Thermal Simulation of Hybrid Drive System Kathiravan Ramanujam Shiva Kumar BM Fluid and Mechatronic Systems

Degree Project Department of Management and Engineering LIU-IEI-TEK-A--11/01145--SE

Thermal Simulation of Hybrid Drive System

Master Thesis By Kathiravan Ramanujam Shiva Kumar BM

Examiner Professor Karl-Erik Rydberg Supervisor Mr. Kristoffer Nilsson TTM, BorgWarner TorqTransfer Systems AB

Linköping University Department of Management and Engineering Fluid and Mechatronic Systems Linköping, June 2011 LIU-IEI-TEK-A--11/01145--SE

ii

Abstract Safety, performance and driving comforts are given high importance while developing modern day cars. All-Wheel Drive vehicles are exactly designed to fulfill such customer requirements. In modern times, human concern towards depleting fossil fuels and cognizance of ecological issues have led to new innovations in the field of Automotive engineering. One such outcome of the above process is the birth of electrical hybrid vehicles. The product under investigation is a combination of all wheel drive and hybrid system. A superior fuel economy can be achieved using hybrid system and optimized vehicle dynamic forces are accomplished by torque vectoring action which in turn provides All-Wheel Drive capabilities. The very basic definition of engineering is ‘to optimize phenomena of physics for the improvement of human lives’. Ideally, Engineers don’t like power losses. But heat generation is inevitable whenever there is a conversion of energy from one form into another. In this master thesis investigation, a thermal simulation model for the product is built using 1D simulation tool AMESim and validation is done against the vehicle driving test data. AMESim tool was chosen for its proven track record related to vehicle thermal management. The vehicle CAN data are handled in MATLAB. In a nutshell, Simulation model accounts for heat generation sources, oil flow paths, power loss modeling and heat transfer phenomena. The final simulation model should be able to predict the transient temperature evolution in the rear drive when the speed and torque of motor is supplied as input. This simulation model can efficiently predict temperature patterns at various locations such as casing, motor inner parts as well as coolant at different places. Various driving cases were tried as input including harsh (high torque, low speed) ones. Simulation models like this helps Engineers in trying out new cooling strategies. Flow path optimization, flow rate, convection area, coolant pump controlling etc are the few variables worth mentioning in this regard.

iii

Acknowledgement It gives us immense pleasure to thank all the wonderful people who helped us in this journey of master thesis.

At the outset, we are greatly indebted to our supervisor Mr. Kristoffer Nilsson, firstly for choosing us to carry out this work and secondly for having patience to answer tonnes of our questions. Big thanks to Mr. Gustaf Lagunoff for his tireless support in the lab.

When we started working on our thesis neither the software tools nor the thesis objectives looked something unusual. All thanks to our Prof. Karl-Erik Rydberg and his support staff at Linköping university for their anticipation of industrial research requirements and preparing students accordingly.

Countless thanks to all our colleagues at BorgWarner TTS. They created very amicable working environment and always had some constructive suggestions for us. Innebandy games on Mondays and Friday fikas made sure that each week started and ended with some fun!

Finally, we extend warm regards to our parents for giving us an opportunity to experience all this and backing us in all our endeavors.

Kathiravan R Shiva Kumar BM

Linköping June 2011

iv

Contents Abstract……………………………………………………………………………………………………………………….. iii Acknowledgement………………………………………………………………………………………………………… iv List of Figures……………………………………………………………………………………………………………….. vii List of Tables…………………………………………………………………………………………………………………. ix List of Symbols……………………………………………………………………………………………………………… x 1. Introduction……………………………………………………………………………………………………… 1 1.1. Design challenges……………………………………………………………………………………….. 1 1.2. Thermal management………………………………………………………………………………… 1 1.3. Objectives…………………………………………………………………………………………………… 1 1.4. All wheel drive and Torque vectoring............................................................. 2 1.4.1. Vehicle architecture of AWD.............................................................. 3 1.4.2. Torque Vectoring................................................................................ 4 1.5. 1D Simulation……………………………………………………………………………………………… 5 1.6. Method............................................................................................................ 5 1.7. De-limitations.................................................................................................. 7 2. Heat transfer…………………………………………………………………………………………………….. 8 2.1. Conduction…………………………………………………………………………………………………. 9 2.1.1. Thermal contact conductance…………………………………………………………. 9 2.2. Convection………………………………………………………………………………………………….. 10 2.2.1. Forced convection…………………………………………………………………………….11 2.2.2. Free or Natural convection………………………………………………………………. 12 2.3. Radiation…………………………………………………………………………………………………….. 13 2.4. Biot number………………………………………………………………………………………………… 13 3. Hybrid drive components – Theory…………………………………………………………………… 15 3.1. Electric motor……………………………………………………………………………………………… 15 3.1.1. Basic working principle……………………………………………………………………. 17 3.1.2. Losses in motor……………………………………………………………………………….. 17 3.2. Planetary gears…………………………………………………………………………………………… 19 3.2.1. Working………………………………………………………………………………………….. 22 3.2.2. Heat generation………………………………………………………………………………. 22 v

3.3. Bearings……………………………………………………………………………………………………… 24 3.3.1. Operating principle………………………………………………………………………….. 24 3.3.2. Heat generation………………………………………………………………………………. 25 3.4. Power electronics……………………………………………………………………………………….. 26 4. System modeling in AMESim……........................................................................... 27 4.1. General............................................................................................................ 27 4.2. Electric motor.................................................................................................. 28 4.2.1. Modeling methodology....................................................................... 29 4.3. Planetary gears and bearings.......................................................................... 31 4.3.1. Modeling of planetary gears................................................................ 31 4.3.2. Modeling of bearings........................................................................... 32 4.4. Cooling circuits................................................................................................ 32 4.4.1. Hydraulic circuit................................................................................... 33 4.4.2. Heat exchanger.................................................................................... 36 4.4.3. Secondary circuit................................................................................. 37 5. Validation............................................................................................................... 39 5.1. Experiments..................................................................................................... 39 5.1.1. Viscous friction..................................................................................... 39 5.1.2. Pressure drop – Hydraulic circuit......................................................... 39 5.1.3. Pressure drop – Flow control valve...................................................... 39 5.2. Sensors............................................................................................................. 39 6. Results.................................................................................................................... 40 6.1. Plots – Drive cycle 1......................................................................................... 40 6.2. Plots – Drive cycle 2......................................................................................... 49 6.3. Plots – Drive cycle 3........................................................................................ 51 7. Conclusion.............................................................................................................. 52 8. Future work............................................................................................................ 53 9. References.............................................................................................................. 54 Appendix...................................................................................................................... 57

vi

List of Figures Figure 1.1: eAWD drive system......................................................................................... 2 Figure 1.2: Layout of FWD based AWD............................................................................. 3 Figure 1.3: Layout of RWD based AWD............................................................................. 4 Figure 1.4: Torque vectoring action.................................................................................. 4 Figure 1.5: Methodology................................................................................................... 6 Figure 2.1: General steps in a Heat transfer problem....................................................... 8 Figure 2.2: Conduction heat transfer through a solid....................................................... 9 Figure 2.3: Velocity boundary layer in convection............................................................ 10 Figure 2.4: Types of convection heat transfer................................................................... 11 Figure 2.5: Thermal boundary layer................................................................................... 13 Figure 3.1: Flow of magnetic flux in a permanent magnet................................................ 15 Figure 3.2: Direction of flux lines – Left hand clasp rule.................................................... 16 Figure 3.3: Relation between magnetic flux lines and direction of current in a coil......... 16 Figure 3.4: Working principle of electric motor................................................................ 17 Figure 3.5: Hysteresis loss.................................................................................................. 18 Figure 3.6: Planetary gear train.......................................................................................... 19 Figure 3.7: Spur gear contact forces.................................................................................. 20 Figure 3.8: Line of contact of spur and helical gears......................................................... 21 Figure 3.9: Helical gear contact forces.............................................................................. 21 Figure 3.10: Viscous friction force vs. Sliding velocity....................................................... 23 Figure 3.11: Schematics of a deep groove ball bearing..................................................... 24 Figure 3.12: Forces acting in a bearing.............................................................................. 25 Figure 4.1: Thermal solid and hydraulic properties icon.................................................... 27 Figure 4.2: Fluid properties of 50% ethylene glycol........................................................... 28 Figure 4.3: Schematic modeling of Electric motor............................................................. 29 Figure 4.4: Thermal sub models used................................................................................ 30 Figure 4.5: Signals and controls sub models...................................................................... 30 Figure 4.6: TFCV0 – Internal flow general forced convection exchange sub model.......... 31 Figure 4.7: TFCV3 – Internal flow convection sub model with two thermal ports............ 31 Figure 4.8: THCV4 – External flow free convective exchange............................................ 31 vii

Figure 4.9: Schematic thermal model in AMESim............................................................. 31 Figure 4.10: Block diagram of cooling circuits................................................................... 32 Figure 4.11: Hydraulic circuit AMESim............................................................................... 35 Figure 4.12: General schematic of a Heat exchanger........................................................ 36 Figure 4.13: Double pipe heat exchanger.......................................................................... 37 Figure 4.14: Schematic sketch of the secondary circuit..................................................... 38 Figure 4.15: Block diagram – Drive oil flow....................................................................... 38 Figure 6.1: Stator winding temperature vs. time............................................................... 40 Figure 6.2: Sump oil temperature vs. time........................................................................ 41 Figure 6.3a: Drive outlet oil – right side temperature vs. time.......................................... 42 Figure 6.3b: Drive outlet oil – left side temperature vs. time........................................... 42 Figure 6.4: Drive outlet oil – average temperature vs. time.............................................. 43 Figure 6.5: Drive inlet oil temperature vs. time................................................................. 44 Figure 6.6: Housing surface temperature – right vs. time................................................. 44 Figure 6.7: Housing surface temperature – left vs. time................................................... 45 Figure 6.8: Housing surface temperature – bottom vs. time............................................ 45 Figure 6.9: ISG inlet coolant temperature vs. time........................................................... 46 Figure 6.10: TV motor inlet coolant temperature vs. time................................................ 46 Figure 6.11: Power electronics coolant outlet temperature vs. time................................ 47 Figure 6.12: Heat Exchanger – coolant out temperature vs. time.................................... 48 Figure 6.13: Pressure variation in the Hydraulic circuit vs. time....................................... 48 Figure 6.14: Flow in Hydraulic circuit vs. time................................................................... 49 Figure 6.15: Stator winding temperature vs. time............................................................ 50 Figure 6.16: Flow in Hydraulic circuit vs. time................................................................... 50 Figure 6.17: Stator winding temperature vs. time............................................................. 51 Figure 6.18: Flow in Hydraulic circuit vs. time................................................................... 51

viii

List of Tables Table 3.1: Gear ratios from Planetary gear........................................................................ 22 Table 3.2: Operating principle of different bearings......................................................... 24 Table 4.1: Values of friction factors for the two flow regions........................................... 34 Table 4.2: Components of hydraulic circuit from AMESim Library.................................... 34

ix

List of Symbols eAWD – Electric All Wheel Drive ATF – Automatic Transmission Fluid TV motor – Torque Vectoring Motor ISG – Integrated Starter Generator CFD – Computational Fluid Dynamics HEV – Hybrid Electric Vehicle Cp – Specific heat of the fluid h – Convective heat exchange coefficient Cdim – Characteristic heat transfer length σ – Stefan – Boltzmann constant IGBT – Insulated Gate Bipolar Transistor FEM – Finite Element Method OEM – Original Equipment Manufacturer FWD – Front Wheel Drive RWD – Rear Wheel Drive DC – Direct current Nu – Nusselt number Pr – Prandtl number Re – Reynolds number AC – Alternating current Bi – Biot number W – Heat generated R - Resistance T – Temperature ΔP – Pressure drop I - Current η – Efficiency

x

Chapter 1 Introduction BorgWarner TorqTransfer Systems AB (formerly known as Haldex Traction Systems) is a leading developer and supplier of All-Wheel Drive systems for passenger cars. The company has R&D centre and manufacturing facility in Landskrona, Sweden. In order to be a reliable and competitive product developer, one has to meet the ever changing market demands. BorgWarner is developing the prototypes of electrical All-Wheel Drive systems (eAWD) for modern day requirements. The company has proven record in producing famous AWD systems such as limit slip couplings (LSC) and electronic limit slip differential (eLSD). LSC systems are purely mechanical drives whereas eAWD is a hybrid electric drive. eAWD works as a rear differential (replacing regular differential) as well as houses a hybrid electric motor and provides the vehicle with All-Wheel Drive capabilities.

1.1 Design Challenges Since the drive system has an electrical motor and mechanical friction sources such as bearing and gears, heat generation and efficient cooling of such parts has to be dealt right from the design stages. To state few other design challenges in a nut shell, the product has to fulfill a set of key factors like efficiency, weight and compactness, cost, durability, energy saving etc.,

1.2 Thermal Management Drive’s thermal management has direct implications on all the design requirements discussed so far. So, this Master thesis is related to thermal management of a Hybrid Electric Drive.

1.3 Objectives The objectives of thermal simulation of the electric hybrid drive system are  To build and validate a simulation model in AMESim.  To account for the hydraulic thermal losses in the system.  To account for cooling of the system. The simulation has to take into account the heat sources present in the system and cooling them accordingly. The drive system in focus is the Prototype 2 of the electric all wheel drive (eAWD) system. This system has a permanent magnet AC synchronous motor (propulsion motor) which propels the vehicle through planetary gears on either side to each wheel. Figure 1.1 shows the arrangement of components of the drive system.

1

Figure 1.1: eAWD drive system The drive systems also includes the function of torque vectoring, a new concept in the field of vehicle dynamics to provide better yaw damping and stability to the vehicle. The principle of torque vectoring is to influence braking and driving torque to either of the wheels so the above parameters are achieved. Torque vectoring is achieved by another motor (TV motor) that controls the output through the planetary gears. All this systems are optimized and controlled by sophisticated control algorithms developed in-house. Power to the electric motors is supplied from a battery. The battery is charged by the engine power through an Integrated Starter Generator (ISG) in the car. The power from the battery is inverted by Power electronics and supplied to the electrical machines. The system is lubricated and force cooled by automatic transmission fluid (ATF). The system has two cooling circuits with ATF cooling the hybrid system and 50 % Ethylene Glycol cooling the power electronics and the ATF through cross flow heat exchanger.

1.4 All wheel Drive and Torque vectoring All wheel control is the strategy where in all the dynamic parameters of four wheels of a vehicle are used intelligently to enhance the handling and performances of the vehicle. Vehicle’s accelerating capability is a function of available traction between tires and ground or road. In brief, there are three sets of control philosophies that can be used for All-Wheel control 1. What to control ? Vertical load acting on the tires to enhance grip so that firm road contact can be obtained. 2

Control parameters: Vertical dynamic components such as suspension and body. 2. What to control ? Slip ratio’s and slip angles of front and rear tires. This helps in increasing longitudinal and lateral forces acting on tires. Cornering acceleration can be controlled with this strategy. Control parameters: Antilock Braking system (ABS), Traction control system and steering system. 3. What to control ? Force (torque) distributed to each wheel. Control parameters: Vehicle’s power train and braking system are employed for this purpose. All wheel drive systems replace the regular differential in the vehicles. Basic advantage of differential is that it can allow outer and inner rear wheels to rotate at different speeds, for example: while taking a turn. If not this differential then excessive tire wear and steering difficulties would have been inevitable. But differential can’t supply different torques to rear wheels in case of slip (loss of traction) in one of the wheels. All-Wheel Drive system comes into picture exactly at this situation. These systems can supply different torques to individual wheels based on the road conditions and requirements. 1.4.1 Vehicle architecture of AWD 1. Front wheel drive based AWD Most passenger All-Wheel drive vehicles fall into this category. Limit slip differential can transfer torque from slipping wheel to non-slipping (more grip) wheel by various methods. In this vehicle architecture, a power transfer unit or a power take-off unit turns rotation motion of primary axle to 90 deg and transfers it to longitudinal propeller shaft. Which then transfers power to secondary axle through torque coupling device. Limit slip couplings can be used to distribute torque between rear wheels as well between the front and rear wheels.

Figure 1.2: Layout of FWD based AWD.[18] 3

2. Rear wheel drive based AWD As seen in the figure 1.3, The engine directly drives rear wheels and mostly engine is mounted longitudinally (North-South power train layout). A transfer case arrangement is used to transfer power to the front driving shaft.

Figure 1.3: Layout of RWD based AWD.[18] 1.4.2 Torque Vectoring[19] Traditional All wheel drive systems focused on distribution of torque to each tire to achieve high cornering, traction and braking performances. There is one more mechanism known as ‘torque vectoring’, which enhances the stability of a vehicle by yaw moment control. Yaw moment

Driving force

Torque vectoring

Braking force

Figure 1.4: Torque vectoring action The beauty of this system is torque vectoring between left and right wheels such that equal magnitude of driving and braking forces are generated on each wheel (as seen in figure 1.4).

4

This helps in controlling yaw moment of the vehicle without regulating engine torque and interfering driver’s commands. This leads to two vehicle dynamic advantages. Firstly, cornering performance can be improved by load optimization between right and left wheel. Secondly, cornering performance can be further improved by load distribution between front and rear wheels.

1.5 1D Simulation Since the problem involves fluid flow and heat exchange, one might think of 3D simulation tool such as CFD. The problem in hand will become too tedious in terms of time and expenses. Using methods such as lumped parameter model, it is possible to use 1D simulation tools to solve the automotive cooling problems. 1D simulation tool is an essential tool for multi-physics simulations. Automotive lubrication and cooling application falls under such a category. One can find subsystems such as coolant flow, heat transfer and mechanical system in a typical vehicle thermal management problems. Using 1D simulation tool like AMESim it is possible to study those subsystems separately and integrate them to understand the behavior of the entire system. AMESim can be applied to replicate the cooling and lubrication system of eAWD. This gives an idea of heat generation sources and how to extract heat from them efficiently. Simulation model helps in sizing the coolant pump, change the convection area, proper dimensioning of the coolant channels, etc., if required. Practically it is impossible to achieve the same level of accuracy as 3D simulation such as CFD or Finite element approach. But 1D simulation offers flexibility in terms of simulation time to try out new heat management strategies. Problem in hand is a typical example for a transient heat transfer. As the name says, lumped system means some of the bodies in the system are lumped together since the temperature is uniform throughout the mass. Lumped system analysis typically has varying temperature in the mass with time but the temperature is considered to be uniform throughout the mass under consideration. So temperature is only a function of time.

1.6 Method The methodology applied to solve this problem involves identifying the heat sources, get to know how they are cooled in the actual system, study of basic theory about heat transfer and losses in heat sources. Modeling the components in AMESim individually. Once the individual simulations are successful, they are put together to get the total model. Model is validated and simulations are performed for different drive cycles. Finally, the simulation and actual results are compared.

5

Problem Definition

Objectives and Project scheduling

Literature Survey and simulation tool studies

Real system data collection

Model conceptualization

Model building in pieces

No

Verification

No

Yes Individual model integration Yes No

Verification

No

Yes No

No Validation Yes Documention and results

Figure 1.5: Methodology[20]

6

1.7 De-limitations Some of the de-limitations of this simulation model are 1. Input values that are within the area of operation of the look-up table can only be processed. 2. Temperature of solid masses in the drive are hard to predict accurately. 3. Accuracy in results is not as close as numerical methods such as FEM or CFD. But results are good for a 1D lumped parameter approach. 4. Simulation takes long time for long run driving cycles. 5. Some components are not considered to simplify the simulation model.

7

Chapter 2 Heat Transfer[1] Before simulating heat transfer phenomena in an advanced software environment like AMESim. It is very important to learn the basic concepts and mathematical relations beforehand. Advanced softwares should only simplify the job of an engineer but not at the cost of knowing the basic facts. “In the hands of poorly educated people, these software packages are as dangerous as sophisticated powerful weapons in the hands of poorly trained soldiers ” A quote from the book ‘Heat transfer and mass transfer’ by Yunus A. Cengel (Page 38)[1]. Heat is a form of energy, the science which deals with the rate of transfer of heat from one system to another because of temperature difference is called heat transfer. There are three modes of heat transfer: 1. Conduction, 2. Convection and 3. Radiation. All types of heat transfers require the existence of temperature gradient. Of course, energy transfer takes place from a medium at higher temperature to a medium at lower temperature.

Figure 2.1: General steps in a Heat transfer problem

8

2.1 Conduction Conduction is the transfer of energy from a particle of higher energy to adjacent particle with lesser energy. Conduction can take place in solids, liquids and gases. In solids, heat transfer always takes place by conduction phenomenon since the molecules remain fixed. Conduction heat transfer rate can be expressed by Fourier’s law of conduction as kA where

t x

k = Thermal conductivity, W/m °C A = Wall area, m2 Δt Temperature gradient, Δx Wall thickness, m

Figure 2.2: Conduction heat transfer through a solid[1] 2.1.1 Thermal contact conductance In many practical cases, heat conduction happens through more than one layer solid. Since the surface contacts are not absolutely smooth there exists some resistance to heat conduction. This resistance per unit surface area is called thermal contact resistance. Inverse of thermal contact resistance is called ‘thermal contact conductance’. In this simulation this parameter plays vital role in evaluating temperature conduction in radial direction of motor. According to Newton’s law of cooling hc A

Tinterface

A & Tinterface = interface contact area & interface temperature difference. hc, which is called convection heat transfer coefficient is here know as thermal contact conductance.

9

2.2 Convection Unlike conduction, convection phenomenon requires fluid motion around the system. Convection heat transfer is not as simple as conduction since it involves conduction heat transfer in addition to fluid movement. It is a fluid property dependent phenomenon depending on fluid properties such as viscosity, thermal conductivity, specific heat and density. One other important variable is fluid velocity. Convection heat transfer is directly proportional to fluid velocity. Convection heat flow rate is given by the equation equation is directly derived from Newton’s law of cooling.

h A

Twall

–T

in Watt. This

where h = convection heat exchange coefficient in W/m2 °C A = heat transfer surface area, m2 Twall = temperature of the surface, °C T = temperature of the fluid at a considerable distance, away from the surface, °C Convection heat transfer coefficient (h) is defined as “the rate of heat transfer between the solid body and fluid per unit surface area and unit temperature difference.”

Figure 2.3: Velocity boundary layer in convection[1] From figure 2.3 we can see that when a fluid is forced over a surface, its velocity comes to zero at the contact area. Reason for this is the viscosity of fluid, it means fluid is sticking to the surface and there is no slip, which is called no-slip condition. Velocity profile as shown in the figure 2.3 is developed because of no-slip condition. Also at the point of contact the temperatures of fluid and surface are same and this is called as no-temperature jump condition. Because of no-slip and no-temperature jump conditions, heat transfer from solid surface to fluid layer happens by pure conduction.

10

There are 2 types of convection , based on method of flow initiation. 1. Free convection 2. Forced convection

Figure 2.4: Types of convection heat transfer[1]

2.2.1 Forced Convection When the flow over a body is influenced by external means, the convection process is categorized as forced convection. Convection heat transfer rate for forced convection is defined as below QForcedConvection = hforced*A*(Twall – T ) Complexity of forced convection lies in evaluating hforced[2] hforced

Nu r, e k Cdim

where Nu is the Nusselt number, k is the thermal conductivity of the fluid and C dim is called characteristic dimension which depends on the surface geometry. Nusselt Number: Nusselt number is the ratio of convection heat flux and conduction heat flux. Heat flux can be defined as the rate of heat transfer per unit time per unit surface area. Nu

h Cdim k

Nusselt number basically shows how convection is enhancing the heat transfer rate in the fluid layer when compared to the conduction in the same layer of fluid. A Nusselt number ‘1’ can be interpreted as heat transfer in the fluid layer is purely by conduction. Nusselt number is a function two dimensionless numbers, the Prandtl number and the Reynolds number. 11

Prandtl Number: r

olecular di usivity of momentum olecular di usivity of heat r

Cp k

Prandtl number is another important dimensionless number in convection analysis. It describes relative thickness of velocity and thermal boundary layers. If Prandtl number is small then it means heat diffusion is stronger than velocity (momentum). Reynolds Number: It’s a very famous non dimensional number and it’s used to differentiate laminar and turbulent flows. It is defined as the ratio of inertia force to viscous force. e

V

c

= Density of the fluid V = velocity of the stream Lc = Characteristic length µ = Absolute viscosity of the fluid When Reynolds number is small the viscous forces in the fluid are more dominant. This means inertia forces are overshadowed by viscous forces. This is the reason for laminar flow at small Reynolds number. Inertia forces in the fluid are proportional to density and velocity of the fluid. So on the same lines, when Reynolds number is high enough inertia forces take lead over viscous forces. This is the reason for random zig-zag flow fluctuation. 2.2.2 Free or Natural Convection This type of convection phenomena happens when the mobility of the fluid takes place because of change in density triggered by heat exchange process. So fluid is not forced by external sources. Heat flow rate equation given as below QFreeConvection = hfree* A * (Twall – T ) Convection exchange coefficient for free convection is gives as hfree

Nu r, r k Cdim

Nusselt number in case of free convection is a function of Prandtl number and Grashof number. Grashof number governs the flow regime in free convection and it is defined as the ratio of buoyancy force to viscous force.

12

Thermal boundary layer: Velocity boundary layer is developed when a fluid flows over a surface. It can be seen that flow velocity adjacent to the surface is zero because of viscosity and flow velocity varies from zero to 0.99 u in the fluid flow regime.

Figure 2.5: Thermal boundary layer[1] Similarly when a fluid flows over a surface the thermal boundary layer is developed as shown in figure 2.4, provided there exists a temperature difference between fluid and surface.

2.3 Radiation As a result of changes in the electronic configuration of atoms a form of energy called electromagnetic waves are emitted from the matter. This is called radiation. Radiation process doesn’t require any medium for energy propagation. Radiation heat transfer is gives as Q= As σ T4 where σ = Stefan- Boltzmann constant = 5.67 e-8 W/m2 As = Surface area T = absolute temperature

2.4 Biot number Biot number is gives the eligibility criterion for lumped parameter model. In order to check if lumped mass method is applicable or not, the following has to checked. Biot number , Lc = Characteristic length definition k= thermal conductivity h = convection coefficient Biot number is defined as the ratio of convection at the surface of the body to conduction within the body. Since temperature is assumed to be uniform in lumped mass the immediate assumption is thermal resistance to the heat conduction is zero. So lumped system approach 13

is exact when Bi= 0 and when Bi >0 it is an approximate model. Since this method itself has few assumption. So lumped system parameter method can be applied only if Bi ≤ 0.1 and this means that temperature variation within body is minor and uniform temperature can be assumed. The software AMESim used in this simulation works on the above principle.

14

Chapter 3 Hybrid drive Components – Theory The hybrid drive system is made of electrical and mechanical components. These components are responsible for generation of power, which is then fed to the wheels to drive the vehicle. The components used are not 100% efficient, they produce some losses. The losses are generated in the system as heat. Each components have different parameters that are responsible for heat generation in them. In this chapter, the working as well as the losses generated in the components are explained. The main components that generate heat in the system are 1. Electric motor. 2. Planetary gears. 3. Bearings.

3.1 Electric Motor It is possible to produce force, induce voltage and current using electromagnetic principles. Basically, all electric machines are based on these principles. All electric motors work on the basic principle of converting electrical energy into mechanical energy. It’s the other way around when they work as generators. Magnetism: Permanent magnets can attract metal object when they are in close vicinity. The reason for this is magnetic field. Magnetic flux lines shown below represents the magnetic field and they are invisible, so the figure 3.1 below helps us in visualizing it.

Figure 3.1: Flow of magnetic flux in a permanent magnet[3]

15

A similar magnetic field is produced when a current flows through the conductor, which is of importance in this study. According to left hand clasp rule, when a conductor carrying current is clasped with the left hand pointing thumb in the direction of current then the fingers point in the direction of flux lines

Figure 3.2: Direction of flux lines – Left hand clasp rule[3]

Figure 3.3: Relation between magnetic flux lines and direction of current in a coil[3] When the current carrying conductor is modified in the form of a coil as shown in figure 3.3, the intensity of magnetic flux lines is increased greatly. As in the case of permanent magnets, the flux line leave north pole and enter the south pole. Magnetic field can be even strengthened by winding the coil around an Iron core since Iron is a better conductor of flux lines than air. Advantage of electro-magnets is that poles can be switched when the direction of current is reversed. AC motors are based on this phenomenon.

16

3.1.1 Basic working principle:

Figure 3.4: Working principle of electric motor[3] Motor of simplest form can be explained using two electro-magnets and a permanent magnet. Figure 3.4A : Current is passed in the coils in such a direction so that north and south poles face against each other in coil 1 and coil 2. A permanent magnet is placed in between the coils such that north and south poles of each magnets face each other. Since like poles repel and unlike poles attract magnet is rotated as shown. Figure 3.4B: Magnetic poles remain the same in both coils. Now north and south poles of both magnets attract each other to give a position like this. Figure 3.4C: In order to continuously rotate permanent magnet, current direction in both the coils is changed after every 1800 rotation of permanent magnet. This would make like poles repel and unlike poles attract. 3.1.2 Losses in Motor Heat generation in a motor is the direct consequence of power loss. Unnecessary heat generation in various parts of the motor leads to failure unless it is efficiently cooled down. There are various kinds of losses in a motor. They are 1. Copper winding losses in the stator 2. Magnetic (core) losses  Eddy current losses  Hysteresis losses 3. Friction losses

17

Copper Winding losses: Motor windings offer some resistance to the flow of electric current . This results in heating of copper windings. This loss is also called as Electric or Joule loss. W = I2 R where W the heat generated, I the Current and R the Resistance of the windings. Current requirement is directly proportional to torque output. So at higher torques winding losses are more. Energy conversion: Electrical energy is converted into heat energy in copper windings. Iron losses or Core losses: 1. Eddy current losses: It’s a common phenomenon in cored conductive materials undergoing a change in magnetic field. Since Iron is experiencing a change in magnetization, a voltage is generated in it and this results in a small circulating current called as eddy current. This can be reduced by using laminated core instead of a solid core. Also laminations are insulated to reduce the losses. These kinds of losses are largely speed dependent, as the speed of the motor increases the losses gradually increase as well. Energy conversion: Electrical energy is converted into heat energy in Iron. 2. Hysteresis loss:

Figure 3.5: Hysteresis loss[4] When Iron undergoes a cycle of magnetization and demagnetization, some residual magnetization is left over because of hysteresis. It is denoted as BR in the figure 3.5. Some energy is lost to magnetize the core back to zero. This is called hysteresis and it depends on the area enclosed by the curve.

18

Frictional Losses: These are also known as mechanical losses. Losses happen due to the conversion of mechanical energy in thermal energy. Losses in the rotating bearings and viscous friction represent a part of losses. Another important loss consumer is windage loss. Power lost in the bearings is proportional to square of motor speed and where it varies as cube of motor speed for windage loss. Stray Losses: Some losses in the motor are very tedious to account for and they are not major either. So they are neglected in this work.

3.2 Planetary gears Planetary gears are epicyclic gear trains used to transmit power. They consist of a sun gear, a ring gear and planet gears mounted on a carrier. A simple representation of planetary gear is given in figure 3.6.

Figure 3.6: Planetary gear train[5] The main advantages of using planetary gear is multiple tooth contact which results in reduced tangential gear force for a gear pair. Also the gear train occupies less space for any possible combination and gear ratio than an ordinary gear train. Well the disadvantage is the complex design and increased bearing forces (Bearing forces also depend on the type of gear being used). The type of gear being used contributes to the load carrying capacity and wear. Generally spur and helical gears are used in planetary gears. The gear contact forces for both the type of gears and their advantages are discussed below. 19

Spur gear The gear teeth is parallel to the axis of the gear. This gear is capable of meshing with the other gear only if they are mounted in parallel shafts. The meshing gear teeth creates a line contact along the involute profile of the teeth. The contact line is along the width of the teeth and is parallel to the axis. The equivalent gear contact force acts on the contact surface of the teeth at an angle above the horizontal called the pressure angle. The equivalent force (F) is then resolved into horizontal and vertical forces. The horizontal force (F t) acts tangentially and it transmits power to the meshed gear. The vertical force or the radial force (Fr) acts as load on the shaft, this force acts on the bearing that supports the shaft. Figure 3.7 gives the representation of gear forces in the case of a Spur gear.

Figure 3.7: Spur gear contact forces[6] Helical gears In this case, the teeth are not parallel to the axis of the gear but are skewed at angle with respect to the axis called the helix angle. The gear teeth can be either right hand or left hand. Gear meshing takes place in both parallel and perpendicular axis. The pinion and gear should have opposite hand teeth for them to mate in parallel axis,. And for perpendicular axis mating, the pinion and gear should have same hand teeth. In the case of helical gears, the contact between a gear pair is line. It starts at one end and moves diagonally along the flank of the teeth. Figure 3.8 compares the line of contact of both spur and helical gears.

20

Figure 3.8: Line of contact of Spur (left) and Helical (right) gears [7] The equivalent gear force (WN) acts perpendicular to the flank of the teeth and it is resolved into three components from two different planes. From the normal plane, radial (W r) and horizontal forces are obtained. The horizontal force in then resolved into tangential (W t) and axial (Wa) forces in the transverse plane. Figure 3.9 represents the resolved contact forces.

Figure 3.9: Helical gear contact forces[8] Helical gears proves to be advantageous than spur because of the high load carrying capacity and smoother action at the expense of the manufacturing cost.

21

3.2.1 Working Planetary gears has three members namely sun gear, ring gear and carrier holding the planets. The three main configuration[11] of planetary gears are 1. One input, one output and one member fixed. The possible combinations and the resulting mechanisms are given in table 3.1. 2. One input, two outputs and no fixed member. The carrier may be input, sun and ring gears are the outputs. In this case, the mechanism is a differential. 3. Two inputs, one output and no fixed member. The sun and ring gear can be inputs and the carrier be the output. The resulting mechanism is a speed combiner. Input Sun Planet Carrier Ring

Output Stationary Result Planet Carrier Ring Gear reduction Ring Sun Overdrive Ring Planet Carrier Gear reduction Table 3.1: Gear ratios from Planetary gear[9]

Direction Same as input Same as input Reverse

High or low gear ratios can be obtained by varying the power input and output to the members of the planetary gear. It gives 1:1 ratio when any two members are held stationary. There are some constraints in the design of the planetary gears. They are 1. Number of teeth:[10] R = 2*P + S where R, S and P are the number of teeth in ring, sun and planet gears respectively. 2. Number of planets:[11] The constant should be an integer. R and S are the number of teeth in ring and sun gear respectively. np denotes the number of planet gears. Gear ratio of planetary gear can be calculated using the relation +S ωC = R* ωR + S* ωS[10] where and S are the number of teeth in ring and sun gear. ωS, ωR, ωC are the angular velocity of sun gear, ring gear and planet carrier. In this drive system, the sun is the input, carrier is the output and the ring gear is either stationary or act as input. 3.2.2 Heat generation Heat generation in planetary gears is because of two types of friction. 1. Mechanical friction 2. Viscous friction The total losses are sum of the losses due to mechanical and viscous friction. 22

1. Mechanical friction: This is caused by the sliding and rolling of gear teeth in contact with each other. This friction depends upon the contact force between mating gears and is maximum when the driving torque or the tangential force is maximum. This heat generated can be calculated from the efficiency of the planetary gears. Efficiency, Loss = 1 – η Heat generated = Loss * Input power Input power ωin* Tin output power ωout* Tout Tin and Tout are in the input and output torque. ωin and ωout are the input and out speed. Mechanical friction is calculated from the input speed and torque from the electric motor and the mechanical efficiency of the gear train. The mechanical efficiency provided by the designer (company) is 0.98. So from the formula given, loss is 0.02* ωin* Tin. 2. Viscous friction:[12] It is generated when two surfaces separated by a liquid slide against each other. The viscosity of the liquid tries to oppose the rotation of the sliding gears. This friction depends upon the viscosity fluid and the sliding velocity of the gears.

Figure 3.10: Viscous friction force vs Sliding Velocity[13] From the figure 3.10, µv is a function of fluid viscosity and V is the sliding velocity of the gear. The friction force acts opposite to the driving force in the gears.

23

3.3 Bearings Bearing are used to support rotating elements with a fixed member or allow relative motion between two members in motion supported on each other. Bearings are classified based on their operating principle.

Figure 3.11: Schematics of a Deep grove ball bearing[14] 3.3.1 Operating Principle The parts of a bearing include the inner race (or inner ring), outer race (or outer ring) and contact or rolling element between them. The inner and outer race are fixed to the members that are to be in relative motion and the contacting element allows the relative motion between them. The operating principle of bearings depend upon the type of contacting element used. They are 1. Mechanical contacting element – Ball and cylindrical roller bearings (Roller bearings) 2. Fluid contacting element – Hydrostatic and Hydrodynamic bearings. 3. Magnetic bearings – Bearing supported by magnetic flux. Contact element – Type Ball roller Cylindrical roller Hydrostatic Hydrodynamic Magnetic

Operating principle Four or two point contact between rolling and fixed elements. Line contact between rolling and fixed elements. A layer of pressurized fluid separates fixed and moving element. Fluid pressurized by external source. A layer of pressurized fluid separates fixed and moving element. Fluid is self pressurized by the rotating element. Relative motion is provided by magnetic levitation between the fixed and moving element. Table 3.2: Operating principle of different bearings

Selection of bearings: The factors responsible in the selection of bearings are namely available space, loads, misalignment, precision, speed, stiffness, quiet running, axial displacement, mounting and dismounting and integral seals.[15]

24

3.3.2 Heat generation In the case of bearings, heat is generated by rolling and sliding friction between metal parts. The factor that is responsible for friction are the loads that the bearing support, the type of contact element used in the bearing and the operating speed. 1. Types of load – Axial, radial, both axial and radial load and moments (all constant or variable). As explained under the section Planetary gears, bearings support the gear mounted on a shaft and the type of gear used induces contact forces which in turn act on the bearings. For example, a spur gear contact force has two components – tangential and radial. The tangential force drives the gear and the radial force acts on the bearing. On the other hand, helical gears has three contact force components – tangential, radial and axial. Here, the bearing supports both the radial and axial forces.

Figure 3.12: Forces acting in a bearing [15] 2. Type of contact element – Ball, cylindrical, fluid and magnetic. Mechanical friction depends upon the wear between metals in contact. Metal wear is high when the contact area between the rolling and fixed elements is high. From table x, the contact area of cylindrical rollers > ball rollers > hydrodynamic > hydrostatic > magnetic. Calculation of heat generated: Heat generated can be calculated from the loss in power due to friction. Power lost = Frictional moment * operating speed. Frictional moment, M = 0.5 * µ * P * d [N.mm][15] where, P = equivalent dynamic bearing load [N] P = X*Fr + Y*Fa Fr and Fa are the radial and axial forces X and Y are the radial and axial bearing factor obtained from product catalogue. d = bearing bore diameter [mm] µ = constant coefficient of friction for bearing – value obtained from product catalogue. 25

In the case of the drive system, the bearings that operate continuously (over the whole operation time) are considered for heat generation. Power loss is calculated by force analysis of the individual bearings.

3.4 Power Electronics Power electronics is another source of heat generation in this investigation. This is the significant heat source in the Ethylene glycol circuit. Power electronics is a must in hybrid vehicles since it is used to control and convert electric power. In HEV, they are mostly used as inverters ( DC to AC ) and DC to DC converters. ower electronic devices are made up of semiconductor components such as I BT’s and diodes. Thermal aspects are an important design consideration for semiconductor based devices since they have operating temperature limits. Switching losses in the IGBT and diodes are the source of heat generation. The heat generated is then transferred to the heat sink, which is normally cooled by some means. In this case, it is cooled by forced convection by 50% ethylene glycol.

26

Chapter 4 System modeling in AMESim In this chapter, basic concepts of modeling the system and the approximations made while modeling are explained. The software has built in components in its libraries containing all the components required to design the circuit. The work from the part of the user is to select the right component and input the necessary data. The user guide from AMESim is helpful in understanding and modeling individual parameters in the system. Also some trials were done to understand the sub models better. Modeling includes sub models from signals and control, thermal hydraulic, thermal and mechanical libraries.

4.1 General Before starting the modeling of the system, the basic requirements like material properties and fluid properties has to be modeled first.

Figure 4.1: Thermal solid and hydraulic properties icon[2] The solid and fluid icon represents properties of the desired solid and fluid. AMESim library has some default solids and fluids which can be selected when it matches with the requirement. But new properties can also be created by knowing the required parameters. The procedure to create new solid and fluid properties are given in the user guide for thermal and thermal hydraulic library of AMESim respectively. For the case of the drive system, new properties like ATF, 50% ethylene glycol, magnets and insulation paper has to be created. Figure 4.2 gives the required properties of 50% ethylene glycol, which is then used in the simulation as a data file. Multiple solids and fluid icons can be used in the same file, they separated by index numbers. The sub models used further in the design has to be mapped with the corresponding index number. In the case of simulation, solid properties of copper, aluminium, carbon steel, iron, magnet and insulation paper are used. Similarly, the fluid properties like ATF, 50% ethylene glycol and air are used in the simulation model.

27

Figure 4.2: Fluid properties of 50% ethylene glycol

4.2 Electric Motor The most critical part for thermal simulation in eAWD is the electric motor. Predicting temperature raise of various parts in the electric motor is dealt with utmost importance. Individual components of the electric motor are built separately and the type of heat transfer between the components are defined. This model finds out the temperature distribution along the radial direction of propulsion motor. Different parts of the electric motor are modeled as thermal mass. Their corresponding heat transfer mode of contact with the subsequent component is shown in figure 4.3. Interaction between the masses is modeled by thermal contact conductance (inverse of thermal contact resistance). Heat generating sources in the motor are identified correctly. They are    

Stator Copper ( Copper loss ) Stator Iron ( Iron loss ) Rotor Iron ( Iron loss ) Air gap ( Windage loss )

Heat (Power lost) generated in the system is supplied in the form of 2D tables in AMESim. This 2D data is obtained from the electric motor supplier and they in turn got those values from Finite element method.

28

Figure 4.3: Schematic modeling of Electric motor 4.2.1 Modeling methodology:[2] The following sub models are used for this simulation THC000 - Thermal Capacity This sub model usually called as mass because user has to supply only the mass as input and of course the index number. As seen in figure 4.4 A, there are four thermal ports. Input to all these ports is heat flow rate W and temperature is calculated by the sub model. Temperature is calculated using the equation below. dT = ∑ incoming heat flow rates / mass specific heat Derivative of energy stored in the mass is dE

∑ incoming heat flow rates

Unused ports can be blocked by using zero heat flow source sub model THHF0 THCC0 – Signal (null) to heat flow This sub model converts input constant value at port 1 into the heat flow rate of same value at port 2. Figure 4.4 B gives the representation of this sub model. THCDC0 – Contact conductive exchange This sub model (figure 4.4 C) requires contact area (mm2 ) and thermal contact conductance (W/m2/ degC ) values as input from the user. As shown in the picture, temperatures are the 29

inputs and heat flow rate is calculated by the sub model using the relation below. dH = contact conductance * area ( temperature difference b/w port 1 & 2 )

Figure 4.4: Thermal sub models used FXYA001 – Function of two inputs defined by ASCII file

Figure 4.5: Signals and controls sub models It’s a duty cycle sub model, it receives speed values in rpm at port 3 and torque values at port 1. So the output is a function of speed and torque, which is the power loss in this simulation. Speed and torque are defined as 1D tables using UDA01 sub model with X1 column containing the time ( sampled at various frequencies) and Y column containing the respective data. Figure 4.5 A is the representation of this sub model. Power loss data is specified in a ASCII data file having 2D table format. The input data must be of monotonic sequence , data interpolation and out of range data must be taken care of. UDA01 – Signal from ASCII data file. It’s a duty cycle sub model wherein output from this is a function of time and data is defined in 1D ASCII file. The representation is given in figure 4.5 B .

30

Figure 4.6: TFCV0 – Internal flow general forced convection exchange sub model

Figure 4.7: TFCV3 – Internal flow convection sub model with two thermal ports

Figure 4.8: THCV4 – External flow free convective exchange Figures 4.6, 4.7 are forced convection sub models and figure 4.8 is a natural convection sub model. The above three work on the principle of convection as explained in chapter 2. The entire thermal model of motor is given in appendix 1. The convection and conduction area values are given in appendix 2.

4.3 Planetary gears and Bearings 4.3.1 Modeling of planetary gears Planetary gears are modeled as a single lumped mass taking into account the mass of sun, planets and the ring gear. This mass is supplied with the heat generated, which is the sum of the mechanical and viscous friction. The gear train weighs 3.88 kg and are made of Plain carbon steel.

Figure 4.9: Schematic thermal model in AMESim

31

The planetary gears are cooled by the oil through forced convection. Modeling involves adding the thermal mass of planetary gear with the convection model and the surface area which the oil cools. The surface area is calculated to be 26422.25 mm2. 4.3.2 Modeling of bearings Bearings are modeled as a lumped mass. The heat generated in all the bearings is put together in a single mass and is cooled by convection. Figure 4.9 gives the schematic sketch of thermal model of bearings in AMESim. The mass of all bearings is 1.6 kg. The total convection cooling area of the bearings is calculated to be 30965 mm2. The calculated losses in bearing are given under appendix 3.

4.4 Cooling circuits Two separate cooling circuits are employed in hybrid drive system as shown in figure 4.10. They are 1. Hydraulic circuit. 2. Secondary circuit.

Figure 4.10: Block diagram of cooling circuits These circuits too are modeled in AMESim. Their purpose and modeling are the details discussed under this section.

32

4.4.1 Hydraulic Circuit Hydraulic circuit is the primary circuit of the drive system. This circuit includes oil flow from and to the main tank. The oil is supplied to the drive system for cooling the heat sources and to lubricate the moving parts. A hydraulic pump supplies oil to the circuit. The circuit has bends, joints, sensors, flow control valves and hoses connecting them to the drive. With the presence of these elements, the resistance in the circuit increases creating a pressure drop across the circuit. Also the fluid friction between the oil and hose causes pressure drop. All these factors have to be taken into account while designing the circuit. The theory behind the resistance in flow across pipes (or hoses), joints, bends comes from basic fluid mechanics. The fluid flow resistance depends upon the factors such as viscosity of the fluid, diameter of the hose, material of the hose, roughness of the flow surface and the flow velocity. The pressure drop Δ for duct flow across any cross section is given by Darcy – Weisbach equation.[16] f

d

V2 2

The variables are the density of the fluid , length of the duct , velocity of the fluid V , the hydraulic diameter (d) and the constant called as the Darcy friction factor (f). The Darcy friction factor, f is a function of type of flow and the relative roughness of the duct. The type of flow is classified as laminar and turbulent, and is determined by the Reynolds number. The equation for Reynolds number (Re)[16] is V d

e

where, µ is the dynamic viscosity of the fluid. In general case, the flow is laminar when Re < 2000 and turbulent when Re > 4000. The region between these two values of Reynolds number is the transient region. Eddy currents are created when fluid flows through the duct. This resistance to flow is caused by the internal roughness of the duct or pipe.[17] The relative resistance is obtained by the relation[16] Absolute roughness ela ve roughness Hydraulic diameter d The value of absolute roughness is a material property and can be obtained from tables for various materials.[16] This parameter has a major influence when the flow is turbulent and has no impact in the case of laminar flow. The friction factor for laminar and turbulent flow are given as follows.

33

Laminar flow e value obtained from Moody diagram[16] – function of Re and relative roughness. Table 4.1: Values of friction factor for the two flow regions

Friction factor (f)

Turbulent flow

This is the basic theory that the user should know to build the hydraulic and secondary circuit. Table 4.2 gives the list of components and the required input data to built the hydraulic circuit. The required data is either measured from the actual component or calculated by simple tests. The data for which the tests are done is explained under chapter 5. Some information is obtained from the manufacturer’s catalogue of the product. Sl. No. 1

Component

Representation

Required Data

Electric motor

1. Input speed

2

Hydraulic pump

1. Displacement volume 2. Efficiency

3

900 bend

1. Diameter 2. Radius of curvature 3. Critical Reynolds number

4

T - bend

1. Diameter at all ports 2. Friction factor at all ports 3. Critical Reynolds number

5

Pipe

6

Restriction

1. 2. 3. 4. 5. 1.

7

Variable flow control valve

8

1. 2. 3. 4. 1. 2. 3.

Filter

34

Diameter Length Relative roughness Wall thickness Young’s modulus of material Cross sectional area of the orifice Characteristic flow rate at full opening of the valve Corresponding pressure drop Critical Reynolds number Min. and max. signal value Characteristic flow rate Corresponding pressure drop Critical Reynolds number

9

Heat exchanger

10

Variable temperature tank Pressure sensor

9

1. 2. 3. 4. 5.

Mass Type of heat exchanger Wall thermal resistance Equivalent area of both liquids Volume of liquid in Heat exchanger 1. Volume of the tank 2. Initial temperature 1. Gain 2. Offset

10

Temperature sensor

1. Gain 2. Offset

11

Signal

1. Constant input control signal (null) Table 4.2: Components of hydraulic circuit from AMESim Library.

The schematic hydraulic circuit built in AMESim is shown is figure 4.11.

Figure 4.11: Hydraulic circuit – AMESim

35

4.4.2 Heat Exchanger Theory: [1] Heat exchangers are devices that are used to cool hot liquid by heat transfer with a cold liquid without mixing the liquids with each other. The process of heat transfer involves convection between the liquid and wall, and conduction through the wall separating them. Figure 4.12 gives a simple representation of heat exchanger. Depending on the direction of flow of the hot and cold liquids, the heat exchangers are classified into 1. 2. 3. 4.

Parallel flow – when flow is parallel and in same direction as in figure 4.12. Counter flow – when flow is parallel but in opposite direction. Cross flow – when the flow direction is perpendicular to each other. Shell and tube – here a shell and a tube is used to transfer liquids. The tube takes multiple pass within the shell to provide better cooling.

Figure 4.12: General schematic of Heat exchanger The rate of heat transfer in a heat exchanger is calculated by the formula T

ΔT is the temperature difference between the hot and cold liquids. is the total thermal resistance across the network. The total thermal resistance in the circuit can be obtained by knowing the heat transfer path. Figure 4.13 shows the representation of a double pipe heat exchanger. Assuming that the hot and cold liquids pass through the inner and outer tubes respectively. The heat transfer path is convection from hot liquid to wall, conduction in wall and convection from wall to cold liquid. The total thermal resistance is given by R = Rconv1 + Rwall + Rconv2 36

Figure 4.13: Double pipe heat exchanger Assuming that the hot and cold fluid flows through the inner and outer tubes, the individual resistances are given by 1 conv1 h i Ai conv2

Wall

1 h o Ao

D ln Do i 2 wall

hi and ho are the convection heat transfer coefficients of hot and cold liquids, A i and Ao are the surface area contacts of liquids with the inner and outer wall, D i and Do are the diameters of the inner and outer wall, Kwall is the thermal conductivity of the wall material, L is the length of the heat exchanger. With the above knowledge and the required data from table 4.2, the values are calculated from the actual component as shown in appendix 4. In the actual case, ATF oil and 50% ethylene glycol are the hot and cold liquids. 4.4.3 Secondary circuit Secondary circuit carries 50% Ethylene Glycol as coolant. A coolant pump is used to circulate the fluid. Temperature sensors are mounted at the required locations in the circuit and a flow sensor is mounted as well. The circuit consists of the following components. 1. 2. 3. 4. 5. 6.

Power electronics of Propulsion motor, TV motor and ISG. Integrated starter generator (ISG). Pump Radiator DC/DC converter Heat exchanger

The circuit is modeled only with power electronics and heat exchanger, as the required data was unavailable for the other components from the OEM. 37

This approximation is made accurate with the help of the sensor data. It can be explained better from the figure 4.14.

Figure 4.14: Schematic sketch of the secondary circuit. Instead of providing flow through a pump, coolant flow is given corresponding to the sensor outputs of flow meter and temperature values (as seen in the left of the figure 4.14). Appendix 5 gives the thermal modeling of secondary circuit in AMESim. Appendix 6 gives the input data to the secondary circuit. The modeling of power electronics is same as that of electric motor. The loss table corresponding to speed and torque is obtained for the three power electronic units (TV motor, ISG and Propulsion motor). The power loss is given as heat source to the heat sink, which is then cooled by 50% ethylene glycol. Heat sink design is important in obtaining the convection area. It is obtained from the supplier’s catalogue. The secondary circuit has temperature sensors after each power electronics modules and after the heat exchanger. This helps in validating the simulation model in a better way. To simplify the simulation model, TV motor is not included in the simulation. The block diagram showing the cooling of the drive system is given in figure 4.15. The simulation model of the total system is given in Appendix 7.

Figure 4.15: Block diagram – Drive oil flow 38

Chapter 5 Validation The simulation model built has to be checked and verified whether it works similar to the actual system or not. This is done by comparing the values for simulation and actual system. Also few other parameters required for the simulation were determined by experiments on the actual system. These details are discussed in this chapter.

5.1 Experiments 5.1.1 Viscous friction In the case of Planetary gears, Viscous friction is calculated by a test done on the eAWD. The maximum torque required to turn the output shaft at no load is calculated. Because at no load, the gears tend to overcome the damping force from the fluid and gear inertia. This torque is then multiplied with the angular velocity of the vehicle to get the loss due to viscous friction. 5.1.2 Pressure drop – Hydraulic circuit This test is to calculate the pressure drop across the hydraulic flow circuit. From figure 4.11, the hydraulic circuit shows that the ATF being pumped is divided into two paths, namely the stator and the rotor. Pressure drop across the two circuits are calculated individually by closing V1 and opening V2 and vice versa. The pressure difference between the pressure sensor value and absolute atmospheric pressure is calculated (since the pressure in atmospheric when the oil drains to the sump). This value of pressure drop is then added to the circuit by introducing an orifice with equivalent pressure drop. The cross section of the orifice can be obtained from pressure drop and the corresponding flow. The orifice gives the actual flow resistance in the circuit. 5.1.3 Pressure drop – Flow control valve From table 4.2, the data for flow control valve requires pressure drop at full opening of the valve. This is calculated by fixing two pressure sensors before and after the valves and the pressure drop is calculated with one valve full open and the other closed.

5.2 Sensors In order to check the functioning of the simulation model, the results of the simulation model has to be compared with the actual system results. Temperature sensors are available in the actual system at various places like copper windings, sump, oil inlet to drive, oil outlet in the left and right side of the drive. Apart from these, additional thermistors were placed on the housing of the drive at places like bottom, top, right and left side. Measured sensor readings were compared with simulation results to complete the validation process.

39

Chapter 6 Results This chapter discusses the outcome of temperature variations at various locations in the drive. Temperature readings are compared between the actual drive system and the thermal simulation model.

6.1 Plots – Drive cycle 1 This is a normal city driving cycle. The initial temperature of all the components of the drive system are at a normal temperature of 18 deg C. The temperatures of various parts of the drive system are given below: Stator winding temperature: The propulsion motor in the drive has an inbuilt temperature sensor which gives the temperature reading of the copper windings in the stator. The below plot shows the comparison of temperature variation between the observed temperature and the thermal simulation copper mass temperature. 80

70

Temperature degC

60

50

40

30

20

10

Actual Simulation 0

500

1000

1500 Time S

2000

2500

3000

Figure 6.1: Stator winding temperature vs. time Comments: The reason for variation in the temperature readings is because of the approximations made in calculating the 2D loss tables, convection areas and heat transfer coefficients. The actual system has one temperature sensor fixed on the copper winding. If ‘n’ number of sensors are fixed in the copper windings at regular interval and the average of ‘n’ temperature readings would be a better comparison with the simulation. This is because the simulation is 40

a lumped model and it assumes uniform temperature distribution throughout the mass, which is not the actual case. When flow was increased during the first 1000 seconds, the cooling was better (simulation model tuning). Sump oil temperature: The sump collects the oil that falls down after cooling the components of the drive. Also the sump retains some amount of oil, the oil below the outlet point in the drive. The temperature of the oil collected is compared with a thermal – hydraulic volume temperature. 40 Actual Simulation

Temperature degC

35

30

25

20

15

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.2: Sump oil temperature vs. time Comments: Temperature sensor in the sump is positioned at the bottom point. But coolant coming from left and right sides after cooling heat generating sources doesn’t mix completely in the sump. It is sucked out to tank before proper mixing takes place in the sump. Whereas in simulation, it is assumed that coolant mixes well in the sump. This explains the difference in two curves shown in figure 6.2. Drive outlet: The oil in the drive is pumped back to the main tank from both the right and left sides. The temperature of those in the actual system is compared with the thermal heat exchange.

41

40 Actual Simulation

Temperature degC

35

30

25

20

15

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.3a: Drive outlet oil – right side temperature vs. time

50 Actual Simulation

45

Temperature degC

40

35

30

25

20

15

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.3b: Drive outlet oil – left side temperature vs. time

42

45 Actual Simulation

Temperature degC

40

35

30

25

20

15

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.4 : Drive outlet oil – average temperature vs. time Comments : In the real system, coolant temperature coming out in the left and right side of the drive is not same. But in the simulation model, coolant temperatures are almost same. Since it is complex to predict flow pattern inside the drive, a better approximation would be to take the average of outlet temperatures. As seen in figure 6.4, temperature profiles are closely matching for simulation and actual case. Oil – Drive inlet: From the hydraulic circuit, the pumped oil in cooled in heat exchanger before entering the drive. The temperature at that point is compared with the oil temperature sensor readings before the T – joint in the hydraulic circuit. The simulation results closely follows the actual reading.

43

36 Actual Simulation

34

Temperature degC

32 30 28 26 24 22 20 18

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.5: Drive inlet oil temperature vs. time Housing temperature: Additional thermistors are added on the outside of the drive housing at places on the right, left and bottom (when seen from behind the vehicle). These temperatures are compared with their respective housing mass in the simulation. 36 Actual Simulation

34 32

Temperature degC

30 28 26 24 22 20 18 16

0

500

1000

1500 Time S

2000

2500

Figure 6.6: Housing surface temperature – right vs. time

44

3000

45 Actual Simulation

Temperature degC

40

35

30

25

20

15

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.7: Housing surface temperature – left vs. time

40 Actual Simulation

Temperature degC

35

30

25

20

15

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.8: Housing surface temperature – bottom vs. time Comments: Housing is getting heated up because of hot coolant and partly due to heat conduction from motor and mechanical heat sources. Since hot coolant flow is not same in left and right side of the drive, the simulation predictions are not quite close. Also, housing is cooled by external air flowing at relative velocity and heat transfer co-efficient in this case is a sensible approximation. 45

Power electronics coolant temperature: The power electronics system has temperature sensors in between each stages of Propulsion motor, ISG and TV motor. The temperatures are compared at each stage with the simulation. Appendix 4 shows the order in which the systems are placed. 34 32

Temperature degC

30 28 26 24 22 Actual Simulation

20 18

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.9: ISG inlet coolant temperature vs. time

34 32

Temperature degC

30 28 26 24 22 Actual Simulation

20 18

0

500

1000

1500 Time S

2000

2500

Figure 6.10: TV motor inlet coolant temperature vs. time

46

3000

34 32

Temperature degC

30 28 26 24 22 Actual Simulation

20 18

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.11: Power electronics coolant outlet temperature vs. time Comments : The heat generating source in power electronics module is I BT’s and diodes. There exists thermal resistance between heat sink and these components. In this simulation thermal resistance is not used because of unavailability of value and modeling complexity. Heat exchanger – coolant outlet temperature: After cooling the power electronics, the coolant passes through heat exchanger cooling the ATF. The outlet temperature of coolant from the actual system is compared with the outlet temperature sensor reading from simulation heat exchanger circuit in the figure 6.12. There is variation in temperature because the starting temperature was set at 18 deg C in the simulation. But the actual system starting temperature is 22 deg C. Pressure in the Hydraulic circuit: There is pressure build up in the circuit as result of flow resistances. The pressure values between the actual and the pressure reading from a sensor after the hydraulic pump in simulation are compared in figure 6.13. The huge variation in actual pressure signal is because of system dynamics while in operation and also the presence of air in the oil circuit.

47

34 32

Temperature degC

30 28 26 24 22 Actual Simulation

20 18

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.12: Heat Exchanger – coolant out temperature vs. time

2 1.8 1.6

Pressure bar

1.4 1.2 1 0.8 0.6 0.4 Actual Simulation

0.2 0

0

500

1000

1500 Time S

2000

2500

3000

Figure 6.13: Pressure variation in the Hydraulic circuit vs. time

48

Hydraulic oil flow: The below figure shows the variation of flow pattern from the hydraulic pump and the flow to the stator and rotor.

Figure 6.14: Flow in hydraulic circuit vs. time

6.2 Plots – Drive Cycle 2 This is a yaw damping drive cycle. In this case, the vehicle is subjected to sharp turns and skids. The functioning of the dynamics of the vehicle is tested. Here, the electric motor is heated up more than the previous case. The initial temperature of all the components are not at a normal temperature, but at an increased temperature and vary with the component.

49

Stator temperature: 115 Actual Simulation

110

Temperature deg C

105 100 95 90 85 80 75 70

0

20

40

60

80 100 Time S

120

140

160

180

Figure 6.15: Stator winding temperature vs. time This temperatures of actual system and simulation model are bit close that the previous drive cycle. Variation at some points are because of the assumption made in 2D loss table. Hydraulic oil flow:

Figure 6.16: Flow in hydraulic circuit vs. time

50

6.3 Plots – Drive cycle 3 This is same as the drive cycle 2 but with increased operation time. Stator temperature: 80 Actual Simulation

Temperature deg C

75

70

65

60

55

0

50

100

150

200 Time S

250

300

350

Figure 6.17: Stator winding temperature vs. time Hydraulic oil flow:

Figure 6.18: Flow in hydraulic circuit vs. time

51

400

Chapter 7 Conclusion The best words to sum up this thesis work and the whole report is ‘Heat up and cool down’. In retrospect, the problem aroused when the hybrid drive encountered heat generating sources like electric motor and mechanical members such as bearings, gears etc. As demanding conditions were tried out on the drive, the temperature curves kept climbing upwards. Thanks to transfer theories, they were kept below threshold limits. Simulation model development was done step by step by analyzing problems in pieces. This approach reduces the number variables to keep an eye on, while simulating. When all the heating sources were modeled satisfactorily, they were connected together to form a complete cooling circuit. This model satisfactorily predicted transient temperature evolution in the drive. Importance was given to radial distribution of temperature i.e. from rotor shaft to drive casing. Thermal contact conductance, convection areas, power losses and coolant flow are the major variables which can make difference to the final result. Some of these variables are hard to figure out, because of the complex nature of heat transfer problems as discussed in the report. CAD models are very important for area calculations. Convection heat transfer coefficients for complex systems like this are being investigated all around the world, so sensible approximations were made from reliable sources. Simulation results followed the same pattern as the actual ones. So from results section, it can be concluded that the model had around 10% errors in temperature prediction. A ESim helps by reducing Engineer’s effort on mathematics and programming, so that physics of the problem can be thoroughly analyzed. By keeping this developed model as base, further strategies related to heat management can be formulized. Example cases: coolant pump optimization, changing the coolant, dimensioning the casing for better free and force convection, oil cooling in tank, flow components optimization, etc., Hopefully, this investigation bolsters the company by reducing the burden of having to produce more prototypes before going into production.

52

Chapter 8 Future work Based on the studies done so far, following works may be investigated in the future: 1. Planetary gears (and bearings) can be modeled using available AMESim sub models instead of lumped mass approach as done in this work. Power loss method can be avoided. 2. A model including the torque vectoring motor can be simulated. 3. New sub models of external surface convection can be created using AMESet or AMECustom and used in simulation. 4. A complete secondary circuit with other components can be built, on obtaining the losses values of the other components. 5. Distributed mass modeling can be implemented. 6. Some innovative (out of box!) heat management strategies can be tested.

53

References 1.

Yunus A. Cengel, Afshin J. Ghajar, “Heat and Mass Transfer: Fundamentals and Applications” Fourth edition, McGraw – Hill, ISBN-13 9780077366643, 2011.

2.

LMS Imagine.Lab AMESim – online reference – help (last access – June 2011)

3.

“http://www.reliance.com/mtr/mtrthr.htm” (last access - May 2011)

4.

“http://www.rficdesign.com/hysteresis-and-eddy-current” (last access - May 2011)

5.

“http://www.mathworks.com/help/toolbox/physmod/sdl/ref/planetarygear.html” (last access - May 2011)

6.

“http://nptel.iitm.ac.in/courses/IIT-MADRAS/Machine_Design_II/pdf/2_7.pdf”(last access - May 2011)

7.

“http://www.mechanics.iei.liu.se/edu_ug/tmme14/utdelat7.pdf” (last access - May 2011)

8.

“http://engr.bd.psu.edu/rxm61/MET210/Lecture/Chapter10_helicalgears.pdf” (last access - June 2011) “http://science.howstuffworks.com/transport/engines-equipment/gear7.htm” (last access - June 2011)

9.

10. “http://woodgears.ca/gear/planetary.html” (last access - June 2011) 11. F.L. Litvin, A. Fuentes, D. Vecchiato and I.G. Perez, “New design and improvement of Planetary gear trains “ F.L. Litvin, A. Fuentes, D. Vecchiato and I.G. Perez, University of Illinois, Chicago, 2004. 12. “http://www.ee.iitkgp.ernet.in/~soumitro/dops/chap1.pdf” (last access - April 2011) 13. “http://www.20sim.com/webhelp/modeling_tutorial/friction/staticdynamicphenome na.htm” (last access - May 2011) 14. “http://www.skf.com/skf/productcatalogue/jsp/viewers/imageViewerJs.jsp?image=0 101f01.gif&file=1_0_1b&maincatalogue=1&lang=en” (last access - May 2011) 15. www.skf.com – Interactive Engineering Catalogue – Rolling Bearings – Principles of bearing selection and application (last access - May 2011). 16. Frank M. White , “Fluid Mechanics”, Fifth edition, 2003, McGraw – Hill, ISBN 0-07240217-2.

54

17. “http://www.pipeflow.co.uk/public/articles/Laminar_And_Turbulent_Flow.pdf” (last access - April 2011) 18. Matthijs Klomp, “ assenger car All-Wheel Drive analysis” 19.

aoru Sawase, Yuichi Ushiroda, Takami iura “ eft-right torque vectoring technology as the core of super all wheel control (S-AWC ” itsubishi technical review , No-18, 2006. 20. http://www.nets.rwthaachen.de/content/teaching/lectures/sub/simulation/simulationSS06/slides/09.pdf (last visit - June 2011)

Additional References 1.

R. Meillier, A. Brizard, R. Resch, “1D Simulation of Lubrication Systems and Heat anagement Strategies“ from ölkreislauf von verbrennungs motoren II, pp 40-65, Expert Verlag, ISBN 978-3-8169-2671-9, 2007.

2.

C.A. Cezario, M. Verardi, S.S. Borges, J.C.D. Silva, A.A.M. Oliveria, “Transient Thermal Analysis of an Induction Electric Motor“ 18th International congress of Mechanical Engineering, Nov 2005, WEG Industries – Motor/UFSC. Brazil.

3.

J. Lindström, “Thermal Model of a Permanent Magnet Motor for a Hybrid Electric Vehicle” Report No. 11R, ISSN: 1401-6176, Department of Electric Power Engineering, Chalmers University of Technology, Göteborg, Sweden, April 1999.

4.

F. Marquez-Fernandez, A. Reinap, Z. Huang, M. Alakula, “Dynamic Evaluation of the Overloading Potential of a Convection Cooled Permanent Magnet Synchronous otor” Department of Industrial Electrical Engineering and Automation, Lund University, Sweden, 2010.

5.

F. Freudenstein, A.T.Yang, “Kinematics and Statistics of a Coupled Epicyclic Spur-Gear Train” echanism and achine theory, Vol.7, pp 263-275, Pergamon Press, 1972 .

6.

J. Angeles, C. Chen, “Virtual-Power Flow and Mechanical Gear-Mesh Power Losses of Epicyclic ear Trains”, Department of Mechanical Engineering and Centre of Intelligent Machines, McGill University, Montreal, 2007.

7.

J. Ottosson, . Alakula, D. Hagstedt, “Electro-Thermal Simulations of a Power Electronic Inverter for a Hybrid Car”, Lund University, 2011.

8.

“http://knowledgebase.nativedynamics.com.au/getArticle?article=020002&topic=2” (last access – June 2011)

9.

“www.engineeringtoolbox.com” (last access - April 2011)

55

10. M. A. Laughton, D. F. Warne, “Electrical Engineer’s eference Book Sixteenth Edition”, Newnes, ISBN 0 7506 46373, 2003. 11. “http://www.kxcad.net/SolidWorks/COSMOSWorks_online_help/analysisbackground /thermalanalysis/thermal_contact_resistance.htm” (last access - May 2011) 12. “http://www.mhtl.uwaterloo.ca/old/onlinetools/airprop/airprop.html“ (last access May 2011) 13. “www.sauer-danfoss.com/stellent/groups/publications/documents/product_litera ture/520l0463.pdf” (last access - April 2011) 14. “http://www.ndfeb-info.com/characteristics.aspx” (last access - May 2011) 15. “http://www.mmcmagnetics.com/ourproducts/main_NdFeB.htm” (last access – May 2011) 16. Erik Nordlander, “Modelling and Validation of a Truck Cooling System”, Department of Vehicular Systems, Linköping University, LiTH-ISY-EX--08/4119—SE, 2008. 17. E. . Booser editor , “Tribology Data Handbook”, CRC Press LLC, ISBN 0-8493-3904-9, 1997. 18. ABB Industrial Systems, Inc. – Technical Guide No. 102, 1997. (last access June 2011)

56

Appendix 1. Thermal model of Electric motor

57

2. Convection and conduction area values in the electric motor Sl. No. 1 2 3 4 5 6 7 8 9

Specification Stator case – oil cooling area End windings – oil cooling area Contact area – Stator ID and air Contact area – Rotor OD and air Contact area – insulation paper and windings Contact area – insulation paper and stator iron Contact area – magnet and rotor iron Rotor – oil cooling area Contact area – Stator case and Housing

Area, mm2 100468.13 141005 53042.65 52703.36 276480 146880 43200 31045.22 15582.3

3. Losses in bearings Sl. No. 1 2 3 4

Location Main shaft end bearings Carrier – shaft bearings Planet gear bearings Output shaft bearings

Bearing Type Needle roller Ball Needle roller Ball

Power loss 89.69 e-5 * NEM * TEM 357.66 e-5 * NEM 58.44 e-5 * NEM * TEM 95.218 e-5 * NEM

4. Details of Heat exchanger Sl. No. 1. 2. 3. 4.

Data Mass Wall thermal resistance Oil convection area 50 % Ethylene glycol convection area

5. Thermal model of Secondary circuit in AMESim

58

Value 0.517 Kg 0.004 K/W 19695.9 mm2 8158.72 mm2

6. Secondary circuit Input: Flow input – 50% ethylene glycol:

5

4.5

Flow lit/min

4

3.5

3

2.5

2

0

500

1000

1500 Time S

2000

2500

3000

2000

2500

3000

Temperature input – 50% ethylene glycol:

32 30

Temperature degC

28 26 24 22 20 18 16

0

500

1000

1500 Time S

59

7. Thermal simulation model

60

8. Thermal contact conductance[1] values for metal contacts

9. Drive cycle 1 - Propulsion motor speed:

5000 4500 4000

Speed rpm

3500 3000 2500 2000 1500 1000 500 0

0

500

1000

1500 Time S

61

2000

2500

3000

10. Drive cycle 1 - Propulsion motor torque:

200 180 160

Torque Nm

140 120 100 80 60 40 20 0

0

500

1000

1500 Time S

2000

2500

3000

1500 Time S

2000

2500

3000

11. Drive cycle 1 - TV motor speed:

900 800 700

Speed rpm

600 500 400 300 200 100 0

0

500

1000

62

12. Drive cycle 1 - TV motor torque:

2.5

Torque Nm

2

1.5

1

0.5

0

0

500

1000

1500 Time S

2000

2500

3000

1000

1500 Time S

2000

2500

3000

13. Drive cycle 1 - ISG speed:

6000 5500 5000

Speed rpm

4500 4000 3500 3000 2500 2000 1500 1000

0

500

63

14. Drive cycle 1 - ISG Torque:

18 16 14

Torque Nm

12 10 8 6 4 2 0

0

500

1000

1500 Time S

2000

2500

3000

15. Drive cycle 1 - Speed input of electric motor that operates the hydraulic pump:

800 700

Motor Speed rpm

600 500 400 300 200 100 0

0

500

1000

1500 Time S

64

2000

2500

3000

16. Program used to vary the frequency of data: function y=separate(signal_name) j=1; y=[]; for i = 1:26500 if mod (i,10) == 0 y(j,:) = signal_name(i,:); y(j,:)= abs(y(j,:)); j=j+1; end end % disp(y); End

65