NREL's Battery Life Predictive Model Helps Companies Take Charge
October 26, 2015
Companies that rely on batteries for enhanced energy efficiency-including electric vehicle (EV) manufacturers, solar and wind energy generation companies, and utilities-need to know how to use batteries most effectively. As investment in large-scale battery energy storage grows, it is also vital to know how long batteries will last in the field.
This spring, the National Renewable Energy Laboratory (NREL) licensed its Battery Life Predictive Model to two leading utility providers in the United States: Southern California Edison (SCE), one of the nation's largest investor-owned utilities serving 14 million customers, and Next Era Energy, a leading clean energy company with revenues totaling around $17 billion. The utilities will use the NREL model to select long-lasting energy storage systems most capable of reliably balancing grid electricity demands.
"Having a robust battery life model is very important to support the cost-benefit analysis and reduce the business risks. NREL's modeling approach enables the life-prediction capabilities that facilitate business decisions," said SCE Electrochemical Storage Evaluation Manager Loic Gaillac.
Because electrochemical batteries lose functionality over time due to various factors, it is important to understand their degradation behavior in order to estimate life-cycle costs accurately for users and to set realistic parameters for manufacturers' warranties. Yet batteries, especially lithium-ion (Li-ion) batteries, are complex electrochemical systems. There are typically several different degradation mechanisms, which simultaneously cause battery energy and power performance to fade. Pinpointing use conditions that lead to premature failure is difficult.
In response, NREL has created the unique Battery Life Predictive Model, which extends standardized laboratory battery-aging datasets to complex real-world scenarios. Unlike empirical models, the NREL model captures separate degradation pathways that can occur depending upon the exact combination of calendar time, environment, and duty cycle. Based on physical degradation mechanisms, the model accurately calculates lifetime for variable temperature and charge/discharge cycling, such as those encountered in real-world vehicle and grid storage applications. The model is also applied in real-time control algorithms that co-optimize battery life and performance.
SCE is looking to incorporate MWh-class energy storage on its distribution circuits, with expected system life of more than 15 years. The storage systems help reduce costs to upgrade the grid when demand grows, and respond quickly to grid fluctuations, such as those that occur as wind and solar power vary with weather conditions.
The model's origins began in 2010, when the U.S. Department of Energy (DOE) Vehicle Technology Office supported NREL researchers to analyze tradeoffs in EV battery systems design. Kandler Smith, a senior engineer in NREL's Transportation and Hydrogen Systems Center, said the research focused on issues such as examining the costs and benefits of using active thermal management systems that provide longer life versus less expensive passive cooling systems.
Other recent projects have allowed the NREL team to refine and validate the methodology. Under funding from DOE's Advanced Research Projects Agency-Energy, NREL is demonstrating that Eaton Corporation can downsize its hybrid electric vehicle (HEV) battery pack by 30% while maintaining the same HEV performance and life. NREL's Battery Life Predictive Model allows Eaton's HEV controls to manage battery performance and lifetime in real time. The Eaton project, along with previous projects with General Motors, have demonstrated that models developed from standard laboratory cell aging tests can accurately predict lifetime for a complete battery systems undergoing complex real-world cycling and variable temperature conditions.
The model has been licensed to a variety of automotive manufactures, EV service providers, and university and laboratory research groups, and there are various ways it is being applied. NREL researchers can take client's usage data and run them in the predictive model; companies can license the software code and do their own analyses; or NREL can conduct the battery aging tests in its labs, analyze the results and develop specific models for the client. The Battery Life Predictive Model is also integral part of the Battery Lifetime Analysis and Simulation Tool (BLAST) and Computer-Aided Engineering for Electric-Drive Vehicle Batteries (CAEBAT) activities at NREL.
"Industry benefits from standardized procedures in evaluating batteries," Smith said, which explains why the Battery Life Predictive Model is gaining traction in a variety of applications.
NREL is a recognized leader in energy storage research and development, spearheading modeling, simulation, and testing activities. The laboratory's innovative and integrated approach to sustainable transportation helps government, industry, and other partners develop and deploy the components and systems needed for market-ready, high-performance, low-emission, fuel-efficient passenger and freight vehicles, as well as alternative fuels and related infrastructure.