Battery Management for Maximum Performance in Plug-In Electric and Hybrid Vehicles

1
Battery Management for Maximum Performance in
Plug-In Electric and Hybrid Vehicles

• P. T. Krein
• Dept. of Electrical and Computer Engineering
• University of Illinois at Urbana-Champaign

2
Acknowledgements

• Thanks to Ryan Kroeze for literature work and
  analysis contributions.
• A version of this presentation was delivered at
  the IEEE Vehicle Propulsion Power symposium in
  September.

3
Outline

• Performance requirements
• Present situation
• Lead-acid cells
• NiMH cells
• Li-ion cells
• Battery management components
• Conclusion

4
Performance Requirements

• Hybrid vehicles
• High power density, meaning
• High charge acceptance for braking
• High power delivery for acceleration
• Cycle life tens of thousands of shallow cycles
• Adequate energy density, but this is secondary
• Wide ambient temperature range
• Electric vehicles
• High energy density
• Fast, reliable charging
• Cycle life thousands of deep cycles

1 cycle/5 miover 100,000 miles
5
Plug-In Hybrids

• Require the power capabilities and cycling
  capabilities of hybrids.
• Benefit from high energy density and good
  recharge properties.
• In other words must satisfy everyone and
  everything.
• This motivates work on hybrid storage that
  combines batteries (high energy density) with
  ultracapacitors (high power density).
• Here we explore the batteries.

6
Present Situation

• EVs and HEVs require thousands of battery cycles
  with minimal degradation.
• Typical strategy derates batteriesuse a narrow
  state of charge (SOC)regime.
• This results in a low effective energydensity
  in exchange for power density.
• Space applications get much more.
• The presentation emphasizes ways to maximize
  battery capabilities

UoSat-5 University of Surrey
7
Present Situation

• NiMH cells today are being usedin about a 15
  SOC range. Reasons are explored here.
• Lead-acid cells provide a similarrange.
• Li-ion cells are more promising.
• Active balancing that worksthroughout the SOC
  range isan important enabler.

8
Lead-Acid Cells

• Operating results from starting-lighting-ignition
  (SLA) batteries.
• Consistent with float operation in telecom.
• Best life results above 85 SOC.
• But the top end involves gassing reactions and
  sacrifices efficiency.
• Energy density is about 35 W-h/kg given 100
  discharge cycles.
• Effective energy density (15) is5.3 W-h/kg.
• Ultracapacitors can do as well.

9
Lead-Acid Cells

• Cells show damage from sulfation when operated at
  lower SOC.
• Present designs should be able to support an SOC
  range of 50 to 100, but only if the batteries
  are stored full.
• Promising future designs are likely to correct
  the most severe damagemechanisms.
• Do not favor HEV and EVapplications except on
  ause, park, charge cycle.

10
NiMH Cells

• Extensive data in preparation for and from
  experience with commercial hybrids.
• Toyota has had fewproblems with Priustraction
  batteries routine replacementhas not been
  required.
• Limited SOC swing about 50 to 65.

11
NiMH Cells

• Given density of 70 W-h/kg for full discharge,
  the effective density is less than 10 W-h/kg.
• The argument can be made that these designs use
  nickel-metal-hydride batteries for the functions
  of ultracapacitors.
• What aspects is this application attempting to
  optimize?

12
NiMH Cells

• At the high end, positive electrode degradation
  and electrolyte loss occurs.
• Positive pressure can transfer hydrogen among
  adjacent cells but amplifies degradation and
  imbalances cells.
• At the low end, the negative electrode
  experiences irreversible oxidation.
• Impedance rises for discharge.

13
NiMH Cells

• High-end effects are minimized if SOC is limited
  well below 80.
• Low-end effects are strong below 20 SOC, but
  performance degrades to some degree below 40
  SOC.
• External active balancing helps maintain
  discharge performance between 20 and 40 SOC,
  and limits degradation above 80.

14
NiMH Cells

• Differential power density is the remaining
  issue. (Here DOD 100 - SOC.)

From Menjak, Gow, Corrigan, Venkatesan, Dhar,
Stempel, Ovshinsky, Advanced Ovonic high-power
nickel-metal hydride batteries for
hybridelectric vehicle applications, in Ann.
Battery Conf. Appl. Advances, 1998, pp. 13-18.
15
NiMH Cells

• The reduction in charge power density as the high
  end has been treated as a limiting factor
  regeneration energy acceptance drops rapidly
  above 60 SOC.
• The SOC range from 20 to 80 can be utilized if
• Active balancing over the whole range prevents
  local limitations from pulling cells out of
  balance between 20 and 40 SOC, and between 60
  and 80 SOC.
• Braking strategy limits charge power at the high
  end.

16
NiMH Cells

• Thus the SOC range from 20 to 80 can be used
  for plug-in operation.
• Increases effective energy density to 42 W-h/kg
  factor of 4 improvement.

Harding Handbook for Quest Batteries, Fig.
3.7.2,available http//www.hardingenergy.com/pdfs
/NiMH.pdf
17
Li-Ion Cells

• Lithium-ion cells in general have much better
  reversibility than other common secondary
  chemistries Energy reversibility can exceed
  90.
• Discharge curves indicate regimes of reduced
  reversibility.

18
Li-Ion Cells

• Experience with laptop computers is showing that
  Li-ion cells degrade under float conditions
  extended operation when held at 100 SOC
  decreases operating life.
• Life testing in telecom applications shows that
  limiting the upper end charge voltage reduces
  degradation substantially.
• The effect is similar to limiting SOC to less
  than 90.

19
Li-Ion Cells

• The curve shown earlier shows rapid imbalance and
  capacity reduction below 20 SOC.
• Key problem cellbalancing no inherent
  mechanism in Li-ion.
• Typical systems useresistive limiters toenforce
  the upper voltage limit.
• Limiters add system nonlinearity that drives
  (lossy) cell balancing at the top end of SOC,

www.popularmechanics.com
20
Li-Ion Cells

• Balancing is more important at the low end, where
  discharge effects begin to pull cells apart.
• In reality, a method is needed that can balance
  over the entire useful SOC range.
• When this is done, the possible range of SOC
  becomes 20 to 90.
• If the cells achieve 200 W-h/kg for 100
  discharge, the effective energy density is 140
  W-h/kg more than triple the best NiMH results.

21
Battery Management Components

• Vehicle system-level control strategy must focus
  on a limited SOC range, as present hybrids do.
• The proven long-life SOC range is considerably
  wider than in present practice.
• Components
• Strategies with active top-end and bottom-end SOC
  limits.
• Active cell balancing over the full range.
• Techniques to limit or mitigate power density
  requirements at extremes.

22
Choices for Limits

• Use established charge sustaining strategies, but
  open the tolerance bands.
• NiMH 50 ? 30 SOC range
• Li-ion 55 ? 35
• Target a daily driving and charging profile.
• Seek to end the day at the low end, ready for
  charging.
• Allow a high SOC pack to decrease slowly during
  the daily drives.
• Adaptive cycle intelligence.

23
Choices for Mitigation

• Divert power demand extremes to ultracapacitors
  but only at the extreme SOC ends.
• This leads to relatively small ultracapacitor
  packs that absorb as little as 10 of a given
  braking energy sequence or deliver just 20 of
  peak acceleration power
• Use resistive brake auxiliarieswhen SOC upper
  limit is reached.

24
Active Cell Balancing

• In Li-ion packs, cell mismatch is not restored by
  altering the charge process alone.
• The cells can be pulled apart at the low end of
  SOC, especially for high power pulses.
• Resistive or switched voltage limiters can only
  function at the high end.
• In HEV applications, there is limited dwell time
  at the high end.
• In EV applications, limiters must follow the SOC
  limit settings.

25
Active Cell Balancing

• Active balancing methods bring cells together
  regardless of SOC.
• Switched capacitor types low energy use,
  efficiency is high as mismatch reduces.
• Switched inductor types drives current to match
  charge in a controller manner.
• Individual cell or monoblock chargers the
  ultimate, but expensive, solution.

26
Discussion

• Present lead-acid cells are comparatively weak
  for plug-in hybrid applications.
• NiMH cells can be used for swings between 20 and
  80 SOC, achieving effective energy densities of
  40-50 W-h/kg in plug-in applications. Based on
  known results from commercial hybrids, this
  should be viable.
• Li-ion cells can be used for swings between 20
  and 90 SOC, achieving effective energy densities
  of 140 W-h/kg or more.

27
Discussion

• All can have efficiency enhanced with
  ultracapacitors as auxiliaries.
• The application in the stated range is predicated
  on active battery management, especially active
  balancing.
• There are commercial Li-ion batteries that have
  been matching the claimed performance specs and
  should be able to perform to the requirements.

28
Discussion

• Is it enough?
• In city driving, a well-designed car needs no
  more than 80 W-h/km (125 W-h/mile).
• At 140 W-h/kg, 100 kg of Li-ion batteries could
  deliver 175 km of all-electric city range.

29
Conclusion

• There is growing knowledge of considerations for
  maximum battery performance in the context of
  plug-in hybrids.
• Li-ion cells should be able to deliver more than
  ten-fold effective energy density improvement
  compared to present hybrid strategies.
• For all cell types, limiting the SOC range is
  vital for longevity.
• Cell balancing to permit arbitrary SOC levels
  also appears to be vital.

30
Questions and Discussion