Optimal Design of a Hybrid Solar Vehicle

Transcription

1 Proceeings of AVEC 6 The 8 th International Symposium on Avance Vehicle Control, August 2-24, 26, Taipei, Taiwan AVEC691 Optimal Design of a Hybri Solar Vehicle I.Arsie, G.Rizzo, M.Sorrentino Department of Mechanical Engineering, University of Salerno, 8484 Fisciano (SA), Italy Keywors: Hybri Vehicle, Solar Energy I. INTRODUCTION In the last years, increasing attention has been spent towars the applications of solar energy to cars. Various solar car prototypes have been built an teste, mainly for racing an emonstrative purposes [1] [2]. Despite of significant technological efforts an some spectacular outcomes, several limitations, such as low power ensity, unpreictable availability of solar source an energetic rawbacks (i.e. increase in weight an friction an aeroynamic losses ue to aitional components), cause pure solar cars to be still far from practical feasibility. On the other han, the concept of a hybri electric car assiste by solar panels appears more realistic [3][4][5][6][7]. In fact, ue to relevant research efforts [8], in the last ecaes Hybri Electric Vehicles (HEV) have evolve to inustrial maturity. These vehicles now represent a realistic solution to important issues, such as the reuction of gaseous pollution in urban rive as well as the energy saving requirements. Moreover, there is a large number of rivers utilizing aily their car for short trips with limite power. Some recent stuies, conucte by the UK government, report that about 71% of UK users reach their office by car, an 46% of them have trips shorter than 2 min, mostly with only one person on boar [9]. The above consierations open promising perspectives with regar to the integration of solar panels with pure -electric hybri vehicles (i.e. tri-hybri cars), with particular interest in the opportunity of storing energy even uring parking phases. In spite of their potential interest, solar hybri cars have receive relatively little attention in literature [7]. An innovative prototype has been evelope at Western Washington University [5][6] in the 9s, aopting avance solutions for materials, aeroynamic rag reuction an PV power maximization with peak power tracking. Other stuies an prototypes on solar hybri vehicles have been presente by Japanese researchers [3][4] an at the Queenslan University [1]. Although these works emonstrate the general feasibility of such an iea, etaile presentation of results an performance, along with a systematic approach to solar hybri vehicle esign, seem still missing in literature. Therefore, suite methoologies are require to aress both the rapi changes in the technological scenario an the increasing availability of innovative, more efficient components an solutions. A specific ifficulty in eveloping an HSV moel relates to the many mutual interactions between energy flows, power-train balance of plant an sizing, vehicle imension, performance, weight an costs, whose connections are much more critical than in either conventional or hybri electric vehicles. Preliminary stuies on energy flows in an HSV has been recently evelope by the authors [11][12]. The current paper presents a more etaile stuy on the optimal sizing of a solar hybri car. The optimization analyses are base on a longituinal vehicle ynamics moel, evelope to account for, besies the impact of weight an costs, also the influence of energy flows an aopte control strategies. II. STRUCTURE OF THE SOLAR HYRID VEHICLE Different architectures can be applie to HEVs: series, parallel, an parallel-series. These two latter structures have been utilize for two of the most iffuse hybri cars in the market: Toyota Prius (parallel-series) an Hona Civic (parallel). Instea, for solar hybri vehicles the series structure seems preferable [7], ue to its simplicity, as in some recent prototypes of HSV [1]. With this approach, the Photovoltaic Panels (PV) concur with the Electric Generator, powere by the, to recharge the battery pack in both parking moe an riving conitions, through the electric noe EN. The electric motor EM can either provie the mechanical power for the propulsion or restore part of the braking power uring regenerative braking (FIG. 1). In this structure, the thermal engine can work mostly at constant power (P AV ), corresponing to its optimal efficiency, while the electric motor EM can reach a peak power P max : Pmax P av (1) FIG. 1 - SCHEME OF THE SERIES HYRID SOLAR VEHICLE PV EN EM In orer to estimate the net solar energy capture by PV panels in real conitions (i.e. consiering clous, rain etc.) an available for propulsion, a solar calculator evelope at the US National Renewable Energy Lab has been use [13], consiering four ifferent US

2 AVEC 6 locations, ranging from 21 to 61 of latitue, base on time series. The calculator provies the net solar energy for ifferent panel positions: with 1 or 2 axis tracking mechanism or for fixe panels, at various tilt an azimuth angles. The most obvious solution for solar cars is the location of panels on roof an bonnet, at almost horizontal position. Nevertheless, two aitional options have been consiere: (i) horizontal panels (on roof an bonnet) with one tracking axis, in orer to maximize the energy capture uring parking moe; (ii) panels locate also on car sies an rear at almost vertical positions. The maximum panel area can be estimate as function of car imensions an shape, with a simple geometrical moel [12]. The energy from PV panels can be obtaine summing up the contribution from parking (p) an riving () perios. While in the former case it is reasonable to assume that the PV array has an unobstructe view of the sky, this hypothesis coul fail in most riving conitions. Therefore, the energy capture uring riving can be reuce by a factor <1. In orer to estimate the fraction of aily solar energy capture uring riving hours (h ), it is assume that the aily solar energy is istribute over h sun hours. A factor <1 is then introuce to account for further egraation ue to charge an ischarge processes in the battery for energy taken uring parking. The net solar energy available for propulsion, store uring both parking an riving moes, can therefore be expresse as: E E h h sun s, p p APV esun (2) hsun h s, p APV esun (3) hsun Where e sun is the average aily energy capture by solar panels in horizontal position. Hereinafter, e sun is assume equal to 4.3 kwh/ay, which correspons roughly to a latitue of 3 in June month. The energy require to rive the vehicle uring the ay can be expresse as function of the average positive power P av an the riving hours h : E h t P t h P (4) av The instantaneous power can be compute starting from a given riving cycle, for assigne vehicle ata, integrating a vehicle longituinal ynamic moel. Require riving energy E epens therefore on vehicle weight an vehicle cross section, which in turn epen on the sizing of the propulsion system components an on vehicle imensions, relate to solar panel area. The contribution of solar energy to the propulsion can be therefore etermine as: E E sun s, p Es, (5) E E The fuel consumption for both conventional vehicle () an HSV can be then compute an compare. Of course, in parallel with fuel saving, corresponing reuction in pollutants an CO 2 emissions with respect to the conventional vehicle is also achieve. III. WEIGHT MODEL A parametric moel for the weight of an HSV can be obtaine aing the weight of the specific components (PV panels, battery pack,, Generator, Electric Motor, Inverter) to the weight of the HSV boy. This latter has been obtaine starting from a statistical analysis of small commercial cars (CC). A linear regression analysis has been performe, consiering weight W (W boy,cc ), power P an vehicle imensions (length l, with w, height h an their prouct V=lwh) for 15 commercial cars, with power ranging from 9.5 kw to 66 kw [12]. In orer to use these ata to estimate the base weight of the HSV (W boy,hsv ), the contribution of the components not present in the series hybri vehicle (i.e. gearbox, clutch) has been subtracte. The CC car boy also inclues other components (thermal engine, electric generator, battery) that will be consiere separately for the hybri car moel; the weight of is estimate as function of peak power, whereas the influence of electric generator an battery has been neglecte (their weights are of course much lower than the corresponing components neee on the hybri car). TA. I RRESSION ANALYSIS FOR COMMERCIAL CAR ODY MASS. # Variables R 2 1 W=k 1 +k 2 P W= k 1 +k 2 P+k 3 l+k 4 w+k 5 h W= k 1 +k 2 P+k 3 V.946 A subtractive term (W) has been introuce to inclue weight savings achievable through the use of aluminum instea of steel for chassis: in this case, aitional costs have been consiere in the cost moel [13]. Thus, the mass of the car boy an of the entire HSV can be expresse as: W W boy HSV, HSV Wboy, CC Pmax, V W P A P PV max PV m m W P boy, HSV max,, g m m P m m N V W max The battery mass m is proportional to the number of moules neee to balance the maximum power of the reference vehicle (i.e. CC), estimate as follows: N, u m EM (6) (7) Pmax P (8) P

3 AVEC 6 where P,u is the nominal power of a single battery moule. IV. COST ESTIMATION In orer to assess the benefits provie by HSV with respect to conventional vehicles, both the aitional costs, ue to hybriization an solar panels, an achievable fuel savings are to be estimate. The aitional cost C HSV can be expresse starting from the estimate unit cost of each component: C HSV c P P c max C EM c C APV cpv N Wc al (9) The last two terms account for: i) reuction in chassis weight if aluminum is use [14] an ii) cost reuction for Internal Combustion Engine in HSV (where it is assume P = P / ) with respect to conventional vehicle (where P =P max ). The aily saving with respect to conventional vehicle can be compute starting from fuel saving an fuel unit cost: S m f CV m f HSV c f,, (1) The pay-back, in terms of years necessary to restore the aitional costs with respect to the conventional vehicle, can be therefore estimate as: CHSV CHSV P (11) n S 3S D For further etails about the meaning an the values of some of the parameters introuce in eqs. 2 through 11, the reaer is aresse to a previous work [12]. V. VEHICLE DYNAMIC MODEL In a previous paper [12], an extene optimization analysis was conucte to investigate the effects of latitue, costs, prices an layout on optimal vehicle structures, in terms of panel area, vehicle imension an weight. The results presente have been obtaine by computing the fuel saving of the HSV with respect to the conventional vehicle with a simplifie approach, assuming average values for fuel consumption in the two cases [12], an average yearly solar ata. This approach, although sufficient to assess the general feasibility of HSV an to unerstan the role of the main variables on costs an energy saving, oes not allow accurate evaluation of the effects of vehicle weight an imensions on inertial an aeroynamic forces uring the riving cycle. Moreover, a more precise analysis is require to analyze the effects of control strategies on energy flows, also consiering seasonal effects on solar energy. In orer to overcome these limitations, a longituinal vehicle moel has been evelope to simulate the ynamic behavior of both HSV an conventional vehicle over a riving cycle. attery, electric motor an generator have been simulate by the ADVISOR moel [15]. VI. ENGINE CONTROL FOR HSV In most electric hybri vehicles, a charge sustaining strategy is aopte: at the en of a riving path, the battery state of charge shoul remain unchange. With a solar hybri vehicle, a ifferent strategy shoul be aopte, since battery can be charge uring parking hours as well. In this case, a ifferent goal can be pursue, namely restoring the initial state of charge within the en of the ay rather than after a single riving path. For this en, the internal combustion engine shoul be operate whenever possible at maximum efficiency, corresponing to power P opt. If the energy require to restore battery charge is lower than the amount corresponing to a continuous use at P opt throughout the riving time h (case ), an intermittent operation can be aopte (cases A1-A2). In case that more energy is require, the internal combustion engine is operate at constant power between P opt an P max (case C). The ifferent operating moes can be escribe by the variable, ranging from to = P max / P opt, as escribe in the table below. TA. II Engine control strategies for HSV. A1 1 P t h A2 1 1 C 1 max P P h t h opt P P t h P opt P opt t h The optimal value is foun by imposing that the energy provie by an PV panels uring the riving hours guarantees a charge sustaining strategy over the whole ay. This conition is expresse as: SOC ay 24h SOC SOC D t SOC p (12) Assuming that the riving scheule, long h hours, is compose of a sequence of ECE-EUDC cycles, eq. (12) can be satisfie by iteratively solving, over one cycle, the following nonlinear equation: SOC p SOCECE (13) N cycles where N cycles is evaluate as function of each moule uration T cycle : h N cycles (14) Tcycle FIGs 2 through 4 show the moel outputs for a selecte HSV configuration, simulate over an ECE-EUDC

4 AVEC 6 riving cycle an controlle accoring to the strategies efine in TA. II. FIG. 2 shows that the peaks in power eman are mainly met by the batteries. FIG. 2 Power contributions for the ECE-EUDC cycle (A PV,H = 4 m 2, P = 13 kw, l = 3.5 m, w = 1.6 m, h =1.4 m). 4 HSV power [kw] FIG. 4 attery state of charge (A PV,H = 4 m 2, P = 13 kw, l = 3.5 m, w = 1.6 m, h =1.4 m) SOC for HSV SOC N cycles p -1 rive -2 gen -3 sun batt Time [s] FIG. 3 Comparison between CC an HSV rpm over the ECE-EUDC cycle (A PV,H = 4 m 2, P = 13 kw, l = 3.5 m, w = 1.6 m, h =1.4 m) Engine spee [rpm] Reference car HSV Time [s] Since, for the selecte case, the optimal is lower than one, the thermal engine can be operate at its highest efficiency in an intermittent way, as shown in FIG. 3 (green line). In the first part (i.e. in the time winow 4-8 s), ue the low-power eman, the is mainly use to recharge batteries. Afterwars, it concurs with batteries to power the vehicle. In case of conventional vehicle (blue line), instea, the works in most cases at partial loas, with higher values of specific fuel consumption. The resulting variation of battery state of charge (SOC) is shown in FIG. 4. Due to the constraine introuce by eq. (13), the final SOC iffers from the initial value by a fraction of the amount of energy store uring the parking hours. It is worth mentioning here that other strategies are possible, in case the coul run uring parking moe too: in that case, the engine can be use to restore battery charge by working always at its maximum efficiency VII. Time [s] PARAMETRIC ANALYSIS In orer to evaluate the energetic benefits of HSV configuration with respect to conventional vehicle, a parametric analysis has been carrie out, ranging the electric generator () power from 6 to 28 KW an the PV area from, corresponing to pure hybri electric vehicle, to 4 square meters. The analysis has been performe by imposing a constant overall max power of 5 kw, therefore a ecrease of power is compensate by an increase of battery moules. In this analysis, vehicle imensions have been kept constantly equal to the reference vehicle ones (l = 3.5 m; w = 1.6 m; h = 1.4 m). The analysis has been aime at comparing the fuel consumption of conventional an Hybri Solar vehicle, along a time horizon corresponing to the riving hours h. FIG. 5 shows the expecte tren of vehicle mass vs. power: as the power is increase, the number of battery moules, neee to maintain the impose maximum power, is reuce. This behavior results in a lighter vehicle since specific power (kw/kg) is greater than specific battery power. Furthermore the figure shows that the introuction of panels results in an increase of mass, almost linearly with PV area. The behavior of the control variable with the power is shown in FIG. 6: accoring to the control strategy aopte, as the power is increase the provies the requeste energy in a shorter time, thus reucing the operation time. Particularly, when the power is lower than 1 kw, the operation at the optimal power (P opt ) oes not guarantee the requeste energy. Thus, the is force to work at a greater power with reuce efficiency (> 1). This behavior results in the fuel savings trens shown in FIG. 7, that evience poor benefits in case of power lower than 1 kw. Furthermore, in case of conventional hybri vehicle (PV area = ), the fuel saving exhibits a negative value as the operation approaches the max power with very low efficiency (= 1.8). This negative tren is improve with the introuction of the panels that compensate for the low efficiency operation by recharging the batteries uring parking hours.

5 AVEC 6 FIG. 5 Vehicle weight vs. electric generator power for ifferent panels area Vehicle Mass [kg] PV Area = m 2 PV Area = 2 m 2 PV Area = 4 m Power [kw] FIG. 6 Control variable vs. electric generator power for ifferent panels area PV Area = m 2 PV Area = 2 m 2 PV Area = 4 m Power [kw] FIG. 7 Fuel savings vs. electric generator power for ifferent panels area Fuel Savings [%] PV Area = m 2 PV Area = 2 m 2-5 PV Area = 4 m Power [kw] As the power is increase over 1 kw, the operates at max efficiency (< 1) an the fuel savings exhibits an almost asymptotic tren. This behavior can be explaine consiering that, ue to the negligible vehicle mass variation, the requeste energy is almost constant an is provie with the same specific consumption in all cases where the control variable is lower than one. The figure also eviences that significant improvements of fuel economy can be achieve by introucing the panels, with max values approaching 25 % of savings with respect to conventional vehicle, in case of 4 m 2. It can also be observe that the benefits achieve by the series hybri vehicle without solar panels, with respect to the conventional vehicle, are relatively limite in this case, if compare with some results achieve by parallel HEV aopting avance control strategies [8]. This result is ue to the limite power aopte for the in the conventional vehicle (5 KW), to the cascae energy losses associate to hybri series structure an to the influence of riving cycle. Concerning the first effect, it is worth mentioning that the gain in fuel economy achievable by hybriizing a meium-high-power car (e.g. P max = 7 kw) can reach up to 2%, which is more than ouble the increase obtaine for the 5 kw vehicle analyze in the current work (see FIG. 7). The influence of the riving cycle is also significant; in case of pure urban cycle, the gain in fuel economy for the same vehicle (5 KW) reaches 25 %. This is ue to both the operation of the conventional vehicle at poor efficiency an the higher benefits of regenerative braking in urban riving. VIII. OPTIMIZATION APPROACH The moels presente in the previous chapters allow to achieve the optimal esign of the HSV via mathematical programming, consiering both technical an economic aspects. The payback is assume as objective function, while esign variables X are represente by electric generator power P, horizontal panel area A PV,H an car imensions (l,w,h). X PX X i 1 NG min (15) G, (16) i The inequality constraints G i express the following conitions: i) Car imensions, length to with an height to with ratios within assigne limits, as aresse by the available atabase on commercial vehicles. ii) PV panels area compatible with car imensions, accoring to the given geometrical moel. A. Optimization analysis The optimization analysis has been carrie out by minimizing the P evaluate by eq. (11). The reference vehicle consiere to evaluate the aily saving (i.e. eq. 1) has the following specifications: P = P max = 5kW; W boy,cc = 9 kg; l = 3.5 m; w = 1.6 m; h = 1.4 m. TA. III summarizes the results applying the optimization criteria efine through eqs. (15) an (16). Case 1, representing a series HEV without solar panels, is consiere as reference, with a payback of 6.74 years with respect to the conventional vehicle. The aition of solar panels (e.g. case 2, A PV,H = 2 m 2 ) allows getting a 13% contribution from solar energy, but results in a

6 AVEC 6 higher payback (from 6.74 to 1.63 years), ue to the actual costs of fuel an panels. TA. IIII Optimization results for ifferent fuel cost an PV technology scenarios. c f # c PV P A PV,H P P [ /kg] [ /m 2 ] [/] [m 2 ] [kw] [yrs] [%] The HSV can represent the optimal solution consiering the occurrence of one (case 3) or all (case 4) of the following circumstances: a) PV cost reuction (by a factor 4); b) fuel cost oubling (by a factor 2); c) panel efficiency increase from.13 to.16. The variations in prices an costs are significant but not unrealistic, consiering actual trens of reuction of solar component costs an increase in oil prices. The moel allows to estimate the optimal vehicle imensions as well: while for case 1 a reuction in imensions with respect to base case has been propose, to reuce vehicle weight an aeroynamic losses, in case 4 larger values are aopte (l=3.85, w=1.73; h=1.39), compatible with the selecte solar panel area (5.85 m 2 ). CONCLUSIONS A comprehensive moel for the stuy an the optimal esign of a solar hybri vehicle with series architecture has been presente. The moel escribes energy flows between horizontal an/or vertical solar panels, internal combustion engine, electric generator, electric motor an batteries, consiering vehicle longituinal ynamics an the effect of control strategies. Vehicle weight is preicte, starting from a atabase of commercial vehicles, consiering the effects of power-train sizing, vehicle imensions an possible use of aluminum. The effects of vehicle imensions on aeroynamic losses an maximum panel area also can be accounte for. The moel preicts the aitional costs with respect to conventional vehicles, an the pay-back. It has been shown that significant savings in fuel consumption an emissions can be obtaine with an intermittent use of the vehicle at limite average power, compatible with typical use in urban conitions uring working ays. This result has been obtaine with commercial PV panels an with realistic ata an assumptions on the achievable net solar energy for propulsion. The future aoption of last generation photovoltaic panels, with nominal efficiencies approaching 35%, may result in an almost complete solar autonomy of this kin of vehicle for such uses. y aopting up to ate technology for electric motor an generator, batteries an chassis, power to weight ratio comparable with the ones of commercial cars can be achieve, thus assuring acceptable vehicle performance. Future evelopments may concern a systematic stuy of optimal configuration for various riving cycles an latitues, also consiering seasonal variations of the solar energy, more accurate stuy of control strategies, incluing possible application of on-boar optimization couple with provisional methos for car loa an solar energy base on Recurrent Neural Network. More etaile moels for component weights an costs, incluing non-linear effects, also can be necessary, as well as further stuies on the interactions between vehicle an propulsion system. The results obtaine by optimization analysis over a ECE/EUDC cycle have shown that the hybri solar vehicles, although still far from economic feasibility, coul reach acceptable payback values if large but not unrealistic variations in costs, prices an panel efficiency will occur: consiering recent trens in renewable energy fiel an actual geo-political scenarios, it is reasonable to expect further reuctions in costs for PV panels, batteries an avance electric motors an generators, while relevant increases in fuel cost coul not be exclue. Moreover, the recent an somewhat surprising commercial success of some electrical hybri cars inicates that there are grouns for hope that a significant number of users is alreay willing to spen some more money to contribute to save the planet from pollution, climate changes an resource epletion. In orer to valiate the moel, a prototype of Hybri Solar Vehicle with series structure is being evelope at DIMEC, within a project fune by EU [16]. REFERENCES [1] Hamma M., Khatib T. (1996), Energy Parameters of a Solar Car for Joran, Energy Conversion Management, V.37, No.12. [2] Wellington R.P. (1996), Moel Solar Vehicles Provie Motivation for School Stuents, Solar Energy Vol.58, N.1-3. [3] Saitoh, T.; Hisaa, T.; Gomi, C.; Maea, C. (1992), Improvement of urban air pollution via solar-assiste super energy efficient vehicle. 92 ASME JSES KSES Int Sol Energy Conf. Publ by ASME, New York, NY, USA.p [4] Sasaki K., Yokota M., Nagayoshi H., Kamisako K. (1997), Evaluation of an Electric Motor an Gasoline Engine Hybri Car Using Solar Cells, Solar Energy Material an Solar Cells (47), [5] Seal M.R. (1995), Viking 23 - zero emissions in the city, range an performance on the freeway. Northcon - Conference Recor IEEE, RC-18.p [6] Seal M.R., Campbell G. (1995), Groun-up hybri vehicle program at the vehicle research institute. Electric an Hybri Vehicles - Implementation of Technology SAE Special Publications n SAE, Warrenale, PA, USA.p [7] S.Letenre, R.Perez, Christy Herig, Vehicle Integrate PV: a Clean an Secure Fuel for Hybri Electric Vehicles, Proc. of Annual Meeting of the American Solar Energy Society, June 21-26, 23, Austin, TX. [8] Arsie I., Graziosi M., Pianese C., Rizzo G., Sorrentino M. (24), Optimization of Supervisory Control Strategy for Parallel Hybri Vehicle with Provisional Loa Estimate, Proc. of AVEC4, Arhnem (NL), Aug.23-27, 24. [9] Statistics for Roa Transport, UK Government, [1] [11]Arsie I., Di Domenico A., Marotta M., Pianese C., Rizzo G., Sorrentino M. (25); A Parametric Stuy of the Design Variables for a Hybri Electric Car with Solar Cells, Proc. of METIME Conference, June 2-3, 25, University of Galati, RO. [12]Arsie I., Marotta M., Pianese C., Rizzo G., Sorrentino M. (25); Optimal Design of a Hybri Electric Car with Solar Cells, Proc. of 1st AUTOCOM Workshop on Preventive an Active Safety Systems for Roa Vehicles, Istanbul, Sept.19-21, 25. [13]Marion. an Anerberg M., PVWATTS An online performance calculator for Gri-Connecte PV Systems, Proc.of the ASES Solar 2 Conf., June 16-21, 2, Maison, WI. [14] [15]urch, S., Cuy, M., Johnson, V., Markel, T., Rausen, D., Sprik, S., an Wipke, K., 1999, "ADVISOR: Avance Vehicle Simulator", available at: [16]Leonaro Program I5//P/PP Energy Conversion Systems an Their Environmental Impact,