehicles is still missing. Thus a global optimization approach for the design and energy management on the vehicle level, with multi-objective criteria (technical, economic and environmental) is needed. The innovative approach in this work is to optimize the vehicle propulsion system considering simultaneously the tank to wheel efficiency of the powertrain and the economic objectives (investment and operating costs), to assess the environmental impacts on the vehicle life cycle perspective. This paper illustrates the optimization methodology for low CO2 emission hybrid electric vehicles. At this stage, the tank-to-wheel efficiency and CO2 emissions are considered as objectives for the optimization. The techno–economic optimizations are performed for two case studies. Firstly the life cycle inventory is written as a function of the parameters of the techno–economic model. In this way, the obtained environmental indicators from the life cycle assessment are calculated for the optimal points. Thus the environmental impacts are evaluated with well-to-wheels perspective. In the second example, the energy management parameters represented by a simple continuous energy distribution function are included as decision variables in the techno–economic optimization. 2. Optimization methodology To optimize the energy efficiency of the vehicle, one has a hybrid electric simulation model. The optimization and simulation tools used for this publication have to fulfil the following requirements:
Enough flexibility to simulate a wide range of conversion technologies with different level of detail.
Integrate a dynamic profile simulation, an estimation of the system performances and the resistance efforts.
Possibility to include alternative fuels for an influence in the use phase.
Define the size of the equipment.
Include economic models to deduce the cost of the equipment.
Include environmental LCA based models to deduce the environmental impacts.
Give operation strategy (control) possibilities.

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

3

Fig. 2.1. Structure for multi-objective optimization, OSMOSE tool.

2.1. Optimization superstructure Multi-objective techniques are used to investigate the effects of the sizing of the powertrain components and to adapt the operation strategy of the vehicles. In this study, multi-objective optimization is performed with the OSMOSE tool ([9]) (Fig. 2.1). The general computational framework has already been described in Gerber and Maréchal [9], where techno–economic optimization is coupled with environmental indicators. The authors studied ‘‘environomic” optimal configurations of geothermal energy conversion systems. After that, the superstructure is adapted for vehicle applications by the introduction of physical vehicle simulation models and vehicle economic and environmental models. After multiobjective optimizations, the authors present in Dimitrova and Maréchal [10] ‘‘environomic” designs for hybrid electric vehicles. These designs are optimal from technical, economic and environmental point of view. ‘‘Environomic” designs of electric vehicles are studied in Dimitrova and Maréchal [11]. The computational structure for optimization contains a physical vehicle simulation model, with dynamic and thermal layouts. The cost equations are written in the economic model. The energy integration model uses the results from the dynamic and thermal flows calculations. The optimizer in OSMOSE is based on a genetic algorithm ([9]). This optimization technique is multi-modal and gives local optimums. The optimization is decomposed into four major parts:
A master multi objective optimization (MOO).
A thermo-economic simulation (TES).
A slave optimization (energy integration – EI), where the energy integration occurs.
A techno–economic evaluation (TEE). The developed model (Fig. 2.1) is a mixed integer non-linear problem (MINLP). It is solved by a decomposition method, using a master-slave algorithm in which decision variables can be grouped in master or slave sets as explained in [12].

The master set of decision variables includes the type and size of the equipment. These variables are used to define a superstructure for the vehicle energy system as illustrated by Fazlollahi and Maréchal [13]. The energy technologies database is presented as a list of available equipment, and the characterizing equipment tags are inputs for the master optimization problem. The master optimization is solved by an evolutionary genetic algorithm – EMOO. The superstructure receives sets of master decision variables and objective functions. Then the efficiency and the economic states are calculated using simulation models of the vehicle dynamic behavior. The cost and the LCA models are also executed. The decision variables can be for design of for energy management. The list of fuels or electricity and operating cycles (time, speed profiles) are requested for the optimization. The output is a proposition of equipment and parameters for energy management. The selected superstructure is in the master level and the results of the optimization are used in the post-processing phase to calculate the objective functions of the master problem as illustrated also in Fazlollahi and Maréchal [13]. The objective functions can be, economic and efficiency or environmental indicators. After the specified iterations by the user, the optimal solutions converged on the Pareto frontier curve. The energy integration model will not be used in the present study. Hybrid-electric vehicles differ also according to the degree of hybridization of the powertrain (Fig. 2.2) and the battery capacity. In the optimization part, the battery capacity is one of the decision variables, so the solutions are classified according to the functional classification in Fig. 2.2. 2.2. Vehicle model The vehicle simulation tool is SIMULINKÒ. The vehicle model is based on mechanical and electrical flows. The thermal layout of the internal combustion engine is constructed from measurement maps and included in the vehicle model. The level of the model is quasi-static. The vehicle is able to follow dynamic profiles

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

4

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

wheels to the energy sources. Proceeding in this manner insures the flexible and fast nature of the simulations. This is an important advantage for an optimization study. Fig. 2.3 illustrates the generic units that are modelled in the vehicle powertrain and the backwards approach to estimate the energy consumption. The different boxes of the model are explained in the next paragraphs. 2.2.1. Driving cycle The input of the model is a predefined discrete time and speed profile. The profile used in this study is the New European Driving Cycle (NEDC).

Fig. 2.2. Functional classification of HEVs in term of degree of hybridization and battery capacity (Guzzella [14]).

2.2.2. Vehicle The vehicle model uses the traction force on the wheels Ft, in order to find the power demand. It takes into account the rolling and aerodynamic resistance, the vehicle mass and the uphill force if driving on a slope according the equation below.

m generated from a library of driving cycles. The model has a loop control structure, linked to the required mechanical power, to follow the dynamic cycle. This control loop is called ‘‘back and forward” and allows, for a given design of the vehicle powertrain to simulate the energy consumption of the vehicle, on the given driving profile. The energy flow is computed backwards from the

dv ðtÞ ¼ F t ðtÞ  F RESISTANCE ðtÞ dt

ð2:1Þ

The resistance force can be divided on four sub-forces: The tire friction:

F r ¼ m  g  cr

ð2:2Þ

where g is the Earth’s gravitational constant in m, m is the mass of the vehicle and cr is the rolling friction coefficient

Fig. 2.3. Quasi-static model of the parallel thermal electric hybrid.

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

5

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

PEM ðtÞ ¼ g

The wheel bearing friction:

F bearing ¼ m  g  l 

d D

ð2:3Þ

where l the bearing coefficient, D the wheel diameter is and the d the bearing bore diameter. The aerodynamic losses:

Fa ¼

1  qa  C a  A  ðv þ v wind Þ2 2

ð2:4Þ

where qa is the air density, C a is the aerodynamic drag, A is the vehicle front area, v is the vehicle speed and vwind is the frontal wind speed. The uphill force:

F uphill ¼ m  g  sinðaÞ

ð2:5Þ

where a is the slope. Using v , F t and the wheel radius one can compute the wheel torque and rotation speed and then determine the power required to move the vehicle. All forces are defined in N. 2.2.3. Transmission A transmission model is used between the wheels and the energy converter to adapt the speed and the torque levels. The model has a manual gear box used for the validations on harmonized driving cycles (NEDC), with imposed gear ratios. For the usage driving cycles, when the gear ratios are unknown, a CVT is used for the estimation of the optimal gear ratio for each point of the drive cycle. 2.2.4. Energy converter The energy converter transforms the energy (chemical or electrical) from the energy storage into the mechanical power. The dynamics of such converters can be complex. Their modelling is simplified using efficiency maps, obtained from measurements in test benches. The electric motor considered in the model is a synchronous AC motor. The electromagnetic equations are not modelled, a black box approach based on motor efficiency being preferred. The inputs of the electric motor are its shaft’s rotation speed wEM and torque T EM . The output is the power demand PEM (Fig. 2.4). Thus, one can write the following equations for the positive and negative traction cases:

Fig. 2.4. Two-quadrant electric motor efficiency map.

wEM ðtÞ  T EM ðtÞ

EM ðW EM ðtÞ;T EM ðtÞÞ

ð2:6Þ

for T EM  0 where gEM is efficiency of the electric machine

PEM ðtÞ ¼ wEM ðtÞ  T EM ðtÞ  gEM ðwEM ðtÞ; T EM ðtÞÞ for T EM  0

ð2:7Þ

The efficiency values gEM ðwEM ðtÞ; T EM ðt ÞÞ for T EM  0 are obtained from the electric motor efficiency map of the QSS toolbox available in Guzzella [15]. As the efficiency of the electric motors is usually not measured in generator mode, it is proposed in Guzzella [14] to approximate it by mirroring the power losses as shown below:

gEM ðwEM  jT EM jÞ ¼ 2 

1

gEM ðwEM ; T EM Þ

ð2:8Þ

The efficiency map is illustrated in Fig. 2.4: The model of the combustion engine is based on the fuel consumption map. The inputs of the model are the engine shaft speed _ fuel . wICE and torque T ICE and the output is the fuel consumption m An example of a fuel consumption map is given in Fig. 2.5. The engine fuel consumption map is not scaled numerically. A new map is imported for each displacement volume. 2.2.5. Battery The battery model comes from Guzzella [15]. As illustrated in Fig. 2.3, the battery input is the power demand and the output is the state of charge. This is later the ratio between the delivered electric charge Q and the nominal battery capacity Q 0 (defined in A h) (2.9).

SoC ¼

QðtÞ Q0

in ð—Þ

ð2:9Þ

Q ðtÞ can be estimated considering that the current is a flow of electric charge. Thus introducing the discharge current IBT , one obtains:

_ QðtÞ ¼ IBT ðtÞ in ðA hÞ

ð2:10Þ

Furthermore, in electrical theory, power is equal to current multiplied by voltage. Thus one can write:

IBT ¼

PBT ðtÞ U BT ðtÞ

in ðAÞ

ð2:11Þ

Thus, in order to link the input PBT to the output SoC, an equation for U BT has to be found. Considering the battery as equivalent circuit

Fig. 2.5. Fuel consumption map of a 2.2 L Diesel engine.

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

6

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

composed by an ideal voltage source U OC in series with an internal resistance Ri and delivering voltage U BT at the load, Guzzella proposed in Guzzella [15] to write (2.12).

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi c11 þ c22  SoCðtÞ c11 þ c22  SoCðtÞ2  U BT ðtÞ ¼ 2 4  PBT ðtÞ  ðc4  SoCðtÞ þ c3 Þ in ðVÞ:

ð2:12Þ

The coefficients c11 and c22 depend only on the battery design and number of cells, but not on operative variables, they can be considered as constant in the time. Thus the battery model can be solved. Knowing the input PBT and the current state of charge SoC, the battery voltage U BT is obtained. Operating limits of the battery are defined bellow. Writing and solving (2.12) for P BT yields:

PBT ðtÞ ¼

U 2OC ðtÞ þ U BT ðtÞ  U OC ðtÞ Ri ðtÞ

in ðkWÞ

ð2:13Þ

The maximal power is derived as:

PBT;max ðtÞ ¼

U 2OC ðtÞ 4Ri ðtÞ

in ðkWÞ

ð2:14Þ

IBT;lim ðtÞ ¼

U OC 4Ri ðtÞ

U OC 2Ri ðtÞ

in ðVÞ

ð2:15Þ

in ðAÞ:

ð2:16Þ

Hr ¼ f ðV; SoCÞ

If the model detects that either (2.15) or (2.16) is reached, the simulation is stopped. Scaling of the battery size: In the model, the battery is sized by varying a variable representing the battery’s energy storage content. This variable is used to compute Q0 and find the operating limits of the battery. The specifications of the performances for simulations are defined below:







Maximum speed 140 km/h. Maximal time for acceleration from 30 km/h to 60 km/h – 7.5 s. Maximal vehicle mass: Case 1 – 1100 kg. Case 2 – 1800 kg. Initial charge of the battery – 80%, final charge of the battery 50%.

2.3. Hybridization strategy A power distribution strategy is put into place at the coupling of the electric motor and the combustion engine torques. This strategy defines the hybrid ratio, which represents the power contribution of the electric side of the powertrain.

Hr ¼

2.3.1. Case study 1 strategy In the first study case, a traditional discontinuous map with the modes is used. The Simulink control block contains 15 switches for positions, as a function of the 3 vehicle speed limits and 3 SoC limits, summarized in Table 2.1. During the multi-objective optimization, the optimizer will choose these limits, to obtain the lowest fuel consumption. The outcomes will be the optimal vehicle strategy, in function of the converter sizes and the storage tank sizes. 2.3.2. Case study 2 strategy In a second step, applied in the second case study, a continuous function, relating the hybridization ratio (the electric motor contribution), the state of charge of the battery and the speed demand is developed:

It is located on the corresponding voltage and current limits:

U BT;lim ðtÞ ¼

optimization strategies and most of them are heuristic based. Their guiding principles are that in the hybrid vehicle the engine should be used when its efficiency is relatively high and the battery charge and discharge should be regulated such that state of charge stays within predefined limits. The operating modes of the powertrain are so mapped, for example, the torque request is expressed as a function of the vehicle speed demand and the state of the charge of the battery. Thus, the strategy of the power distribution, between the thermal engine and the electric motor, depends on the SoC of the battery and the vehicle speed requirement.

PEM PEM;max ¼ PTOT PEM;max þ PICE;max

in ð%Þ

ð2:18Þ

where V in (m/s) is the vehicle speed and SoC (–) is the state of charge of the battery. Thus, the parameters relating the state variables of the electric motor usage can be optimized for minimal energy consumption and minimal size of the powertrain components. In the multiobjective optimization problem, these are the variables for control. The energy management strategy consists to use electric drive at low speeds and high state of charge (SoC) of the battery. The hybridization ratio decreases at high speeds and low SoC in order to save the battery from overload. A S-curve function suits to archive these requirements and is proposed in (2.19).

 Hr ðSoCÞ ¼ 3 þ

c1  SoC  ðSoC min þ SoC max Þ SoC min þ SoC max

3 ð2:19Þ

The behavior of such shape is illustrated in Fig. 2.6, where SoCmin ¼ 0:3 and SoCmax ¼ 0:8. During the simulation, if the SoC is high, the energy management computes a high hybridization ratio and drains the battery, so its SoC will tend to the central plateau of the S-curve (Fig. 2.6). On other hand, if the SoC is below the plateau (Fig. 2.6), the strategy computes a low ratio of hybridization in order to recharge the battery during the regenerative braking. The strategy has a stabilizing effect on the state of charge. The parameter c1 influences the height of the plateau. To extend the

ð2:17Þ

where PEM;max in (kW) is the maximum power of the electric machine and PICE;max in (kW) is the maximal power of the internal combustion engine. The degree of hybridization is defined in (2.17), as the ratio between the electric power and the total power. The benefits – i. e. the reduction of the energy consumption, but also the additional cost associated to the hybridization – increase with the degree of hybridization. To maximize energy benefit for associated hybridization investment, one needs an accurate energy management strategy. There are various possibilities to develop energy

Table 2.1 Powertrain discontinuous energy distribution strategy.

SoC P SoC high SoC P SoCmedium SoC P SoC medium SoC < SoC high SoC P SoC low SoC < SoC medium SoC < SoC low

v < v low

v P v low v < v medium

v  v medium v < v high

v  v high

3

7

11

15

2 1 1 1 0

6 5 5 5 4

10 9 9 9 8

14 3 3 3 12

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

7

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

Fig. 2.6. Hybridization ratio strategy: powertrain energy distribution strategy based on continuous function.

Fig. 2.7. Influence of the c2 parameter on the hybridization ratio: (a) after rotation with a large stretch factor c2 , and (b) after rotation with a small stretch factor.

strategy to the vehicle speed, the S-curve is rotated in the space around the upper vertical axe. The speed axis is normalized using the maximal speed V max of the driving cycle. The parameter c2 is related to the stretch of the rotation. Fig. 2.7 illustrates the relation.

where

CostPowertain ¼ Cost Electric

2.4. Economic model

with

2.4.1. Vehicle cost The cost of the vehicle is computed for each run, as a function of the energy converters and energy storage devices size and efficiency. Only the vehicle cost is optimized, the other costs are directly calculated from the vehicle cost. The vehicle cost is given by the powertrain cost and the body cost depending on power-train architecture configuration and size. The vehicle cost during the use phase is only the cost of energy consumption for the LCA functional unit – 150,000 km. In this study the maintenance cost is neglected.

CostElectric

Cost Vehicle ¼ Cost Powertain þ Cost Body

ð2:20Þ

motor

þ Cost Storage

CostThermal

CostStorage

motor

¼ 30

engine

system

þ CostThermal

ð€Þ  PEM kW

¼ 15

Engine

ð2:21Þ

system

ð€Þ  PICE kW

ðkWÞ

ð2:22Þ

ðkWÞ

ð2:23Þ

  € ¼ 600 h kW 0:2477log ðbatspecifmass ðbat typeÞþ0:5126Þ

 qbat

ðkWhÞ

ð2:24Þ

Considering that qbat is the battery capacity and bat_type is the battery type.

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

8

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

And

Cost Body ¼ 17:3  car shell mass ðkgÞ  3905:4 ð€Þ

ð2:25Þ

The cost of the powertrain (electric motor (2.22), thermal engine (2.23), battery (2.24)) comes from the literature [16] and is related to the size of the components. In this study, the high voltage battery is considered for the life of the car component and no replacement cost is assumed for a specific number of charge and discharge cycles. The vehicle body consists only in the car shell. The linear correlation (2.25) takes into account the prices of the parts, the manufacturing cost of the vehicle shell and includes the sales margin of the carmaker. For each calculation, a new vehicle mass is calculated, updated with the mass of the defined powertrain. 2.4.2. Customer investment cost The customer investment cost (in €) is defined below using (2.20). This is the price to buy the vehicle (2.26).

Cost Customer Inv estment ¼ CostVehicle þ Cost CO2 Emissions

ð2:26Þ

where the cost due to the CO2 emissions considering the Bonus/ Malus system is defined in Table B.1 2.4.3. Operating cost The operating cost considers the cost to drive the vehicle and to pay the liquid fuel and the electricity to charge the battery (2.27).

Cost Operating ¼ Cost Fossil

Fuel

þ CostElectricity

ð2:27Þ

where

Cost Fossil

Fuel

¼ CostFuel=L 

Fuel Consumption  150; 000 100

ð2:28Þ

With CostFuel in (€/L), Fuel Consumption in (L/100 km), and 150,000 km the life time of the vehicle. And

Cost Electricity ¼ Cost Electricity=kW 

Electricity Consumption  150; 000 100 ð2:29Þ

computing time. The functional unit for the LCA vehicle study is to transport passengers across on 150,000 km for 10 years [19]. The inventory in the production phase is composed from a hybrid electric Peugeot 3008Ò vehicle combined with unitary processes from the Eco InventÒ database. The use phase corresponds to the energy consumption of the vehicle. The inventory for the corresponding ‘‘energy carrier” production comes from the Eco InventÒ database (Fig. C.1). The maintenance and the end of life phases are represented by average technology car values from the Eco InventÒ database. 3. Results: Application on a hybrid electric vehicle 3.1. Problem definition – Optimization definition A hybrid vehicle with multiple propulsion systems can be operated independently or together. The model contents are the electric machine, battery, thermal engine and fuel tank, with two possible fuels – diesel and gasoline. The objective is to size the components of the hybrid powertrain – the converters and the storage tanks, regarding the fuel consumption and the cost objectives. At the same time, the environmental impact, GWP 100 y, from the Well-to-Wheel perspective and during the vehicle life is assessed for the solutions of the Pareto curve. The environmental model is executed to reach one run of the optimization and allows evaluating the environmental impacts of the proposed techno/economic optimal solutions. The decision variables are defined in Table 3.1 and in Table 3.2. The optimization problem is defined as minimization of the fuel consumption and the cost of the vehicle. 8 9 constraints > > < = minðFuel consumptionðxÞ; CostVehicle ðxÞÞ;s:t: dynamic v ehicle model; > > : ; cost model

ð3-1Þ

This study refers to the 4 main impacts categories, used from the automotive industry: GWP 100 years, acidification potential, eutrophication potential and ozone depletion (Table C.1). In the OSMOSE Superstructure, the CML_01_short method [18] is chosen because it has just 4 categories and this is allowing an optimal

where dynamic vehicle model is described in §2.2, cost model in §2.4 and x are the decision variables belonging to the defined range of the decision variables X decision variables in Table 3.2. The idea of the hybrid electric vehicle design is to operate the combustion engine only on the good efficiency range. According to the velocity demand and the SoC of the battery, the electrical powertrain is insuring the propulsion with relatively good efficiency (Fig. A.1). The electric machine in generator mode is producing electricity stored in the battery. The energy recovery function allows reducing the size of the battery. The thermal engine serves to extend the autonomy and to recover the power demand, as range extender mode. The REX block is designed for minimum 20 kW and for theoretical top speed of 140 km/h. The electric motor is adapted by a factor of 3.5 which allow reaching maximal torque of 250 Nm and maximal power output of 37 kW. The electric motor is linked to the wheel with a fixed gear ratio. The gear ratio of 3.5 allows a top speed of around 150 km/h. Thanks to its large maximum torque at small rotating speeds, at low speeds (around 60 km/h) the vehicle has a rather good acceleration potential. The batteries can deliver approximately 11 kW h and are charged to about 30% of energy, when the simulation starts.

Table 2.2 Fuel prices for France.

Table 3.1 Decision variables for design.

With CostElectricity in (€/kW), Fuel Consumption in (kW/100 km), and 150,000 km the life time of the vehicle.The fuel prices cited above are defined in Table 2.2. 2.4.4. Mobility cost The cost is presented from customer perspective and defines the total cost of mobility (2.30). The total mobility cost is the sum of the investment and the operating cost on the functional unit of the system – 150,000 km (2.30).

Cost Mobility ¼ CostCustomer Inv estment þ Cost Operating

ð2:30Þ

2.5. Environmental model

Description

Costs (€)

Hybrid electric propulsion system components:

Range

Unit

Vehicle use in France 2013 [17] Electricity household Gasoline Diesel

0.14269 (€TTC/kW h) 1.645 (€/L) 1.451 (€/L)

Electric machine Battery Li-Ion capacity Thermal engine Fuel

37 [6.5–26] 43 [Gasoline–Diesel]

kW kW h kW (–)

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

9

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx Table 3.2 Decision variables for control. Hybrid electric propulsion system conditions:

Range

Unit

SoC_high SoC_middle High-speed Middle speed Low speed

[85–95] [25–85] [25–33] [15–25] [0–15]

% % % % %

3.2. Case study 1 – Multi objective optimization results for a hybrid electric vehicle with different fuels 3.2.1. Optimization results The solutions of a two objectives optimization converged on a Pareto Frontier optimal curve (Fig. 3.1), representing the trade-off between the fuel consumption and the cost. With a vehicle body mass of 750 kg and as a function of the powertrain components design, one obtains solutions between 2.2 L/100 km and 0.6 L/100 km of fuel consumption. The customer investment cost, composed of powertrain cost and the CO2 taxes (bonus/malus) as per equation (2.26), varies between 7500 € and 12,000 €. The fuel consumption is influenced by the hybridization ratio, expressed through the battery size and the thermal engine efficiency (gasoline or diesel engine). The powertrain cost is strongly influenced by the battery capacity, with a proportional coefficient of 600 (€/kW h) as per equation (2.24). The algorithm converges on battery size solutions from 11 kW h to 26 kW h. For a Li-Ion battery, with an energy density of 90 Wh/ kg, the battery mass varies between 120 kg and 288 kg. The whole vehicle mass (body + powertrain) is between 950 kg and 1300 kg. The minimal fuel consumption, for the car shell mass of 750 kg, is 0.6 L/100 km. The solutions in Fig. 3.1 can be organized in two zones, depending on the hybridization ratio:
Full Hybrid Electric Vehicles (HEV), higher fuel consumption zone.
Range Extender (REX) – Plug-in Hybrid Electric Vehicles (PHEV), lower fuel consumption zone. In the HEV zone, for fuel consumption solutions of around 2 L/100 km, the advantage of the diesel efficiency is visible. For urban HEV vehicle of around 1000 kg with fuel consumption target of 2 L/100 km, a powertrain with small diesel engine of 30 kW could be interesting. Two cylinder engine of 800 cm3 can be an industrial solution for that. But taking into account the future emissions standards and also the diesel after treatment efforts for the automotive industry, the induced cost for such diesel

Fig. 3.2. Correlation between the fuel consumption and the battery energy.

engine presents a disadvantage in comparison to a small gasoline engine. So the recommended industrial solution, with large production volume for HEV, is hybrid electric powertrain with small, two cylinder gasoline engine of around 30 kW. In the PHEV and REX vehicle domain, for the same battery energy capacity, the advantage of the diesel engine from consumption point of view is decreasing (Fig. 3.2). The electric powertrain is insuring the mechanical power, and the thermal engine is operating in his good efficiency range. The engine is also used to support the accelerations phases of the cycle. One point in each zone is selected (ID89 and ID9) and the design details are in Table 3.3. Fig. 3.2 shows the relation between the battery energy capacity and the fuel consumption for both fuels. With the increasing battery capacity, the thermal part of the powertrain is less used. The efficiency of the powertrain increases and the fuel consumption decreases. In the HEV zone, for fuel consumption solutions around 2 L/100 km, the advantage of the diesel efficiency is visible, because of the most important operating part of the thermal engine. With the increase of the electric operating mode (battery capacity), the fuel consumption gap between the gasoline and diesel engine is decreasing. The minimal fuel consumption asymptote, for the body mass of 750 kg, is 0.6 L/100 km.

3.2.2. Comparison with an ICE commercial vehicle Fig. 3.3 displays the cost structure for each point – ID89 for 2 L/100 km HEV and ID9 for REX vehicle. A commercial urban vehicle – Peugeot 107Ò – with a small gasoline engine and a thermal powertrain is introduced for comparison (Table 3.5). The environmental bonus is applicable for all solutions. The vehicles are emitting less than 105 gCO2/km. With an increase of the hybridization ratio, the powertrain cost increases strongly – from around 10% of the cost of a conventional vehicle, to almost 50% for the REX vehicle. This tendency is due to the increasing high voltage (HV) battery cost (Table 3.4). Table 3.3 Multi objective optimization results – details of the Pareto points.

Fig. 3.1. Multi-objective Pareto curve – hybrid vehicle: fuel consumption to powertrain cost.

Characteristics

ID 89

ID 9

q_batt (kW h) Battery type Battery mass Fuel type Fuel consumption (L/100 km) Emissions CO2 (gCO2/km) CO2 emissions cost (bonus) Battery cost (€) Vehicle mass (kg)

15.5 Li-Ion 172 Gasoline 1.88 45.00 5000 9300 970

23.3 Li-Ion 259 Gasoline 0.83 20.00 7000 13,980 1060

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

10

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

Cost structure [€] FU 150,000 km Peugeot 107 thermal powertrain

ID 89 HEV

GWP [eq. kg CO2] elec consumpon fuel consumpon vehicle subsystems producon

ID 9 REX

25,000 18,000 16,000 14,000 12,000 10,000 8,000 6,000 4,000 2,000 -

20,000 15,000 10,000 5,000 0 -5,000 -10,000 total environmental cost [CO2 taxes]

total investment cost for powertrain

total investment cost for vehicle

total operang cost on FU

total cost of mobility

Fig. 3.3. Cost analysis of the selected Pareto curve solution points and real urban car.

Let’s consider that the customer is buying the entire car and the purchase price that he has to invest at the beginning varies from around 10,150 € for a conventional car, to 17,460 € for an REX vehicle, with a small gasoline engine (Fig. 3.3). One can consider that the operating cost, for a thermal vehicle and for the functional unit of 150,000 km, is equal to the investment cost, around 10,000 €. For the HEV, due to its low fuel consumption of less than 2 L/100 km, the operating cost is reduced by more than 50% to 4640 €. For the REX vehicle, the operating cost is even lower reaching only 3100 € and representing 30% of the conventional vehicle operating cost. The fuel consumption is responsible for 2/3 of the cost and the electricity consumption from the grid for 1/3 of the REX vehicle operating cost. The vehicle is considered being charged at home in France, with a cost of 0.14 kW h for the grid electricity. One has to notice that, in this study, the maintenance cost is neglected. The total cost of mobility with a small urban vehicle, after 150,000 km, is almost the same for all discussed solutions, around 20,000 €. The conventional vehicle with thermal propulsion still represents the highest mobility cost. The hybrid electric vehicle presents an advantage of 1000 €, in comparison to the conventional vehicle. The REX vehicle has the intermediate mobility cost – 20,558 €. The REX vehicle solution presents the advantage of having extremely low CO2 emissions- only 20 gCO2/km, especially in its use phase. These powertrains are technological solutions for the European automotive industry to achieve the strict CO2 emissions regulations. The cost structure depends on the vehicle class, customer usage and place of use. The environmental analysis takes into account this sensitivity. Fig. 3.4 shows the Well-to-Wheel global CO2 impact, based on the category GWP 100 y, for the REX vehicle. The production phase of the vehicle and the use phase are contributing to the major GWP impact. The electricity consumption contributes to the GWP of 7387 equivalent kg CO2 in Germany and is 7 times higher than Table 3.4 Orders of magnitude for powertrain configuration and cost for urban car. Cost structure

Peugeot 107Ò thermal

ID 89 HEV

ID 9 REX

Body cost (€) High voltage battery cost (€) Electric machine cost (€) Thermal engine cost (€) Powertrain cost (€) Vehicle mass (kg) Thermal engine power (kW)

8700 0 0 750 1250 790 50

9070 9030 1110 645 11,055 970 43

9069 13,980 1110 300 15,390 1060 20

France

Germany

Fig. 3.4. Environmental analysis for a functional unit 150,000 km and from the life cycle perspective, based on GWP 100 years indicator, point 9- REX vehicle.

Table 3.5 Orders of magnitude for powertrain configuration and fuel consumption for urban car. Cost structure

Peugeot 107Ò thermal

ID 89 HEV

ID 9 REX

El machine power (kW) Battery capacity (kW h) Electric consumption (kW h) Fuel type Fuel consumption (L/100 km) CO2 emissions (g/km)

– 0 0 Gasoline 4.3 99

37 15.5 0 Gasoline 1.88 45

37 23.3 5 Gasoline 0.83 20

the one in France. This is due to the different electricity production mixes between France and Germany. The French electricity mix is 77% based on nuclear, while for the German one, 45% comes from coal (Fig. C.1). The environmental impact of the use phase of the REX vehicle strongly depends of the electricity mix of the country where the vehicle is driven. 3.3. Case study 2 – Results for a hybrid electric vehicle and continuously basic energy distribution function – NEDC In this optimization case, the environmental model is deactivated. A D-class vehicle with a nominal mass of 1600 kg (see Table 3.6) is studied. This type of vehicle has typically an extraurban use. This case illustrates the consideration of the parameters of the proposed continuous S-function (§2.3.2) for the basic energy distribution function in the optimization structure. A hybrid vehicle with multiple propulsion systems can be operated independently or together. The model contents are the electric machine, battery, thermal engine and fuel tank, with diesel fuel. The thermal electric hybrid powertrain model characteristics are given in Table 3.6. The optimization problem is defined as:

Table 3.6 D-class vehicle characteristics. Sub-system

Characteristic

Value

Vehicle Gear box Engine

Nominal mass (kg) CVT efficiency (–), [20] Displacement (L) Number of cylinder Deceleration Fuel cut-off Type Density (kg/L) Lower heating value (MJ/kg) Power (kW) 17.5 kW/unit Li-Ion Capacity (kW h)

1660 0.84 1.4 4 Yes Diesel 0.84 42.5 130 2

Fuel

Electric motor Supercapacitors Battery

7

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

11

Table 3.7 Decision variables for energy distribution function. Decision variable for control

Range

SoC strategy parameter c1 (–) SoC strategy parameter c2 (–)

[1.8–2.4] [0.1–1.0]

8 9 constraints > > < =   min gpowertrain ðxÞ; CostVehicle ðxÞ ; s:t: dynamic v ehicle model; > > : ; cost model ð3-2Þ where gpowertrain in (%) is the powertrain efficiency and Costvehicle is the vehicle cost in (€), and x are the decision variables belonging to the defined range of the decision variables X decision variables in Table 3.7. The decision variables for the energy distribution function, the S-Curve coefficients, are defined in Table 3.7: The strategy parameters c1 and c2 are optimized for every calculation iteration of the decisions variables. The decision variables for design are the same as in the previous optimization problem, related to the size of the powertrain components. The simultaneous optimization of the coefficients c1 and c2 gives the optimum energy strategy distribution and results in the optimum size of the equipment. After each iteration of the model, the mean powertrain efficiency in traction is calculated as:



gpowertrain ¼ mean

Pwheel Pfuel þ Pbattery þ PSC

 ð3-3Þ

where Pbattery and PSC are respectively the battery and the super capacitors powers and Pfuel is the power stored in the fuel, Pwheel is the power transmitted on the wheels. All powers are defined in (kW). The vehicle cost is recomputed for each iteration of the decisions variables. The vehicle cost is defined in Eq. (2.20). The degree of hybridization is defined in (2.17), as the ratio between the electric power and the total power. A continuous function, relating the hybridization ratio (the electric motor contribution), the state of charge of the battery and the speed demand is developed (2.18). Thus, the parameters relating the state variables of the electric motor usage can be optimized for minimal energy consumption

Fig. 3.6. Hybridization strategy on NEDC for D class Vehicle.

and minimal size of the powertrain components. In the multiobjective optimization problem, these are the variables for control. The basic energy management strategy consists to use electric drive at low speeds and high state of charge (SoC) of the battery. The hybridization ratio decreases at high speeds and low SoC in order to save the battery from overload. A S-curve function from (2.19) suits to archive these requirements. During simulation, if the SoC is high, the control computes a high hybridization ratio and drains the battery, so its SoC will tend to the central plateau of the S-curve. On other hand, if SoC is below the plateau, the strategy computes a low ratio of hybridization in order to recharge the battery during the regenerative braking. The strategy has a stabilizing effect on the state of charge. The parameter c1 influences the height of the plateau. To extend the strategy to the vehicle speed, the S-curve is rotated in the space, around the upper vertical axe; this rotation is representing the c2 coefficient. The speed axe is normalized using the maximal speed Vmax of the driving cycle. The parameter c2 is related to the stretch of the rotation. The optimal solution is obtained for c1 = 2 and c2 = 0.5. From the hybrid distribution strategy in Fig. 3.5, it is visible that the stability hybridization ratio is 22%- this is where the plateau is situated. The powertrain is strongly electrified (Fig. 3.6) in the urban part of the NED Cycle. The thermal engine is supporting the accelerations in the urban zone and especially the three last accelerations in the extra urban zone (Fig. 3.6).

Fig. 3.5. Electric power distribution for D Class vehicle.

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

12

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

Table 3.8 CO2 emission and cost results – ID 1. Parameters

Optimal design point (NEDC)

Emissions (gCO2/km) Powertrain efficiency (%)

83 32

SoC strategy parameter c1 (–) SoC strategy parameter c2 (–)

2.3 0.5

ICE cost (€) Electric motor cost (€) Battery energy cost (€) Super capacitors cost (€) Nominal car shell cost (€)

825 3900 4156 700 31,167

Total vehicle cost (€)

40,748

Table 4.1 CO2 emission comparison between calculated vehicle definition and similar market vehicles. Range extenders (urban class)

ID 9 REX

B-class REX (BMW) [21]

Engine power (kW) Fuel consumption (L/100 km) CO2 emission (g/km)

20 0.8 20

28 0.6 13

D-class hybrid electric vehicles

ID 1 Plug in HEV

D-class serial HEV [The central] [22]

Engine power (kW) Fuel consumption (L/100 km) CO2 emission (g/km)

50 3.0 83

120 3.6 95

The D class vehicle solution reaches 32% powertrain efficiency and consumes 3.1 L/100 km. The total vehicle cost is around 41,000 euros. The optimal results are summarized in Table 3.8. For NEDC, the optimal hybridization ratio is 22%. This ratio is a balance between the vehicle mass and the efficiency of the conversion chain. One can see in Fig. 3.6 that with the selected strategy, the urban part (between 0 and 850 s) of the NEDC is in the major time supplied by the electric part of the powertrain (except for the pick power demands of 20 kW). The power demands till 10 kW are fully covered in electric mode. In the extra-urban part of NEDC (800–1200 s), the thermal powertrain is activated for power demands up 15 kW and contribute with 50% to the propulsion. During the two last accelerations of the cycle, the ICE participates with respectively 70% and even 100% for the last one. This is also related to the state of charge of the battery – one can notice that it is diminishing in the extra urban part of the cycle – from 800 s. The battery is recharged during the strong deceleration in the end of the cycle. This case study illustrates the possibility to optimize the parameters of a continuous function for energy distribution between the components of the vehicle powertrain. The proposed S-function suits well the to the hybrid electric vehicles behaviors during a given drive cycle profile. Its advantage is that it is a very simple model for a basic energy management strategy for simulation studies. The function is not predictive – the drive (time/speed) profile has to be known before. The function is limited to the applied vehicle modelling approach.

4. Conclusion This paper presents an integrated methodology for vehicles design considering, on a holistic way, the parameters for design and macro level operating strategy. The cost-efficiency trade-off is assessed and the integration of the life cycle inventory in the techno–economic optimization structure allows estimating the environmental impacts of the optimal points. The methodology is

applied on low CO2 emissions vehicles and performed for two case studies. In the first case study, an optimization of a hybrid electric vehicle, with around 1000 kg of mass, for urban mobility is studied. The demonstration of the simultaneous configuration of the powertrain, with sizing of the components, and the assessment of the economic and the environmental impacts is done. The obtained optimum solutions are for an urban car range between 2.2 L/100 km and 0.6 L/100 km of fuel consumption. All these solutions are, with low tank-to-wheel CO2 emission – less than 50 gCO2/km. In the case of purchase in France, these vehicles benefit from an economic bonus, given by the government. The customer investment price for the powertrain is between 7500 € and 12,000 €. The use phase of the PHEV and REX is sensitive to the place of use. The environmental impact GWP 100 years, takes into account the well-to-wheel CO2 emission and shows the environmental impact of the different production electricity mixes. Considering the customer perspective, the total mobility cost for all the solutions (conventional ICE, HEV, PHEV, REX) for the life time of 150,000 km is really close (between 20,000 and 21,000 €). On the price point of view, there is, in the present time, no real advantage in the acquisition of a hybrid vehicle. However, the GWP impact is much higher with the conventional ICE vehicle. The pen-

Fig. A.1. Electric machine operating map, point ID 9.

Table B.1 CO2 emissions cost, environmental cost, supported by the French government for 2013 [17]. Emissions (CO2(g)/km)

(+)Bonus/() Malus (€)

6 20 > 20 and 6 50 > 50 and 6 60 > 60 and 6 90 > 90 and 6 105 > 105 and 6 135 > 135 and 6 140 > 140 and 6 145 > 145 and 6 150 > 150 and 6 155 > 155 and 6 175 > 175 and 6 180 > 180 and 6 185 > 185 and 6 190 > 190 and 6 200 > 200 and 6 230 > 230

+7000 (€) +5000 (€) +4500 (€) +550 (€) +200 (€) 0 (€) 100 (€) 300 (€) 400 (€) 1000 (€) 1500 (€) 2000 (€) 2600 (€) 3000 (€) 5000 (€) 6000 (€) 6000 (€)

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

13

Fig. C.1. Electricity mix production (a) France, (b) Germany (Eco InventÒ, 2013).

ID 9 is compared with serial B-class range extender vehicle. The REX solution from the optimal Pareto curve ID9 has very close characteristics in terms of equipment size and performances compared to the commercial REX vehicle. The techno–economic optimization delivers valid solutions and shows the advantage to present in details the techno–economic and environmental performances of the vehicle. In the domain of D-class vehicles with hybrid electric powertrain, the ID 1 which is a plug-in hybrid is compared with serial D-class hybrid electric vehicle non plug-in Hybrid. Due to the difference in the hybridization ratio, the ICE engine in the serial vehicle is much bigger and supports the major part of the driving cycle. Thus the obtained fuel consumption is higher by 0.6 L/100 km. The heuristic energy management distribution based on modes and described by the state of the art is replaced by a simple continuous energy distribution function. Its optimal set up is defined in the optimization approach for vehicle design (case study 2). The methodology described in this paper presents an interest for the systematic analysis and the understanding of the techno–economic trade-off for hybrid electric vehicles. The continuous energy distribution function can be adapted for every hybrid vehicle design, without having existing vehicle. This allows analyzing and comparing different designs for hybrid vehicles. In a future work, the environmental impact can be introduced as an optimization objective to research optimal environomic designs. Acknowledgments

Fig. C.2. Environmental analysis on functional unit 150,000 km and life cycle perspective, based on CML_01_short method [18], point 9- REX vehicle.

I would like to thank EPFL for the support of the research. I would like to thank Nicolas Amalric for the technical support. Appendix A

etration of the hybrid vehicles on the market depends on the orientation taken by the policy makers in the transportation sector. The second case study researches the optimal design of a 3 L/100 km hybrid electric vehicle for D segment (1600 kg), using continuous energy management function. Thus the parameters for design and operating strategy are considered simultaneously to define the optimal powertrain equipment size. Table 4.1 summarizes the low CO2 emission vehicle solutions and compares the obtained emissions results and fuel consumption with vehicles from the market with similar definition. The

See Fig. A.1. Appendix B See Table B.1. Appendix C See Figs. C.1 and C.2 and Table C.1.

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071

14

Z. Dimitrova, F. Maréchal / Applied Energy xxx (2015) xxx–xxx

Table C.1 Listing of impact categories implemented in cml01_short [18]. Category

Definition

Unit

Acidification potential, average European Climate change, GWP 100a

H2CO3 formation in oceans and on land by dissolving CO2 in water Significant and lasting change in the statistical distribution of weather patterns over an extended period of time (100 years) Environmental response to addition of nitrates and phosphates. (CML01) – Mass of algae, phytoplankton – Biodiversity desirable fish species Ozone depletion

kg SO2-eq kg CO2-eq

Eutrophication potential, generic

Stratospheric ozone depletion, ODP steady state

References [1] ExxonMobil. Report outlook of energy – a view to 2040. ; 2012 [accessed 13.02.13]. [2] ERTRAC. ERTRAC report 2011, ERTRAC research and innovation roadmaps. Brussels, Belgium: ERTRAC; 2011. [3] Thiel C, Schmidt J, Van Zyl A, Schmid E. Cost and well-to-wheel implications of the vehicle fleet CO2 emission regulation in the European Union. Transport. Res. Part A: Policy Practice 2014;63:25–42. [4] Bishop JDK, Martin NPD, Boies AM. Cost-effectiveness of alternative powertrains for reduced energy use and CO2 emissions in passenger vehicles. Appl Energy 2014;124:44–61. [5] Ribau JP, Silva CM, Sousa JMC. Efficiency, cost and life cycle CO2 optimization of fuel cell hybrid and plug-in hybrid urban buses. Appl Energy 2014;129 (15):320–35. September. [6] Caton JA. The thermodynamic characteristics of high efficiency, internalcombustion engines. Energy Convers Manage 2012;58:84–93. June. [7] Finesso R, Spessa E, Venditti M. Layout design and energetic analysis of a complex diesel parallel hybrid electric vehicle. Appl Energy 2014;134:573–88. [8] Millo F, Rolando L, Fuso R, Mallamo F. Real CO2 emissions benefits and end user’s operating costs of a plug-in Hybrid Electric Vehicle. Appl Energy 2014;114:563–71. [9] Gerber L, Maréchal F. Environomic optimal configurations of geothermal energy conversion systems: application to the future construction of Enhanced Geothermal Systems in Switzerland. Energy 2012;45:908–23. [10] Dimitrova Z, Maréchal F. Environomic design of vehicle integrated energy systems-application on a hybrid electric vehicle energy system. CET 2014;39:475–80. http://dx.doi.org/10.3303/CET1439080. [11] Dimitrova Z, Maréchal F. Environomic design of vehicle energy systems for optimal mobility service. Energy 2014. http://dx.doi.org/10.1016/j.energy. 2014.09.019.

kg PO4-eq

kg CFC-11-eq

[12] Weber C. Multi-objective design and optimization of district energy systems including polygeneration energy conversion technologies. Ph.D. thesis. Ecole Polytechnique Fédérale de Lausanne. Lausanne, Switzerland; 2008. [13] Fazlollahi S, Maréchal F. Multi-objective, multi-period optimization of biomass conversion technologies using evolutionary algorithms and mixed integer linear programming (MILP). Appl Therm Eng 2013;50:1504–13. [14] Guzzella Lino, Sciarretta Antonio. Vehicle propulsion system, 3rd ed. Berlin, Germany; 2013. [15] Guzzella L. QSS toolbox. ; 2013 [accessed 01.07.13]. [16] Guzzella L, Amstutz L. The QSS toolbox, technical report. Zurich, Switzerland: IMRT; 2005. [17] French Government. Sustainable development statistics. ; 2013 [accessed 03.06.13] [in French]. [18] Gerber L, Gassner M, Maréchal F. Systematic integration of LCA in process systems design: application to combined fuel and electricity production from lignocellulosic biomass. Comput Chem Eng 2011;35:1265–80. [19] Richet S, Tonnelier P, Martin S. Project ENVironment – Life Cycle Assessment (LCA) – approach and application at PSA Peugeot Citroën, French, unpublished internal report, Vélizy, France; 2012. [20] Nissan. CVT Nissan technology. ; 2014 [accessed 20.10.14]. [21] BMW. The clean automobile. ; 2015 [accessed 06.04.15] [in French]. [22] The Central. Technical data of the Peugeot 508 ; 2014 [accessed 17.09.14] [in French].

Please cite this article in press as: Dimitrova Z, Maréchal F. Techno–economic design of hybrid electric vehicles and possibilities of the multi-objective optimization structure. Appl Energy (2015), http://dx.doi.org/10.1016/j.apenergy.2015.09.071