Investigation of the Idea of Active Suspension System Application in Hybrid Electric Vehicles

Transcription

1 Invetigation of the Idea of Active Supenion Sytem Application in Hybrid Electric Vehicle Seyedmohen Hoeini, Ruhed Abdolah, and Amir Khani Abtract - Thi paper decribe the idea of application of active upenion ytem in hybrid electric vehicle (HEV). For the invetigation of thi idea, a methodological approach for imultaneou imulation of power train and upenion ytem i preented. In thi approach, the hybrid electric power train and the active upenion ytem imulation are integrated and the exchange of power and data between two ytem i imulated. Conidering the effect of power train imulation input, introduced a driving cycle on the upenion diturbance, two combined road-peed diturbance are generated and are applied to the upenion ytem. Furthermore, energy interaction of the AS and power train ytem i modeled for both conventional and hybrid electric power train. The imulation reult how the ability of the approach for invetigation of the active upenion ytem power demand a well a fuel conumption, emiion, and the ride characteritic of the hybrid electric vehicle equipped with active upenion ytem. Keyword: Active Supenion- Hybrid Electric, Vehicle- Simultaneou Simulation I. Introduction Vehicle upenion ytem i reponible for driving comfort and afety a the upenion carrie the vehicle body and tranmit all force between body and road [1]. In order to poitively influence thee propertie, active component are introduced. Active component enable the upenion ytem to adapt to variou driving condition. By adding a feedback controlled actuator, driving comfort and afety are coniderably improved compared to upenion etup with fixed propertie [2]. Although auto indutry ha witneed coniderable advancement in active upenion (AS) ytem technology in recent year, the main barrier ahead of widepread application of thi ytem in conventional vehicle i it high energy demand and cot. In addition, application of AS ytem in conventional vehicle encounter ome problem that may affect it performance. In conventional vehicle, the power required for the active upenion ytem i powered by the vehicle engine. Due to it dynamic, the internal combution engine (ICE) require time to adjut to the time-varying load impoed by AS. If thee load vary fater than the engine can adjut, then the Seyedmohen Hoeini i with the Faculty of Engineering, Kington Univerity. ( .m.kimia@gmail.com) Ruhed Abdolah i with Bradford Univerity. ( Bradford_uk0@yahoo.com) Amir Khani i with Univerity of Tehran, Tehran. ( P.alomeh48@gmail.com) upenion will, temporary, not function properly while the vehicle itelf may heitate. If, on the other hand, the load get reduced too fat, vehicle urge may occur [3]. Beide, AS load may hift the engine operating point to inefficient region and increae it intant emiion and fuel conumption. Hybrid Electric Vehicle (HEV), on the other hand, are projected a one of the olution to the world need for cleaner and more fuel-efficient vehicle. HEV employ two (ore more) energy torage ource and the aociated energy converter to generate the power required to drive the vehicle and to operate on-board acceorie. Mot typically, the architecture of thee vehicle include an internal combution engine with an aociated fuel tank and an electric machine with it aociated battery (energy torage) ytem. In the HEV, unlike the conventional vehicle, the ub ytem may be powered by either mechanical combution engine or the electrical power bu ource. Thi can be epecially helpful when one of the power ource can t provide the load due to it performance contraint. In thi paper, the idea of application of AS upenion ytem in HEV i introduced. The motivation of thi tudy i the poible advantage of thi application over the conventional one. For thi tudy, the modeling, imulation, and control of the active upenion ytem and the hybrid electric power train are firtly decribed eparately. Simulation of the AS i performed baed on a even DOF model aociated with peudo kyhook control approach. In addition HEV imulation i preented uing a fuzzy control cheme. An approach for the imultaneou imulation of two ytem i then propoed and the application of AS in HEV i invetigated. In the firt tage and a a preliminary goal, providing a imulation tool in which the upenion and power train ytem are imulated imultaneouly i conidered. In the imulation media, the data and energy exchange between two ytem i modeled. The problem of time tep for the coherent imulation of two ytem i dicued. In addition, combination of the road irregularitie and the vehicle peed pattern impoed by the driving cycle i conidered. Moreover, computer imulation are conducted and ytem imultaneou imulation reult are tudied. AS ytem power conumption in both bumpy and random terrain are calculated conidering vehicle peed pattern. Furthermore, the ICE performance in either conventional and hybrid electric power train ytem i examined while it i experiencing AS ytem load. Beide, the conventional and the hybrid electric vehicle fuel conumption and emiion reult with and without AS load are compared.

2 II. Active Supenion Modeling and Simulation The regular even degree of freedom model (Fig. 1) i ued for the upenion ytem imulation. Thi model i a proper choice for ride quality tudy a it capture the effective degree of freedom employed in the ride quality evaluation, according to ISO 2631 tandard criteria i.e. body bounce, pitch and roll acceleration. It i aumed that body and wheel diplacement are meaured from their tatic equilibrium poition. Tire-road holding force i alo etimated a a nonlinear function uing the following equation: In order to expand the kyhook control law to the even DOF model, it i aumed that the body i upended at it corner by kyhook damper. Conidering thi cenario, the even DOF model ha been divided to four eparate quarter model (Figure 3). The logic behind thi control deign i that, by reducing the tranmitted force from upenion ytem to the body at it corner, the total tranmitted force to the body i reduced. Thi reduction, conequently, will reult in reduction of body bounce, pitch and roll acceleration. Conidering thi inference, the body motion i controlled at it corner uing the kyhook trategy where the active upenion actuator force have been calculated at the body four corner. Comparing the equation of motion of the kyhook model and the upenion quarter model (Equation 2 and Equation 3), the active upenion actuator force can be etimated uing equation 4. Figure 1: Supenion ytem even DOF model. F holding = k t (x u -x 0 ) if (x u -x 0 ) Δ tat F holding = 0 if (x u -x 0 ) Δ tat The upenion pring tiffne in the active model ha been decreaed up to half of it original value in the paive model. Supenion damper in the active model are alo replaced with active actuator. III. Active Supenion Control In the active upenion ytem, a feedback controlled actuator i ued to improve the driving comfort and afety. The active component enable the upenion ytem to adapt to variou driving condition. Trade off between the ride comfort and the vehicle tability ha been driving force for advancement in automotive upenion [2, 4 and 5]. One of the main objective of the active upenion deign in thi work i to improve ride quality. The other objective i to keep the tire load within an acceptable limit. For control force calculation in thi tudy, the well known kyhook control trategy patented by Karnop [6] i employed. We call thi force quai-kyhook force to ditinguih the difference between a real kyhook upenion and the active upenion ytem conidered in thi work. A the name implie, the kyhook configuration ha a damper connected to ome inertial reference in the ky, a hown in Figure 2. Uing a kyhook configuration, the tradeoff between reonance control and high-frequency iolation, common in paive upenion, i eliminated [7]. Figure 2: Skyhook model Figure 3: Supenion ytem quarter model

3 x = [ k ( x ) + C ( x x u ) + F] (2) M x u = [ ku ( x ) + C ( x ) + M.. u 1 [ k ( ) (. x Ckyhook x F] x = + )] (3) M.. x u = M 1. [ K ( ) ( ) (. x kt xu x0 Ct xu x0 u )] V. HEV Power Train Control Deign A parallel HEV incorporate two power drive including ICE and electric motor (EM). Therefore, it i the reponibility of the parallel HEV control trategy to determine how to ditribute the driver required torque between the ICE and EM. The HEV control trategy i aimed at everal imultaneou target uch a minimization of the fuel conumption (FC) and exhaut emiion (HC, CO and NOx). Unfortunately, thee aim are often in conflict with each other i.e. the minimum fuel conumption doe not necearily reult in the minimum emiion, implying the need for a trade off olution.. F = C kyhook. x (4) IV. HEV Power Train Simulation In thi ection, the power train imulation i preented to obtain an acceptable etimation for the fuel conumption and exhaut emiion of conventional and hybrid electric vehicle. The power train imulation method ued in thi tudy incorporate a backward-facing approach [8]. Backward-facing approach i hown in Figure 4 a a chematic diagram. Thi approach doe not require any model for driver behavior. Intead, a driving cycle i traced by the vehicle. Vehicle required traction force and velocity are then tranlated into the required torque and velocity that mut be provided by the vehicle wheel and other component uptream. Thi calculation approach carrie backward through the driveline againt the tractive power flow direction until the fuel ue and/or the electrical energy ue that would be neceary to meet the trace i computed. Backward approach i greatly dependent on the internal combution engine (ICE) pecific fuel conumption (SFC) and emiion map. Figure 4: Backward facing imulation chematic The parallel configuration of the HEV ha been conidered for thi tudy. In addition, ADVISOR (Advanced Vehicle Simulator) i employed for the backward imulation of both conventional and HEV. However, the kernel of the HEV modeling and imulation i the control deign dicued in the next ection. The bottom line for control trategy i that the vehicle mut follow the driver requet. Thi mean that the total torque delivered by the ICE and EM mut be determined uch that the driver torque requet (from brake and accelerating pedal) are atified conitently. The driver requet i equivalent to the driving cycle. Therefore, control trategy mut perform uch that the driving cycle i tracked adequately. Another aim for control trategy i to maintain or enhance vehicle performance like grade ability, acceleration, etc. In thi tudy, PNGV paenger car contraint [9] are ued to enure that the vehicle performance i not acrificed during the trade off olution. One of the other contraint for control trategy are remaining charge-utained. Thi contraint ha been introduced to force the battery SOC to recover it initial value at the end of the driving cycle. In thi work, the difference between final SOC and initial SOC i forced not to exceed 0.5% of the initial SOC. Fuzzy control approach i employed to devie HEV control deign in thi tudy. The main objective of the fuzzy controller i to caue the ICE to work in the vicinity of it optimal operating point. The optimal operating point of the engine are determined baed on ICE parameter at the current vehicle peed to minimize intantaneou fuel conumption and emiion. At any particular point in time, the peed of rotation for the ICE i determined baed on the power train configuration and the current gear ratio. Thi i the peed at which the intantaneou optimization i performed. For the current peed, all poible torque that the ICE can provide are conidered. Then the fuel conumption (FC) and emiion (HC, CO and NOx) for all torque at the current peed are taken from the engine map and the following cot function i calculated for all of thee point: in which all variable are normalized to the correponding target value limit. The target value of the fuel conumption i defined by the deigner, while for the emiion, the tandard of tailpipe emiion are ued. Figure 5 illutrate the optimal and maximum torque curve veru engine velocity. (5)

4 Table 1 FLC Rule Bae Figure 5: Optimum engine torque Figure 6 depict the fuzzy controller chematic diagram. A it can be een, the controller input are required tractive torque and batterie tate of the charge. The controller output i alo EM torque determined baed on the controller input and rule bae. VI. Simultaneou Simulation of HEV Power Train and Active Supenion Sytem Vehicle power train and active upenion ytem are tudied in a novel coherent imulation approach. In thi approach, the vehicle power train imulation i integrated with the active upenion imulation a hown in Figure 7. In thi figure, data and energy flow are hown. Uing thi approach, power train emiion and fuel conumption a well a the vehicle dynamic characteritic including the ride comfort parameter and the tire-road holding force i calculated imultaneouly. Figure 7 how the chematic diagram of thi approach. Figure 6: Hybrid Power train Control Strategy The rule bae i defined by a et of 9 rule, a lited in table 1. The general logic behind thee rule i the idea of load leveling, in which the EM i ued to ait or generate, while it i running the ICE near optimal operating point. In fact the optimal operating point of the ICE i varied baed on SOC contraint. For intance, conider a cae (the lat rule in rule table) when the required torque i above the optimal operating point. Let u aume that the SOC i high. We would like to bring the ICE operating point near the optimal operating point (for that peed). Thi would mean a lower torque output by the ICE than what i required to meet the driver demand. Thi require that the EM to be run a a motor to make up for the remaining torque provided that there i enough battery charge. Since we do have ufficient charge in thi cae, the ICE i allowed to operate near the optimal operating point. Figure 7: Schematic diagram of imultaneou imulation of power train and active upenion ytem. A it i een in thi figure, the problem of imultaneou imulation i divided into three ection including environmental input, vehicle dynamic, and output. Each ection i dicued eparately in detail in the following. The environmental input of the upenion and power train ytem are quite different. Supenion ytem diturbance incorporate the road irregularitie including random or bumpy ignal, while the power train ytem diturbance i a predefined vehicle peed pattern, called driving cycle. In addition, upenion ytem input depend trongly on vehicle peed which i determined by the driving cycle. Therefore, the upenion ytem diturbance mut be tuned continuouly according to vehicle peed. In thi tudy, two et of road-cycle input are contructed for the imultaneou imulation purpoe. In the firt etup named Bumpy-ECE, the vehicle trace the ECE driving cycle while it pae over a

5 bumpy road. In thi cae, the ICE engine i tarted in the cold condition and the emiion ampling i tarted immediately after the engine tartup. In the econd imulation etup named Rand-Cont, the vehicle trace a contant km/hr peed while it pae over a mooth random road. In thi cae, the IC engine i tarted in the hot condition and the emiion ampling i tarted after engine i reached the teady condition. Figure 8, 9, and 10 how the ECE driving cycle, bumpy road, and the combined bumpy-ece road diturbance, repectively. In addition, the upenion and the power train are two ytem with different dynamic. Power train dynamic, in a backward facing approach, i motly baed on a low dynamic uing the lookup table repreented motly by algebraic equation while the upenion dynamic i introduced by much fater differential equation. The acceptable numerical time tep i therefore crucial in the imultaneou imulation of thee two ytem. In order to tackle thi problem, the power train ytem time tep ha been fined up to an acceptable active upenion time tep level. Furthermore, the active upenion and power train ytem have interaction in term of energy. The active upenion required power i demanded from the power train ytem continuouly. In the conventional vehicle, thi power load i impoed on the ICE while in the HEV it may be demanded form the electric power bu. In thi tudy, energy interaction of AS ytem with conventional and hybrid electric power train ha been conidered eparately. In the conventional vehicle, the active upenion load (Equation 6), i modeled a an extra torque added up to the engine required torque. In the hybrid electric vehicle, on the other hand, thi load i conidered a an additional electrical acceorie load requeted from the electrical power bu. P act = F V = C. V.( V V ) (6) act. act kyhook u Figure 8: Tuned bumpy-ece diturbance Figure 9: Tuned bumpy-ece diturbance VII. Simulation Study and Reult Analyi The advanced vehicle imulator (ADVISOR) [8] and MATLAB-SIMULINK are ued for imulation tudy. Simulation are performed for both conventional and hybrid electric vehicle. Uing two power train configuration and two upenion option, four poible combination of the power train and upenion ytem i.e. hybrid electric or conventional vehicle with active or paive upenion have been conidered for imulation. Performing computer imulation, the vehicle dynamic repone a well a the power train emiion and fuel conumption have been extracted. In addition, the body bounce and pitch acceleration a well a the tire road holding force are computed by imulation. Furthermore, the impact of active upenion load on the combution engine a well a the energy torage ytem ha been examined. Figure 11 depict the random road diturbance. In addition, Figure 12, 13, and 14 how the body bounce acceleration, upenion travel, and tire-road holding force repectively. Figure 11: Random road profile Figure 10: Tuned bumpy-ece diturbance

6 Thee reult how that, uing the peudo-kyhook active upenion ytem, a coniderable decreae in body acceleration i occurred. Thi confirm the devied AS ytem capability in improving the ride quality. Figure 12: Body bounce acceleration Figure 13: Supenion travel Figure 14: Tire holding force 1: Conventional, random+ contant 2: Conventional, bumpy+ ECE 3: HEV, random+ contant 4: HEV, bumpy+ ECE Figure 15: Emiion and fuel conumption reult Figure 15 depict the emiion and fuel conumption reult for four vehicle-road-driving configuration. A it can be een in thee Figure, AS ytem load ha caued an increae

7 of the fuel conumption and emiion in both conventional and hybrid electric vehicle. However, thi increae i more coniderable for the Bumpy-ECE road-cycle input. In addition, Figure 16 and 17 depict the AS ytem intant power demand for different road-cycle input. A it can be een in thee Figure, the AS ytem intant power demand for the ECE-Bumpy input i greater than the AS intant power demand for the Rand-Cont road cycle input Moreover, AS intant power demand increae dratically a the vehicle velocity increae in Bumpy-ECE road-cycle input a hown in Figure 17. Figure 18: ICE required torque in conventional vehicle with ECE-bumpy input Figure 16: Power conumption of AS-random road Figure 19: ICE required torque in HEV vehicle with ECE-bumpy input Figure 17: Power conumption of AS-bumpy-ECE Furthermore, Figure 18 and 19 how the ICE required torque for both the conventional and hybrid electric vehicle where an ECE-bumpy input i applied. A it can be een in thee Figure, the intant torque impoed on the ICE by AS ytem i quite large at ome pecific condition in the conventional vehicle. In thi cae, the ICE operation will reult in a more fuel conumption and emiion. In addition, the engine might not be able to provide the required torque at ome pecific point that may affect the performance of the AS ytem. However, uing an HEV configuration, the intant torque impoed on the ICE by the AS ytem i reduced coniderably, leading to more efficient condition. The required torque i alo within the ICE maximum torque limit, providing the neceary AS ytem power at all intant condition. VIII. Concluion In thi paper, a methodological approach i preented for invetigation of the idea of the application of active upenion ytem in hybrid electric vehicle. In thi approach, the active upenion and power train imulation are integrated and a imultaneou imulation tool i developed for energy tudy. Four configuration including the conventional and hybrid electric vehicle a well a the ECE-bumpy and Contant-random input are conidered. The imulation reult including the ride quality parameter a well a the fuel ue and emiion how the effectivene of the approach for thi tudy. The reult reveal that the intant torque impoed on the ICE by AS ytem i quite large at ome pecific condition in the conventional vehicle. However, uing an HEV configuration, the intant torque impoed on the ICE by the AS ytem i reduced coniderably, leading to more efficient condition.

8 Reference [1] Reimpell, J. and Stoll, H., 1996, The Automotive Chai: Engineering Principle. [2] Ficher, D., Iermann, R., 2004, "Mechatronic emi-active and active vehicle upenion", Control Engineering Practice, 12, [3] Mrad, R. B., Faoi, S. D., and Levitt, L. A., 1998 "A Polyniminal-algebric method for nontationary TARMA ignal analyi- Part II: Application to modeling and predicting power conumption in automobile active upenion ytem", Signal Proceing, 65, [4] Goncalve, F. D., 2001 "Dynamic Analyi of Semi-Active Control Technique for Vehicle Application", M.S.C Diertation, Virginia Polytechnic Intitute and State Univerity. [5] Mounir, M., Kamal, M., Wolf, Jr., and Joeph, A., 1982, Modern Structural Analyi, New York. Van Notrand Reinhold Company. [6] Karnopp, D., Croby, M. J., 1974, "Sytem for Controlling the Tranmiion of Energy Between Spaced Member". United State Patent #3,807,678. [7] Yi, K., Hedrick, K., 1993, "Dynamic Tire Force Control by Semiactive Supenion", Journal of Dynamic Sytem, Meaurement, and Control, 115, [8] Wipke, K. B., Cuddy, M. R., and Burch, S. D., 1999, " ADVISOR 2.1: a uer-friendly advanced Power train imulation uing a combined backward/forward approach", IEEE Tran. Vehicular Technology, 48 (6), [9] Moore, T.C. and Lovin, A.B., 1995, "Vehicle deign trategie to meet and exceed PNGV goal", Electric and Hybrid Vehicle Implementation of Technology SAE Special Publication SP-1105, [10] Schafer, F. and Bahuyen, R. V., 1995, Reduced Emiion and Fuel Conumption in Automotive Engine.