,
STOCKHOLM SWEDEN 2020
Tank-to-Wheel Energy
Breakdown Analysis
XU YU
Xu Yu <yuxu@kth.se> Vehicle Engineering
KTH Royal Institute of Technology
Place for Project
Gothenburg, Sweden
Examiner
Prof. Mikael Nybacka
KTH Royal Institute of Technology
Supervisor
Prof. Mikael Nybacka
KTH Royal Institute of Technology
Dr. Simon Klacar
In early design phase for new hybrid electric vehicle (HEV) powertrains, simulation is used for the estimation of vehicle fuel consumption. For hybrid electric powertrains, fuel consumption is highly related to powertrain efficiency. While powertrain efficiency of hybrid electric powertrain is not a linear product of efficiencies of components, it has to be analysed as a sequence of energy conversions including component losses and energy interaction among components.
This thesis is aimed at studying the energy losses and flows and present them in the form of Sankey diagram, later, an adaptive energy management system is developed based on current rule-based control strategy. The first part involves developing energy calculation block in GT-SUITE corresponding to the vehicle model, calculating all the energy losses and flows and presenting them in Sankey diagram. The second part involves optimizing energy management system control parameters according to different representative driving cycles. The third part involves developing adaptive energy management system by deploying optimal control parameter based on driving pattern recognition with the help of SVM (support vector machine).
In conclusion, a sturctured way to generate the Sankey diagram has been successfully generated and it turns out to be an effective tool to study HEV powertrain efficiency and fuel economy. In addition, the combination of driving pattern recognition and optimized control parameters also show a significant potential improvement in fuel consumption.
Keywords
Under den tidiga utvecklingsfasen av nya elektrifierade drivlinor for hybridapplikationer (HEV) används simulering för uppskattning av fordonets bränsleförbrukning. För dess drivlinor är bränsleförbrukningen i hög grad kopplad till drivlinans verkningsgrad. Även om drivlinans verkningsgrad inte är en linjär prokukt av komponenternas verkningsgrad behöve rden analyseras som en sekvens av energiomvandlingar, inklusive förluster och energipåverkan mellan komponenter.
Detta examensarbete syftar till att undersöka energiförluster och flöden samt presentera dessa i form av sankey diagram. Senare utvecklas ett anpassningsbart energihanteringssystem baserat på nuvarande regelbaserad kontrollstrategi. Den inledande delen involverar utvecklandet av energianalys i GT-SUITE som motsvarar fordonsmodellen, beräkningar av totala energiförluster och flöden samt presentation av dessa i ett sankey diagram. Den andra delen innefattar optimering av energihanteringssystems kontrollparametrar enligt olika representativa körcykler. Den tredje delen involverar utveckling av anpassningsbara energihanteringssystem genom användning av optimala kontrollparameterar baserad på detektering av körbeteende med hjälp av SVM ( stödvektormaskin).
The master thesis was carried out at Powertrain Engineering Department of China Euro Vehicle Technology.
First of all, I would like to express my sincere appreciation to my supervisor Simon Klacar and Enrico Fichera from CEVT, for for the valuable technical support and patient instructions during the whole thesis project. I would also like to thank Mikael Nybacka, my academic supervisor for his countless suggestions and pratical help, without which it is quite tough for me to complete the thesis.
Secondly, I want to thank Håkan Sandquist, my team manager at CEVT, for his continuous encouragement and unreservedly life guidance. I also want to thank all the teammates in powertrain strategy team, for their technical assistance.
Thirdly, I would like to thank all the restaurants in Lindholmen, for offering so much good food, which these six-month experience more tasty.
CEVT China Euro Vechile Technology AB
CMA Compact Modular Architecture
ICE Internal Combustion Engine
BEV Battery Electric Vehicle
HEV Hybrid Electric Vehicle
REEV Range Extended Electric Vehicle PSH Power Split Hybrid
SI Spark Ignition
CI Compression Ignition
TDC Top Dead Center
BDC Bottom Dead Center
BLDC Brushless DC Motor
PMSM Permanent Magnet Synchronous Motor EMF Electromotive Force
AC Aternative Current
DC Direct Current
IGBT Insulated-gate Bipolar Transistor
MT Manual Transmission
CAE Computer Aided Engineering
GTI Gamma Technologies
OEM Original Equipment Manufacturer
GUI Graphical User Interface
PID Proportional Integral Derivative
BMEP Brake Mean Effective Pressure FMEP Friction Mean Effective Pressure BSFC Brake Specific Fuel Consumption
EMS Energy Management System
DHT Dedicated Hybrid Transmission
SOC State of Charge
WLTC Worldwide harmonized Light duty Test Cycle DOE Design of Experiment
SVM Support Vector Machine
PCA Principal Component Analysis
1 Introduction
1
1.1 Company presentation . . . 1 1.2 Background . . . 3 1.3 Goal . . . 4 1.4 Procedure. . . 4 1.5 Limitation . . . 42 Theory
5
2.1 Hybrid Vehicle Powertrain Topology . . . 52.2 Powertrain Components and Energy Losses . . . 9
2.2.1 Internal Combustion Engine . . . 9
2.2.2 Electric Motor . . . 13 2.2.3 Inverter . . . 18 2.2.4 Clutch . . . 20 2.2.5 Transmission . . . 20 2.2.6 Battery . . . 23 2.3 Powertrain Simulation . . . 25
3 Vehicle Model
29
3.1 GT-SUITE. . . 293.2 Vehicle Modelling in GT-SUITE . . . 30
3.2.7 Transmission . . . 36
4 Energy Breakdown Analysis
37
4.1 Energy Calculation . . . 374.2 Sankey Diagram . . . 46
5 Energy Management System
49
5.1 Control Strategy . . . 495.2 Control Parameter Optimization . . . 53
6 Adaptive Energy Management System Based on Driving
Pattern Recognition
57
6.1 Representative Driving Cycles Control Parameters Optimization . . . . 586.2 Driving Pattern Recognition . . . 59
6.2.1 Support Vector Machine . . . 59
6.2.2 Driving Pattern Feature Extraction . . . 63
6.2.3 SVM Model Training . . . 66
6.3 Implementation of Adaptive Energy Management System . . . 66
6.4 Simulation Result . . . 67
6.5 Driving Pattern Recognition Based on Preknown Routine . . . 67
6.6 Simulation Result . . . 69
7 Powertrain Efficiency Analysis with Sankey Diagram
72
8 Conclusions
75
8.1 Summary . . . 75Introduction
As shown in Figure 1.0.1, stricter regulations on fuel consumption and emission have been imposed or proposed worldwide. In this context, developing a novel technology for lower fuel consumption and emission for vehicles is becoming an urgent task in automotive industry.
Therefore, hybrid electric vehicles came into being in response to the requirements of the era by combining the advantages of conventional vehicles with pure electric vehicles in energy saving and emission reduction. Because of their significant advantages, hybrid electric vehicles are highly valued and gradually scaled up.
It can be seen from Figure 1.0.2 that hybrid electric vehicle can save up around 60% of fuel and reduce about 40% of CO2emissions.
1.1
Company presentation
China Euro Vehicle Technology AB, hereby referred to as CEVT, is an innovation centre for the Zhejiang Geely Holding Group. CEVT is fully owned by the Zhejiang Geely Holding Group. The group encompasses a range of renowned global car brands with different characteristics and audiences, shown in Figure 1.1.1. The consumer car brands include Geely Auto, Lynk & Co, Volvo Cars, Polestar, Proton and Lotus. Commercial vehicle brands include London Electric Vehicle Company (previously known as London Taxi) and YuanCheng Auto.
Figure 1.0.1: Vehicle emission and fuel consumption regulation in different market [1]
(a) Fuel economy and fuel economy change for HEV in comparison to conventional vehicle
(b) CO2emissions and CO2emissions change for HEV in comparison to conventional vehicle
or all-wheel drive and combustion engine or hybrid propulsion. The platform was named Compact Modular Architecture (CMA) and debuted 2017 in the Volvo XC40 and the Lynk & Co vehicles. No production is conducted at CEVT since they are an innovation centre. CEVT’s main objective is to develop automotive technology for future demands.
Figure 1.1.1: Zhejiang Geely Holding Group structure
1.2
Background
1.3
Goal
The aim of this thesis is to develop a solid understanding of component and system losses, being able to quantify them with a correct methodology and present them in Sankey diagram. The ultimate goal is to be able to optimize current energy management system in order to enhance system efficiency using Sankey diagram as an analysis tool.
1.4
Procedure
The whole thesis project can be split into four substeps shown below:
1) Literature review on powertrain simulation, components and system losses. 2) Analyse complete vehicle model and develop energy calculation model in
GT-SUITE.
3) Extract simulation data and establish Sankey diagram in e!Sankey.
4) Optimize and modify current energy management system control strategy. 5) Using Sankey diagram to analyse different energy management system control
strategy.
1.5
Limitation
Theory
2.1
Hybrid Vehicle Powertrain Topology
In today’s world, the shortage of fossil fuel and global warming are two main challenging issues that automotive industry is facing. To reduce the fuel consumption and emission of traditional vehicle with only internal combustion engine (ICE), battery electric vehicle (BEV) and hybrid electric vehicle (HEV) have been come up with. So far, BEV and HEV have been demonstrated as effective solutions for fuel saving and emission reduction. However, in case of BEV, the driving range is usually limited because the insufficient energy density of current battery technology, meaning that BEV cannot meet the requirement of long-distance travelling. Also, due to the limitation of charging speed and number of charging infrastructure, HEV is still under intense development and manufacturing and research is conducted on comparison between HEV and BEV [3] [4].
A HEV powertrain includes both an ICE and at least one electric machine to drive the vehicle. The combinations of electric machines, gearboxes, and ICE are quite flexible, resulting in a large number of possible hybrid powertrain toypologies. One way to classify different hybrid powertrain topologies is based on powertrain configuration: series hybrid, parallel hybrid and complex hybrid.
electric vehicle (REEV). Since the ICE is decoupled completely from wheels, the efficiency which is related to speed and torque, can be chosen freely in order to achieve the optimal working point. Figure 2.1.1 shows a typical series hybrid powertrain configuration.
Figure 2.1.1: Series hybrid powertrain configuration[5]
The chemical energy in the fuel is converted into mechanical energy by ICE, which is converted into electric energy by generator later on. The generator charges the battery and battery supplies the electric energy to traction motor. In this way, a lot of energy conversions occur which has a negative effect on the overall powertrain efficiency.
The series hybrid is more suitable for buses and city driving because of its high performance of stop-and-go. While it is not so applicable when it comes to highway driving because of higher conversion losses and larger electric machine at high speed [6].
2) Parallel Hybrid: Unlike the series hybrid powertrain, the ICE is connected to wheels mechanically through a gearbox. Moreover, the electric motor is also connected to the wheels. The torque supplied by electric motor and ICE are combined by means of a torque coupling device and then the torque is transferred to wheels (shown in Figure 2.1.2).
Similar to series hybrid powertrain. the working point of the ICE in parallel powertrain can be chosen freely, i.e. the speed of the ICE is chosen with the gearbox and the torque with the electric motor. The parallel hybrid powertrain can operate on three modes: pure electric mode, pure ICE mode and a mixed mode when both ICE and electric motor work together to meet high power demand and improve the ICE efficiency.
Figure 2.1.2: Parallel hybrid powertrain configuration[5]
conversion losses because of the mechanical connection between ICE and wheels. However, this is achieved at the expense of a larger ICE, whereas smaller and less powerful electric motor is used. Parallel hybrid powertrain is not as suitable for frequent stop-and-go driving condition as series hybrid powertrain is.
3) Complex Hybrid: Complex hybrid (shown in Figure 2.1.3) is also known as power split hybrid (PSH) because one of the key components in this kind of powertrain is the power split device, which is usually a planetary gear. Complex hybrid powertrain is a mix between series and parallel hybrid powertrain.
Figure 2.1.3: Complex hybrid powertrain configuration[5]
Complex hybrid powertrain takes both the advantages of series and parallel hybrid powertrain. When it comes to frequent urban stop-and-go condition, the vehicle can switch to series mode while in high speed condition and it can be easily switched back to parallel mode. Moreover, the size of both electric motor and ICE can be reduced. However, complex hybrid powertrain requires more complicated control and structure that might increase the cost.
Table 2.1.1 summarizes the pros and cons of three different configurations.
Table 2.1.1: Comparison of hybrid powertrain configurations
Powertrain Advantage Disadvantage
Series Hybrid
·Smaller ICE
·Packaging space saving·Excellent transient response ·Fuel consumption and emission reduction
·Larger electric machine ·Multiple energy conveersions Parallel Hybrid ·Further fuel consumption and emission reduction
·Smaller ICE and electric machine
·Higher cost
·Demanding packaging space Complex Hybrid ·Flexibility on driving modes
·Further fuel consumption and emission reduction
· Higher cost · Control complexity
· Demanding packaging space
Table 2.1.2: Different types of hybridization vehicle Type of hybrid DOH
Micro <5% Mild 5% to 10% Full hybrid 10% to 75% Pure electric vehicle 100%
degree of hybridization (DOH).
Degree of hybridization is related to the power supplied by ICE (PICE) and electric
motor (PEM). The ratio between installed electric power and total installed power
(power of ICE and electric motor) is known as degree of hybridization shown in Equation 2.1.
DOH = PEM PEM + PICE
· 100% (2.1)
According to the degree of hybridization, the hybrid vehicle can be defined as Table 2.1.2.
1) Micro Hybrid: In a micro hybrid powertrain, the electric machine usually does not provide additional torque to the vehicle, nevertheless, it is typically used for the stop-and-go functionality of the vehicle and to power the vehicle accessories. The electric machine is relatively small, usually up to 2.5 kW, and works with 12 V [7].
[8].
3) Full Hybrid: Full hybrid powertrain has full electric drive capability where electric motor provides at least 40% of the ICE power as additional torque. Full hybrid powertrain also contains main features of mild hybrid powertrain, as the electric motor can be used to power the vehicle and the accessories, but at the expense of bigger battery and electric machine.
2.2
Powertrain Components and Energy Losses
A hybrid vehicle powertrain combines conventional powertrain components, i.e.,ICE and transmission, with electric components, electric motor, power electronics and battery. To do the energy break down analysis of a hybrid powertrain, it is essential to understand the energy losses from single component’s perspective.
2.2.1
Internal Combustion Engine
The internal combustion engine (ICE) is a heat engine which converts chemical energy of fuel into mechanical energy.
There are two main types of ICEs, spark ignition(SI) engines, where the ignition of fuel is done by spark and compression ignition (CI) engines where highly compressed air rises the temperature igniting the fuel injected. Both SI and CI engine can be designed to run in either four strokes or two strokes of the piston. In this thesis work, the four-stroke SI engine will be investigated.
In a typical SI engine, the piston moves up and down in a cylinder and transmits power through a rod and crank mechanism which convert the reciprocating motion of the piston to the rotary motion of crankshaft as shown in Figure 2.2.1. When the piston is at the top dead center (TDC) position and bottom dead center (BDC) position, the cylinder volume is minimum and maximum respectively. The volume swept out by the piston, the difference between the maximum and the minimum volume, is called the displacement or swept volume Vd. The ratio of maximum geometric volume to
minimum volume is the geometric compression ratio rc.
Figure 2.2.1: Motion inside engine cylinder [9]
Figure 2.2.2: Four-stroke cycle[10]
1) Intake Stroke: The piston starts at TDC and stops at BDC, meanwhile the inlet valve opens, which intakes fresh air-fuel mixture into the cylinder.
2) Compression Stroke: Both valves close and the air-fuel mixture inside the cylinder is compressed. At the end of the compression stroke, combustion is initiated.
3) Combustion Stroke (or Expansion Stroke): The gases with high temperature and pressure push the piston from TDC down to BDC and force the crank to rotate.
4) Exhaust Stroke: The burned gases exit the cylinder, the outlet valve opens, since cylinder pressure is higher than the exhaust pressure along with the piston moving towards TDC, the burned exhaust gases are swept out.
in a typical internal combustion engine are categorized three main terms: Exhaust enthalpy loss, coolant loss and friction & pumping loss and they are visualized in Figure 2.2.3.
Figure 2.2.3: Energy losses flow in ICE [11]
While in [12], the losses are divided as four different components, which are friction loss, heat loss to surroundings, exhaust enthalpy loss and coolant loss shown in Figure 2.2.4.
However, the sum of the energy goes over 100% since the friction loss is double counted. Because the friction loss is mainly resulted from mechanical contact hence producing a large fraction of heat which is carried away by coolant.
Therefore, the main concerning losses of engine can be split into exhaust enthalpy loss, coolant loss, friction & pumping loss and heat loss to surroundings.
1) Exhaust Enthalpy Losses: During the exhaust stroke, the exhaust gases leave the engine through exhaust system. Usually the temperature of exhaust gases are very high temperatures. Consequently, these gases with high heat content are carried away as, meaning a considerable amount of energy is wasted to the ambient with the exhaust gas. In general, approximately 30% of the fuel energy is lost through the engine exhaust [13].
Figure 2.2.4: Energy losses fraction in ICE [12] Equation 2.2
˙
QL= ( ˙ma+ ˙mf)× cP × (Ti− Tamb) (2.2)
Where ˙QLis energy lost in exhaust gas (kJ/h), ˙mais air consumption (kg/h), ˙mfis
fuel consumption (kg/h), cP is specific heat of exhaust gas (kJ/kgK), Tiis exhaust
gas to calorimeter inlet temperature (◦C) and Tambis ambient temperature (◦C).
2) Coolant Losses: Engine combustion increases the temperature of the gases. Consequently the heat inside the cylinders are transferred through the cylinder walls, cylinder heads, pistons and valves to the coolant by convection. The heat is later carried away by the coolant which accounts for around 25% of the energy losses [15].
Especially when the engine is running at part-load condition, where the heat absorbed into the coolant is greater. At low speed and load condition, the coolant heat transfer rate is about 2 to 3 times the brake power [10].
friction in the main journal bearing of the crankshaft and frictions in piston skirt, piston rings and connected rod.
The auxiliary component losses which refer to the work used to drive engine auxiliary are also included in friction losses. The auxiliary components are only those needed for a normal engine operation, such as oil pump, water pump, fuel pump and sometimes generator.
Pumping losses are the difference in work done to pump out the exhaust gas and the work done to intake the air-fuel mixture, usually defined as the difference between cylinder pressure and atmospheric pressure integrated over the volume of the intake or exhaust stroke [16]. Hence the higher the intake pressure, less will be the pressure difference between intake and the less pumping losses. These three terms sum up as total friction & pumping losses and are eventually dissipated as heat.
4) Heat Losses to Surroundings: The main fraction of the excessive heat generated in the Engine is convected into coolant and counted in coolant loss. However, since the engine is operating at a very high temperature compared to the ambient temperature, there is still a certain amount of the heat that is dissipated directly to the surroundings.
2.2.2
Electric Motor
Apart from ICE, the electric motor is the other important power source in BEV and HEV. Unlike ICE, the electric motor converts electric energy to mechanical energy as an output. On the current automotive market, several types of electric motors have been widely applied. These widely used electric motors are Brushless DC Motor (BLDC), Permanent Magnet Synchronous Motor (PMSM) and Three Phase AC Induction Motor.
1) Brushless DC Motor: BLDC (shown in Figure 2.2.5) consists of two main parts: rotor and the stator. The rotor is the rotating part and made of permanent magnets while the stator is the stationary part consisting of stacked steel lamination with windings in the slots. The current flowing through windings creates an electromagnetic field which makes the rotor to rotate.
Figure 2.2.5: Structure of BLDC motor [17]
hence it is maintenance free. There are many favourable characteristics of BLDC motors, such as high starting torque, high efficiency around 95-98%, and longer lifetime which are suitable for high power density traction application. Thus the BLDC motors are the most preferred motors for the electric vehicle.
2) Permanent Magnet Synchronous Motor: A PMSM (shown in Figrue 2.2.6) is also composed of a rotor with permanent magnets and a stator with windings which is supplied with three phase AC current typically. In general, the PMSM is similar to BLDC motor not only in structure but also traction characteristics like high power density and high efficiency. The main difference between BLDC motors and PMSM is the back EMF (electromagnetic field) of PMSM is sinusoidal while the back EMF of BLDC motors is trapezoidal. A PMSM usually has higher torque and efficiency hence PMSM is available for higher power ratings compared to BLDC motor which makes PMSM to be the best choice for high performance applications like cars, buses. Currently, PMSM is chosen by most automotive manufacturers such as Toyota Prius, Chevrolet Bolt EV, Ford Focus Electric, Nissan Leaf, and BMW i3, etc.
Figure 2.2.6: Structure of PMSM [18]
to this EMF a current flows through the rotor conductor bars. Next, the rotor will start to rotate in the same direction according to Lenz’s law.
Induction motor is favorable for performance oriented electric vehicles due to its cheap cost and high performance. Tesla Model S is a good example showing the high performance capability of induction motors. By using induction motors, it can easily achieve higher power and faster acceleration. Also there will be no demagnetization when the motor is running continuously for a long time at high power and high temperature.
Figure 2.2.7: Structure of three phase AC induction motor [19]
A comparison study between these electric motors is done in [20], and pros and cons of each type of electric motor are listed in Table 2.2.1.
Table 2.2.1: Comparison among electric motors Advantage Disadvantage BLDC ·High power density
·Good overheating capability
·High cost ·Torque ripple
·DemagnetizationPoor field weakening
PMSM
·Smooth torque output ·High efficiency
·High torque
·Good overheating capability
·Higher cost
·Demanding packaging space
Induction Motor
·Good dynamics ·High speed possibility ·Cheap
·High duration
·Complex control ·Power lag
·Low efficiency at high load
1) Copper losses: The windings used in electric machines are not ideal which means that the resistance of the windings cannot be neglected . According to Joule’s Law, when the current flows through the windings, a certain amount of the energy is dissipated in the form of heat, which be can be expressed as Equation 2.3.
CopperLoss = I2· R (2.3)
Where I(A) is the current flowing over a conductor and R(Ω) is the resistance of the conductor.
The winding resistance will increase as the temperature goes up such that 25 ◦C increase in wire temperature will result in 10% rise of the resistance. Meaning that with the same amount of current, the copper losses will increase by 10% [21].
2) Iron losses: Iron losses or core losses mainly consist of two parts: hysteresis losses and eddy currents losses.
zero), there is still flux in the material. The phenomenon is known as Retentivity. To make that flux zero, a cohesive force is applied, which results in hysteresis losses.
Figure 2.2.8: B-H curve of typical magnet material [22]
Eddy current losses are the result of Farady’s law, whenever a core is rotating in a magnetic field, there will be a EMF induced in the wingdings. The induced EMF causes circulating currents which is also known as eddy current to flow in the core. The power losses caused by these circulating currents is called eddy current losses.
In AC machine, whereas the magnetic field is sinusoidal, the iron loss is usually described by Steinmetz Equation 2.4 [21]:
PF e= khf ˆBα+ kef2Bˆ2 (2.4)
Where the first term is the hysteresis loss and the second is the specific eddy current loss. ˆB is the peak value of flux density and β, kk, αand keare constants
determined by curve fitting from the manufacturer’s data.
While in BLDC motors, the variation of flux in the stator core is not sinusoidal. Thus an alternative approach for iron loss is given by Equation 2.5 [20]:
PF e= khf ˆBα+ ke 2π2 ( dB dt )2 RM S (2.5)
determined by curve fitting from the manufacturer’s data.
3) Mechanical losses: The mechanical losses of electric machine can be divided into two parts: friction losses and windage losses (airgap loss). Friction losses are mostly from the rotor bearings and lubrication. While windage losses are due to the air resistance during the rotation of the rotor.
Friction losses in bearings can be expressed by the Equation 2.6 [20]:
Pf r ≈
3
2nrGrotN × 10
−3 (2.6)
Where nr is the number of bearings, Grotis the rotor weight, and N is the speed
(rpm).
And windage losses are determined by Equation 2.7 [20]:
Pwind ≈ 2Dout3 LN
3× 10−6 (2.7)
Where Doutis the outside rotor’s diameter (m), and L is the rotor’s length (m).
4) Stray losses: The stray losses are mainly produced by electromagnetic flux leakage in the windings, tank, core, core clamping plates, magnetic shields, etc. They also involve the distortion of magnetic flux caused by a load current which creates eddy current circulating in windings, frame, tank flanges, etc. Although this portion of losses is due to eddy current, but it is usually referred to stray loss. The generation of the stray loss is also called “skin effect” where the eddy current tends to flow closer to the surface of the conductor.
Stray losses are usually difficult to determine. In general practice, stray losses are considered as 1% of the total load power output.
2.2.3
Inverter
HEV is usually a bi-directional, meaning that the converter can transfer power from both direction.
This power converter is usually made up of IGBTs (Insulated Gate Bipolar Transistor) (shown in Figrue 2.2.9).
Figure 2.2.9: Simplified Inverter Structure[23]
The energy losses occur in inverter can be divided into two different types: conduction losses and switching losses. Conduction losses are generate when the switch is in the on state, while switching losses appear during on and off transient sate of switch.
1) Conduction Losses: Conduction losses are due to switch voltage drop when conducting current, i.e., when the switch is in the on-state. Generally, the voltage drop consists of a constant on-state voltage drop V0 and a resistive voltage drop
due to Ron. Hence the power losses can be formulated as Equation 2.8.
Pon(t) = vonion= (V0+ Ronion) ion= V0ion+ Roni2on (2.8)
Where von is the total voltage drop of the switch when on and ion is the current
flowing through the switch.
losses presenting in [23] [24].
In general, the switching losses are smaller compared to the conduction losses at low switching frequency, and vice versa.
2.2.4
Clutch
Clutch is a mechanical device which can engage or disengage during vehicle operation in order to connect and disconnect the engine power output to the following transmission system. Therefore, the vehicle can perform starting, gear changing and stopping with engine running on. A structural diagram of a typical disc clutch is shown in Figure 2.2.10. The torque is transferred from the engine flywheel and pressure plate to a friction plate, then converted to the primary (input) shaft of the gearbox. The pressure plate is normally spring-loaded, and when the clutch is disengaged, the spring pressure is relieved.
Figure 2.2.10: Clutch structure [25]
Since the disc clutch exploits friction for torque transfer, the main losses of the clutch are the heat losses during the engagement and disengagement of the clutch. A lot of studies investigated slipping duration and energy dissipation. In [26], a mathematical model is developed to calculate heat generation during vehicle launch. At the meantime, different friction models are investigated [27].
2.2.5
Transmission
set). It bridges the speed differences between the engine and the driving wheels while engine power is still transmitted. Since the operating range of a ICE cannot cover the torque and speed requirement in all driving scenarios, these components work together to overcome this problem, making it possible to provide the driver of the vehicle with the possibility of maximum exploitation of the available power.
1) Gearbox: Gearbox can be further divided into manual transmission (MT), automatic manual transmission (AMT), automatic transmission (AT) and continuously variable transmission (CVT). MT requires driver’s engagement for fixed gear ratios (permanent-mesh gears) change. Based on manual transmission, AMT is developed by electronic control of gear shifts. While the AT is composed of torque converter and compound planetary gear sets which operates automatically. Fixed gear ratios apply in the above-mentioned three transmission. However, CVT allows step-less change in gear ratio within a limited range with the help of belt (steel/rubber) running on variable-diameter pulleys.
2) Differential: To allow speed differences between the driven wheels during cornering while maintaining torque distribution, a differential is applied. The differential gear is a planetary gear set where the wheel axles are the sun gear shafts and the differential cage is the planet carrier (shown in Figure 2.2.11).
Figure 2.2.11: Differential structure [28]
4) Power Split Device: In some hybrid vehicles, a power split device connects the ICE and two electric machines to make sure the the vehicle can operate in different hybrid modes. It is a usually planetary gear set where one electric machine is connected to the sun gear, while the other electric machine to the ring and the ICE is connected to the carrier. The ring finally drives the differential gear to the wheels.
4) Driveshaft: A drive shaft with universal joints delivers torque between two powertrain components that are located in different positions in the vehicle according the configuration.
Gearbox, differential, final drive and power split device can be generalized into gear sets. Hence they share the common types of energy losses.
1) Mechanical Losses: The mechanical losses in transmission gear set can be further divided into sliding losses, rolling losses, and bearing losses.
• Sliding Losses: When the gears are meshing, two surfaces slide against each other. Even the contacting surfaces seem to be smooth in macro scale, the roughness still exist in micro scale (shown in Figure 2.2.12). This unevenness will result in surface wear as well as frictional heat losses.
Figure 2.2.12: Surface roughness in Micro and Macro level[29]
• Rolling Losses: When the two gears roll against each other, a pressure is built up in the lubricant. Hence rolling losses are generated by the hydro dynamic pressure forces.
order to predict and calculate the bearing losses.
2) Churning Losses: When considering splash lubrication of a gear set, churning losses are generated from the resistance of the liquid. Related experiments show that churning losses at high speed account for more than half of the total energy losses of the gear transmission [31].
The churning losses can be subdivided into drag losses and pocketing losses according to [32].
• Drag Losses: A rotating gear set is usually fully or partially immersed in oil, as shown in Figure 2.2.13. The gear set is then subjected to drag forces induced along the direction of flow on the periphery and faces of the gears, creating drag losses.
Figure 2.2.13: Drag forces on surface [29]
The drag losses depend on many factors, such as peripheral velocity, gear module, oil submersion depth and also the amount of oil mist [29].
• Pocketing Losses: When two adjacent teeth compress the oil filled cavities and squeeze the oil out, the pocketing losses are generated. Hence the power needed to disperse the oil caught in the gear cavities is defined as pocketing losses. A detailed model used to calculate pocketing losses is introduced in [33].
2.2.6
Battery
powertrain. The working principle of a battery is very simple: a potential difference occurs between two different active materials (electrodes) immersed in an electrolyte, converting chemical energy into electrical energy. There are numerous types of battery with different characteristics. Figure 2.2.14 shows the characteristics of the commonly used batteries. The selection of battery should accord to the requirements and applications of the vehicles. The energy density (Wh/kg) is of more interest about driving range while power density (W/kg) is more focused on when it comes to the performances (acceleration, maximum speed) of the vehicle.
Figure 2.2.14: Specific Energy vs. Specific Power for batteries[34]
The battery usually has high efficiency. For example, Li-ion has 99 percent charging efficiency, and the discharge loss is small. However, since the electrochemical reactions take place inside the battery involving the active materials which will result in an internal resistance of the battery. Hence, a portion of energy of battery is dissipated in forms of heat due to the internal resistance. The losses happen during both charging and discharging.
The losses can be described by a simple model with an EMF in series with a resistance shown in Figure 2.2.15. And the losses can be expressed as Equation 2.9 according to Joule’s Law.
Ploss = Rbatt· i2batt (2.9)
Where Plossis losses power in battery, Rbattis battery internal resistance and ibatt is the
Figure 2.2.15: Battery losses model [35]
2.3
Powertrain Simulation
The design of a powertrain is becoming more and more challenging due to harsher requirements on emission and fuel economy while maintaining the driveability. Thanks to the fast development of computational power, precise and complex simulation can be applied and conducted within powertrain design process. It is an efficient and cost-effective approach for powertrain development at every stage. In general, powertrain simulation aims to investigate how the changes of components, controls, boundary conditions or vehicle can affect on fuel economy, performance, drivability and other customer oriented attributes. Based on this, several modeling approaches are commonly used and they are categorized by the energy flow direction: forward simulation and backward simulation.
1) Backward Simulation: Using a backward simulation is nothing else but answering the question “Assuming the vehicle met the required trace, how must each component perform? [36]”
(a) Vehicle level
(b) Powertrain level
Figure 2.3.1: Backward simulation scheme[37]
Backward simulation often relies on efficiency maps that are based on torque and speed data, and the efficiency are usually created by real world steady state testing. This makes the calculation much simpler than forward approach (usually requires lookup tables) [38], hence the simulation can therefore run faster over relatively larger time steps.
There are also some limitations of backward simulation. Backward simulation uses the assumption that the power demand is always met. It can be sometimes unrealistic, for example, during a hard acceleration event, the needed power can exceed the capabilities of the powertrain, consequently generating unrealistic simulation result. Moreover, the efficiency maps used in backward simulation are mostly created by steady state testing, meaning dynamic effects are neglected. Therefore this approach is often used as preliminary estimation of the fuel consumption or emissions of the vehicle. Also a driver model is absent meaning that there are no brake and throttle signal in the model, making it more difficult for dynamic system simulation and control system development.
(a) Vehicle level
(b) Powertrain level
Figure 2.3.2: Forward simulation scheme[37]
Once the torque and speed of the engine have been determined, fuel consumption or emission can then be calculated from the engine map. Therefore, the torque is passing forward from engine to the wheels through the powertrain, with the same direction of the power flow as reality (shown in Figure 2.2(b)).
The forward simulation is more favorable in terms of hardware development and control simulation. Thanks to the driver model, some quantities in a physical powertrain such as brake signal, throttle signal and torque output from engine, become measurable. Hence vehicle controllers using these quantities can be implemented and tested in simulations. In addition, the dynamic models can be used in this kind of simulation, for example, ICE transient effects can be taken into account by using detailed 1D or 3D fluid-dynamic models. Last but not least, forward simulation can guarantee that the power supplied to wheels will never exceed the power capability of the powertrain.
Vehicle Model
3.1
GT-SUITE
GT-SUITE is a CAE software focusing on vehicle industry developed by Gamma Technologies (GTI), which allows the user to model and simulate different vehicles. GT-SUITE is commonly used by automotive OEMs in development and calibration of both subcomponent level and the entire vehicle powertrain level.
The graphical user interface (GUI) is GT-ISE and it is component-based modelling methodology. Built-in templates of various subsystems are available in template library of the software. Attributes of the subsystem can be filled and modified according to the real subsystem. It is also capable of different kinds of powertrain, including conventional, HEV and BEV powertrain. Thus, one can create any vehicle powertrain that can be simulated for various conditions efficiently.
3.2
Vehicle Modelling in GT-SUITE
A complete hybrid electric vehicle powertrain is modelled in GT-SUITE. The overall configuration is shown in Figure 3.2.1. It is a highly integrated powertrain architecture providing the possibility to simulate different configurations within one model by using different case setups. For example, by only activating electric machine P1 and P2, the overall powertrain becomes conventional series-parallel complex hybrid powertrain. While by only activating electric machine P3 and P4, the powertrain becomes parallel hybrid in front axle and series hybrid in rear axle. The activation and deactivation of electric machines can be easily done by changing on/off state to the related clutches in different case set up.
Figure 3.2.1: Overall hybrid electric powertrain model
In this thesis work, the configuration with electric machine P1 and P2 will be analyzed. The electric machine P1 is connected to ICE’s flywheel directly and a clutch links the bridge between electric machine P1 and P2 making it possible to switch between series hybrid configuration and parallel series hybrid. Detailed components modeling will be explained as follows.
3.2.1
Driver
transmission gear number, clutch pedal position to make sure the vehicle follows the speed target from different drive cycles.
Figure 3.2.2: Driver model
3.2.2
Vehicle
The vehicle model (shown in Figure 3.2.3) includes subsystems such as axle, tire, suspension, brakes, differential, propeller shafts, environment and road template.
Aerodynamic force, rolling resistance force and axle friction force are approached by a fitting retarding force expressed in Equation 3.1
F = A + Bv + Cv2 (3.1)
Where retarding force coefficient A, B, and C correspond to rolling resistance force, axle friction force, and aerodynamics force.
Figure 3.2.3: Vehicle model
3.2.3
Engine
The engine modelled is a four-stroke gasoline engine with the displacement of 1.477 liters. The engine block used in the model is a map-based model which is fed with several engine maps from measurement.
-Mechanical Output Map (see Figure 3.2.4): The map specifies the mechanical output (torque, BMEP or power) as a function of engine speed (RPM) and percent accelerator position (0 - 100%).
-Engine Friction Map (see Figure 3.2.5): The map specifies engine friction (FMEP) as a function of engine speed (RPM) and engine load (BMEP or torque).
Figure 3.2.5: Engine friction map
-Fuel Consumption Map (see Figure 3.2.6): The map specifies brake specific fuel consumption (BSFC) as a function of engine speed (RPM) and engine load (BMEP or torque).
Figure 3.2.6: Fuel consumption map
3.2.4
Electric Machine
comes to modelling in GT-SUITE. Both P1 and P2 are set to be brake power controlled. Again, several electric machine maps from measurement are used to describe the characteristics.
-Electromechanical Conversion Power Loss Map : It specifies electric machine power loss as a function of shaft speed and brake torque. As discussed in previous section, the power loss of the motor converting electrical power to mechanical power (motoring mode) or mechanical power to electrical power (generating mode) are composed of two parts: electromagnet losses and mechanical losses. In addition, the inverter losses are also counted in. However, in this map, the mechanical friction losses are not counted. Instead, the friction torque is neglected since the friction losses in electric machine are rather small and also due to the inadequate information of friction losses.
- Maximum Brake Torque and Minimum Brake Torque Map (see Figure 3.2.7).
(a) P1 brake torque map
(b) P2 brake torque map
Figure 3.2.8: HV auxiliary load model
3.2.5
HV Auxiliary Load
The high voltage auxiliary load is not a main interest in powertrain system point of view, hence it is modelled simply by a constant power consumer with 7kw, shown in Figure 3.2.8.
3.2.6
Battery
The battery model defines resistive or Thevenin electrical-equivalent battery models, consisting of open-circuit voltage, internal resistance which is shown in Figure 3.2.9.
Figure 3.2.9: Battery model used by GT-SUITE
3.2.7
Transmission
The transmission model (shown in Figure 3.2.10) includes the necessary clutches that changes hybrid mode controlled by supervisory energy management system (EMS) and a gearbox with different efficiencies in different gear ratio.
Energy Breakdown Analysis
4.1
Energy Calculation
In GT-POST, although it can provide the results of some commonly used energy losses directly, however, it is limited in single component level which means it is not possible to get the exact amount of energy flowing within two or more components. In addition, GT-POST can only consider the total energy dissipated or consumed throughout the whole simulation, ignoring the energy flow direction. For instance, in HEV powertrain, one important feature is that the electric machines can operate as either motor or generator, the energy losses should be split in two parts for understanding of the energy flows better since the efficiency of being a motor or generator can be significantly different even though it is the same machine. Hence it is necessary to set up an energy calculation block in GT-SUITE model which allows to extract all the necessary data to analyze the energy flow of the powertrain.
1) Engine: As described in Section 3.2.3, the engine model is fed with friction map, fuel consumption map and brake torque map. Hence, the instantaneous engine friction power, fuel consumption rate and brake power can be extracted directly. Therefore, the energy losses as well as useful energy can be calculated as the equations in Table 4.1.1.
Table 4.1.1: Equations for engine energy calculation Engine energy calculation
Total Fuel Energy
Ef uel =
∫tf
t0 LHV × ˙Qdt
where LHV is the lower heating value of fuel and ˙Qis the fuel consumption rate.
Engine Brake Energy
Ebrake =
∫tf
t0 Pbrakedt
where Pbrake is engine brake power.
Engine Friction Loss
Ef riction =
∫tf
t0 Pf rictiondt
where Pf riction is engine friction power.
Engine Heat Loss
Eheat = Ef uel − Ebrake− Ef riction
the direction of energy flow should be determined which means a logic control should be created in order to detect whether the electric machine is operating as a motor (when brake power is positive) or generator (when brake power is negative). In addition, for electric motor P1, when it is used as electric motor, it can either be the starter for the engine or the boost motor for the vehicle. Moreover, in GT-POST, only the mechanical brake power can be derived directly, the electric power flowing in or out of the electric machine needs to be calculated indirectly. The energy losses as well as useful energy of both electric machines can be calculated as the equations in Table 4.1.2 and 4.1.3 respectively.
3) Battery: As described in Section 3.2.6, the battery is modelled by SOC-dependent open circuit voltage and resistance. Hence, the instantaneous loss power and terminal power can be extracted directly. However, the energy flow in battery is also bidirectional, similar to electric machine, the direction of energy flow should be determined by a logic. The energy losses as well as terminal energy of battery can be calculated as the equations in Table 4.1.4.
Table 4.1.2: Equations for P1 energy calculation P1 energy calculation P1 Boost Energy EP 1Boost = ∫tf t0 PP 1Boostdt
where PP 1Boostis the boost power request to P1, which can be extracted from control block.
P1 Starter Energy
If PP 1Boost = 0and PP 1brake> 0 PP 1Starter = PP 1brake
else PP 1Starter = 0 EP 1Starter =
∫tf
t0 PP 1Starterdt
where PP 1Startis the energy used to crank the engine when P1 is acting as a starter and PP 1brake
is P1 brake power. P1 Motoring Losses If PP 1brake > 0 PP 1motoringloss = PP 1loss else PP 1motoringloss= 0 EP 1motoringloss= ∫tf t0 PP 1lossdt
where PP 1lossis loss power of P1.
P1 Motoring Input Energy
PP 1Electric = V1× A1 If Pp1Electric > 0 PP 1M otrotingInput = PP 1Electric else PP 1M otrotingInput = 0 EP 1M otrotingInput = ∫tf t0 PP 1M otrotingInputdt
where V1is the terminal voltage of P1, A1 is the terminal current of P1 and PP 1M otrotingInput
is the input electric power when P1 is motoring.
P1 Generating Output Energy
PP 1Electric = V1× A1 If PElectric < 0 PP 1GeneratingOutput = PP 1Electric else PP 1GeneratingOutput= 0 EP 1GeneratingOutput=− ∫tf t0 PP 1GeneratingOutputdt
where V1is the terminal voltage of P1, A1 is the terminal current of P1 and PP 1GeneratingOutput
is the output electric power when P1 is generating.
P1 Generating Input Energy
Table 4.1.3: Equations for P2 energy calculation P2 energy calculation P2 Motoring Energy If PP 2brake > 0 PP 2M otoringOutput = PP 2brake else PP 2M otoringOutput = 0 EP 2M otoringOutput = ∫tf t0 PP 2Bbrakedt
where where PP 2brakeis P2 brake power.
P2 Motoring Losses If PP 2brake > 0 PP 2motoringloss = PP 2loss else PP 2motoringloss= 0 EP 2motoringloss= ∫tf t0 PP 2lossdt
where PP 2lossis loss power of P2.
P2 Motoring Input Energy
PP 2Electric = V2× A2 If PP 2Electric > 0 PP 2M otrotingInput = PP 2Electric else PP 2M otrotingInput = 0 EP 2M otrotingInput = ∫tf t0 PP 2M otrotingInputdt
where V2is the terminal voltage of P2, A2is the terminal current of P2 and PP 2M otrotingInput
is the input electric power when P2 is motoring.
P2 Generating Output Energy
PP 2Electric = V2× A2 If Pp2Electric < 0 PP 2GeneratingOutput = PP 2Electric else PP 2GeneratingOutput= 0 EP 2GeneratingOutput=− ∫tf t0 PP 2GeneratingOutputdt
where V2is the terminal voltage of P2, A2is the terminal current of P2 and PP 2GeneratingOutput
is the output electric power when P2 is generating.
P2 Generating Input Energy
Table 4.1.4: Equations for battery energy calculation Battery energy calculation
Battery Charging Energy
If Pterminal < 0 Pcharge= Pterminal else Pcharge = 0 Echarge =− ∫tf t0 Pchargedt
where Pterminal is battery terminal power.
Battery Charging Loss
If Pterminal < 0 PcharLoss = PDissipated
else PcharLoss = 0 EcharLoss =∫tf
t0 PcharLossdt
where PDissipated is battery dissipated heat power.
Battery Discharging Energy
If Pterminal > 0 Pdischarge= Pterminal else Pdischarge = 0 Edischarge = ∫tf t0 Pdischargedt
Battery Discharging Loss
If Pterminal > 0 PdisLoss = PDissipated else PdisLoss = 0 EdisLoss = ∫tf t0 PdisLossdt
5) Vehicle: Similar to transmission, the energy flow and dissipation in vehicle level can be divided into two direction as well. According to the vehicle parameters (e.g., axle friction coefficient, aerodynamic coefficient, and friction coefficient) in Section 3.2.5, the relevant energy losses in axle friction, rolling friction and aerodynamic friction can be derived. The energy losses and energy flow can be calculated as the equations in Table 4.1.6.
6) Interaction Energy: Besides the above-mentioned component-level energy flow and losses, the energy interaction among different powertrain components, especially when it comes to hybrid electric powertrain where there are a lot of different energy transformations.
Table 4.1.5: Equations for transmission energy calculation Transmission energy calculation
Transmission Input Energy During Regenerating
If PP 2brake < 0 PT ranRegInput = PT ranOutShaf t else PT ranRegInput = 0 ET ranRegInput = ∫tf t0 PT ranRegInputdt
where PT ranOutShaf tis transmission output shaft power.
Transmission Loss During Regenerating
If PP 2brake < 0 PT ranRegLoss = PT ranLoss else PT ranRegLoss = 0 ET ranRegLoss = ∫tf t0 PT ranRegLossdt
where PT ranLossis transmission loss power.
Transmission Output Energy During Regenerating
If PP 2brake < 0
PT ranRegOutput = PT ranInShaf t
else PT ranRegOutput = 0 ET ranRegOutput =∫tf
t0 PT ranRegOutputdt
where PT ranInShaf tis transmission intput shaft power.
Transmission Input Energy During Propulsion
If PP 2brake > 0
PT ranP roInput = PT ranInShaf t
else PT ranP roInput = 0 ET ranP roInput =
∫tf
t0 PT ranP roInputdt
Transmission Loss During Propulsion
If PP 2brake > 0
PT ranP roLoss = PT ranLoss
else PT ranP roLoss = 0 ET ranP roLoss =∫tf
t0 PT ranP roLossdt
Transmission Output Energy During Propulsion
If PP 2brake > 0
PT ranP roOutput = PT ranOutShaf t
else PT ranP roOutput = 0 ET ranP roOutput = ∫tf t0 PT ranP roOutputdt Clutch Loss EClutchLoss = ∫tf t0 PClutchLossdt
Table 4.1.6: Equations for vehicle energy calculation Vehicle energy calculation
Rolling Loss During Regenerating
FRoll = A× v PRoll = FRoll× v If PP 2brake < 0 PRollReg = PRoll else PRollReg = 0 ERollReg =∫tf t0 PRollRegdt
where A is retarding force coefficient related to rolling force and v is vehicle speed, PRollis
vehicle rolling loss power and PRollRegis vehicle rolling loss power during regeneration.
Aerodynamic Loss During Regenerating
FAero = B× v2 PAero = FAero× v If PP 2brake < 0 PAeroReg = PAero else PAeroReg = 0 EAeroReg = ∫tf t0 PAeroRegdt
where B is retarding force coefficient related to aerodynamic force, PAerois vehicle aerodynamic
loss power and PAeroRegis vehicle aerodynamic loss power during regeneration.
Axle Loss During Regenerating
FAxle = C× v3 PAxle = FAxle × v If PP 2brake < 0 PAxleReg = PAxle else PAxleReg = 0 EAxleReg = ∫tf t0 PAxleRegdt
where C is retarding force coefficient related to axle friction force, PAxleis vehicle axle loss
power and PAxleReg is vehicle axle loss power during regeneration.
Rolling Loss During Propulsion
FRoll = A× v PRoll = FRoll× v If PP 2brake > 0 PRollP ro = PRoll else PRollP ro= 0 ERollP ro= ∫tf t0 PRollP rodt
Aerodynamic Loss During Propulsion
FAero = B× v2 PAero = FAero× v If PP 2brake > 0 PAeroP ro = PAero else PAeroP ro= 0 EAeroP ro= ∫tf t0 PAeroP rodt
Axle Loss During Propulsion
While P1 is acting as generator, the electric energy generated can be either used to charge battery or supplied to P2 directly.
Moreover, when the discharging energy from battery should be divided into three parts, electric energy supply to P1, electric energy supply to P2, and electric energy supply to auxiliary. While when it comes to charging of the battery, the electric energy can be either from P1 or P2.
When it comes to the mechanical energy on the drive shaft, it is composed of engine mechanical output, P1 boost mechanical output and P1 starter mechanical output energy.
These energy can be calculated as the equations Table 4.1.7.
6) Others: There are some more energy needed to be calculated in order to enclose the energy loop with powertrain system. From the perspective of vehicle system, the useful energy, or the ultimate goal of a vehicle energy system, is the kinematic energy which moves the vehicle. In addition, it also acts as a bridge linking the energy flow between propulsion and generation process since the kinetic energy is also maximum potential regeneration energy.
Meanwhile, usually the final SOC of the battery will not remain the same of initial SOC even if charging sustaining control strategy is adopted, leading to energy deviation. When the final SOC is lower than initial SOC, it means a part of inherent energy in the battery is extracted which needs to be taken into account, and vice versa.
This energy can be calculated as the equations Table 4.1.8.
Table 4.1.7: Equations for interaction energy calculation Interaction energy calculation
P1 Starter Energy
EStarter2Shaf t = Ebrake+ EP 1Boost+ EP 1Starter + EP 2M otoring− ET ranP roInput− EClutchLoss
where EStarter2Shaf tis the additional mechanical energy to drive shaft from P1 when it is
cranking the engine.
EStarterLoss = EP 1Starter− EStarter2Shaf t
where EStarterLossis the energy used to starting the engine but not converted to mechanical
energy to the drive shaft.
P1 Electric Energy to P2
EP 1ChargeP 2= EP 1GeneratingOutput+ EP 2GeneratingOutput− Echarge
where EP 1ChargeP 2is electric energy supplied to P2 as motor and from P1 as a generator.
Battery Discharge to P1
EBatt2P 1= EP 1M otrotingInput
where EBatt2P 1is the battery discharged electric energy to P1 as motor.
Battery Discharge to P2
EBatt2P 2= EP 2M otrotingInput− EP 1ChargeP 2
where EBatt2P 2is the battery discharged electric energy to P2 as motor.
Battery Discharge to Auxiliary
EBatt2Aux =
∫tf
t0 PAuxdt
where EBatt2Auxis the battery discharged electric energy to auxiliary load and PAuxis auxiliary
load power.
P2 Charge Battery
EP 2ChargeBatt= EP 2GeneratingOutput
where EP 2ChargeBattis electric energy generated from P2 and supplied to battery.
P1 Charge Battery
EP 1ChargeBatt= EP 1GeneratingOutput− EP 1ChargeP 2
where EP 1ChargeBattis electric energy generated from P1 and supplied to battery.
Drive Shaft Mechanical Energy
EEng2shaf t = Ebrake− EP 1M otrotingInput
EShaf t = EStarter2Shaf t+ EEng2shaf t + EP 1Boost
where EEng2shaf tis engine mechanical output energy going to drive shaft directly and EShaf t
is the overall mechanical energy transmitted to drive shaft.
Mechanical Brake Loss
EM echanicalBrake =
∫tf
t0 PM echanicalBrake
Table 4.1.8: Equations for other energy calculation Other energy calculation
Kinetic Energy or Potential Regeneration Energy
EKinetic = ET ranP roOutput− ERollP ro− EAeroP ro− EAxleP ro
SOC Deviation Energy
ESOCdev = ∆SOC× Cmax× VAvg
where ∆SOC is the difference between initial SOC and final SOC and VAvgis the average battery
terminal voltage.
4.2
Sankey Diagram
The Sankey diagram was invented more than 100 years ago by the Irish engineer Riall Sankey to analyze the thermal efficiency of steam engines. Since then, it has become an important tool in identifying energy flow and losses for complicated systems. The Sankey diagram is also a powerful tool for visualizing vehicle energy consumption and losses both in component level and overall powertrain level. Sankey diagram is used to show the energy flow direction and quantity (the ratio between each other) within the system. The width of the arrow or line is proportional to the value, so the larger the arrow, the larger the flow. In each process stage, the flow arrows or lines can be combined together, or separated to different paths. Different colors is used to distinguish different categories in the chart, or to indicate the transition from one stage to another.
The most obvious feature of the Sankey diagram is that the sum of the flow branches before and after a process stage are always the same, meaning that energy balance is always maintained.
Energy Management System
5.1
Control Strategy
The powertrain includes three power units: ICE, electric machine P1, and electric machine P2. The power request to each power unit is determined according to the energy management strategy (EMS). Different power split methods determine the diversity of energy flow paths of a hybrid electric powertrain.
The working mode of a hybrid electric vehicle is generally classified according to the energy flow path. Figure 5.1.1 shows the various energy flow paths.
(a) Pure electric drive mode (b) Series hybrid mode
(c) Parallel hybrid mode (d) Regeneration braking mode
Figure 5.1.1: Energy flow paths (red: traction flow blue: regeneration flow black: no flow)
the rule introduced by FEV [39] is applied, with following features.
1) ICE On & Off : When the vehicle is driving in low load condition with a certain amount of SOC such as urban driving condition, it is sufficient to run the vehicle in pure electric mode. This helps to avoid the low engine efficiency area of the engine. One control parameter is defined as k1, which specifies the power
threshold of engine on/off state. The power demand of ICE is composed of battery charging power demand and wheel demand power. As shown in Figure 5.1.2, if the sum of power demand from battery and wheel is lower than k1, the
ICE will be switched off to avoid the low engine efficiency area. On the other hand, if the power demand exceeds electric motors’ capability or the battery SOC is below a certain level, the ICE will switch on again. Therefore, k1 requires the
dependency on battery SOC.
Two more control parameters are defined: k25 and SOCReserve. k25 is the factor
determining dynamic k1based on battery SOC as described in Equation 5.1 while SOCReserve is the minimum SOC level for the vehicle to operate in pure electric
drive mode.
k1 = k25× (SOCActual− SOCReserve) (5.1)
Where SOCActual is the actual battery SOC.
Figure 5.1.3 shows that k25is also the slope of k1 vs SOC curve, the higher k25the
higher battery charging power demand at a certain SOC.
Figure 5.1.3: Power demand for pure electric driving depending on battery SOC 2) Battery Charging: Apart from regenerative braking, when the engine is on
whereas the vehicle is running on hybrid mode, the battery will be charged by engine. The charge power will depend on three control parameters: SOCReserve, k23, and k24, shown in Equation 5.2.
PRequest = k23− k24× (SOCActual − SOCReserve) (5.2)
Where k23 is defined as power request to the engine for battery charging at SOCReserve level and k24is the slope which modifies the charging power request
depending on percentage SOC deviation.
Figure 5.1.4 indicates the higher k24 the higher battery charging power demand
at a certain SOC.
Figure 5.1.4: Battery charging power request depending on battery SOC
electric drive mode. Two control parameters are defined to decide whether the vehicle will run on series hybrid mode or parallel hybrid mode. The decision will be based on demanded power and vehicle speed. k6defines power threshold for
switching from serial to parallel hybrid mode and k7defines speed threshold for
switching from serial to parallel hybrid mode. When the power demand or the vehicle speed is lower than the threshold, the electric motor can fulfill the power demand to drive the vehicle. While as the power demand or vehicle speed exceeds the threshold, the engine will engage in the propulsion to compensate inadequate power or torque from electric motor.
As Shown in Figure 5.1.5, k6 together with k1, fully split the engine to three
different working mode areas to optimize the engine efficiency.
Figure 5.1.5: Engine mode shift
4) Series Mode Load Point Shift: When the vehicle is running on series mode, ideally the engine will work at the point where the efficiency is maximum. However, in order to smoothen engine power demand in transient driving cycles, load point shift parameter k5is used to enable the shift of engine power to higher
and lower power. Thereby, the actual engine power is defined as Equation 5.3
P = (PDemand− PEf f max)× k5+ PEf f max (5.3)
Where PDemand is the demanded power and PEf f max is the power in operating
In general, the control logic can be described by flow chart in Figure 5.1.6.
Figure 5.1.6: Operating mode shift logic
5.2
Control Parameter Optimization
With the GT-SUITE built-in DOE feature, a series of DOE were done to investigate the optimal combination of control parameters for a certain driving cycle.
WLTC (shown in Figure 5.2.1), coming into force in 2017, is developed with the aim of being used as a global test cycle across different world regions. WLTC driving cycle is divided into four parts with different average speeds: low, medium, high and extra high. Each part contains a variety of driving phases, stops, acceleration and braking phases. It is more representative for vehicle’s daily use based on real world driving data.
Hence, the control parameters will be optimized using WLTC driving cycle as a reference. The initial SOC was set as 0.5, in order to maintain charge, the optimization target is to find the proper values of control parameters that achieves the lowest fuel consumption with final SOC as close to 0.5 as possible. The DOE results are represented in Figure 5.2.2.
The optimal solutions is marked in red in the figures. The corresponding optimal control parameters and default control parameters are listed in Table 5.2.1.
Figure 5.2.1: WLTC driving cycle speed profile
Figure 5.2.2: DOE result for WLTC
Table 5.2.1: Default and optimized control parameters
Control Strategy k5 k6(kW) k7(km/h) k23(kW) k24(kW/%) k25(kW/%)
Default 0.5 40 80 25 0.2 2
driving cycle, apart from WLTC, UDDS (shown in Figrue 5.2.3) and a real-world test driving cycle (shown in Figrue 5.2.4) consisted of different driving conditions and are used as input of the simulation.
Figure 5.2.3: UDDS driving cycle speed profile
Figure 5.2.4: Real-world driving cycle speed profile
Table 5.2.2: Fuel consumption (L/100km) comparison
Driving cycle Default control parameters Optimized parameters Improvement
WLTC 1 0.991 0.90%
UDDS 1 0.989 1.11%
