Enhancing Vehicle Road Efficiency by Regenerative Braking

ENHANCING ROAD VEHICLE EFFIENCY BY REGENERATIVE BRAKING A M Walker, M U Lampérth and S Wilkins Mechanical Engineering Department Imperial College of Science, Technology and Medicine, London, UK ABSTRACT Regenerative braking is an established feature of rail vehicles. The subject is under development in the road vehicle industry through recent developments in Hybrid Electric (HEV) and pure Electric Vehicles (EV). Up to 30% of the overall energy demand can be satisfied by energy saved through regenerative braking, significantly improving a vehicle’s overall efficiency. In addition, the brake force requirement on friction brake modules is reduced and, in-turn, their size can be reduced. The paper discusses an overall approach to brake system, powertrain and energy storage system components. Computer codes have been developed to simulate component performance and examine control strategies, which demonstrate significant reduction in friction brake use. A notable finding has been the improved energy retention by the incorporation of ultracapacitors. NOMENCLATURE A cw E g h m Vehicle front area (m2) Coefficient of aerodynamic drag Energy (W.h) Gravity constant (9.81 m/s2) Height (m) Mass (kg) t v η µ ρ Time (s) Speed (m/s) Efficiency Coefficient of rolling resistance Density / Density of air (kg/m3) 1 INTRODUCTION Hybrid Electric (HEV) and pure Electric Vehicles (EV) use more energy efficient technology than traditional vehicle technologies. They are also increasingly being seen as the answer to stringent emissions regulations in North America, Europe and Japan. Most major automobile manufacturers are involved in HEV and EV development programmes, with a few early production platforms and countless prototype vehicles already on the road. Although the prime mover and powertrain configuration is unique to each model, a common feature of all these vehicles is regenerative braking. The approach to vehicle energy efficiency, and incorporation of regenerative braking, requires that powertrain, brake, and energy storage components no longer be addressed separately. 2 REGENERATIVE BRAKING 2.1 Electrical Braking Retaining the vehicle’s inertial energy during deceleration, and saving for the next acceleration phase, is termed as "regenerative” braking. When returned, this energy reduces the demand on the power train. The vehicle’s overall efficiency is improved such that up to 30% of the overall energy demand can be satisfied through energy recaptured during regenerative braking (1). The level of energy recapture depends on the capacity and efficiency of the electric traction motor, which operates as a generator during deceleration, and the capacity and efficiency of the power control and energy storage system, connected to the motor. Only in exceptional circumstances (where sustained over power results in very high temperature) is the motor unable to sustain the regenerative energy flow. The critical component in most cases is the energy storage system, and the requirement for rapid charge acceptance during heavy braking. Improvement, in braking performance, of a regenerative braking system can only be achieved by development, in terms of efficiency and response, of energy storage devices. 2.2 Friction Braking Where a regenerative braking system is used, a friction brake system is still required, even if seldom employed. A brake control architecture is required for interaction of the components and communications. This brake control architecture might be similar to those in existence on rail vehicles (2), but is not addressed here. As regenerative braking assumes the burden of most brake actuation, the brake force required from friction brake modules is reduced, and a reduction in physical size is possible. 2.3 Mechanical Integration To integrate the friction brake and electric traction motor, particularly in axle-less vehicles (such as low floor buses), combination units are now under development (3), (4). 3 ENERGY STORAGE 3.1 Energy Conversion Multiple energy conversions (i.e. Ekinetic → Eelectric → Ekinetic) are present in vehicle operation. The vehicle’s movement can be described as an energy storage mechanism – the energy of which is converted to electricity or other while the vehicle is at rest. Under this description, braking is the means by which the vehicle’s kinetic energy is returned to the on-board electrical or mechanical storage devices. The powertrain is the means by which the on-board stored energy is converted into kinetic energy (i.e. vehicle movement), along with additional requirements to overcome losses. 3.2 Energy Density / Power Density Energy storage in chemical (fuel) benefits from a high energy density, illustrated in Fig. 1, but features an undesirable conversion process. Traditionally, only a small level of electrical energy storage is required in the vehicle for restarting the prime mover. EV technology depends fully on electrical energy storage, and these limitations have hindered vehicle range to date. Fig. 1 Energy density (W.h/kg) in mechanical, chemical and electrical forms (5). Hydrogen includes the containment. 3.3 Electrical Energy Storage Device 3.3.1 Battery Development Practical EV development requires batteries that provide performance comparable with conventional vehicles, and at a comparable cost. Although cost effective, standard lead acid battery technology provides a very limited range. Table 1 Battery Development Objectives for Vehicle Applications (6). Energy Density Power Density Useful Life Cost (W.h/kg) (W/kg) (year) ($/kW.h) Vehicle Application 200 400 10 100 Requirements With relatively high charge acceptance, lead acid technology has acceptable power density, but poor energy density. Lithium-ion polymer technology is closer to providing acceptable energy and power density. However, cost is still a significant issue (Lead acid 120-150 $/kW.h, Lithium-ion 200 $/kW.h, and Nickel/metal hydride 200-350 $/kW.h), although work is proceeding on this (7), (8). Table 2 Operational Properties of Existing Electrical Energy Storage Devices (9). Energy Storage Energy Power Energy Cycle Life Self System Density Density Efficiency (Num of Discharge (W.h/kg) (W/kg) (%) Cycles) (%/48 h) Lead acid (10) 35-50 150-400 >80 500-1000 0.6 Nickel/cadmium (10) 40-60 80-150 75 800 1 Nickel/metal hydride 70-95 200-300 70 750-1200+ 6 (10) Lithium-ion (10) 80-130 200-300 >95 1000+ 0.7 Ultracapacitor (11) 2.5 >1270 (12) 92- 98 10,000hr na. (12) Ultracapacitor cycle life is under average operating temperature of 35 ˚C, and voltage of 2.3 V 3.3.2 Charge Acceptance and Ultracapacitors Further development of battery technology for vehicle applications has focused on high charge and discharge rates. The rapid, and efficient, charging of the battery, the number of cycles per life, and full operation on low state-of-charge is important. Electrochemical battery life is dictated by a limited number of charge-discharge operations. An alternative storage medium for absorbing and releasing large amounts of energy quickly is the double-layer ultracapacitor. Ultracapacitors have a very high rate of charge acceptance and discharge characteristic (12), which is appropriate in the vehicle application for quick acceleration and energy absorption from braking. Energy storage based on 2000-3000F ultracapacitors assists regenerative braking better than batteries by buffering the energy generated (through high current charging). However, a high discharge rate cannot be sustained for a long period, rendering a low energy density. The energy in the powertrain (i.e. when the storage unit is used for lower power output) still requires optimising through a battery, although the million-plus cycle life of the ultracapacitor greatly extends the battery life. Ultracapacitors cost 270 $/kW.h (from a projected $30 per 2700F unit). A city bus with a dieselelectric powertrain and an ultracapacitor has demonstrated 24% vehicle power from regenerated brake energy (13). 4 SIMULATION The selection of powertrain components, the power requirements and the optimisation of control systems involve complex trade-offs, such as performance versus energy consumption and cost (14), (15), (16), range, emissions and acceleration. Comprehensive, computer-based simulation models, which accurately represent the power train and power requirements, are helpful in examining the powertrain configuration as these parameters can be evaluated in a systematic approach and sensibly matched to the vehicle application. 4.1 Simulation Program The programming approach is modular, such that components can be respecified and retested without rebuilding the entire system. Evaluation may be performed under any user defined driving cycle, or embedded schedules, such as the FUDS (Federal Urban Driving Schedule) or the FHDS (Federal Highway Driving Schedule). There are three modes of simulation. 4.1.1 Non-feedback Mode The non-feedback mode is the most commonly used where the simulation starts from a given drive cycle (time step, velocity, road gradient) and vehicle configuration (powertrain specifications, component sizes etc.). For every time step the energy demand is calculated and then the performance in terms of fuel consumption, driving range and emissions are computed. 4.1.2 Cycle Feedback Mode Cycle feedback is similar to the non-feedback mode, except that the maximum capacity of the powertrain components dictates the vehicle performance, as in reality, in attempting to satisfy the simulated driving cycle. In this case, the vehicle velocity-time profile will stray from the pre-defined profile, which can be seen by overlaying the velocity profiles, see Fig. 2. Fig. 2 Performance deterioration with an undersized component. 4.1.3 Powertrain Feedback Mode The powertrain feedback mode is designed to determine the sizing of the powertrain components using feedback from the energy demand of the driving cycle. Hence, from a given vehicle arrangement and drive cycle, the program sizes the different powertrain components in order to match the requirements in terms of power and required range. This proves to be a particularly useful feature where parametric studies are required for a certain design. 4.1.3 Vehicle Dynamics Equations For the simulation, the program uses vehicle dynamics equations for the calculation of the energy demand for every time step in the driving cycle. These equations include: Etractive = EKinectic + EPotential + Erolling + Eaero [1] where kinetic energy during a time step is E Kinetic = 1 m (vE2 − vS2 ) 2 [2] change in potential energy is ( E Potential = mg h E − hS ) [3] losses due to the rolling resistance are ( ) E rolling = mgµ v E − v S ∆t [4] and losses due to aerodynamic drag, calculated by integrating over the time step, are c Aρ t 3 E aero = w ∫ v( t ) dt 2 t − ∆t [5] When Etractive > 0, the powertrain provides energy to the motor to drive the wheels, given by E Powertrain = Etractive ηmotor + E DriveTrain [6] where EDriveTrain represents the energy losses due the inertia of the drivetrain and friction. In Equation [6] ηmotor is the motor efficiency, which is a function of motor speed and torque calculated from the motor map. When Etractive < 0, braking operates, and energy is recovered and fed back into the batteries, given by E Powertrain = E tractiveη regen − E DriveTrain [7] where ηregen is the regeneration efficiency (also a function of motor speed and torque, determined from the motor map). 4.1.4 Battery Model Batteries are modelled using a “hydrodynamic two-tank model” (17). This is a validated analogy to the electrochemical reaction occurring within the battery. The procedure takes into account the battery time history in order to estimate the actual state of charge. Other energy storage devices are similarly modelled. 5 HYBRID H.G.V. EXAMPLE Based on limited driving, the full traction and braking loads on a Hybrid H.G.V. truck have been investigated by simulation in a Non-Feedback Mode. This Hybrid vehicle has a Series Hybrid topology, with electrical rather than mechanical connection between the powertrain and the wheels, illustrated in Fig. 3a. The Parallel Hybrid topology, illustrated in Fig. 3b, with mechanical and electrical powertrain connection, has proportionately less electrical traction power. The Series topology benefits from electrical braking more than the Parallel. Fig. 3a Series Hybrid Electric Vehicle Systems Layout Fig. 3b Parallel Hybrid Electric Vehicle Systems Layout 5.1 Drive Cycle A drive cycle has been selected to be demanding on both traction power and braking power, as traction power is the determining factor in the vehicle configuration (18), (19). To date, no commonly used drive cycles exist, which demand heavy braking due to gradients, so a new drive cycle has been rendered. The selected route is part of Edinburgh’s inner ring road – the streets Drum Brae North and Drum Brae South. Heavy braking was implemented due to the steep gradients involved, illustrated in Fig. 4. The cycle features an altitude climb of 50 m and descent over the 2 km street length, and has the same altitude at start and finish to ensure no overall energy change. The curvature of the Drum Brae streets is negligible, allowing even distribution on wheels. 15 10 5 0 1 21 41 61 81 101 121 141 161 181 201 -5 -10 -15 Time [s] Speed [m/s] Gradient [%] Fig. 4 Drive Cycle on Drum Brae North and South, Edinburgh. 5.2 Vehicle Speed The vehicle speed was ramped to 10 m/s (36 km/h) prior to imposing the drive cycle. The speed was then maintained constant over the duration of the cycle, illustrated in Fig. 4. This ensured substantial, high-current discharge from the energy storage during ascent, and a rapid, high-current charge during descent. 5.3 Vehicle Configuration Two configurations were implemented, with different electrical braking capacities. Availability of electrical braking depends on available charging capacity of the energy storage device. A “low charge acceptance” requires comparatively more use of the friction brake. A “high charge acceptance” allows more use of electrical braking and requires comparatively less use of the friction brake. This partially corresponds to the difference between Hybrid topologies. A high charge acceptance vehicle features more electric powertrain (the Series) than a vehicle with low charge acceptance (the Parallel). 5.3.1 Configuration A – “Low Charge Acceptance” parallel hybrid vehicle The general characteristic of “low charge acceptance” was achieved through a standard traction motor and a standard energy storage device. Without an alternative means for “dumping” excess regenerated energy (such as a brake resistor), electrical braking is effectively useless when the energy storage device is “full” (i.e. has 100 % state-of-charge). 5.3.2 Configuration B – “High Charge Acceptance” series hybrid electric vehicle “High charge acceptance” was achieved through a standard traction motor and a larger energy storage device. This energy storage device comprised a standard-sized battery and, importantly, an ultracapacitor, in order to attain rapid, high current charging, which the battery was unable to sustain. 5.3.3 Control Strategy In both cases, the control strategy used to implement powertrain optimisation was based on the state-of-charge of the energy storage device. The braking control strategy ensures that brake applications are achieved initially through electrical braking, with friction braking only employed where the electrical braking application is insufficient. Configuration A B Table 3 Vehicle Configurations. Charge Acceptance Motor Size (kW) Energy Storage Device Size Battery (Ah) Ultracapacitor (F) Low 90 180 – High 90 180 50+ 5.1 Results 100 80 60 Power [kW] 40 20 0 -20 1 21 41 61 81 101 121 141 161 181 201 -40 -60 -80 -100 Time [s] Required Power [kW] Recovered Energy [kJ/s] Friction Brake [kW] Fig. 5 Performance for Configuration A over drive cycle. Table 4 Total Brake Energy for Configuration A and B over Drive Cycle. Configuration Charge Total Brake Energy over Electric Brake Friction Brake Acceptance Drive Cycle (W.h) Energy (W.h) Energy (W.h) A Low 600 285 315 B High 600 580 20 6 CONCLUSIONS In the Series Hybrid Electric vehicle, braking through the electric motor is the norm, and the requirement for friction braking becomes greatly reduced. The level of this reduction is dependent on the charge acceptance capacity of the energy storage device employed. A new, heavy-braking drive cycle has been rendered to facilitate simulation of Hybrid Electric and pure Electric Vehicles. A Hybrid H.G.V. truck, with a Series Hybrid topology, has been simulated driving over this drive cycle at constant velocity. With a low charge acceptance configuration, due to a standard energy storage device, the friction brakes were employed regularly. With a high charge acceptance level, due to the inclusion of an ultracapacitor with the standard battery, the friction brakes were rarely used. Therefore, friction brake modules on Hybrid Electric vehicles can be reduced in size to service a reduced requirement, and a control architecture should be developed to optimise the combined use of electric and friction braking. 7 REFRENCES (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16) Mitsubishi data sheet: “Motor Vehicle Brake Energy Conservation System – III”, 2001 Knorr-Bremse data sheet: “Distributed EP-Brake Controller EP2002”, 2001 Meritor data sheet: “Axles Designed for Ultra-Low Floor City Buses”, 2001 Irisbus data sheet: “Civis: New System of Electric Propulsion”, 2001 G. Zavitsanakis, P. Lambrou, N. Economidis, P. Lagias, “Energy Efficiency Simulation of a Hybrid Vehicle for Battery Use Optimisation”, UG report Imperial College, 1996 U.S. Advanced Battery Consortium, 1998, <http://www.uscar.org> E. Forouzan, B. Jay, “Improved, Low-Cost Lithium Polymer Battery System for Electric and Hybrid Vehicle Applications”, Electrosource, Inc. promotion material, 2001 D. Clerc, M. Fay, L. Wang, “High-Capacity Rechargeable Batteries Using Tin-Based Electrodes”, T/J Technologies, Inc. promotion material, 2001 A. Pal, D. Dareing, “Regenerative Braking Systems with Energy Capture for Heavy Duty Vehicles – A Research Objective”, GATE Center, University of Tennessee, Knoxville, 2001 D. A. J. Rand, R. Woods, R. M. Dell, “Batteries for Electric Vehicles”, Taunton Research Studies Press, 1998 Northeast Advanced Vehicle Consortium, 2000, <http://www.navc.org> Maxwell data sheet EPCOS data sheet: “Optimised Starting – Economic Driving”, 2001 C. Leontopoulos, M. R. Etemad, K. R. Pullen, M. U. Lamperth “Hybrid Vehicle Simulation for a Turbo-Generator Based Powertrain”, Journal of Automobile Engineering, Proc IMechE Part D –Vol 212 Part D pp 357-368. 1998 K. R. Pullen, M. R. Etemad, M. U. Lampérth, “Description of a Performance Simulation Methodology and Software for Electric and Hybrid Vehicles”, Paper 97EL036, 30th ISATA Proceedings – Electric, Hybrid and Alternative Fuel Vehicles, Italy, 1997 M. U. Lampérth, “Hybrid Vehicle Simulation”, MSc Imperial College, University of London, 1996 (17) M. U. Lampérth, K. R. Pullen, M. R. Etemad, “Development of an Analysis Software to Allow Electromechanical Battery Modelling for Electric and Hybrid Vehicles” Paper 97EL035, 30th ISATA Electric, Hybrid and Alternative Fuel Vehicles, Florence, 1997 (18) M. U. Lampérth, K. R. Pullen, “How size and Performance of Hybrid Electric Vehicle Components are Influenced by Acceleration Patterns”, SAE Technical Paper Series 1999-01-2909 Future Transport Technology, California, 1999 (19) M. U. Lampérth, K. R. Pullen “Impact of Acceleration Patterns on Size and Performance of Hybrid Electric Vehicle Components”, 32nd ISATA – Electric, Hybrid and Alternative Fuel Vehicles, Florence, 1999 View publication stats