An Impedance-Based BMS to Identify Bad Cells Rengaswamy Srini Srinivasan Bliss G. Carkhuff

Transcription

1 An Impedance-Based BMS to Identify Bad Cells Rengaswamy Srini Srinivasan Bliss G. Carkhuff (443)

2 Impedance-Based T internal, R internal, SOC and SOH Note: This slide provides an overview of battery impedance and how we use it to measure anode temperature, cathode temperature, electrolyte resistance, state of charge, state of health, etc. For more details, see references below and in the next slide -Z'' / SOC 87% 80% 70% 60% 50% 40% 30% 20% 10% 1 khz 20 Hz (SoC-Insensitive) Temperature Sensitive Z' / Sensitive to SOC and Temperature 0.8 Hz Methodology: Srinivasan et al., Electrochimica Acta, Vol. 56: , Year 2011 Perturbing the cell with a smallamplitude ac current Results in an ac voltage Measure the amplitude and phase shift Compute the impedance Impedance at any one frequency between 40 Hz to 100 Hz measures anode temperature (T a ) Impedance at any one frequency between 10 Hz to 20 Hz measures cathode temperature (T c ) T a and T c can be estimated using phase shift values only High-frequency X-axis intercept corresponds to electrolyte resistance, R s Impedance at low-frequency (<10 Hz) is sensitive to temperature and state of charge 2

3 Impedance-Based T internal, R internal, SOC and SOH 3 R. Srinivasan et al., Instantaneous Measurement of the Internal Temperature in Lithium-Ion Rechargeable Cells, Electrochimica Acta, 2011, 56, R. Srinivasan, Monitoring Dynamic Thermal Behavior of the Carbon Anode in a Lithium-ion Cell Using a Four-probe Technique, Journal of Power Sources, 2012, 198, R. Srinivasan and B. G. Carkhuff, Empirical analysis of contributing factors to heating in lithium-ion cells: Anode entropy versus internal resistance, Journal of Power Sources, 2013, 241, R. Srinivasan et al., The Five Modes of Heat Generation in a Li-Ion Cell under Discharge, Journal of Power Sources, 2014, 262, R. Srinivasan and L. Srinivasan, Graphitic carbon anode temperature excursions reflect crystallographic phase transitions in lithium-ion cells, Journal of Power Sources, 2015, 293, R. Srinivasan et al., US Patent No. 7,544,294 B2, 30 June 2009, Battery Health Monitor. R. Srinivasan et al., US Patent No. 8,961,004 B2, 24 February 2015, Battery Phase Meter to Determine Internal Temperatures of Lithium-Ion Rechargeable Cells Under Charge and Discharge. (Also see: (WO 2012/ A1 published 26 April, 2012; Europe: 2,630,687; Japan: 5,840,693.) R. Srinivasan and B. G. Carkhuff, US Patent No. 9,331,507 B2, 3 May 2016, Control Apparatus and Method for Conducting Fast Battery Charge.

4 Overview BMS based on current, voltage and surface temperature monitors have struggled to provide safety Some have single-frequency (1 khz) impedance sensor that serves little or no purpose in providing cell or battery safety BMS based on multi-frequency impedance have a much better chance to ensure thermal safety and electrical efficiency 4

5 The 2009 Incident: BMS with Voltage, Current and Surface Temperature Monitors Electrical and Thermal Data Battery Voltage (V) and Current (A) Time (hour) Temperature ( 0 C) What purpose did the voltmeter, ammeter and thermocouple serve? 5

6 What does a voltmeter tell you? 6 Cell Voltage (V) % SOC V V V 30% SOC V V V ± V 10% SOC V Cell 40A Cell 40B Cell 40C V ± V V ± V Time (minute) 83% SOC V V V New 3-cell battery ± V Cell voltages in a new battery are nearly identical

7 What does a voltmeter tell you? % SOC V V V ± Cell Voltage (V) % SOC V V V ± <5% SOC V V V ± Time (minute) Cell 11 Cell 30A Cell 30 Cycle-aged 3-cell battery Cell voltages in a cycled battery are nearly identical 7

8 8 What does a voltmeter tell you? Cell Voltage (V) Cell Voltage (V) Cell 1 Cell 2 Cell 3 Bad Cell Cell 5 Cell 6 83% SOC V V V V V V ±4.73 mv Time (minute) Time (minute) Cell 1 Cell 2 Cell 3 Bad Cell Cell 5 Cell 6 Bad Cell: Twice over-discharged to 0.01 V 30% SOC V V V V V V ±3.68 mv Cell voltages in a battery with a bad cell are nearly identical

9 What does a voltmeter tell you? Nothing much, unless you drain the battery Cell Voltage (V) Cell Voltage (V) Cell 1 Cell 2 Cell 3 Bad Cell Cell 5 Cell 6 83% SOC V V V V V V ±4.73 mv 30% SOC V V V V V V ±3.68 mv Cell 1 Cell 2 Cell 3 Bad Cell Cell 5 Cell 6 10% SOC V V V V V V ±11.34 mv Time (minute) 2.0 Low Voltage Cutoff = 6x2.7 V Time (minute) Cell voltages in a new battery are nearly identical until the battery is fully discharged

10 What does Surface-Mounted Sensor Identifies? Surface Temperature ( 0 C) Bad Normal T = 1 ºC Battery Voltage (V) Time (minute) Cell Temperatures in a battery with a bad cell are nearly identical

11 Protecting Armors for Battery Safety? AMMETER VOLTMETER THERMOCOUPLE 11 Archaic instruments do not help protect Li-ion batteries

12 The 2016 Incident: Had Battery Voltage and PackTemperature Monitors in 3 seconds 12 Archaic instruments did not help protect this Li-ion battery

13 What does a thermocouple tell you? Temperature ( 0 C) Discharge Current Surface Temperature Current (A) Discharge Time (second) T Surface response-time is slow

14 What does a thermocouple tell you? 14 Temperature ( 0 C) Ah Cell is being overcharged Surface Temperature Current Capacity Charge Time (second) Voltage Current 2.5 Ah T surface is <32 ºC, indicating everything is fine! Voltage (V), Current (A) or Capacity (Ah)

15 What does a thermocouple tell you? 70 Rate of discharge and charge = 1C Temperature ( 0 C) Surface Temperature Voltage Battery Voltage (V) Discharge 20 Charge (CC part) Time (hour) Srini; 10/14/2015 T surface is <40 ºC therefore safe? 15

16 A Practical Impedance-Based BMS Non-Invasive and Autonomous Ageing due to cycle life and calendar life Cell screening and matching Before and after constructing a battery Identifying internal defects caused by over-discharge, over-charge, etc. Thermal safety 16

17 Sensitivity: Measurements based on bench-top impedance meter (Solartron) - Imaginary (milli-ohm) Cell #1 Cell # 2 Cell # 3 70 Hz (A) New Hz - Imaginary (milli-ohm) Cell # 4 Cell # 5 Cell # 6 Cell # 7 Cell # 8 Cell # 9 70 Hz (B) Aged Hz 1.6 khz Real (milli-ohm) 1.2 khz Real (milli-ohm) - Imaginary (milli-ohm) khz (C) One Overcharged Cell # 4 Cell # 5 Cell # 6 Cell # 7 Cell # 8 Cell # 9 Over-discharged (twice to 0.01 V) Over-charged (twice to 4.6 V) - Imaginary (milli-ohm) Cell # 4 Cell # 5 Cell # 6 Cell # 7 Cell # 8 Cell # 9 Over-discharged (twice to 0.01 V) 70 Hz 70 Hz (C) One Over-discharged Hz Hz Real (milli-ohm) 1.2 khz 1.2 khz Real (milli-ohm)

18 Miniaturized Impedance-Based BMS BMS 16-Cell Multiplexer BMS for 16 cells, 55-Ah battery Frequency range: 2 Hz to 1 khz Anode and cathode temperatures Anode, cathode and electrolyte resistance SOH Cell Voltage 18

19 Perspectives of Impedance-Based BMS Far simpler than using thermocouple for T surface 19

20 Identify Matched Cells within Seconds BMS Output (Arbitrary Unit) New Cell #3 = Cell #1 = Cell #2 = Imaginary (milli-ohm) Cell #1 Cell # 2 Cell # 3 70 Hz Hz Time (second) 1.6 khz Real (milli-ohm) The three cells in the battery are matched within ±0.5 % 20

21 BMS Identifies Mismatching in Aged Cells BMS Output (Arbitrary Unit) Cell #1 = Cell #6 = Cell #2 and #3 = Cell #4 = Imaginary (milli-ohm) Imepdace data shows the cells are not matched Cell # 4 Cell # 5 Cell # 6 Cell # 7 Cell # 8 Cell # 9 70 Hz Calendar Aged (6 months) Hz 2520 Cell #5 = Time (second) 1.2 khz Real (milli-ohm) In the 6-cell battery, calendar ageing drove the cell apart by ±2% 21

22 BMS Output of the Five Calendar-Aged Cells Five Calendar-Aged, but Otherwise Normal Cells BMS Output (Arbitrary Unit) Battery Voltage (V) Time (minutes) 22

23 One of the Six Cells was over-discharged: The BMS identifies the bad cell 2520 Normal Cell Bad Cell 26 BMS Output (Arbitrary Unit) Battery Voltage (V) BMS Output (Arbitrary Unit) Battery Voltage (V) Time (minute) Time (minute) 23

24 Why anode temperature is more important? Samsung SDI F26 (2.6-Ah Li-ion) behavior during overcharging 44 5 Temperature ( 0 C) Voltage Anode Temperature Current Capacity Gas release Surface Temperature Current Voltage (V), Current (A) or Capacity (Ah) Charge Time (second) 24 T surface misses overcharging completely, but not T anode Srinivasan et al., Journal of Power Sources, 241, pp ; Year: 2013

25 Continuous Discharge-Charge at 1C Rate T internal can be much hotter than T surface Temperature ( 0 C) T internal T surface Battery Voltage (V) Discharge 20 Charge (CC part) Time (hour) Srini; 10/14/2015 T internal >> T surface when the current is continuous 25

26 Srinivasan et al., Journal of Power Sources, 262, pp ; Year: 2014 Surface temperature can be misleading 26 Temperature ( 0 C) Surface Anode Cathode Discharge Time (second) T Surface response time is slower Current (A)

27 Fast Charge through Feedback-Controlled Closed-Loop 27 Charging current profile is semi-autonomously determined by the charger: based on the user-set limits on T internal and Cell Voltage Temperature ( 0 C) Current T internal T surface Charge Time (minute) Voltage Voltage (V) or Current (A) Example shown: Conventional charging time is 150 minutes Fast charger-enabled time is 95 minutes

28 Impedance-Based SOH Estimation Electrolyte Resistance (milli ohm) 28 (A) Nomalized Ah-Capacity (Ah) (B) Cycle Number Cycle Number 28

29