Method for determining a control observer for the SoC

• US 10,338,146 B2
• Filed: 01/17/2014
• Issued: 07/02/2019
• Est. Priority Date: 01/23/2013
• Status: Active Grant

• Current Assignee: AVL List GmbH (List Privatstiftung)
• Original Assignee: AVL List GmbH (List Privatstiftung)
• Inventors: Hametner, Christoph, Jakubek, Stefan, Suljanovic, Amra
• Primary Examiner(s): Sofocleous, Alexander
• Assistant Examiner(s): Lozada, Alfredo Bermudez
• Application Number:
  US14/762,719
• Publication Number:
  US 20150362559A1
• Time in Patent Office: 1,992 Days
• Field of Search: 702 63
• CPC Class Codes: G01R 31/367 Software therefor, e.g. for battery testing using modelling or look-up tables
• CPC Class Codes: G01R 31/3828 using current integration
• CPC Class Codes: G01R 31/3842 combining voltage and current measurements
• CPC Class Codes: G01R 31/392 Determining battery ageing or deterioration, e.g. state of health
• CPC Class Codes: Y02E 60/10 Energy storage using batteries

Abstract

In order to create a control observer for any battery type in a structured and at least partially automated manner, first, a nonlinear model of the battery, in form of a local model network including a number of local, linear, time-invariant, and dynamic models, which each have validity in specific ranges of the input variables, is determined from the measuring data of a previously ascertained, optimized experimental design via a data-based modeling method. For each local model (LM_i) of the model network determined in this manner, a local observer is then determined. The control observer (13) for estimating the SoC then results from a linear combination of the local observers.

Claims

1. A method for determining a control observer (13) to estimate a state of charge (SoC) of a battery (10), comprising the following method steps:

specifying an initial excitation signal of the battery (10), wherein the excitation signal comprises a time sequence of a plurality of design points defined by a load current (1) and an SoC value;

specifying an initial model of the battery (10) having a model output (y) and model parameters (ψ

);

determining an optimized experimental design in form of an optimized time sequence of design points by use of a model based design of experiments method on the basis of a Fisher information matrix (I_FIM), defined by

2. The method according to claim 1, wherein the optimized experimental design is determined by minimizing a result function

3. The method according to claim 1, wherein each ith local model (LM_i) of the local model network (LMN) comprises a validity function (Φ

_i) and a model parameter vector θ

, wherein the model parameter vector (θ

_i) includes all model parameter for the ith model (LM_i) and the validity function (Φ

_i) defines the validity range of the ith local model (LM_i), and a local estimate of output variable ŷ

_i(k) of the ith local model (LM_i) is determined at the point in time k from ŷ

_i(k)=x^T(k)θ

_i, wherein x(k) references a regression vector, which comprises present and past inputs and outputs, that the global model output ŷ

(k) is determined from the linear combination by weighting the local model outputs via the validity function (Φ

_i) according to

4. The method according to claim 3, wherein for each ith local model (LM_i) a state vector z_i(k)=A_iz(k−

1)+B_iu(k) is applied, wherein a system matrix A_i and a input matrix B_i result from the determined model parameter vector (θ

_i), and the total system state results by weighting the local states with the validity function (Φ

_i) in

5. The method according to claim 4, wherein the estimated state {circumflex over (z)}(k), which comprises the estimated SoC, results from the equation

6. The method according to claim 1, wherein at least one input variable of the local model network (LMN) is scaled by at least one parameter (C_akt, R_akt) for a state of health (SoH) of the battery (10).

7. The method according to claim 3, wherein at least one model parameter of the local model network (LMN) is scaled by at least one parameter (C_akt, R_akt) for a state of health (SoH) of the battery (10).

8. The method according to claim 6, wherein the observer (13) for the SoC is extended by an observer (14) for the parameter (C_akt, R_akt).

9. The method according to claim 8, wherein the observer (14) for the SoH minimizes an observation error (e) between the measured output variable (y) and the estimated output variable (ŷ

) of the nonlinear battery model with the aid of a specified quality measure for the observation error (e) in relation to the parameter (C_akt, R_akt).