1. Introduction
Practical studies are important for learning in engineering education. For effective learning, theoretical knowledge should be supported by practice in the closest way to reality. Today, practical education is carried out with computer-based programs and experimental studies in laboratories. Although simulation programs support theoretical knowledge, they are not sufficient for effective learning. On the other hand, laboratory training provides the closest training to reality, while improving the hand skills of the engineering students. It is possible to show the components of a system practically in laboratories with separate modules. However, this method is not enough to understand the whole system. Therefore, a compact test setup is needed for special systems. There is a need for a test setup in laboratories to make the EV technology, which has become widespread these days, more understandable.
Electric vehicles (EVs), which were invented towards the end of the 19th century, have been subject to more demand than fossil fuel vehicles thanks to their ease of use and comfort. However, in the middle of the 20th century, interest in EVs gradually decreased due to the mass production of internal combustion engines (ICE), their low cost and long driving distances. In the middle of the 20th century, with the effects of the world economic crisis and the emergence of global warming problems, the search for alternative energy sources became necessary. Thanks to the rapid development of technology in these years, countries became interested in EVs as a solution to reducing their dependence on oil and reducing carbon emissions [
1].
In the last decade, automotive companies have not been indifferent to the interest in EVs, and have launched hybrid vehicles to overcome the range problem of EVs. With the increasing interest in EVs and the development of battery technology, pure EVs have also started to be produced. However, the energy demand increases day by day because of the intense interest in EVs. According to the International Energy Agency (IEA) 2020 World Energy Outlook Report, while the energy demand of the transportation sector was 51 TWh in 2018, it is expected to be 1551 TWh in 2040. Considering that the demand for electrical energy will increase in other sectors similarly to the transportation sector, it is not sustainable to meet the demand with alternative energy sources. In a statement made in the June 2020 IEA report seeking to ensure the energy supply-–demand balance, the importance of designing more efficient vehicles in the electricity sector was stated [
2]. In addition, the global energy crisis nowadays affects the entire transportation sector. This crisis threatens new investment plans in the EV sector with the increase in energy costs. Therefore, in order to expand the interest in EVs, the cost of travel with EVs must be reduced. To reduce energy consumption in EVs, first, it is necessary to design more efficient and improved EV components. In addition, optimized driving algorithms can require less energy consumption. Thus, by increasing the driving efficiency, people can travel longer ranges with a limited energy capacity.
EVs contain several components, such as a motor, a battery, a power electronics driver, a transmission mechanism, and other equipment. Generally, the electrical energy stored in the battery is converted into mechanical energy by the electric motor, and the drive is performed by transferring it to the wheels of the car thanks to the powertrain. The speed and acceleration of the vehicle depend on the traction power, vehicle mass, rolling resistance, aerodynamic drag, hill resistance, and the physical specifications of the vehicle.
Considering the above issues, simulation models, such as equation-based models [
3], forward models [
4], car-following models [
5] and predictive models [
6], have been designed for a better understanding of EV dynamics. Besides this, six different drivetrain systems have been presented in Ref [
7]. A comparative study showed that among these systems, the single-level geared topology is the most favored drivetrain system for electric motor drives, with a wide speed range and high maximum speed in EVs. Additionally, considering this drivetrain system, the requirements of customers and the motor characteristics of EVs are presented in the study. In another study conducted during the years of increased interest in EVs, DC, induction, reluctance and permanent magnet motors, recommended for EVs, were compared. As stated in that study, DC motors, which initially attracted great attention due to their ease of driving, have been replaced by AC motors, thanks to the development of motor drivers and effective control methods [
8]. Additionally, the advantages of DC motors, brushless DC (BLDC) motors, permanent magnet synchronous motors (PMSMs), line-start PMSMs, induction motors (IM), and switched reluctance motors (SRM) have been investigated. It was reported that the motors including magnets are highly efficient, and their performances are more suitable for EVs [
9,
10,
11].
On the other hand, inner- and outer-rotor-type motors have been used, depending on the transmission topology. While inner-rotor electrical motors are used in topologies such as fossil fuel vehicle designs, outer-rotor electrical motors are preferred in in-wheel designs. In a study in which an outer-rotor permanent magnet motor was designed and tested for EVs, it was emphasized that this motor was highly efficient, and more suitable for EVs [
12].
Another important component studied in EVs is the battery, which is the energy source required for movement. EVs, in contrast to fossil fuel vehicles, can convert the mechanical energy into electrical energy and store it in their batteries for reuse. The mechanism by which energy recovery takes place is called regenerative braking, and regenerative compression stabilizes the drive. These involve controlling the electrical motor in the generator region during compression or braking mode operation. The bidirectional flow of electrical energy has great significance for energy management in EVs. In recent studies, highly efficient energy management systems using regenerative braking have been used [
13,
14]. In such studies, the maximization of the state of charge (
SoC) was investigated [
15]. The
SoC cannot be measured directly; therefore, it is estimated by many different mathematical methods using some measurable or sensible variables. These methods are classified into four groups. One of these methods uses the direct measurement of the battery voltage and its impedance. Another is the bookkeeping estimation method, which calculates the
SoC by integrating the discharging current. The other is an adaptive system based on the neural network. The most commonly used one is a hybrid of these methods, which, when compared to the others, gives more accurate results, as presented in [
16].
For validating the results obtained by the simulations, there are many methods of testing the vehicle components. The V-models used for the developing and testing of vehicles in the automotive industry are as follows, listed from low to high in terms of accuracy; Model-in-the-Loop (MIL), Software-in-the-Loop (SIL), Hardware-in-the-Loop (HIL), and Vehicle-in-the-Loop (VIL). In general, it is acceptable to complete the MIL-SIL-HIL test processes in the development of the vehicles [
17]. In the literature, there are some studies employing these methods on fossil fuel vehicles, considering the car driver’s behavior, road slope, and load factor and road type [
18,
19]. In addition, all of the EVs’ components have been analyzed in detail, since EVs are used in wide speed ranges, variable temperatures, different road conditions, and with different driving characteristics [
20,
21,
22,
23,
24,
25,
26]. In [
27], motor, battery, and energy management systems are investigated using the HIL method to derive more information with high accuracy in relation to the vehicle’s dynamics with real components.
Nowadays, there is an increasing interest in EVs. There are many new components, both mechanical and electrical, in an EV, and each of them should be examined separately and tested under real road conditions with minimal assumptions. However, in the engineering education given in many universities, application laboratories have not yet been developed with this ability. For this reason, only theoretical educational support is offered to students. The experiential learning theory, an innovative approach, was presented by Kolb in 1984. According to Kolb, this learning style includes experience, perception, cognition and behavior. First, in concrete learning, the student encounters a new experience or reinterprets an existing experience. Then, in reflective observation, the learner personally reflects on his/her experience. After this, the learner reaches the abstract conceptualization stage by forming new ideas or modifying existing abstract ideas based on their reflections. Finally, in the active experimentation stage, the student applies the new ideas to see if there are any changes in the next experience. In this study, an EV test system was designed to provide the convergence of the learning cycle for students who have completed the previous stages. With this test setup, students can observe the practical uses of the theoretical knowledge they have acquired about EVs.
It is difficult to understand vehicle dynamics by performing separate tests on the components of EVs. Unfortunately, testing all the components in a real environment is unsafe and involves high costs. Therefore, a test setup should be designed to observe the effects of all parameters under real-world driving conditions. Within the scope of this study, a test setup has been prepared to be used in the laboratory for use by both academicians in their studies and engineering students seeking to obtain learning experience about vehicle dynamics, considering criteria such as cost, safety, portability and flexibility. The test setup consists of a battery, a traction motor, a servomotor, drivers, a user interface, a Programmable Logic Controller (PLC), a PC and sensors. Firstly, maintenance-free, inexpensive, reliable and high-performance 12 V gel batteries, which are frequently used in light EVs, were selected for the energy storage and source. Similarly, a BLDC motor with an outer-rotor structure, as used in many EVs for maintaining high torque, was included in the design. The BLDC motor driver was used to control the speed and torque through a PLC. The servomotor, connected to the same shaft, was used to represent the forces acting on the EV. In addition, a torque sensor was placed on the shaft to observe the torque developing in the traction motor. Thanks to the human–machine interface (HMI) connected to the PLC, the user could change the vehicle parameters, road slope and driving profiles as desired. In addition, it was possible to monitor the status of variables on the HMI screen during the test. After that, when the test was completed, the test data were saved to an external disk or on a computer hard drive. In addition, this device has a flexible structure, and it can be programmed using the available software, such as Python, by making a connection with the PLC and the computer. In this way, it is possible to test complex algorithms on the system, depending on the users’ level of knowledge. Moreover, the test setup was designed modularly so that different motors and their drives could be tested if desired.
In this study, a new test platform for EVs is presented to test both the motor and power electronics components with full integrity under real road conditions. In this way, an application platform suitable for all kinds of research, from the most basic level of knowledge to the most advanced research topics, has been obtained. Testing both expert systems and learning-based control algorithms under real test conditions is possible thanks to this platform. On the other hand, it also allows the testing of regenerative brake-assisted driving cycles to increase the driving range and energy efficiency of EVs. Thus, the energy efficiency between driving with a regenerative brake and driving directly can be easily observed. In this context, there is a need for an EV test system that is suitable for undergraduate student, as well as being low-cost and including basic EV components. For this reason, an EV simulator has been designed with the advantages and innovations mentioned in this study.
3. Designing of the Test Setup
In specialized fields, educators want their students to not only learn or memorize the theory, but also learn how to use their theoretical knowledge in practice. An applied course is indispensable for consolidating the theoretical knowledge and achieving more learning that is permanent. A test setup can provide undergraduate students with an important learning experience. For this reason, there is a need for test systems that will prepare the student for the future, and make the subject clear and understandable. In this study, considering the EV components, an experimental test setup is designed and implemented in accordance with an engineering education curriculum.
The proposed test setup consisted of a battery group including the management system, BLDC motor, BLDC motor driver, servo motor, servo motor driver, energy analyzer, torque sensor, PLC, and HMI, together with all the communication hardware and protocols. The devices included in the test setup and their intended uses are briefly shown in
Table 1. An industrial HMI is preferred for operating the system, making settings, observing, and recording the data obtained. Firstly, the test was conducted by selecting the EV parameters, the desired references and the type of the driving cycle. Then, while the PLC unit was sending the speed reference data to the BLDC driver, it was also sending the torque data to the servo motor driver, depending on the driving profile and the physical characteristics of the vehicle. Thus, power transfer occurred between the BLDC motor and servo motor coupled on the same shaft. Thanks to the bidirectional BLDC motor driver, the BLDC machine worked as both a motor and a generator. When the BLDC machine works as a motor, energy flows from the batteries to the motor. On the contrary, when it works as a generator, the energy recovered by the wheels is stored in the batteries. In our system, six serially connected 12 V-60 Ah batteries were as the energy source, as per laboratory conditions. An HIOKI PW3390 model power analyzer was included in the system for monitoring and recording all data obtained from the current and torque sensors. On the other hand, the whole system can be managed and motorized through a computer, and was observable for the analysis of the results. Moreover, due to its flexible features, we could incorporate different driving profiles, different motor types and complex algorithms. The general scheme of the system is given in
Figure 5a.
The servo system applying the forces acting on the EVs is fed by a three-phase AC network. When the BLDC machine functions as a motor, the servo motor works as a generator, and the produced energy is converted into heat energy due to the braking load resistor that protects the servo driver. All system components were combined on a set. This is shown in
Figure 5b.
To control the test system and observe the data, the training–control interface is illustrated on the touch panel. Through the parameter input menu shown in
Figure 6a, students can set the vehicle’s parameters, the driving cycles and the regenerative braking mode.
Figure 6b shows the expert menu of the test system wherein the servo system and interface settings are changed. In this page, the servo motor can be adjusted independently in the torque and speed modes. The servo control button can also use the training test setup as a manual motor test unit. In addition, the PID parameters can be updated on this page to avoid torque fluctuations of the servo system and to ensure torque accuracy in accordance with the torque reference.
As shown in
Figure 7a, it can be seen that after the necessary data are applied, automatic driving is started, and the system data can be displayed on the data view page. Reference torque-speed and instantaneous torque-speed characteristics can be observed via this page during automatic operation, and the obtained data are saved to the Universal Serial Bus (USB) memory at the end of the driving cycle. In addition, a graph of these recorded data can be viewed instantaneously through the HMI page shown in
Figure 7b.
The braking system’s working principle is shown in
Figure 8. Firstly, the vehicle and road parameters are entered using the HMI. Then, the driving is started and the required torque is calculated. If the torque demand is negative, it can be checked whether the regenerative mode is active. If the regenerative operation mode is active and the
< 90%, the regeneration coefficient is calculated and the brake distribution is performed according to this coefficient. If the regenerative operating mode is passive or
< 90%, the regeneration coefficient is considered zero and all braking is done by the mechanical system. In addition, if there is no negative torque demand, the reference speed is set and these processes continue until the driving cycle is finished.
4. Experimental Studies and Discussion
In this study, several experimental case studies have been composed to show the effects of light EV parameters, and different road slope and energy recovery parameters, on operation of an EV. In addition, EV parameters related to the design are selected by considering the constraints in the experimental set. The vehicle parameters and their values under the test conditions are given in
Table 2.
Many driving cycles or profiles have been created by different countries to test vehicles. One of these profiles is the New European Driving Cycle (NEDC), which is widely used in Europe. Here, the NEDC, consisting of the urban driving cycle (ECE-15) and the extra-urban driving cycle (EUDC), is used for the approval of light vehicles. In the ECE15 driving cycle used for the urban test, the average speed is defined as 18.7 km/h, the maximum speed is 50 km/h and the driving distance is 4 km, and it is completed in approximately 193 s. The vehicle’s speed in the ECE-15 is converted to rotational velocity depending on the wheel radius. According to the vehicle’s parameters, shown in
Table 2, the forces acting on the EV according to the ECE 15 speed profile are calculated based on the EV dynamic equations, and the required torque values are determined.
Figure 9 shows the speed and required torque references of the EV on a straight road for the ECE-15 profile.
4.1. Case Studies
The short driving range is one of the most important problems of EVs. To overcome this problem, optimal energy use is necessary. In addition, the availability of energy recovery in EVs requires the most effective management of energy. The most effective parameters for energy consumption in EVs are the physical specifications of the vehicle, the slope of the road and the speed of the vehicle. For students to understand the effects of these parameters more easily, five different experiments have been set up. The parameters to be applied for these cases are given in
Table 3.
The students participating in the research can easily perform the experiments via the steps given in
Table 4. Thus, students obtain experimental data according to the desired parameters.
ECE-15 is preferred as a driving profile for the speed reference in all experimental studies. The reference speed data are divided into three regions to determine the dynamic behavior of the vehicle. Firstly, the acceleration regions of the EV are labeled 2, 6 and 10. Secondly, the regions wherein the vehicle moves at a constant speed are defined as parts 3, 7, 11 and 13. Lastly, the vehicle slows down in regions of 4, 8, 12 and 14. As shown in
Figure 10, the BLDC machine’s speed is very close to the reference throughout all experiments. The differences due to driving conditions are too small to be considered. Therefore, only one velocity graph from the experiment is shown in
Figure 10. Afterward, the students are asked to compare and evaluate the data from the experiment, such as calculated torque, real torque, instantaneous power, and energy parameters.
4.2. Case Study 1
In case study 1, using the ECE-15 driving profile, our aim is to compare the energy changes and torque demands of the EV during speed changes and when in motion at a constant speed on a flat road without regenerative breaking.
The test graphs are given in
Figure 11, and the observed results of this experiment are evaluated as follows:
The total energy consumption is 26,076 Joules;
Toque demand and energy demand are directly proportional;
The torque demand regions, from the largest to the smallest, are acceleration, constant speed, and deceleration. In the deceleration regions, the motor torque is zero;
In the constant speed regions, the higher the speed, the bigger the torque demand;
Although energy consumption occurs during acceleration and constant-speed motion, no energy is consumed in the deceleration region;
In regions where the reference torque is negative, only mechanical braking is applied.
4.3. Case Study 2
In the second case study, unlike the first case study, our aim is to observe the effects of energy recovery by performing regenerative braking.
The test graphs are given in
Figure 12 and the observed results from this experiment are as follows:
The total energy consumption is 20,319 Joules. However, 2883 Joules was recovered thanks to the regenerative braking;
The torque demand regions, from the largest to the smallest, are acceleration, constant speed, and deceleration. In the deceleration regions, the motor torque is negative;
While energy is consumed during acceleration and constant-speed motion, energy is recovered in the deceleration regions;
While the maximum possible energy recovery is realized, mechanical braking is also applied to adhere to the reference speed;
A decrease is observed in the energy consumption graph in the deceleration regions where energy recovery occurred;
In the regions showing energy recovery, the power is negative. In other words, the energy flows from the electric motor to the battery.
4.4. Case Study 3
Case study 3 aims to observe the effects of the forces acting on the EV under inclined road conditions. To make this parameter more understandable, two different experiments are designed to represent uphill and downhill. In this study, a road with a constant positive slope is tested.
The test graphs are given in
Figure 13 and the observed results from this experiment are as follows:
The total energy consumption is 31,758 Joules, although energy recovery occurs. However, the energy recovery here is lower compared to case study 2, but it is 2580 J;
On an uphill road, the vehicle requires more torque than on a flat road. Conversely, less mechanical braking is required to decelerate.
4.5. Case Study 4
For this test experiment, case study 4, in contrast to case study 3, downhill road driving is performed.
The test graphs are given in
Figure 14, and the results from this experiment are as follows:
There is a very big decrease in the total energy consumed by the vehicle on the downhill road, and the total energy is 8758 Joules. In addition, 3532 Joules is produced as energy recycling and mechanical braking are mostly used;
On a downhill road, the vehicle needs less torque than on a flat road. Conversely, more mechanical braking is required for deceleration.
A high level of energy recycling has been achieved as well. However, the amount of energy lost in mechanical braking is greater.
4.6. Case Study 5
Another parameter that affects the energy requirements in EVs is the weight of the vehicle. In this experiment, it is assumed that the other characteristics of the vehicle do not change, and 50 kg of load is added to the mass of the vehicle.
The test graphs are given in
Figure 15, and the results of this experiment are as follows:
The total energy consumption is 22,966 Joules. The amount of energy recovered is 2496 Joules;
Compared to experiment 2, more toque is demanded at the times of both acceleration and deceleration.
4.7. Experimental Result and Discussion
In EV systems, the regenerative braking system can recover kinetic energy and potential energy to the battery through an electric machine. The main factors influencing the regenerative braking control of EVs are road slope and vehicle weight. The net force acting on the vehicle is shown in Equation (15). Since the aerodynamic properties and speed of the vehicle are similar in all experiments, the aerodynamic friction force does not change. On the other hand, depending on the slope of the road, the uphill resistance force acting on the vehicle varies between negative and positive values. According to Equation (13), the net force acting on the vehicle increases as the vehicle moves uphill. Likewise, the net force acting on the vehicle decreases as the vehicle moves downhill.
In this study, five case studies have been carried out under varying load, road slope and regenerative braking conditions. The results are shown in
Table 5. Firstly, the no-load condition was tested on a flat road without regenerative braking. As a result of the experiment, the total energy was 26,076 J, and the maximum power demand was 590 W. Unlike in the first case, energy recovery was achieved with regenerative braking. Thus, an approximately 22% gain was achieved thanks to regenerative braking. On the other hand, unlike case study 2, the road slope was changed. When the road slope was increased by 3%, the total energy increased by 56%, while the amount of energy recovery decreased by 10%. Conversely, when the road slope was reduced by 3%, the total energy was reduced by 56% and the recovered energy increased by 22%. On the other hand, the fifth case involved adding a load of 50 kg to the conditions of case 2 to observe the effects of the load. As a result of the 25% load increase, there was a 13% increase in total energy. The difference in energy recovered between each state is small because of the motor driver’s limited energy recovery. As a result, the experimental results support the EV’s dynamic equations.
Many case studies performed on this test platform, which was developed to improve engineering education, are presented in a separate section, and the results obtained for each study case are re-evaluated. Vehicle weight, load conditions, road conditions, different driving cycles, useful braking or direct driving situations were examined separately, and the findings are presented under the relevant case study. In particular, the effect of regenerative braking on energy consumption was emphasized for different driving cycles.
6. Conclusions
A test platform for the EVs has been designed and implemented for undergraduate and graduate students of the Faculty of Engineering in this study. The proposed experimental test setup reveals that the whole system’s dynamics can be safely explained to student in practice by following the desired driving cycle road conditions and vehicle parameters, while the designed set emulates the operation of a real EV for each configuration. In this context, firstly, practical knowledge of EV components was presented, and five case studies were designed to observe the impacts of EV weight and road slope parameters on energy recovery. Finally, a survey was conducted with the students who completed the case studies, and their answers are included in the paper.
The energy recovery cases focused on comparing three main variables: the effect of the slope, the effect of energy recovery, and the effect of vehicle weight. The experimental results show that the energy consumption is reduced by 22% in cases of energy recovery on a straight road. On the other hand, when considering uphill and downhill cases, it can be concluded that a 3% road gradient has about a 56% impact on the total energy consumption in the regenerative mode. In contrast, if the total weight of the vehicle increases by 25%, there is an increase of approximately 13%. Considering the forces acting on the EV, the theoretical knowledge supports the experimental results.
On the other hand, the results of the surveys conducted with the students who completed the case studies are discussed. Students’ feedback shows that they believed that the comprehensive teaching method, which combines the basic concepts of theoretical knowledge with hands-on education, can help them better comprehend any EV system and develop their professional knowledge and skills. Considering the development of EV systems, the test platform can be updated to perform advanced experiments. From this, we can develop a philosophy of design that will enable us to learn not only with the basic systems discussed within this training, but also with the unusual and special requirements that continue to arise as student education becomes more sophisticated.
Since the test system can be controlled by software using a computer, many algorithms and driving techniques can be included. In future studies on EVs, artificial intelligence and machine learning algorithms will be applied on the test system presented in this manuscript.