Electric Bicycle Using Batteries and Supercapacitors D. M. Sousa♣, P. J. Costa Branco♠, J. A. Dente♥ ♣ DEEC AC-Energia /CAUTL, Instituto Superior Técnico, TU Lisbon ♠,♥ DEEC AC-Energia, Instituto Superior Técnico, TU Lisbon Av. Rovisco Pais, 1 – 1049-001 Lisboa Lisboa, Portugal ♣ ♠ Tel.: +351 – 21 841 74 29, +351 – 21 841 74 32, ♥ +351 – 21 841 74 35. Fax: +351 – 21 841 71 67. E-Mail: ♣
[email protected], ♠
[email protected], ♥
[email protected] URL: http://www.ist.utl.pt
Keywords «Electric vehicle», «Energy storage», «Supercapacitor», «Power converters for EV», «Electrical drive».
Abstract In this paper, a traction system useful for an autonomous Electric Vehicle of individual use is described. The developed system is constituted in a first approach by two different power sources: one is constituted by batteries or by fuel cells, and the other by supercapacitors. This paper describes a technical solution joining and accomplishing the usage of two energy storage systems in the same traction system. In the developed system, the supercapacitors run as element that store energy temporarily and that can be used to retrieve energy. Starting from the functional characteristics of typical electrical vehicles and characterization of a typical routing profile, the energy consumption is obtained. In order to characterize and design the system, this is described in detail, namely the supercapacitors models, the battery, the power converters and the implemented strategy of control. According to the obtained results, a control strategy that allows an effective management of the stored energy in the system regarding the vehicle’s optimal functioning and increasing its autonomy is also presented and discussed. Based on experimental and simulation results, the advantages and disadvantages of the proposed solution are presented.
Introduction In the modern societies, the increasing needs of mobility means sometimes increasing the number of vehicles circulating. Ambient concerns, as for instance local pollutant emissions for the atmosphere, influence also, in nowadays, the technical decisions related with all kind of vehicles. In this context, new alternatives to the existing internal combustion engines are mandatory. So, vehicles with electric propulsion seem to be an interesting alternative [1, 2, 3]. Starting from this context, this research describes a solution that was developed and studied to be applied in electric vehicles of individual use as bicycles. The solution proposes the combination of two sources of energy, batteries and supercapacitors, and two DC-DC converters. On board, batteries and supercapacitors store the energy. Anyway, the proposed topology considers that fuel cells should be used in two ways: replacing the set of batteries or to charge the batteries and the supercapacitors.
As it is well known, in the typical electric traction systems the batteries drive the high currents and in the worst situation drive the current peaks demanded by the load. As it is well known, this type of operation decreases strongly the autonomy of the vehicles for individual use. The continuous and random operation of electrical vehicles requires and claims for systems improving the autonomy and the performance of the available ones. In this situation, a solution to improve the battery behaviour and its time life is to replace temporarily the battery by another power source or, as in the developed solution, to supply the system using other power source when undesired and transient situations occur [4]. In this case, the load is supplied by the complementary energy source avoiding, at least, deep discharges of the battery. The adopted solution uses supercapacitors, which drive the peaks of power required by the load.
Requirements of the system The first step in order to project the system is to establish the objectives of the work according to the energy consumption and the performance of the vehicle for individual use. To estimate the power required by this type of vehicles, we have considered that the forces applied to the vehicle are, as represented in figure 1, the following: Fa = M ⋅a
(1)
F g = M ⋅ g ⋅ sin θ
(2)
F air =
1 ⋅ ρ ⋅ M ⋅ C D ⋅ A f ⋅ v (t ) 2 2
(3) (4)
F r = M ⋅ g ⋅ C R ⋅ cos θ
Where: Fa is the resulting force; Fg is the gravitational force; Fair is the air friction force; and Fr is the wheels friction force; parameter ρ is the air density (1.29 kg/m3); Af is the frontal area of the vehicle; CD is the air friction coefficient (tipically 0.9 for a scooter and 0.8 for a bicycle); and CR is the wheels friction coefficient (usually between 0.008 and 0.014).
F air =
1 ⋅ ρ ⋅ M ⋅ C D ⋅ A f ⋅ v (t )2 2
F r = M ⋅ g ⋅ C R ⋅ cos θ F g = M ⋅ g ⋅ sin θ
Fig. 1: Forces applied to the vehicle
Considering that the vehicle runs with a speed v, the power required by the system is: PVE = P a + P g + P air + P r ⇔ 1 ⇔ PVE = M ⋅ a (t ) ⋅ v(t ) + M ⋅ g ⋅ v(t ) ⋅ sin θ + ⋅ ρ ⋅ M ⋅ C D ⋅ A f ⋅ v(t )3 + M ⋅ g ⋅ C R ⋅ v(t ) ⋅ cos θ 2
(5)
Assuming that the vehicle speed is equal to the angular speed ω of wheels with radius R, torque of the traction system can be estimated as: T VE = T a + T g + T air + T r ⇔ 1 ⇔ T VE = M ⋅ a(t ) ⋅ R + M ⋅ g ⋅ R ⋅ sin θ + ⋅ ρ ⋅ M ⋅ C D ⋅ A f ⋅ v(t )2 ⋅ R3 + M ⋅ g ⋅ R ⋅ C R ⋅ v(t ) ⋅ cos θ 2
(6)
Anyway, to analyze and compare the performance of different vehicles for individual use, operating conditions in terms of speed and autonomy should be used. So, based on the Portuguese standards (NP EN 1986-1), a typical urban cycle (Figure 2), repeated 10 times, with the total duration of 1180 s should be fulfilled by the traction system in terms of torque and speed and by the power sources on board in terms of energy stored.
Fig. 2: Profile of the used urban cycle To fulfill the cycle above and taking into account the physical dimensions of a bicycle or a scooter, the nominal power required by this type of system stays in the range of 2 kW to 2.5 kW. So, based on the premises and conditions above, a traction system is proposed aiming the autonomy, efficiency and performance of this type of vehicles.
Global system The main elements constituting the global system are two power converters, two energy storage systems (in the basic implementation, batteries and supercapacitors) and the traction motor [5]. With the proposed solution, the most important objective is to increase the capacity of storing energy and vehicle autonomy, avoiding deep discharges of the batteries.
In order to achieve this goal, the global topology represented in figure 3 was investigated. iSC'
L
iSC
iC
ia
idc VSC
Vdc
Va
DC-DC 1
DC-DC 2
Fig. 3: Global system
The power converters The proposed topology uses two power converters. Their main functions are: • •
The power converter DC-DC 1 (operating as buck or book converter (figure 4), in agreement with the level of charge of the supercapacitors) transfers energy from the supercapacitors to the battery. The DC-DC 2 converter adjusts the supply voltage of the traction motor (in this case, a DC motor) to control its speed.
L
iSC'
iSC S2
VSC
S1
iC idc
Vdc
LOAD
Fig. 4: Converter topology of the used DC-DC power converters
The traction motor The implemented traction system is based on a DC motor, which dynamic behaviour can be represented by:
dif U f = r f ⋅ i f + L f ⋅ dt d ia + k ⋅φ ⋅ω U a = r a ⋅ i a + La ⋅ dt dω J ⋅ dt = k ⋅ φ ⋅ ia − T Load
(7)
To project and analyse the system, knowledge of motor parameters is mandatory. In particular, the electrical time constant and the starting current allowed by the system, which according to experimental tests are, respectively:
τa =
L a ≈ 30 ms ra
i start ≈
Ua ≈ 20 A ra
(8)
(9)
The supercapacitors model The nominal voltage of each supercapacitor available is lower than the rated voltage of typical electric traction systems (12 V or 24 V, for instance) [6]. Therefore, in order to fulfill the rated voltage of these systems, it is mandatory to connect supercapacitors in series and in parallel modes. Anyway, to investigate the dynamic behaviour and the performance of the global system it is important to know the supercapacitor model, which electric equivalent model is represented in figure 5.
Fig. 5: Equivalent model of a supercapacitor The supercapacitor model is constituted by an inductance L, a resistance Ri and an impedance Zp connected in serie. The impedance Zp can be calculated using the expression: Z p ( jω ) =
τ coth( jωτ ) ⋅ C jωτ
(10)
To the available supercapacitors (2.5 V; 200F ± 30% (at 20ºC)), the parameters of the model were obtained by electrochemical impedance spectroscopy, which simulated and experimental spectra are shown in figure 6 [7, 8]. 50
- Imag(Z) [mOhm]
40 30 Exp Simul
20 10 0
24,5
25,2
24,7
24,2
24,1
24,4
25,1
26,3
28,3
31,7
35,9
37,5
-10 Re(Z) [mOhm]
Fig. 6: Spectra of a supercapacitor impedance
Experimental and Simulation Results In a first approach to the problem, the supercapacitors were connected in parallel with the battery, as represented in figure 7 [9, 10]. To study and analyse the performance of such system, both experimental and simulation results were obtained. The implemented model (simulated using the @Matlab/Simulink) includes dynamic models to the used battery and supercapacitors [11, 12, 13]. ia ib Rb
isc Rsc Va
VB
VSC
Fig. 7: Circuit connecting in parallel the supercapacitors and battery The electrical equations representing the circuit are in a first approach the following:
i a = i SC + ib v = v SC − i ⋅ R SC = vb − ib ⋅ Rb d i SC = −C ⋅ v SC dt
(8)
A prototype of this circuit was implemented in the laboratory, with a set of batteries (12V, 7Ah each one), a DC motor and a five supercapacitors in series (Ctotal = 200/5 = 40 F). To a random load diagram, the experimental and simulation results obtained are shown in figures 8 and 9. From these results, it is important to point out that current peaks are driven by the supercapacitors, thus avoiding deep discharges of the batteries. Furthermore, when the motor is braking the supercapacitors are charged. 30.00 25.00
CCorrente urren t [A] (A )
20.00 15.00 10.00 5.00 0.00 0
60
120
180
240
300
360
420
-5.00 -10.00
TTempo im e (s)[s] LoadCarga
Bateria Battery
SC
Supercapac.
C urrent ( A )
Fig. 8: Experimental results
T im e (s) Load
Battery
Supercapac.
Fig. 9: Simulation results A reasonable agreement between the experimental results and the simulation ones is observed, leading that the assumption that the developed models constitute a good approach and the circuit behaves as foreseen analytically. Anyway, the above operating principle is only valid if an effective control of the energy transit between the supercapacitors and the battery is reached.
The control strategy The first approach to the problem of energy management is based on the calculation of an average value of the current iDC demanded by traction system. The average current should be supplied by the main power supply, which can be the set of batteries or a fuel cell [3, 14, 15]. The difference between the
instantaneous value of the load current ia and the average value of iDC will be the current supplied by the supercapacitors, iSC. When the current iSC is positive and the total voltage of the supercapacitors is higher than the energy availability, the break level Vα, a duty-cycle is applied to the semiconductor S1 (figure 4) running the DCDC 1 converter as a boost converter. On other hand, if the supercapacitors does not have energy available, that is, the set of supercacitors is discharged or their voltage is under the break level, converter DC-DC 1 is switched off and the main power supply supplies the traction system. For negative values of iSC and if the supercapacitors do not have the maximum load Vβ, converter DC-DC 1 runs as buck converter. When the supercapacitors voltage reaches the value Vβ , converter DC-DC 1 remains in its stand-by mode. Anyway, if a fuel cell is used as the main power supply, condition iDC
