Improved SOH Estimation through Coulomb Counting

Transcription

1 Position Statement for BMS workshop at the 2011 PHM Society Conference Montreal, Quebec Improved SOH Estimation through Coulomb Counting Nick Williard, Wei He, Michael Osterman, Michael Pecht Objective: To demonstrate a novel method of estimating state of health within the framework of the coulomb counting technique College Park, MD

2 The (CALCE) formally started in 1984, as a NSF Center of Excellence in systems reliability. One of the world s most advanced and comprehensive testing and failure analysis laboratories Funded at.$6m by over 150 of the world s leading companies Supported by over 120 faculty, visiting scientists and research assistants Received NSF innovation award in 2009 CALCE Overview 2

3 Capacity (Ah) Decline in Capacity vs. Cycle Number Cycle Number When using capacity as the metric for degradation, we could define SOH as : 100% State Of Health Where the SOH at cycle c is equal to the capacity Q at cycle c over the capacity at the beginning of life, expressed as a percentage State of Health (SOH) refers to the general decline in battery performance with usage or aging. SOH can be characterized by capacity fade, power fade, or increase in internal resistance. Monitoring SOH is critical for performing condition based maintenance and for mitigating failure. In an electric vehicle, a SOH indicator would be analogous to a check engine light, which informs the user that maintenance or battery replacement is required when some degradation threshold is crossed. 3

4 Electrochemical Impedance Spectroscopy (EIS) Parameters from EIS are fit to an equivalent circuit model. Estimations can be improved though filtering methods such as particle filter and Kalman filter. Determining SOH Internal DC Resistance R V ( I ( 2 2 V I 1 ) 1 ) Internal DC resistance can be measured by applying 2 current pulses at I 1 and I 2 and then measuring the voltage change at each pulse. This feature has been shown to have a linear relationship to capacity Coulomb Counting Coulomb counting refers to the continuous monitoring of current that inters and leaves the battery. By integrating current with time, the capacity can be calculated. Whenever the battery is fully discharged to it s cutoff voltage the maximum capacity can be calculated and compared with the maximum capacity at the beginning of life then: 100% 4

5 Effect of Partial Discharges on Coulomb Counting Capacity Measurements If a particular discharge is cut-off midway though it s operation the, observed capacity will be lower due to the decreased about of time spent in operation. This decrease in capacity is not due too degradation mechanisms but rather the time spent in operation Voltage (V) Discharge was cut off at 3.4V 4 Capacity = 1.06Ah Time (hours) Capacity (Ah) Idt Current (A) Voltage (V) Current Voltage Discharge was cut off at 3.9V Capacity = 0.24Ah Time (hours) Current (A) 5

6 Shallow Charging Example The plot below shows the measured capacity of a battery that under went shallow charging for 900 cycles. The charge profile was then changed to cycle from a completely charged state to a completely discharged state. It can be seen that little degradation occurred during the shallow charging cycles. 1 Cycled between 2.7 and 4.2V Capacity (Ah) Cycled between 2.7 and 3.7V Cycles 6

7 Nature of Capacity Fade Assumptions Discharging under a constant condition will produce a smooth continuous trend in capacity fade Different conditions lead to different rates in capacity fade When all other conditions are constant, changes in capacity can be attributed solely to degradation dq Rate of degradation = when all discharge conditions are held constant dc Observed capacity measurements Equivalent Capacity Capacity Idealized trends in capacity fade Increasing cut-off Voltage Cycles 7

8 For Discharge Profiles That Do Not Reach V discharge When a capacity is recorded that does not reach the end cut-off voltage it can not be used to infer SOH. In this case it must be compared to a previous capacity value that was calculated at a similar cut-off voltage. When comparing capacities from similar cut-off voltages, the effective loss of capacity can be attributed to declining SOH rather than depth of discharge. When two discharges are recorded that have similar cut-off voltages the equivalent capacity can be calculated by: Where: c is the cycle number, Q similar are the capacity values that were recorded at similar cut-off voltages, and c similar are the cycle numbers associated with the similar capacity values. 8

9 Equivalent Capacity (Case Study) A battery underwent charge/discharge cycling at a constant discharge current Every cycle the battery was charged to it s fully charged state Every 15 cycles the discharge cut-off voltage was randomly changed Cut-off Voltage The change in cut-off voltage resulted in variations in the observed capacity for each cycle This test simulates the situation where a user decides to re-charge their battery before it has reached a completely discharged state 9

10 Results % Error Of Q equivalent to Q max # of cycles between recalibration The red points show the observed capacity for each cycle The blue points show the equivalent capacity for each cycle (estimation of Q max ) The green points are cycles that were fully discharged indicating the true value of Q max The percent error between the equivalent capacity and the next observed Q max is shown with respect to the number of cycles between recalibration. The model was able to run for over 600 cycles with less than a 10% error of the maximum capacity 10

11 Conclusions By calculating an equivalent capacity, the maximum capacity can be estimated for partial discharges allowing for SOH to be updated for every cycle by: 100% Frequent SOH estimations using equivalent capacity can improve maintenance strategies and mitigate failure. 11

12 Alcatel-Lucent Aero Contol Systes Agilent Technologies American Competitiveness Inst. Amkor Arbitron Arcelik ASC Capacitors ASE Astronautics Atlantic Inertial Systems AVIC AVI-Inc Axsys Engineering BAE Systems Benchmark Electronics Boeing Branson Ultrasonics Brooks Instruments Buehler Capricorn Pharma Cascade Engineering CAPE China Celestical International Channel One International Cisco Systems, Inc. Crane Aerospace & Electronics Curtiss-Wright Corp CDI De Brauw Blackstone Westbroek Dell Computer Corp. DMEA Dow Solar DRS EW Network Systems, Inc. EIT, Inc. Embedded Computing & Power EMCORE Corporation EADS IW France EMC A special thanks to our research sponsors! Emerson Advanced Design Ctr Emerson Appliance Controls Emerson Appliance Solutions Emerson Network Power Emerson Process Management Engent, Inc. Ericsson AB Essex Corporation Ethicon Endo-Surgery, Inc. Exponent, Inc. Fairchild Controls Corp. Filtronic Comtek GE, GE Healthcare General Dynamics, AIS & Land Sys. General Motors Guideline Hamlin Electronics Europe Hamilton Sundstrand Harris Corp Henkel Technologies Honda Honeywell Howrey, LLP IBM Intel Instituto Nokia de Technologia Juniper Networks Johnson and Johnson Johns Hopkins University Kimball Electronics L-3 Communication Systems LaBarge, Inc Lansmont Corporation Laird Technologies LG, Korea Liebert Power and Cooling Lockheed Martin Aerospace Lutron Electronics Maxion Technologies, Inc. Microsoft MIT Lincoln Laboratory Motorola Mobile Digital Systems, Inc. NASA National Oilwell Varco NetApp ncode International Nokia Siemens Nortel Networks NOK AG Northrop Grumman NTSB NXP Semiconductors Ortho-Clinical Diagnostics Park Advanced Product Dev. Penn State University PEO Integrated Warfare Petra Solar Philips Philips Lighting Pole Zero Corporation Pressure Biosciences Oracle Qualmark Quanterion Solutions Inc Quinby & Rundle Law Raytheon Company Rendell Sales Company Research in Motion Resin Designs LLC RNT, Inc. Roadtrack Rolls Royce Rockwell Automation Rockwell Collins Saab Avitronics Samsung Mechtronics Samsung Memory S.C. Johnson Wax Sandia National Labs SanDisk Schlumberger Schweitzer Engineering Labs Selex-SAS Sensors for Medicine and Science SiliconExpert Silicon Power Space Systems Loral SolarEdge Technologies Starkey Laboratories, Inc Symbol Technologies, Inc SymCom Team Corp Tech Film Tekelec Teradyne Textron Systems The Bergquist Company The M&T Company The University of Michigan Tin Technology Inc. TÜBİTAK Space Technologies U.K. Ministry of Defence U.S. Air Force Research Lab U.S. AMSAA U.S. ARL U.S. NSWC, NAVAIR U.S. Army Picatinney/UTRS U.S. Army RDECOM/ARDEC Vectron International, LLC Vestas Wind System AS Virginia Tech Weil, Gotshal & Manges LLP WesternGeco AS Whirlpool Corporation WiSpry, Inc. Woodward Governor 12