NIH PROJECT

Design and Implementation of a State-of-charge meter for Lithium ion batteries to be used in Portable Defibrillators Ramana K.Vinjamuri 08/25/2004 Under direction of Dr. Pritpal Singh

Outline   

   

BACKGROUND PROCEDURE (experimental setup) MEASUREMENTS AND ANALYSIS FUZZY LOGIC MODELING IMPLEMENTATION IN MC68HC12 (micro controller) CONCLUSIONS FUTURE SCOPE

BACKGROUND

Portable defibrillators Today portable defibrillators are considered as sophisticated devices by FDA (Food and Drug Administration). As a trend towards the widespread deployment of portable defibrillators in the hands of non-medical or non-technical personnel increases, there exists a need for a simple procedure to ensure that it will operate properly when needed.

Portable defibrillators According to the FDA the major cause of defibrillator failure was improper care of the rechargeable battery . The effective operation of a portable defibrillator depends critically on the condition of the battery which are defined by State-ofCharge and State-of-Health.

Chemistry of Li ion batteries Reactions that occur at Electrodes Positive LiMO2 → Li 1-xMO2 + x Li + + xe Negative C + x Li + +xe → Li x C Overall LiMO2 + C → Li x C + Li 1-x MO2

Features of Li-ion batteries    

Higher Energy density Higher voltage Long operating time Compact

Definitions SOC denotes the remaining pulses in a battery pack in one discharge cycle SOH represents the remaining number of cycles (charge-discharge) that can be obtained from a battery pack in its entire life. When the battery pack is new it is said to have 100% SOH. As the battery ages SOH eventually decreases.

Battery Interrogation Techniques Efficient battery interrogation techniques are required for determining the state-ofcharge (SOC) of a battery. The three basic methods are: 1) Coulomb counting 2) Voltage delay and 3) Impedance method

TYPICAL NYQUIST PLOT OF ELECTRO CHEMICAL CELL Z’

Capacitive behavior

Diffusion

Anode Cathode

Rs

10 mHz

1kHz

Inductive behavior

100Hz 0 inductive tail

Z”

Equivalent Circuit for this Cell Ranode

Rcathode

RS

L

Canode

Ccathode

Using AC impedance for determination of SOC

Research by J. P.Fellner At Air force laboratory, OH [1]

Using AC impedance for determination of SOC

Research by J. P.Fellner At Air force laboratory, OH [2]

Using AC impedance for determination of SOC

Research by Dr. Pritpal Singh [3]

Using AC impedance for determination of SOC 200

60 60 400

Research by J. P.Fellner At Air force laboratory, OH [2]

Introduction to Fuzzy Logic In fuzzy logic, a quantity may be a member of a set to some degree or not be a member of a set to some degree. The boundaries of the set are fuzzy rather than crisp. A fuzzy system is a rule-based mapping of inputs to outputs for a system.

Two approaches in Fuzzy Logic 



Mamdani Approach: Uses membership functions for both input and output variables Sugeno Approach: Output membership

functions are “singletons” (zero order) or polynomials (first order).

Example: Two input, two rule Fuzzy Model n1 m1

Rule1

F1

S1 m2 Rule2

n2

F2

S2

Sugeno type of inference

PROCEDURE

Li-ion battery pack 



This Li ion battery pack consists of 12 cells connected in series parallel (4s3p configuration) Effective voltage of the battery pack is 16.8 volts(4.2 volts per cell)

Charge profile The profile that we have adopted is A constant current charging of 2.5 A till the battery voltage is 16.6172 v A constant voltage charging of 16.6 v till the charge current drops below 100mA

Discharge profile The profile suggested by Medtronic/ Physio Control was Continuous discharge of 1.4 A and a discharge of 10 A for every 5 minutes for a period of 5 s

Discharge profile

Load current profile

Voltage recovery profile

Apparatus 







For discharge -- Electronic load 6063B from Agilent Technologies For the impedance and the voltage recovery measurements--Solartron 1280B,which is Potentiostat /Galvanostat /FRA For charge --Centronix BMS2000, The Battery Management System For different temperatures Tenney Environmental oven

Battery pack, EC Load and Solartron

EC-Load and Oven

Software 





To control the Electronic Load the software is HP VEE To view and plot the impedance data its Zview and Zplot respectively To view and plot the voltage recovery profiles data its Corr view and Corr ware

Software control

Test process 









Constant current discharge at 1.4A for 5 minutes, monitoring the voltage of the battery pack Constant current discharge at 10 A for 5 seconds, monitoring the voltage of the battery pack Repeat this process for a total of 1100 seconds which includes three 10 A discharges EIS (Electro chemical Impedance spectroscopy) measurement over frequency range of 1Hz-1KHz Repeat above four steps until end of discharge is reached (2.5V/cell)

Test process

MEASUREMENTS AND ANALYSIS

Impedance measurements

Nyquist plot

Impedance measurements

Bode plots

Monotonic variation of the voltage recovery profiles with SOC

Comparing the First and the Last pulse

Analysis

Minimum voltage curves  Difference voltage curves 

Minimum voltage curves 

The locus of the minimum voltages of every pulse in one cycle forms one curve corresponding to Cxx in the graph

One Pulse

Minimum voltage curves 



The locus of the minimum voltages of every pulse in one cycle forms one curve corresponding to Cxx in the graph The above means the set of all As in figure shown

For battery pack at room temperature

Difference voltage curves 

The locus of the difference between the maximum and minimum voltages of every pulse in a cycle forms a curve Cxx in the figure.

One pulse

Difference voltage curves  

Voltage Difference=B-A The locus of the difference between the maximum and minimum voltages of every pulse (B-A) in a cycle forms a curve Cxx in the figure.

For battery pack at room temperature

FUZZY LOGIC MODELING Two models 1. To predict SOC –Remaining pulses (implemented) 2. To predict SOH –Cycle number (theoretical model)

Fuzzy Logic Modeling       

Inputs: Maximum voltage and Minimum voltage Output: Pulses remaining Type of mem. functions: Trapezoidal Type of inference : Sugeno No. of rules : 12 4 mem. Functions for Max. voltage 3 mem. Functions for Min. voltage

Membership Functions for Input1

Membership Functions for input2

Training error (0.95425)

Testing error (0.99126)

Surface plot

Fuzzy Logic Modeling       

Inputs: Maximum voltage and Minimum voltage Output: Cycle Number Type of mem. functions: Trapezoidal Type of inference : Sugeno No. of rules : 12 2 mem. Functions for Max. voltage 6 mem. Functions for Min. voltage

Testing error (2.6554)

Training error (2.565)

Surface plot

IMPLEMENTATION IN MC68HC12 (micro controller)

Implementation in MC68HC12 (micro controller) Features of HC12:  On-Chip A/D conversion (any voltage between 0-5 volts;0-00H and 5-FFH )  Instruction Set with Fuzzy Logic instructions (ability to implement trapezoidal and triangular mem. functions)

Step down circuit Voltage of the battery pack is stepped down to be given as input to HC12

R=511 K Ohms Op Amp=LMC60 42 AIN

Flow chart of the main program

Timing Diagram

Experimental setup

Results Display showing 21 pulses remaining

LCD display

Average error=+/-2 pulses

Stem Plot

Summary 







Impedance and Voltage recovery profiles collected for battery packs at room temperature and 00C Battery characteristics were analyzed and Minimum voltage curves and Difference voltage curves were developed Based on the voltage recovery profiles a good Fuzzy Logic Model was obtained to predict the SOC of the battery pack at room temperature with a minimum error as low as 0.9 Implemented on Micro Controller HC12 with a very low error of +/-2 pulses

Future scope 



This model can be extended to estimate the SOC of the battery packs at different temperatures An SOH meter that can predict the cycle number can also be developed provided, sufficient data is collected for the battery packs at different temperatures

Publications 1. Pritpal Singh and Ramana Vinjamuri, Xiquan Wang and David Reisner “FUZZY LOGIC MODELING OF EIS MEASUREMENTS ON LITHIUM-ION BATTERIES”. EIS’04

2.

Pritpal Singh and Ramana Vinjamuri, Xiquan Wang and David Reisner.” Analysis on Voltage recovery profiles and

Impedance measurements of High Power Li ion batteries”. 41 st Power sources conference,2004

References 1.

2.

3.

J.P.Fellner and R.A. Marsh “Use of the pulse current and AC impedance characterization to enhance Lithium ion battery maintenance”, Electrochemical society proceedings volume 99-25 J.P.Fellner, G.J.Loeber, S.S.Sadhu “Testing of lithium ion 18650 cells and characterizing/predicting cell performance” Journal of Power sources conference 81-82(1999) P. Singh, Y.S. Damodar, C. Fennie, and D.E. Reisner, “Fuzzy Logic-Based Determination of Lead Acid Battery State-of-Charge by Impedance Interrogation Methods”Procs. EVS-17, Montreal, Canada, Oct 15-18, 2000