Systems and methods for estimation and prediction of battery health and performance

Patent metadata

Field	Value
Publication number	US-2019187212-A1
Application number	US-201916276322-A
Country	US
Kind code	A1
Filing date	Feb 14, 2019
Priority date	Nov 21, 2016
Publication date	Jun 20, 2019
Grant date	—

Abstract

Systems and computer-implemented methods are used for analyzing battery information. The battery information may be acquired from both passive data acquisition and active data acquisition. Active data may be used for feature extraction and parameter identification responsive to the input data relative to an electrical equivalent circuit model to develop geometric-based parameters and optimization-based parameters. These parameters can be combined with a decision fusion algorithm to develop internal battery parameters. Analysis processes including particle filter analysis, neural network analysis, and auto regressive moving average analysis can be used to analyze the internal battery parameters and develop battery health metrics. Additional decision fusion algorithms can be used to combine the internal battery parameters and the battery health metrics to develop state-of-health estimations, state-of-charge estimations, remaining-useful-life predictions, and end-of-life predictions for the battery.

First claim

What is claimed is:

1 . A method for analyzing an energy storage device, comprising:
applying a signal to an energy storage device and measuring a response of the energy storage device to the applied signal as active information; collecting passive information about the energy storage device; developing one or more internal parameters for the energy storage device responsive to the active information, the passive information, and one or more learned state models for the energy storage device; updating the one or more learned state models responsive to the one or more internal parameters, the active information, and the passive information; and processing the one or more learned state models to determine one or more health conditions for the energy storage device.

2 . The method of claim 1 , further comprising communicating at least one of the one or more health conditions to a user, a related computing system, or a combination thereof.

3 . The method of claim 1 , wherein the passive information includes at least one of a temperature, a voltage, and a current of the energy storage device.

4 . The method of claim 1 , wherein the active information is derived from rapid AC impedance measurements.

5 . The method of claim 1 , wherein the one or more internal parameters include at least one of a constant phase element exponent, an electrolyte resistance, a charge transfer resistance, and an ohmic resistance.

6 . The method of claim 1 , wherein the one or more health conditions include one or more present health conditions including at least one of a state-of-charge estimation for the energy storage device and a state-of-health estimation for the energy storage device.

7 . The method of claim 1 , wherein the one or more health conditions include one or more future health conditions including at least one of a remaining-useful-life prediction for the energy storage device and an end-of-life prediction for the energy storage device.

8 . The method of claim 1 , further comprising:
performing geometric-based parameter identification responsive to the active information relative to an electrical equivalent circuit model to develop geometric parameters; performing optimization-based parameter identification responsive to the active information relative to the electrical equivalent circuit model to develop optimized parameters; combining the geometric parameters and the optimized parameters to develop one or more new internal parameters for the energy storage device; and updating the one or more learned state models responsive to the one or more new internal parameters.

9 . The method of claim 8 , wherein performing the optimization-based parameter identification further comprises deriving one or more of a constant phase element exponent, an electrolyte resistance, an ohmic resistance, and a charge transfer resistance from the active information and performing a nonlinear optimization of each parameter estimated by minimizing a selected objective function.

10 . The method of claim 1 , further comprising:
performing two or more analysis processes using the one or more internal parameters to develop two or more health metrics corresponding to each analysis process; and determining the one or more health conditions for the energy storage device by performing a decision fusion algorithm for combining the two or more health metrics from the two or more analysis processes.

11 . The method of claim 10 , wherein:
one analysis process of the two or more analysis processes comprises using the one or more internal parameters in a particle filter (PF) analysis to develop a PF capacity metric, a neural network (NN) analysis to develop an NN capacity metric, and an auto regressive moving average (ARMA) analysis to develop an ARMA capacity metric; and further comprising performing a capacity decision fusion algorithm for combining the PF capacity metric, the NN capacity metric, and the ARMA capacity metric to develop an overall capacity health metric for inclusion in the decision fusion algorithm as one of the two or more health metrics.

12 . The method of claim 10 , wherein:
one analysis process of the two or more analysis processes comprises using the one or more internal parameters in a particle filter (PF) analysis to develop a PF available power metric, a neural network (NN) analysis to develop an NN available power metric, and an auto regressive moving average (ARMA) analysis to develop an ARMA available power metric; and further comprising performing an available power decision fusion algorithm for combining the PF available power metric, the NN available power metric, and the ARMA available power metric to develop an overall available power health metric for inclusion in the decision fusion algorithm as one of the two or more health metrics.

13 . The method of claim 10 , wherein:
one analysis process of the two or more analysis processes comprises using the one or more internal parameters in a particle filter (PF) analysis to develop a PF pulse resistance metric, a neural network (NN) analysis to develop an NN pulse resistance metric, and an auto regressive moving average (ARMA) analysis to develop an ARMA pulse resistance metric; and further comprising performing a pulse resistance decision fusion algorithm for combining the PF pulse resistance metric, the NN pulse resistance metric, and the ARMA pulse resistance metric to develop an overall pulse resistance health metric for inclusion in the decision fusion algorithm as one of the two or more health metrics.

14 . The method of claim 10 , wherein:
performing the two or more analysis processes comprises:
performing a capacity analysis to develop an overall capacity health metric;
performing an available power analysis to develop an overall available power health metric; and
performing a pulse resistance analysis to develop an overall pulse resistance health metric; and
performing the decision fusion algorithm comprises using a weighted average to combine the overall capacity health metric, the overall available power health metric, and the overall pulse resistance health metric to determine the one or more health conditions.

15 . A method for analyzing an energy storage device, comprising:
applying a signal to an energy storage device and measuring a response of the energy storage device to the applied signal as active information; collecting passive information about the energy storage device; developing one or more internal parameters for the energy storage device responsive to the active information, the passive information, and one or more state models for the energy storage device; and performing one or more what-if scenarios using one or more aging models to predict one or more future internal parameters of the energy storage device to characterize any changes to the one or more internal parameters as time progresses.

16 . The method of claim 15 , wherein the one or more what-if scenarios use assumed future operational conditions derived from previously observed operational conditions.

17 . The method of claim 15 , wherein using the one or more aging models further comprises:
performing a relevance vector machine algorithm using Bayesian inference to obtain possible solutions for a specific parameter of the one or more internal parameters when the energy storage device is new and an aging parameter indicating how the specific parameter will change over time; and performing a regression analysis to predict a future value for the specific parameter.

18

Assignees

• Battelle Energy Alliance Llc

Inventors

• Garcia Humberto E
• Christophersen Jon P

Classifications

• G06N20/00

  Machine learning · CPC title

• G06N3/04

  Architecture, e.g. interconnection topology · CPC title

• G05B23/0283

  Predictive maintenance, e.g. involving the monitoring of a system and, based on the monitoring results, taking decisions on the maintenance schedule of the monitored system; Estimating remaining useful life [RUL] (preventive maintenance, i.e. planning maintenance according to the available resources without monitoring the system G06Q10/06) · CPC title

• G01R31/392

  Determining battery ageing or deterioration, e.g. state of health · CPC title

• G01R31/382

  Arrangements for monitoring battery or accumulator variables, e.g. SoC · CPC title