Comparison of Battery Life Across Real-World Automotive Drive-Cycles 7th Lithium Battery Power Conference Las Vegas, NV Kandler Smith, Matthew Earleywine, Eric Wood, Ahmad Pesaran November 7-8, 2011 NREL/PR-5400-53470 NREL is a national laboratory of the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, operated by the Alliance for Sustainable Energy, LLC. NREL is a national laboratory of the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, operated by the Alliance for Sustainable Energy, LLC.
Motivation • Overcome barriers to clean, efficient transportation o Electric-drive vehicles • Maximize life, minimize cost of electric drive vehicle batteries (alt: maximize income) • Quantify systems-level tradeoffs for plug-in hybrid vehicle (PHEV) batteries o 3000-5000 deep cycles o 10-15 year calendar life at 35°C o $300/kWh at pack level (2014 target ~ 70% reduction) 2
DOE’s Computer-Aided Engineering of Batteries (CAEBAT) Program Integrating Battery R&D Models Physics of Li-Ion Battery Systems in Different Length Scales Electrode Scale Cell Scale Charge balance and transport Electronic potential & Electrical network in current distribution composite electrodes Heat generation and Particle Scale Module Scale Li transport in electrolyte transfer phase Electrolyte wetting Li diffusion in solid phase Thermal/electrical Pressure distribution Interface physics inter-cell configuration Particle deformation & fatigue Thermal management Structural stability Safety control Atomic Scale System Scale Thermodynamic properties System operating Lattice stability conditions Material-level kinetic barrier Environmental conditions Transport properties Control strategy Challenge: How to perform life-predictive analysis for “what-if” scenarios untested in the laboratory (V2G, charging behavior, swapping, 2 nd use, …) 3
Factors in Vehicle Battery Aging Cell Design Environment Duty Cycle • Chemical • Thermal • System design • Electrochemical o geography o vehicle • Electrical o thermal management o excess power & system ($) energy @ BOL ($) • Manuf. uniformity o heat generation o system controls o defects • Humidity • Driver • Vibration o annual mileage o trips/day o aggressiveness o charging behavior – charges/day – fast charge 4
Factors in Vehicle Battery Aging Cell Design Environment Duty Cycle • Chemical • Thermal • System design • Electrochemical o geography o vehicle • Electrical o thermal management o excess power & system ($) energy @ BOL ($) • Manuf. uniformity o heat generation o system controls o defects • Humidity • Driver • Vibration o annual mileage o trips/day o aggressiveness o charging behavior – charges/day – fast charge (Not considered) 5
Simulation Approach Vehicle drive cycles • 782 speed vs. time traces • Charging assumptions Battery power profile • SOC(t), Heat gen(t), etc. Vehicle e.g., Cyc_4378_1 Model PHEV10 Opp. Chg 6
Simulation Approach Vehicle drive cycles • 782 speed vs. time traces • Charging assumptions Battery power profile • SOC(t), Heat gen(t), etc. • Vehicle Thermal management assumptions Model Battery stress statistics Battery • T(t), Voc(t), ∆DODi, Ni, … Thermal Model 7
Simulation Approach Vehicle drive cycles • 782 speed vs. time traces • Charging assumptions Battery power profile • SOC(t), Heat gen(t), etc. • Vehicle Thermal management assumptions Model Battery stress statistics Battery • T(t), Voc(t), ∆DODi, Ni, … Thermal Model Life Battery Life Model NCA/Graphite Capacity Minneapolis Houston 10 C Phoenix 15 C 20 C 30 C 25 C Years 8
Life Model Approach Battery aging datasets fit with empirical, yet physically justifiable formulas Calendar fade Cycling fade • SEI growth (partially • active material structure suppressed by cycling) degradation and • Loss of cyclable lithium mechanical fracture • a 1 , d 1 = f(∆ DOD,T,V oc ) • a 2 , e 1 = f(∆ DOD,T,V oc ) Relative R = a 1 t 1/2 + a 2 N Resistance Relative Q = min ( Q Li , Q active ) Capacity Q Li = d 0 + d 1 t 1/2 Q active = e 0 + e 1 N Enables life predictions for untested real-world scenarios 9
Acceleration Factors • Arrhenius Eqn. • Describe a 1 , a 2 , b 1 , c 1 as f( T , V oc , Δ DoD ) E 1 1 T exp a R T ( t ) T ref • Combined effects assumed multiplicative • Tafel Eqn. V oc ( t ) V ref F V exp R T ( t ) T ref • Wöhler Eqn. DoD DoD DoD ref 10
Acceleration Factors Resistance growth during storage Data: Broussely, 2007 • Arrhenius Eqn. E 1 1 T exp a R T ( t ) T ref • Tafel Eqn. V oc ( t ) V ref F V exp R T ( t ) T ref • Wöhler Eqn. DoD DoD DoD ref 11
Acceleration Factors Resistance growth during cycling Data: Hall, 2006 • Arrhenius Eqn. E 1 1 T exp a R T ( t ) T ref • Tafel Eqn. V oc ( t ) V ref F V exp R T ( t ) T ref • Wöhler Eqn. DoD DoD DoD ref 12
Acceleration Factors Capacity fade during cycling Data: Hall, 2006 • Arrhenius Eqn. E 1 1 T exp a R T ( t ) T ref • Tafel Eqn. V oc ( t ) V ref F V exp R T ( t ) T ref • Wöhler Eqn. DoD DoD DoD ref 13
Vehicle & Battery Assumptions PHEV10 PHEV40 All-electric range, km 16.7 67 Total vehicle mass, kg 1714 1830 Vehicle PHEV10: Electric motor power, kW 40 43 50% ∆ DOD at BOL IC engine power, kW 77 80 80% SOC max Useable power, kW 44 48 Useable energy, kWh 2.67 11.48 Battery Maximum SOC 80% 90% PHEV40: Electrical 1 Minimum SOC at BOL 30% 30% 60% ∆ DOD at BOL Minimum SOC at EOL 13% 10% 90% SOC max Excess energy at BOL 100% 67% Excess power at BOL, 10% SOC 43% 43% Heat transfer area - cells-to-coolant, m 2 1 3 Battery Heat transfer area - pack-to-ambient, m 2 Thermal 2, 3 1.2 2.9 Heat transfer coeff. - pack-to-ambient, W/m 2 K 2 2 1. EOL condition = 75% of BOL nameplate 1C capacity remaining 2. Heat generation rate at 2/3 of EOL resistance growth 14
Life Variability with Real ‐ World Drive Cycles • Matrix of analytic scenarios Drive Cycles 1 Vehicles • 782 Real-World drive • PHEV10 sedan cycles from Texas Dept. • PHEV40 sedan of Transportation Thermal Management 2 Charging Profiles 3 • Fixed 28 o C battery temperature* • Nightly charge (baseline) • Limited cooling (forced ambient air) • Opportunity charge • Aggressive cooling (20 o C chilled liquid) 1. Average daily driving distance of Texas dataset is 37.97 miles/day. This paper assumes 335 driving days and 30 rest days per year, scaling the Texas dataset to US- equivalent average mileage of 12,375 miles/year. 5 th and 95 th percentile daily driving distances from the Texas dataset are 99.13 and 4.87 miles/day, respectively. ** Level I charging rates. 2. A constant ambient temperature of 28 o C was assumed for all thermal simulations, representative of typical worst-case hot climate in Phoenix, AZ. Under battery storage * Worst ‐ case hot climate, Phoenix Arizona ~28 o C conditions, this effective ambient temperature causes similar battery degradation as would daily and annual temperature variations for a full year in Phoenix. 3. Charging at Level I rate of 1.5 kW. 15
Results • Variability in PHEV battery life with real ‐ world drive cycles • Impact of thermal management • Impact of opportunity versus nightly charging 16
Expected Life – PHEV10 Nightly Charge • Different daily driving distances and battery charge/discharge Phoenix Climate Constant 28 o C histories result in a distribution of expected battery life outcomes • Here, life expectancy across 782 driving cycles in a hot climate is 7.8 to 13.2 years • Key assumptions: o Graphite/NCA chemistry o End ‐ of ‐ life condition: 75% remaining capacity (of initial nameplate) o 80% SOC max o 30% SOC min @ BOL Opportunities for V2G, 2 nd use? 17
Expected Life – PHEV10 vs. PHEV40 Nightly Charge Phoenix Climate Constant 28 o C 86% of driving cycles > 10 mi/day 34% of driving cycles > 40 mi/day 18
Three battery thermal management scenarios illustrated for an example driving cycle Driving State e.g. Cyc_4378_1 PHEV10 Opportunity Charge 2) Limited Cooling Battery Temperature (forced 28 o C ambient air) 1) Isothermal (baseline case, battery fixed at 28 o C) 3) Aggressive (forced 20 o C chilled liquid) (time shown here is initialized to start of first driving trip of the day) 19
Recommend
More recommend
