Method for determining critical operating states in a fuel cell stack

Patent metadata

Field	Value
Publication number	US-10145900-B2
Application number	US-201314398038-A
Country	US
Kind code	B2
Filing date	May 2, 2013
Priority date	May 3, 2012
Publication date	Dec 4, 2018
Grant date	Dec 4, 2018

Abstract

The invention relates to a method for determining critical operating states in a fuel cell stack, consisting of single cells connected in series, wherein a low-frequency current or voltage signal is applied to the fuel cell stack, the resulting voltage or current signal is measured and the distortion factor thd is determined. According to the invention, the weighted sum of a term dependent on the membrane resistance Rm and a term dependent on the distortion factor thd is used to determine an indicator THDAdryout correlating with the drying out of the fuel cell membranes of the fuel cell stack, the membrane resistance Rm being detected by impedance measurement.

First claim

The invention claimed is:

1. A method for determining critical operating states in a fuel cell stack, consisting of individual cells connected in series, wherein a low-frequency current or voltage signal is applied to the fuel cell stack, the resulting voltage or current signal is measured and the distortion factor thd of the signal is determined, wherein the parameters thd dif0 and thd dif1 as well as the fluctuations fd(V) in the measured voltage curve are used for determining an indicator THDA liquid correlating with impermissible water accumulations and droplet formations on the membranes of the fuel cell stack, wherein thd dif0 and thd dif1 respectively concern linear combinations of the distortion factors of current and voltage, and wherein thd dif0 comprises the component of the 2nd harmonics and thd dif0 the component of the 3rd harmonics.

2. The method according to claim 1 , wherein the indicator THDA liquid correlating with impermissible water accumulations and droplet formations on the membranes of the fuel cell stack is determined according to
THDA liquid =α 0 ·ƒ(α 1 ·abs(thd dif0 )+α 2 ·abs(thd dif1 ))+α 3 ·ƒd ( V ),

wherein α 0 ,α 1 ,α 2 ,α 3 are system-dependent parameters.

3. The method according to claim 1 , wherein a simplified electrical equivalent circuit of the fuel cell stack can be used for additionally determining an Indicator SoH correlating with the ageing of the fuel cell stack, which equivalent circuit at least considers the ohmic resistances R 1 , R 2 of the cathode side and the anode side, as well as the double-layer capacitances C 1 , C 2 on the anode side and the cathode side, wherein an equation system for the variables R 1 , R 2 , C 1 , C 2 is set up, which is determined by impedance measurements in at least three measuring frequencies and is used for calculating the indicator SoH.

4. The method according to claim 3 , wherein the indicator SoH correlating with the ageing of the fuel cell stack is determined from the parameters C 1 , C 2 of the double-layer capacitances according to

SoH
=

α
1

·

(

100
·

C
1

0.75
⁢

⁢

C
1

_

-

100
3

)

+

α
2

·

(

100
·

C
2

0.75
⁢

⁢

C
2

_

-

100
3

)

,

wherein the parameters disregard the ohmic resistances of the cathode side and anode side R 1 , R 2 , α 1 +α 2 =1 applies and C 1 and C 2 concern starting values of a new fuel cell stack.

5. The method according to claim 3 , wherein three measuring frequencies are selected for the calculation, in which the impedance curve of the simplified equivalent circuit coincides substantially with the real impedance curve of the fuel cell stack.

6. The method according to claim 1 , wherein a stoichiometric undersupply of the cathode side of the fuel cell stack is determined by the combined occurrence of a rising distortion factor thd according to the indicator THDA liquid and a deviation of internal resistance R i from a reference value of the internal resistance according to the indicator THDA low media .

7. The method according to claim 1 , wherein an artificial neural network (Artificial Neural Network) ANN is used for the additional determination of an indicator avg-min correlating with the deviation of the minimum cell voltage from the average cell voltage of the fuel cell stack, measured quantities derived from the distortion factor analysis THDA and impedance values derived from the real and imaginary part of the applied current and voltage signal are used as input quantities of the network, wherein the neural network is trained by means of signals from the Individual cell voltage measurements for determining the Internal network parameters.

8. The method according to claim 7 , wherein a double-layer feed-forward Artificial Neural Network FFANN is used for simulating the indicator avg-min correlating with the minimum of the cell voltage of an Individual cell of the fuel cell stack, and the neural network is adjusted to the measured values detected by means of individual cell voltage measurement by means of a training function.

9. The method according to claim 7 , wherein the quantity of training data is expanded in a modular manner by calculation results from physical models.

10. The method according to claim 8 , wherein the training function is the Levenberg-Marquardt training function.

Assignees

• Avl List Gmbh

Inventors

• Prenninger Peter
• Renner Katharina

Classifications

• H01M8/04582

  of the individual fuel cell · CPC title

• H01M8/04589

  of fuel cell stacks · CPC title

• H01M8/04537

  Electric variables · CPC title

• H01M8/04559 Primary

  of fuel cell stacks · CPC title

• H01M8/04552

  of the individual fuel cell · CPC title