Mechanical Engineering Journal

123456789 Bulletin of the JSME Mechanical Engineering Journal Vol.2, No.1, 2015 Electric bus system with rapid charging at every bus stop using renewable energy Takeshi KAWASHIMA* *Kanagawa Institute of Technology 1030 Shimo-ogino, Atsugi-shi, Kanagawa 243-0292, Japan E-mail: kawashima@eng.kanagawa-it.ac.jp Received 4 June 2013 Abstract In order to prevent global warming, the emissions of greenhouse gases must be reduced. In public transportation systems, electric buses can help to achieve this. However, the deployment of electric buses has been limited to public corporations as a result of the weight of the required energy storage devices. One approach to addressing this problem is rapidly charging the electric bus at every bus stop, thereby reducing the required energy storage and weight. In addition, charging buses using electrical power generated from renewable energy sources could further reduce emissions of greenhouse gases. For such a system, the required energy for each bus route must be estimated and the storage device designed to minimize the weight of the bus. In this study, the feasibility of the proposed system is confirmed by a demonstration experiment using a converted electric minivan. Then, a simulator to calculate the energy consumption of a full-size bus is developed by extrapolating the parameters used in a simulator for the electric minivan, which were validated experimentally. A trial bus-mounted storage device for the route is also designed. Key words : Environmental engineering, Public transportation, Renewable energy, Automobile, Solar power, Electric bus, Boosting charge, Simulator 1. Introduction As a result of the severe accident that occurred in the Fukushima Daiichi nuclear power station after the Great East Japan earthquake and tsunami on March 11, 2011, many people have begun to consider the best way of providing a safe and secure electrical power supply for Japan. Many have suggested that further promotion of renewable energy is required. In addition, the feed-in tariff system for renewable energy was started in 2012, marking the beginning of practical and widespread deployment of renewable energy. However, the electrical power supply from thermal power plants, which burn petroleum or coal, has increased, while that from nuclear power plants has stagnated, and the emission of greenhouse gases such as carbon dioxide has increased. Therefore, achieving the greenhouse gas reduction targets of the Kyoto protocol, which was adopted in 1997 at the 3rd Conference of United Nations Framework Convention on Climate Change in Kyoto to prevent global warming, is doubtful in Japan, even though it was the host country. The prevention of global warming is an urgent issue, and I believe that public bus transportation must take the lead in reducing carbon dioxide emissions. In this work, I focus on a bus system in which buses are rapidly charged at every bus stop, thus allowing the weight of the storage device mounted on the bus to be reduced because of the lower mileage requirement. Furthermore, electricity generated from renewable energy sources near each bus stop could be used to drive the bus to realize a zero-emission public transport system. Particular attention has been focused on the distribution of bus stops to effectively utilize widely distributed and low density renewable energy sources, such as solar energy. In this system, the time of a short stop at a bus stop is assumed to be 20 seconds, during which passengers get on and off. Therefore, the development of a charge-boosting system that charges the bus with sufficient electricity for it to Paper No.13-00085 J-STAGE Advance Publication date : 25 December, 2014 1

arrive at the next bus stop is one of the most important tasks for successful implementation of the proposed system. Additionally, reducing the weight of the storage device mounted on the bus is also a key issue. In this paper, we first describe a demonstration experiment using a model electric bus with current collectors that was converted from a minivan with a gasoline engine, and a model bus stop with contact wires, a power feeding system, and photovoltaic modules. Second, we show that the results of the demonstration experiment are also valid for the proposed bus system. Third, we develop a full-size electric bus simulator by extrapolating the data from the converted electric minivan and use the results to determine the minimum capacity of the storage device mounted on the bus. Then, the specifications of the bus stop are designed. 2. Proposed electric bus system Solar panels Contact wires Current collectors Motor Controller Storage device Controller Maximum power point tracker Storage device Fig. 1 Illustration of the electric bus system rapidly charged at every bus stop using electricity generated from a solar energy source. Figure 1 shows an illustration of an electric bus system that is rapidly charged at every bus stop using electricity generated from a solar energy source. Carbon dioxide emissions can be reduced to about 50% by changing conventional vehicles with internal combustion engines to electric vehicles. However, the mileage of a single charge is only about 150 km owing to the constraint of battery weight, even if a lithium-ion battery is used. Electric vehicles, especially large electric vehicles, are less attractive than conventional vehicles, because the energy density of lithium-ion batteries is approximately 1/100 that of gasoline. As a result, the use of electric buses has been limited to public corporations. Additionally, environmental considerations have led to a re-examination of railway transportation, reflecting a modal shift. This has been applied to buses, more specifically, trolley buses. However, most trolley bus routes have been discontinued because (1) the cost of laying overhead contact wires is high, (2) tall vehicles cannot go through areas where overhead contact wires pass, and (3) the adverse impact on the landscape. In Japan, trolley buses are currently operated on only two lines in the long tunnels of the Kurobe-Tateyama Alpine route, where other vehicles are not permitted and where environmental damage is considered to be the greatest problem for maintaining the landscape. Based on these considerations, the intermittent charging of buses at every bus stop has been proposed (Fujioka, 2006). The bus in the proposed system combines the merits of an electric vehicle with those of a trolley bus. In this system, the electric bus would carry a motor and a small storage device; for example, lithium-ion batteries or capacitors. The storage device need only store sufficient energy for the bus to travel to the next bus stop. The small 2

capacity of the storage device is one of the primary features of this system. Current collectors are also installed, as in a trolley bus. When the bus stops at a bus stop, the storage device is rapidly charged by electricity supplied from the overhead contact wires installed at the bus stop. Charging continues for approximately 20 s, during which time passengers get on and off the bus, and then several further seconds during which the bus accelerates away from the bus stop. The charged electric bus then travels to the next bus stop as an independent electric vehicle. To achieve zero carbon dioxide emissions, this bus must run on electricity generated from renewable energy sources. Renewable energy sources are low density and widely distributed, and therefore, many power plants are required for the effective generation and these need to be distributed; a citywide network of bus stops fits this description. In other words, the bus stop is a suitable candidate at which to generate renewable electricity. Thus, an electric bus system rapidly charged at every bus stop using electricity generated from solar energy source is proposed (Kawashima and Fujioka, 2008), in which the electric power that charges the bus is generated by photovoltaic modules installed near each bus stop. The effectiveness of this system was confirmed by demonstration experiments using a golf cart with current collectors powered by capacitors. A photograph of the model bus and model bus stop is shown in Fig. 2. The voltage of the capacitor was stabilized by a DC-DC converter. The golf cart was able to repeat the cyclic operation of running for approximately 100 s as an independent electric vehicle, stopping for charging for 20 s, and then accelerating with electricity supplied from the contact wires for approximately 2 s. Moreover, the required area of photovoltaic modules at every bus stop was also estimated for the actual operation of the fixed-route bus. Next, the lithium-ion battery and capacitor were compared as potential storage devices for this system, which requires a boosting charge. Both were found to have advantages and disadvantages and could be used for the proposed system. Additionally, methods for calculating the capacity of these storage devices were proposed. The validity was confirmed by a demonstration experiment using a single-passenger electric vehicle, which was driven by two in-wheel motors and powered by a lithium-ion battery or capacitor (Kawashima, 2010). A photograph of the model bus and model bus stop is shown in Fig. 3. The overhead contact wires are 6 m long. A similar transit system, known as a contact-wireless trolley bus or a hybrid trolley bus has been introduced. This system, in which capacitors and pantographs are mounted on the bus, is currently operated in Shanghai, China. The pantographs are expanded to charge capacitors when the bus stops at the bus stop. After charging, the pantographs are compressed and the bus departs. The bus travels to the next bus stop as an independent electric vehicle. The main difference between the Shanghai system and that proposed here is that, in the latter, electricity is supplied to the bus not only while it is stationary at the bus stop but also while it accelerates in departing from the bus stop to reduce stoppage time. Furthermore, in the system we propose, electrical power is generated from renewable energy sources through distributed power plants located near bus stops. In Japan, a hybrid electric bus that was equipped with both an internal combustion engine and an electric motor was tested between the international and domestic terminals of Haneda International Airport in 2008 (Ministry of Land, Infrastructure, Transport and Tourism (online), 2008). The bus was charged by a non-contact feeding system. However, it was charged in the garage, not on the road while passengers were getting on and off. In Sapporo, Japan, two types of hybrid battery trams were tested during the winter of 2007 2008. The trams were equipped with lithium-ion batteries (Ogasa, 2010) or nickel-hydrogen batteries (Akiyama, 2008), and ran on electrical power supplied by an overhead contact wire. The batteries were charged when the tram was in contact with an overhead contact wire, and the trams ran on power stored in the batteries when they were not in contact with the overhead contact wire. The purpose of the Sapporo system, which was different from that of our proposed system, was to diminish the cost of facilities such as overhead contact wires in case the tracks had to be extended. Storage devices for electric vehicles have been studied by many research groups. A lot of lithium-ion cells must be connected in series in order to drive an electric vehicle because of the high voltage required. In this case, the cell voltages must be balanced for the best performance. In recent research, the active cell balancing method has been proposed to improve cell performance (Einhorm, et al., 2011). Moreover, an energy management technique using a capacitor is proposed to extend the lithium-ion battery cycle life (Wang, et al., 2011). 23

Overhead contact wires Model bus stop DC-DC converter Collectors Capacitor for bus stop Capacitor for bus Model bus Fig. 2 Model bus based on a golf cart and model bus stop with boosting charge system (Kawashima and Fujioka, 2008). Guides for collectors Collectors Overhead contact wires Model bus stop Model bus DC-DC convertor Fig. 3 Model bus based on single-passenger electric vehicle and model bus stop with boosting charge system (Kawashima, 2010). 3. Demonstration experiment using a minivan converted into an electric vehicle In order to confirm the feasibility of the proposed system, a minivan with a gasoline engine was converted into an electric vehicle as the model bus. The model bus stop was built accounting for the necessary boosting charge from renewable energy. Then, the demonstration experiment was conducted. 3.1 Model bus We used a minivan, which is able to change the number of passengers, in order to analyze the effect of passenger number, as it was not possible to run an electric bus on our campus. Finding an electric minivan was difficult, so a minivan with a gasoline engine was converted to an electric vehicle. The selected minivan was a STEPWGN manufactured by Honda Motor Co., Ltd. First, the gasoline engine along with the injection device, transmission, exhaust pipe, muffler, and fuel tank were removed. Next, an AC induction motor (AC24LS: peak and continuous torques of 87 Nm and 36 Nm at 4000 rpm, respectively; peak efficiency of 85%; and weight of 40 kg) with a gearbox (AT1200: gear ratio of 12 and weight of 18 kg) for a passenger vehicle manufactured by Azure Dynamics Inc. was installed. Additionally, two drive shafts were customized to drive the front wheels. A digital motor controller (DMOC445: 15 kg in weight) for the geared motor manufactured by Azure Dynamics Inc. was also installed. Next, 88 lithium-ion battery cells (each with a capacity of 30 Ah) were mounted in the rear cargo space and arranged in 2 parallels of 44 in series with a total nominal voltage of 165 V. Two current collectors were installed on the roof to feed 24

from overhead contact wires installed at the model bus stop. The current collectors, lithium-ion batteries, motor controller, and motor were connected by electrical cables. An air compressor was installed for the brake booster. Finally, a low voltage 12-V circuit was connected from the lead battery. A photograph of the converted minivan is shown in Fig. 4. 3.2 Model bus stop The model bus stop with overhead contact wires is also shown in Fig. 4. The length of the overhead contact wires was kept at 6 m in order to supply electricity to the model bus while the bus starts and accelerates. The feeding time is extended by use of these long contact wires. A pick-up guide is installed at the approach end of each contact wire to ensure easy contact with the current collector. An enlarged picture of the ends of the contact wires is shown in Fig. 4. The feeding electricity is generated by photovoltaic modules rated at 10.7 kw installed on the roof of our school building. The generated electricity is stored in 98 lithium-ion battery cells (the capacity of each cell being 40 Ah), which are arranged in 2 parallels and 49 series. The electricity is then supplied to the model bus through a DC-DC converter manufactured by Nippon Stabilizer Industry Co., Ltd. (JSLGM200-200D-K/UN; the maximum voltage and current were 200 V and 96 A, respectively) and the overhead contact wires. A block diagram of the experimental facility using the converted minivan as the electric bus being rapidly charged at every bus stop using renewable energy is shown in Fig. 5. 3.3 Result of the demonstration experiment First, the feasibility of the proposed system is confirmed by the demonstration experiment using the converted minivan. In the experiment, the minivan drove twice around our school building (about 420 m), stopped at the model bus stop for 20 s to charge, and then started and accelerated with electricity supplied from the contact wires. At the bus stop, the minivan stopped near the approach end of the contact wires and accelerated to the other edge like a trolley bus. The electricity supply turns on automatically when contact between the contact wire and collector is confirmed, and turns off automatically when the collector reaches the other edge. This running pattern was repeated 4 times. The voltage of the mounted battery, output current from the battery, input current to the motor driver, and current from the collector were measured at 10-ms intervals by a portable data logger, NR-2000, produced by Keyence Co. The revolution speed of the motor, pedal position, consumption current of the motor driver, and motor torque were stored at 1-s intervals on a laptop computer connected to the motor driver. This procedure was repeated by changing the number of passengers. As this experiment was run on the campus, the motor driver was set to "Maximum Range," thus limiting the maximum power of the motor to 11.3 kw. A sample of the experimental results is shown in Figs. 6 and 7 for the case of six occupants, i.e., a driver and five passengers. Figure 6 shows the time series of data stored in the portable data logger for one running pattern, i.e., the battery voltage, current from the battery (negative values indicate the charging current), current to the motor driver (negative values indicate the regeneration current), and current from the collector. Figure 7 shows the time series of data measured by the motor driver for one running pattern, i.e., the vehicle velocity, pedal position, consumption current of the motor driver, and motor torque. Figure 6 shows that the minivan repeated acceleration and deceleration using the regenerative brake. Focusing on the currents from the collector and battery (circled in the figure), it can be seen that the current from the collector charges the battery, and it is consumed by the motor driver for acceleration of the model bus. Although the current from the collector decreases because the internal resistance of the battery increases during charging, the maximum current is recovered. This means that the feeding time is shortened and the capacity of the storage device can be reduced. Therefore, the effectiveness of feeding during the departure of the bus, which is one of the main features of the proposed system, has been demonstrated. Lastly, it should be noted that the charging current of the battery decreases because of the increase in the internal resistance during prolonged use. From Fig. 7, the pedal operation is seen to be ON-OFF like, and the maximum velocity is about 30 km/h. Data for each run, i.e., the length of the run, the energy supplied from the battery, and the energy stored in the battery, were almost the same. From the average of 4 runs, the minivan ran for 79.5 s, stopped at the model bus stop for 20 s for charging, and accelerated with electricity supplied from the contact wires for 6.1 s. The average energy supplied from the battery during one running pattern is 428 kj and the average stored energy in the battery including the regenerated energy by the regenerative brake is 489 kj; i.e., the energy charged is greater than the energy consumed. Therefore, this 25

demonstrates that the minivan can maintain this running pattern, and thus, the efficacy of the proposed electric bus system. Overhead contact wires Photovoltaic modules Collectors DC-DC converter Model bus stop Model bus converted into electric vehicle Li-ion battery Fig. 4 Model bus based on the minivan converted into an electric vehicle and the model bus stop with a boosting charge system using solar energy. Photovoltaic modules (10.7 kw) Contact wires DC Motor (15 kw at 4000 rpm) Li-ion battery (80Ah, 184 V) Controller Power conditioner Commercial power source DC-DC converter (Output: 200 V, 96 A) Collectors Li-ion battery (60 Ah, 165 V) (a) Feeding system for model bus stop (b) Storage and drive system for model bus Fig. 5 Block diagram of the experimental facility using the converted minivan as the electric bus being rapidly charged at every bus stop using solar energy. 26

Voltage [*10V], Current [A] 100 80 60 40 20 0-20440 460 480 500 520 540 560 Time [s] -40 Battery Voltage -60 Battery Current -80 Driver Current Collector Current -100 Fig. 6 Experimental results in the case of six occupants based on data stored in a portable data logger. Velocity [km/h], Pedal [%], Current [A], Torque [Nm] 100 80 60 40 20 0-20440 460 480 500 520 540 560 Time [s] -40-60 Velocity Pedal Position -80 Driver Current -100 Motor Torque Fig. 7 Experimental results in the case of six occupants based on data measured by the motor driver. Next, we continued the experiment by changing the number of occupants. The results in the cases of one, four, and six occupants are listed in Table 1. The average running time without feeding per running pattern, average time of feeding from contact wires, average energy supplied from the battery, average energy stored in the battery including the regeneration energy, as well as the gross vehicle mass are listed. The vehicle mass including the battery mass was 1473 kg; the mass of one passenger was assumed to be 65 kg. That is, the gross vehicle mass change was 21.1%. The listed values are the average of four running patterns. The values in parentheses are the percentage increase, assuming the value for the vehicle with one passenger to be 100%. Table 1 Experimental results (effect of changing the number of occupants). Number of occupants 1 4 6 Running time without feeding [s] 78.3 (100.0%) 77.9 ( 99.5%) 79.5 (101.5%) Time of feeding through contact wires [s] 24.4 (100.0%) 24.2 ( 99.2%) 26.1 (107.0%) Energy supplied from battery [kj] 376. (100.0%) 417. (110.9%) 428. (113.8%) Energy stored in battery [kj] 482. (100.0%) 480. ( 99.6%) 489. (101.5%) Gross vehicle mass [kg] 1538. (100.0%) 1733. (112.7%) 1863. (121.1%) 27

From Table 1, it can be seen that the running times of the electric vehicle in the three cases are similar regardless of the number of occupants, although the time of feeding from the contact wires increases in the case of six occupants. Thus, a very similar running pattern is achieved regardless of the number of occupants. From these similar running patterns, it can be observed that the energy consumed (i.e., the energy supplied from the battery) increases as the vehicle mass increases. The relationship between the consumption energy and the gross vehicle mass is shown in Fig. 8. The solid line is the regression line passing through the origin and was obtained by the least squares method. It can be expected that the relation is proportional when the electric bus is modeled by a simple single-particle model that is not affected by the mass, except for the inertia term in flatland running. Consumption Energy [kj] 500 450 400 350 300 250 200 150 100 50 0 Experimental result Approximated line 0 500 1000 1500 2000 Gross vehicle mass [kg] Fig. 8 Effect of vehicle mass on energy consumed. 4. Simulator for the electric bus The main feature of the proposed electric bus system is the reduction in the required capacity of the mounted storage device, as it only needs to store the electricity needed to reach the next bus stop. This can lead to a reduction in the weight of the electric bus, which is currently the main obstacle to the practical deployment of electric buses. In order to fully exploit this feature, it is necessary to estimate the energy consumed between bus stops and to ascertain the minimum capacity of the mounted storage device, with which the bus is able to continually operate. To determine this, a simulator to calculate the energy consumption of the electric bus was developed. In this study, the simulation parameters for the full-size electric route bus were extrapolated from those of the converted electric minivan, the validity of which was confirmed by the experiments. 4.1 Design of electric bus simulator The input for the simulator is distance, pitch of slope, maximum and minimum velocities, and the torque rate (i.e., the position of the accelerator pedal) for each running pattern. The running pattern consists of acceleration, running at a constant speed, and deceleration, and is divided across the bus route by signals and bus stops. The output is the bus position, the energy supplied from the battery, and the energy charged to the battery including the regenerative energy. Therefore, the characteristics of the motor controller follow that of the converted electric minivan. Although an accurate simulation model can be conducted including dynamic characteristics of an electrical system, such as the motor and the battery, the number of parameters increases and increasing amounts of time and effort must be expended. Therefore, based on the experimental results described above, the electric bus model was made to be single mass particle, as shown in Fig. 9. The equation of motion for the model is as follows: 2 d x m F R ( C K) v mg sin (1) 2 dt where m is the mass of the bus; F, the driving force calculated from the motor torque considering the gear ratio (12:1); R, the running resistance; C, the damping coefficient; K, the back-emf (electromotive force) constant; v, the velocity of the bus; and, the inclined angle of the slope. 28

Fig. 9 Model of the electric bus. These parameters were identified using the experimental acceleration, coasting, and deceleration running test results. Figure 10 shows a sample result of acceleration running test with the simulation result for the case of three occupants. The vehicle s velocity and the motor torque simulated using the identified parameters agreed with the experimental results. Therefore, the validity of the model and identified parameters has been confirmed. In this simulation, the torque was calculated using the measured accelerator pedal position. Velocity [km/h], Torque [Nm] 100 90 80 70 60 50 40 30 20 10 0 Velocity (Exp.) Torque (Exp.) Velocity (Sim.) Torque (Sim.) 0 2 4 6 8 10 Time [s] Fig. 10 Acceleration running test result with the simulation result using the identified parameters in the case of three occupants. Then, a simulator for the minivan converted into an electric vehicle was constructed using the identified parameters. A sample of the simulation results are shown in Fig. 11, with the experimental results under the same condition described in Section 3.3 for the case of an occupant. The torque was also calculated using the measured accelerator pedal position. The vehicle s velocity and the motor torque estimated by the developed simulator agreed with the experimental results. Therefore, the validity of the constructed simulator has been confirmed by the model experiment using the converted electric minivan. The simulation parameters for the full-size electric bus were derived by extrapolating the data for the converted electric minivan. The weight of the full-size bus was assumed to be 15000 kg and the weight of the converted electric minivan was about 1500 kg. That is to say, the weight of the full-size bus is about ten times that of the converted electric minivan. Therefore, the parameters for the full-size electric bus model were obtained by multiplying the parameters of the converted electric minivan by ten. For simplicity, it was assumed that the accelerator pedal is manipulated discretely; i.e., only three torque rates of 100%, 50%, and 0% are set. When driving straight, the torque rate was set to 100%, and during turning and slow running on corners or in bus terminals, the torque rate was set to 50%. However, the torque was limited by the set maximum power. In the converted electric minivan, the regenerative brake was also manipulated by the accelerator pedal. Therefore, it was assumed that the regenerative brake is manipulated in an ON-OFF like manner, and the 29

minivan decelerates by a constant torque. However, the regenerative brake torque was also limited by the set maximum power of regeneration. Furthermore, it was assumed that the minivan continues to decelerate due to the constant torque by the combined use of the regenerative brake and the foot brake at the velocities less than 11 km/h, because the regenerative brake torque gradually weakens according to the velocity at these velocities. It was assumed that the storage device consists of lithium-ion batteries, and the bus is operated using the energy between 40-75 % of the battery capacity, i.e., a state of charge (SOC) of 40 75% for the boosting charge. The SOC shows the charging condition of the battery in percentage. It was assumed that the efficiency of the gearbox was constant at 85%, the efficiency of the motor with the controller in both the driving and the regenerative braking was also constant at 80%. In addition, it was assumed that the lowest velocity at the signal was 10 km/h, because the bus does not stop every time, and the lowest velocity was 0 km/h at the bus stop. 80 Velocity [km/h], Torque [Nm] 60 40 20 0 320 340 360 380 400 420-20 Time [s] -40 Velocity (Exp.) -60 Torque (Exp.) Velocity (Sim.) -80 Torque (Sim.) Fig. 11 Results estimated by the simulator and measured by the experiment in case that the minivan converted to electric vehicle was drove around our school building with an occupant. 4.2 Simulation results In this study, the bus route from "Atsugi Bus Center" to "Tobio Housing Complex" by way of "Kanagawa Institute of Technology" was selected. This route includes four steep slopes. They are respectively 410, 335, 85, and 170 m in length, have average gradients of 2.68, 2.98, 5.88, and 4.10%, and are 3.76, 6.94, 7.38, and 7.78 km from the departure bus depot. Both way of the route was divided into 84 running patterns, and the distance, the pitch of the slope, the maximum and minimum velocities of each running pattern were input into the simulator before running the simulation. The results are shown in Table 2. Although the input travel distance was 16610 m, the distance calculated by the simulator was 16442.9 m, and thus the relative error was 1.01%. This confirms the accuracy of the simulator. Although the travel time set by the bus operator is 22 minutes for the outward journey and 23 minutes for the return journey, the calculated travel time without the stoppages was 2808.8 s, so the error was about 10%. It was found that the maximum power should be adjusted to the gradient during climbs if the electric bus is expected to run similarly to a fixed-route bus with an internal combustion engine. 5. Design of bus-mounted storage device and charging facility at bus stop From Table 2, the total energy consumption was 95.5 MJ for the round trip. Therefore, a charge of 2.33 MJ is needed at each bus stop. If it is assumed that the charging voltage is 600 V at the bus stop and the charging time, including during acceleration, is 22 s, a charging current of 177 A is required. Furthermore, the lithium-ion battery cells mounted on the bus can be charged to 3C, which means that a current of 180 A is possible if the capacity of each battery cell is 60 Ah. The maximum voltage is 4.1 V and the rated voltage is 3.8 V. Under these conditions, the bus-mounted storage device can be designed as 146 60-Ah battery cells connected in series, which results in a total capacity of 120 MJ and corresponds to the energy consumption energy for 1.26 round trips. Thus, this confirms that the capacity and weight of the mounted storage device in the proposed electric bus can be reduced. 210

Table 2 Simulation results for the bus route. Number of Energy for Energy for Time [s] Distance [m] Bus Stop Driving [MJ] Regeneration [MJ] 0 126.4 390.1 2.299-0.340 1 168.4 620.2 1.888-0.491 2 206.6 810.3 1.591-0.419 3 295.9 1200.6 3.235-0.825 4 330.9 1360.6 1.348-0.355 5 419.9 1880.8 3.923-0.928 6 468.2 2180.8 2.267-0.541 7 565.8 2671.0 3.605-0.803 8 669.7 3156.3 4.050-1.055 9 719.9 3476.4 2.349-0.541 10 833.8 4242.2 8.384-0.919 11 889.8 4627.3 2.617-0.541 12 956.6 4922.5 2.483-0.649 13 1035.3 5337.6 3.325-0.839 14 1094.3 5687.8 3.757-0.641 15 1168.4 6197.9 4.792-0.728 16 1242.6 6578.0 3.028-0.755 17 1303.1 6938.1 3.788-0.753 18 1373.8 7378.5 5.240-0.541 19 1439.2 7723.5 4.448-0.338 20 1489.7 7953.7 3.223-0.276 21 1616.8 8693.8 5.043-1.364 22 1667.7 8924.0 1.548-0.861 23 1718.6 9179.3 1.227-1.325 24 1795.8 9679.4 1.907-1.437 25 1862.6 10029.5 2.407-0.715 26 1910.8 10349.5 2.016-0.591 27 1982.9 10799.7 2.348-0.807 28 2025.9 11074.7 1.691-0.633 29 2131.5 11685.0 3.801-1.335 30 2185.0 11905.1 1.666-0.605 31 2265.2 12435.3 2.829-0.774 32 2380.0 13145.5 3.606-2.083 33 2460.8 13595.6 3.339-0.767 34 2577.0 14105.8 4.191-1.056 35 2646.6 14556.1 3.365-0.800 36 2684.2 14741.3 1.552-0.409 37 2779.7 15331.4 4.270-0.960 38 2808.7 15441.4 0.929-0.239 39 2915.8 15961.6 4.302-1.114 40 2980.6 16241.8 2.355-0.613 41 3019.7 16441.9 1.669-0.439 Total 127.700-32.204 211

The change of SOC of the lithium-ion battery mounted on the bus is shown in Fig. 12 considering the charged energy at each bus stop. The bus departs the depot with an SOC of 60%. The SOC then increases because the distances between bus stops are short. The SOC drops near 4 km and 8 km from the bus depot, because the bus must ascend steep slopes. Although the SOC halfway to the bus terminal is less than 50%, it is recharged to 60%, which is the condition with which it departed, as a result of power regeneration on the return route. Therefore, this demonstrates that continuous operation is possible using the estimated battery capacity. In addition, the discharged or charged energy in one round trip was 127.7 MJ, which is almost equal to the total capacity of the mounted batteries. Therefore, the bus can operate for at least 3000 roundtrips if the life of the lithium-ion battery, i.e., the cycle durability, is 3000 cycles of full charging and discharging. For example, the cycle durability of lithium-ion battery LIM40 produced by GS Yuasa Corporation is over 3000 cycles according to the catalog. One of the demerits of lithium-ion battery is that the charging capacity decreases as discharge and charge is repeated. However, it would be expected that the bus of the proposed system could operate more than 3000 roundtrips because the mounted batteries are charged and discharged in the part of the capacity, i.e., in the SOC of 50 64%. SOC [%] 64 62 60 58 56 54 52 50 48 0 4 8 12 16 Distance [km] Fig. 12 Change in the state of charge. Next, the bus stop is considered. With respect to the DC-DC converter, the maximum voltage of the contact wire is 600 V and the maximum current is 180 A, and thus, the DC-DC converter is set to limit the voltage to 600 V and the current to 180 A, and to change between constant-current and constant-voltage functions automatically. The storage device is connected to the photovoltaic modules installed near the bus stop and the commercial power source through the power conditioner, as shown in Fig. 5, so power shortages from the photovoltaic modules in bad weather or at night are compensated for by a commercial power source, and surplus power generated by the photovoltaic modules is fed into the commercial power source. The capacity of the storage device installed on the bus stop is designed considering the relationship between charging power and supplied power. The capacity must be several times the boosting charge energy to the bus in order to prevent large currents being supplied by the commercial power source. The device can be expected to function as a power supply in an emergency. The rated generating power of the photovoltaic modules is estimated by assuming the energy of the discrete boosting charge to the bus is smoothed by the batteries and the commercial power source. In addition, it is assumed that the photovoltaic module with a nominal maximum output of 1 kw generates 1000 kwh in a year (National Institute of Advanced Industrial Science and Technology (online), 2009), and the bus is operated in 15-min intervals from 06:00 to 23:00; i.e., the bus stop supplies electrical power to the bus 69 times. Under these assumptions, the bus stop supplies a power of 161 MJ per day. Therefore, each bus stop requires photovoltaic modules rated at 16.3 kw. It needs the installation area of 102-163 m 2 under the assumption that the conversion efficiency is 10-16 %. This area is too large to install it on the roof of the bus stop. Further improvements in the conversion efficiency of photovoltaic modules would be needed to spread the proposed system. 212

6. Conclusions First, the reliability of the proposed electric bus system, in which the bus is charged rapidly at every bus stop using solar energy, was confirmed by the demonstration experiment. In the experiment, a minivan converted into an electric vehicle and fitted with current collectors was used as a model bus, and the contact wires, the power feeding system, and the photovoltaic modules were fitted to a model bus stop. It was shown that the proposed bus is capable of continuous operations. One of the main characteristics of the proposed electric bus system is the low weight of the storage device on the electric bus. In order to determine the capacity of the storage device to minimize this weight, a simulator calculating the energy consumption between bus stops was developed by extrapolating the parameters obtained from the converted electric minivan. Then, the minimum capacity of the storage device mounted on the bus and the required capacity of the photovoltaic modules to be installed near each bus stop were calculated. Acknowledgements This work was supported by the High-Tech Research Center Project for Private Universities through a matching fund subsidy from MEXT (Ministry of Education, Culture, Sports, Science and Technology, Japan), 2007-2011. References Akiyama, S., Low Floor Battery-driven LRV (SWIMO), Journal of the Japan Society of Mechanical Engineers, Vol.111, No.1075 (2008), pp.468-469 (in Japanese). Einhorm, M., Roessler, W. and Fleig, J., Improved Performance of Serially Connected Li-ion Batteries with Active Cell Balancing in Electric Vehicles, IEEE Transactions on Vehicular Technology, Vol.60, No.6 (2011), pp.2448-2457. Fujioka, I., Intermittent Power Supply Type Electric Vehicle System and the Electric Vehicle, Japanese Patent Disclosure P3768982 (2006). Kawashima, T. and Fujioka, I., New Public Transportation System with Bus Charged Intermittently at Every Bus Stop Using Green Energy (Model Experiment Using Golf Cart), Journal of Environment and Engineering, Vol.3, No.2 (2008), pp.374-384. Kawashima, T., Basic Research on a Novel Zero-Emission Public Transportation System (Design of Charge-Boosting System for an Electric Bus System Charged at Every Bus Stop), Journal of Environment and Engineering, Vol.5, No.1 (2010), pp.168-182. Ministry of Land, Infrastructure, Transport and Tourism (MLIT), Press release on 6 February, 2008 (online), available from <http://www.mlit.go.jp/kisha/kisha08/09/090206_3_.html> (in Japanese), (accessed on 12 December, 2014). National Institute of Advanced Industrial Science and Technology (AIST), The electric power generation in actual environment (online), available from <https://unit.aist.go.jp/rcpvt/ci/about_pv/output/irradiance.html> (Last update date: 24 March, 2009) (in Japanese), (accessed on 12 December, 2014). Ogasa, M., Technology of Interoperability of LRVs between Electrified lines and Non-Electrified Lines, or between Railway Main Lines and On-Street Lines; Running Results of a Contact-Wire/Battery Hybrid LRV on JR Lines, Journal of the Japan Society of Mechanical Engineers, Vol.113, No.1104 (2010), pp.846-849 (in Japanese). Wang, L., Collins, EG. and Li, H., Optimal Design and Real-Time Control for Energy Management in Electric Vehicles, IEEE Transactions on Vehicular Technology, Vol.60, No.4 (2011), pp.1419-1429. 213