ELEVES: A NEW SOFTWARE TOOL FOR ELECTRIC VEHICLE MODELING AND SIMULATION

Vehicle simulation software is essential to vehicle design and develo p ment. Most current vehicle software simulations, however, require that com p onents b e s p ecified p rior to running the simulations. Within this context, this p a p er p resents a simulator offering to users flexi b ility enough to add new com p onent and use different level of accuracy. Thus a first version of a new software tool, in order to assist in develo p ing technologies for Electric Vehicle (EV), is descri b ed. The ELEctric VEhicle Simulator (ELEVES) is currently b eing develo p ed as a research p ro j ect in the University of Technology of Belfort-Mont b éliard (UTBM) b y our la b oratory Systems and Trans p ortations (SeT). This simulator is develo p ed on the b ase of the Com p onent Hy b rid Dynamic N ets (CHD N ). As a pp lication, the p ro p osed software tool has b een used to modeling and simulation of four motorized wheels electric vehicle. Finally, a com p arison b etween the results o b tained b y ELEVES with those o b tained b y Matla b / Simulink software tool has b een included.


INTRODUCTION
Transportation is a primary source of energy consumption and pollutant emissions worldwide [1].In response to concerns over environmental degradation, several government agencies, laboratories and the major automotive manufacturers proposed a progressive increasing number of automobiles to be Zero Emission Vehicles (ZEVs). Until now, only electric, low emission hybrid vehicles, and fuel cell powered vehicles are able to response to these regulations [1] - [2]. Historically, during the development of the EV, their power train configurations have been the same as those of the conventional internal combustion engine type vehicle (ICEV), which are merely based on a single motor with a mechanical transmission system. Only in the recent years and due to the progress on electric motors and motor controller technology, alternative modern configurations have been considered. One of the latest configurations is known as motorized wheels, which consists of either two or four motors that are integrated into each drive wheel and can be controlled independently. This configuration was considered one of the best design alternatives to the Ford Ecostar power train. It has also been used by Nissan in their concept vehicle, known as FEV (Future Electric Vehicle). The motorized wheels configuration will be adopted for both the General Motors electric car and the hybrid electric vehicles in the very near future [3].
The aim of this work is to present a software tool that has been developed in our laboratory especially for simulating the in-wheel electrical vehicles. The adopted methodology in the modeling process allows to our software, to easily include other architectures of EV (hybrid EV, mechanical transmission for example).
We start this paper by displaying the limits of existing EV simulators. Then, a description of the global algorithm of ELEVES is established. A description of the principles and the applied hypotheses, to model the different vehicle units, are explained. So, the paper highlights the modeling and simulation of four motorized wheels EV as a typical application of the tool. Finally, in order to check the validity of the obtained results, a comparison study between ELEVES and Matlab/Simulink software is presented.

LIMITS OF EXISTING SIMULATORS
Several software of simulation of the dynamic and energetic performances of the vehicles was carried out since of tens of years in the research centers of manufacturers or in the university laboratories. We count some of the most known as ADAMS/Car, SmartHEV and ADVISOR. Adams is a motion simulation solution for analyzing the complex behavior of mechanical assemblies. So, it allows testing virtual prototypes and optimizes designs for performance, safety, and comfort [4]. The HEV design software, SmartHEV, is a flexible and robust model of steady state HEV operation [5]. These modeling tools focus on only a portion of the analysis of the vehicle, and don't treat the transient and unbalanced regimes. For ADVISOR, it was developed as an analysis tool, and not a design tool. Its component models are quasi-static, and should not be used alone to predict phenomena with very small time scales. ADVISOR, however, has been linked with other tools that do perform dynamic analysis capability. Tools like SABER for electrical systems modeling and ADAMS/Car for vehicle dynamics have been linked with ADVISOR [6] - [7]. In spite of the obtained new possibilities, this tool multiplicity penalizes generally the user by adding a supplementary complexity and cost in terms of time and money. So, for these reasons one can say that it's more attractive to have a unique tool entirely dedicated to electric vehicle modeling and simulating problems as well in static studies as in dynamic ones.

ELEVES STRUCTURE
ELEVES is an EV component design tool written in Delphi programming language under the environment of Windows. It uses algorithms currently under development at the University of Technology of Belfort Montbéliard (UTBM). This simulator uses the component hybrid dynamic nets (CHDN) [8] - [9] to represent the models and the interaction within them. The main advantage of ELEVES over those mentioned earlier, it is that it allows achieving a high degree of modularity, in order to be able to model new types of component and integrate them easily within the full vehicle model. This modularity includes the vehicle dynamic and its electrical devices (sources, electrical machines and power converters) at different levels of accuracy. Modeling the vehicle by ELEVES involves modeling its different components and the interaction between them. We will study in this article a dynamic model of which we suppose that the vehicle starts of the stop position until it reaches the steady state. During its race, it changes velocity, takes turn and climbs a slope. Moreover, we can treat critical or particular cases like operation in degraded mode and unbalanced regimes. There are also a number of secondary aims, in particular, achieving a good accuracy, providing different levels of modeling for a given type of component and finally, creating a database on EVs.

Fig.1 Block diagram of ELEVES
The graphical user interface of ELEVES shown in Fig. 1, allows selecting the vehicle configuration and its type of electrical motorization but also to select them energy supply including the associated converter.
The three types of configurations actually treated by ELEVES are the integral traction, the traction and the propulsion. In the integral traction all wheels are motorized wheels, but only the front wheels are driving in the traction, and only the rear wheels propel the vehicle in the propulsion configuration. In all these cases, the steering applies only on the front wheels. The user has the possibility to select between the DC motor, the induction motor and the Permanent Magnet Synchronous Motor (PMSM). The electric motor is driven by a power electronics component that converts the fixed DC voltage available from a source into a variable voltage. Although the structure of ELEVES is primarily conceived for Battery EV, it regards future energy sources like fuel cell and the associated storage device as for example the ultracapacitors. The principal algorithm of ELEVES is illustrated in Fig. 2. The adopted specification of different components performances are linked to design and to parameters control, on which the user can act in order to get a desired performance.

FOUR MOTORIZED EV MODEL
In this section, the governing differential equations of motion, the analytical expression for the kinematics and dynamics of the vehicle model have been derived. The analyses, which are presented separately, have been carried out for both dynamic and electric components to clarify the understanding of the vehicle performance. We show also the CHDN model for both electrical and dynamic components of our system in order to achieve the needed simulation.

Dynamic Model
With reference to the model, illustrated in Fig. 3, and by neglecting the effect of roll angle of the sprung mass about the roll axis and the effect of camber angle, the equations of motion of the vehicle are [13]: M are respectively the total forces acting in X and Y directions and the total yawing moment. Although there are many ways to obtain the mathematical model, a procedure based on Lagrange equations will be followed here. Neglecting the rotational kinetic energy of the wheels the kinetic energy of the vehicle is as follows: (3) and the equations of motion can be obtained by the relation: Using a linear tire model, the front and rear cornering forces can be expressed as the product of the cornering stiffness (  We can obtain the three governing equations of motion by finding the resultant of forces acting on vehicle in y x, directions and the total yawing moment about the Center of gravity CG , then substituting in the system of equations (Eq. 5) with taken in consideration the cornering forces and the sideslip angles. The longitudinal, lateral and angular accelerations of vehicle are represented as follow.
are the tractive forces of motors, res F is the resistant forces which including the aerodynamic drag force a F , climbing force c F and rolling force r F . They have the following expressions: In the set of equations (9) px C is drag coefficient, f A is the cross section area, ρ is air density, cg V is vehicle velocity at the CG, g is the acceleration gravity, p α is the grade angle of the road ad rr C is the rolling resistance coefficient.
By applying a driving force to a tire which has a certain slip ( s ), the longitudinal coefficient ( μ ) is influenced by the wheel sideslip angle ( r f α α , ). With reference to Fig. 4, we can see that the peak value of longitudinal characteristic decreases with the increasing of the sideslip angle.
Moreover, the longitudinal slip can be defined for the four wheels as: Where w R is the wheel radius, ω is the angular velocity of the in-wheel motor and t u is the linear speed at which the contact zone moves on the ground and can be written for the four wheels as [3]:

Electric Motor
For the wheel direct drive, especially with in wheel motors, the Permanent Magnet Synchronous Motor (PMSP) is the reasonably, if not only possible solution, due to its relatively low mass which contribute to reduction of unsprung mass and to better drive comfort [14]. The PMSM is supplied by a battery b E through a DC/AC converter developed in [8] - [9]. The drive system model can be described by the following electrical and mechanical equations:

SIMULATION RESULTS
In order to analyze the dynamic characteristics of the vehicle and to study its main performance, the developed model of EV has been implemented. The simulated system is illustrated in Fig. 7.
With an aim of comparing our results with those obtained by existing software, we implemented our system under Matlab/simulink software. The comparison will be presented progressively in this section. Fig. 7 The simulated system The four motors are controlled by a sinusoidal reference current ref I fixed at the amplitude of (100 A) through an inverter supplied by DC bus b E of (540 volt). The parameters of the PMSM are given in Table I. The main geometrical data and inertial properties of the vehicle and wheels are shown in Table II.

.m kg
The vehicle is assumed to start running at a variable speed from 0 second to reach its steady state speed at 7 second and then go into a turn, as a front steering angle o 7 = δ shown in Fig. 8. The resulting output for the vehicle body velocities is shown in Fig. 9 and Fig. 10. It may be seen that the longitudinal speed of vehicle remains at the steady state until the turn begins and the friction starts to dissipate energy, resulting in a lower speed at the completion of the turn. The lateral vehicle body velocity increasing with the steering as does the angular velocity. The Fig. 12 shows the difference in the tractive forces of motors during the completion of turn. This also clearly seen on the speed of motors (Fig. 13) and on the resistant torque of motors (Fig. 14). Where the speeds of motors on the left side and the right side diverge as the left wheels slow down and the right wheels speed up to traverse the turn. The forces of traction, the speeds of PMSMs and the resistive torques to the movement of the vehicle obtained, using the Matlab/Simulink software, are presented on the figure (15). In the same way, we see that these three variables are similar to those obtained by ELEVES. The vehicle body states may be integrated to obtain the absolute vehicle position for simulation output. Since the vehicle body states are measured in the vehicle coordinate frame, a transformation is necessary prior to integration. By integrating the system mentioned in (Eq. 2), we have the absolute position of vehicle. The Fig. 16 illustrates the distance traveled by the vehicle in inertial reference frame ( Y X , ). It is identical with that obtained by Matlab/Simulink as shown in Fig. 17.

CONCLUSION
In this paper, we have presented the first version of a developed simulation tool for predicting the dynamic and energetic performances of electric vehicles. This tool, which is called ELEVES, has a high degree of modularity, in order to be able to model by component the vehicle system whose architecture is becoming increasingly complex. It has also several degrees of accuracy regarding to the modeling level. The simulation results of a four motorized wheels EV have been compared with those obtained with Matlab/Simulink software. This study shows a good agreement between the two software tools. However, knowing that ELEVES is dedicated entirely to electrical vehicles and it offers a graphical programming more practice than that of Matlab/Simulink, one can say that it is better in the area of dynamic electrical vehicles simulating for both design and control targets.