The International Space Station (ISS) Program depends on models to predict battery performance. It is critical that predictions are accurate because only simulations can determine if planned operations will maintain sufficient energy in the batteries. Extensive ground-based testing has provided a database of battery pack performance as a function of temperature, depth of discharge, and charge-control methodology. Battery voltage is then calculated from current, temperature, age, and depth of discharge. This model has been continuously improved, making use of additional battery data and modeling of cycle life versus depth of discharge. This technique has proven capable of predicting performance, but it does not allow for insight into electrochemical processes. For this reason, Glenn researchers, along with a University of Akron team, investigated a first-principles model: one based on the fundamental laws of physics that govern the behavior of electrochemical batteries.
A NASA grant was awarded to a University of Akron team, led by Dr. Tom T. Hartley, to develop a fractional calculus approach to dynamically model nickel-hydrogen battery cells on the basis of charge-discharge cycle data. The model used a real-time observer structure to determine instantaneous damage from life-cycle data and to establish an optimal charging profile to minimize damage.
Advanced control system using an advanced real-time observer.
A control system with a real-time observer for an electrochemical cell incorporates a hybrid model into a microprocessor-based simulation and is automated to track parameters as a battery changes during its lifetime. In addition, the algorithm can correct the initial parameters provided to the real-time observer, thereby effectively “learning” the battery system. Measured variables from an actual battery are input to the microprocessor simulation to force the dynamic states of the real-time battery observer (simulation) to converge to the corresponding states in the actual battery. The feedback is crucial since the actual system is highly nonlinear and the dynamic states represent a collection of many actual states. The observer can be used as part of the general decision-making process. If the observer predicts that certain dynamic states have exceeded the nominal values, the higher level of intelligence can determine if the actual system has become damaged in some way. The advanced controller uses the data provided by the real-time observer to maintain battery performance as the parameters change with time. This requires a continuous change of controller gains as the effects of physical changes in the battery are updated by the advanced real-time observer.
Faraday’s Law and variations of the Butler-Volmer equation enabled the modeling of stored charge, diffusing charge, and electrode voltage. The model was enhanced using on-orbit ISS reconditioning data because the battery started out fully discharged, which allowed the model to encompass behavior over the entire operating range. Last, battery current and state-of-charge were used as input to verify the model against normal ISS cycling telemetry (ref. 1). Results showed that this first-principles-based model predicted voltage with an accuracy similar to that of the data-based model in use today.
Model data at 0 °C; cell voltage versus time at beginning and end of life, for three depths-of-discharge: 20, 35, and 60 percent.
Long description of figure 2.
In addition, modeling different resistances in charge and discharge improved predictions. However, modeling the batteries at the individual cell level or at the half battery level did not improve voltage predictions markedly in comparison to modeling the ISS battery as a whole. Model accuracy is affected by insight into battery characteristics, such as ampere-hour capacity, which can only be measured by performing reconditioning. This dedicated test requires a battery to be temporarily taken offline and discharged.
Data-based and first-principles-based model predictions of battery voltage versus time for ISS cycling on day 90 of 2004.
Long description of figure 3.
Although the nickel-hydrogen battery has been considered to be the industry standard for low-Earth-orbit missions, lithium-ion battery technology with its improved energy density has been baselined on recent missions. Lithium-ion cell testing is being used to evaluate the effects of temperature, end-of-charge voltage, and depth of discharge on the performance of this new battery chemistry. The lithium-ion database will provide input to the advanced control system custom tailored with the lithium-ion model parameters for cell chemistry, cell-to-cell balancing, dynamics, and damage modeling. As test batteries succumb to failures, the control system can be tuned to assist in extending the charge/discharge cycle life. A follow-on program is planned to validate the benefits of this damage-control methodology on battery performance and life.
Find out more about Glenn’s research:
Electrochemistry Branch:
http://www.grc.nasa.gov/WWW/Electrochemistry/
Power and In-Space Propulsion:
http://www.grc.nasa.gov/WWW/5000/pep/
Glenn contacts:
Anthony G. Jannette, 216-433-2818, Anthony.G.Jannette@nasa.gov; and Thomas B. Miller, 216-433-6300, Thomas.B.Miller@nasa.gov
Authors:
Anthony G. Jannette, Michelle A. Manzo, and Thomas B. Miller
Headquarters program office:
Exploration Systems
Programs/Projects:
Power Systems R&T, ISS, CEV
Last updated: October 16, 2006
Responsible NASA Official:
Gynelle.C.Steele@nasa.gov
216-433-8258
Point of contact for NASA Glenn's Research & Technology reports:
Cynthia.L.Dreibelbis@nasa.gov
216-433-2912
SGT, Inc.
Web page curator:
Nancy.L.Obryan@nasa.gov
216-433-5793
Wyle Information Systems, LLC
