European Conference on Nanoelectronics and Embedded Systems for Electric Mobility ecocity emotion 24-25 th September 2014, Erlangen, Germany An Insight into Active Balancing for Lithium-Ion Batteries Federico Baronti, Roberto Roncella, Roberto Saletti University of Pisa, Italy federico.baronti@unipi.it
Presentation Outline Introduction Charge imbalance and usable capacity Overview of common balancing circuits Modeling and Comparison of main balancing topologies Example of a highly efficient cell-to-cell balancing circuit Summary
Introduction When does charge imbalance matter? More battery cells serially connected Large-format Li-ion batteries Battery cells cannot be overcharged What are the root causes of charge imbalance? Intrinsic: variations in the manufacturing process, which lead to static differences in cell capacities and self-discharge rates (SDR) Extrinsic: temperature gradients in the battery pack, which cause dynamic variations in cell self-discharge rates and ageing What is the effect of charge imbalance on the usable capacity of the battery?
Charge Imbalance and Usable Capacity Battery w/ 3 series-connected cells w/ equal capacities, but different state-of-charge 100 % Initial condition Full discharge Full charge Full discharge 0 % Cell 1 Cell 2 Cell 3 Cell 1 Cell 2 Cell 3 Cell 1 Cell 2 Cell 3 Cell 1 Cell 2 Cell 3 Discharge interrupted by Cell 3 Still energy in Cell 1 and Cell 2 Charge interrupted by Cell 2 Cell 1 and Cell 3 not fully recharged Usable capacity is reduced by the maximum SOC unbalance
Passive vs Active Balancing Balancing is usually performed at the end of charge Full charge Passive balancing Cell 1 Cell 2 Cell 3 Active balancing Cell 1 Cell 2 Cell 3 Cell 1 and Cell 2 discharged to the Cell 3 level Energy in excess is dissipated into heat Cell 1 Cell 2 Cell 3 Cell 3 is recharged by Cell 2 Energy in transferred between cells
Common balancing circuits Bleeding resistor Switching capacitor Multi-winding transformer Switching matrix DC/DC converter Simple X Zero efficiency X Speed vs heat Self balancing X Low efficiency X Very slow Self balancing Moderate efficiency X Bulky High efficiency Fast X Complex J. Gallardo-Lozano, E. Romero-Cadaval, M. I. Milanes-Montero, and M. A. Guerrero-Martinez, Battery equalization active methods, J. Power Sources, vol. 246, pp. 934 949, Jan. 2014.
Active Modeling A balancing circuit seen as (N+1)-port network, which includes one or more DC/DC converters whose input/output can be connected to any network s port The balancing topology determines the input/output of the converter Balancing topology DC/DC converter # Input Output Eff. Cell-to-Null N Any cell Heat 0 Pack-to-Cell Pack 1 Pack Any cell η port Cell-to-Pack 1 Any cell Pack η Cell-to/from-Pack 1 Any cell/ Pack Pack/ Any cell Cell-to-Cell 1 Any cell Another cell η η F. Baronti, R. Roncella, and R. Saletti, Performance comparison of active balancing techniques for lithium-ion batteries, J. Power Sources, vol. 267, pp. 603 609, Dec. 2014.
Modeling (cont.) Computation of the balancing time and energy losses to equalize the charge of all the cells Optimum balancing strategy for each topology, which minimizes losses Comparison methodology Balancing current kept the same for all the balancing topologies Statistical analysis to account for random distribution of the charge imbalance given a maximum SOC mismatch For each active topology, we compute: F time as the balancing time divided by the that achieved by passive balancing in the same charge imbalance condition F loss as the related energy lost by the battery divided by the that dissipated by passive balancing in the same charge imbalance condition
< F loss > Comparison Results Case-study: Module w/ 10 LiFePO 4 cells, cell voltage = 3.344 V Maximum SoC mismatch = 10 % 100000 random trials 2.0 1.5 Cell to Cell Cell to Pack Pack to Cell Cell to/from Pack 1.0 0.5 0.0 0.0 0.2 0.4 0.6 0.8 1.0 Cell-to-Cell outperforms all the other topologies h Pack-to-Cell can dissipate more energy on average than passive balancing if the converter is poorly designed (η<0.5, in this case study)
Example of Cell-to-Cell Implementation DE N-1 = - 1 2 C V h 2 2 ( -V l ) h Buck DE 0 = -h Buck h Boost DE N-1 Scap C V h V l 1 DE 0 =h Boost 2 C V h 2 2 ( -V l ) V cap F. Baronti et al. High-Efficiency Digitally-Controlled Charge Equalizer for Series-Connected Cells based on Switching Converter and Super-Capacitor in IEEE Trans. on Ind. Informat., 2013
Buck/Boost Converter Design Model-based design Component selected to maximize efficiency minimizing Eloss E loss = E static + E switching + E Joule Hysteretic control of the current flowing in and out the SCAP
Buck/Boost Converter Prototype 5 F, 5.5 V SCap Control logic implemented on external boards
Buck/Boost Converter Test Experimental set-up for measuring the converter efficiency Automatize the experiment execution Control Logic Emulates the source/sink cell measuring the energy flows
Buck/Boost Converter Efficiency Measured efficiency of a complete energy transfer η Buck η Boost V in and V out : voltage of the source and sink cell respectively Δv cap = V h - V l voltage swing over the SCAP V h = V out - 200 mv Cell-to-cell energy transfer efficiency up to 90 %!
Balancing algorithm Balancing algorithm SOC-OCV map for a NMC cell ΔOCV~10 mv ΔSOC ~ 1% N = 11
Example of Balancing Procedure INITIAL STATE AFTER BALANCING Charge Imbalance from 18 % to 2 % +16% usable capacity F. Baronti, et al., Experimental validation of an efficient charge equalization system for Lithium-ion batteries, in Proc. ISIE 2014
Summary How energy is transferred between the battery cells in active balancing has a significant impact on the achievable performance of the balancing circuit Cell-to-Cell is the best way to implement active balancing, as far as energy losses and balancing time are concerned Cell-to-Cell can effectively be implemented using a Buck/Boost converter and a supercapacitor to temporary store the energy during a cell-to-cell transfer
http://umji.sjtu.edu.cn/personal/chbma/ieeetces/ Thanks for your attention!