Complex Modeling of Li-Ion Cells in Series and Batteries in Parallel within Satellite EPS Time Dependent Simulations Patrick Bailey, ENNEAD, LLC Aerospace Space Power Workshop April 16-19, 2012 Manhattan Beach, CA 1
Dr. Patrick G. Bailey ENNEAD, LLC P.O. Box 201 Los Altos, CA 94023-0201 ennead@sbcglobal.net www.padrak.com/ennead/ 14 Years Nuclear Reactor Safety (USAF, LANL, EPRI) 25 Years Lockheed (Martin), LMSSC Retired, Available, Enthusiastic BS, UC Berkeley PhD, MIT 2
Abstract A presentation is made of the advanced models and various results that have been obtained to simulate complex Lithium Ion (LiIon) battery behavior within any satellite Electric Power Systems (EPS). The battery cell behavior is modeled by the publically available Quallion Lithium Ion battery cell model, whose voltage behavior is defined to be a complicated function of the cell current, the cell temperature, and the cell state-of-charge. A battery is defined to be composed of a series of cells, whose individual properties in each cell at any given time may be different. The overall EPS battery is then composed of a number of such batteries connected in parallel. The simulation model allows the properties of each individual cell in the overall EPS battery to be individually different (for example, each cell at a different state-of-charge and at a different temperature), and each cell can be degraded or dropped out of the battery at any given time. Simulation results are shown for the cases of both charging and discharging, to illustrate the effects of varying the temperature between cells, and the effects of changing the state-of-charge between cells. Studies are also included that show the effects on the overall EPS battery voltage when the individual battery cells are not fully charged (to 100% state-of-charge) during recharging or during rebalancing. In addition, results are shown as cells are degraded and dropped from operation. These models and results are very important for complex EPS simulations and predictions. The inability of cells stacked in series to fully rebalance during recharging can lead to battery voltages lower than planned or designed for, and can result in EPS performance that is much less than expected or desired. Such battery models are planned to be included in the Power Tools Suite (PTS) system of codes and tools used at Lockheed Martin Space Systems Company. These models are already being used in the Satellite EPS Transient Code (Sat-Tran) that has been independently developed by for satellite EPS transient simulation, operation, validation, and prediction. 3
The Problem Need for accurate computer simulations of EPS Time-Dependent Behavior over Mission Life Many Solar Array Cell Types, Char.s, Models Many Battery Cell Types, Char.s, Models Many EPS Designs (architectures, batt. domin., etc.) Need capabilities for each, including: Sizing / Proposals Design / with and without Margins On Orbit Verification and Planning 4
The Solution Lumped Parameter Models, No Fast C/L/Z Transients Accurate for 10 second time steps and above Many Simple Spreadsheets (no generalizations) A Few General Simulation Packages: MATLAB & Simulink [Simplistic, Proprietary] Lockheed Martin Power Tools Suite (PTS) [Proprietary] ENNEAD Time Dependent Simulations (TDS) [Available] Models and Results for Verification and Validation Detailed Documentation 5
EPS Simulation - PTS Time-Dependent Simulation Code for EPS Detailed Dynamic Simulations For Proposals, Sizing, Design Detailed Battery, Solar Array, Non-Linear Models Lumped Parameter Models (1 minute time steps) Excel VB Macros (Same as C++ or FORTRAN) 80,000+ Lines of Code User Friendly Interfaces Easily Expandable Many IECEC Papers and SPW Presentations: www.padrak.com/pts_pgb/ [Publically released] 6
EPS Simulation PTS (IECEC 2004) 7
EPS Simulation PTS (IECEC 2004) 8
EPS Simulation - PTS (IECEC 2004) Battery Model Examples: -0.6 38.000 36.000-0.4 34.000 32.000-0.2 1.000 1.100 1.200 1.300 1.400 1.500 1.600 1.700 30.000 28.000 26.000 24.000 0 0.2 Battery Voltage 22.000 20.000 0.4-0.5-0.4-0.3-0.2-0.1 0 Discharge Rate (C Fraction) 0.1 0.2 0.3 0.6 10 0.2 0.4 1 0.8 0.6 Depth of Discharge 0.6 0.8 1 1.2 Solar Array Model Examples: * Other EPS components have similar models! 9
EPS Simulation - PTS (IECEC 2011) 10
EPS Simulation Using Excel No s/w application license Ease of use Ease of expansion, debugging, verification Use of VB Function Macros (e.g. Get V from many models) Stacked functions for various EPS components/types Function V_Batt_G(soc, curr, temp, age) as Double Code V_Batt_G = The Result End Function 11
EPS Simulation - TDS Time-Dependent Simulation Code for EPS Detailed Dynamic Simulations Developed Independently within Detailed Battery, Solar Array, Non-Linear Models Lumped Parameter Models (1 minute time steps) Excel VB Macros No Computer App. License Needed Complex Solar Array and Battery Designs User Friendly Interfaces Easily Expandable Available Now for Various Applications 12
Solar Array Simulation Simple Wing: One Cell. N cells in series. M strings in parallel. Many such wings. General Wing: Same cells per string. Many strings in parallel. Variable number of cells per string, and strings. Many such wings. Allows multiple spectral cells for higher SA. 13
Battery Simulation Simple Battery: One Cell. N cells in series. M stacks in parallel. Many such batteries. General Battery: Any cells per stack. Many stacks in parallel. Variable number of cells per stack, and stacks. Many such batteries. Allows detailed SOC calculations. Allows detailed chg. rebalancing simulation Different: Types, SOCs, Currents, Temperatures, etc. 14
Detailed Battery Simulation Different Cell Types, Different SOCs, Different Temps Variable # Cells per Stack, and Variable # Stacks 15
Detailed Battery Sim. Looks Easy Not! Example: Given Battery Charging Current Calculate Current Split into Stacks Calculate Cell Properties Calculate Cell New SOCs Calculate Cell New Impedances Compare Cell Old to New Impedances Iterate if Necessary Converge to Current Splits and Cell New SOCs 16
Sample Detailed Battery Simulations Sci-Fi LiIon Battery Cell V vs. SOC, 5 Ah Capacity, 2.5 V Full, with V/ I and V/ T Corrections (Enlarged to show effects in the results) 30 minute simulation time (1 min. T steps) Simulation Cases: (Max SOC = 110%) 1 Constant Current Charge 2 Constant Current Discharge 3 Constant Current Cycles 4 - Sample Orbit Current Cycles Can Cause Cell Dropouts Anytime 17
Chosen Battery Cell Model V Cell @ I=0 Sci-Fi User Chosen Cell Model (Normalized) 18
Chosen Battery Cell Model ( V/ I, V/ T) 19
Chosen Battery Model Same Cell Types, Different SOCs, Same Temps 20
Chosen Battery Cell Initial SOCs User Chosen Stack, Cell, and SOCs 21
Given Time Dependent Input Data 22
Case 1 Constant Charge 23
Case 1 Constant Charge 24
Case 1 Constant Charge 25
Case 1 Constant Charge 26
Case 2 Constant Discharge 27
Case 2 Constant Discharge 28
Case 2 Constant Discharge 29
Case 2 Constant Discharge 30
Case 3 Const. Current Cycles 31
Case 3 Const. Current Cycles 32
Case 3 Const. Current Cycles 33
Case 3 Const. Current Cycles 34
Case 4 Sample Orbit Current Cycles 35
Case 4 Sample Orbit Current Cycles 36
Case 4 Sample Orbit Current Cycles 37
Case 4 Sample Orbit Current Cycles 38
Conclusions Excel allows easy, fast, large, accurate simulations. EPS Time Dependent Simulation is needed for sizing, proposals, design mods, and on-orbit validations. PTS (LM) and TDS are available for general EPS use. TDS includes detailed (cell) SA and Battery modeling. Battery cells need to be modeled for individual SOC and other parameter effects. Battery cell not full recharging and cell drop-outs are very important in EPS design and use. Time Dependent Simulations can predict EPS behavior. 39
Q/A Dr. Patrick G. Bailey ennead@sbcglobal.net www.padrak.com/ennead/ 40