# US 2019/0245223 A1 · Controlled gas conditioning for a reaction gas of a fuel cell

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2019/0245223 A1
- Document ID: US20190245223A1
- Status: Published application
- Kind: A1
- Application number: 16/318,338
- Filed: 2017-07-17
- Published: 2019-08-08
- Earliest claimed: 2016-07-20
- Assignees: AVL List GmbH
- Applicants: Avl List GmbH
- Inventors: Christoph Kügele; Stefan Jakubek; János Kancsár
- Web page: https://patentgrep.com/patent/US20190245223A1

## CPC

- H01M 8/04089: of gaseous reactants
- H01M 8/04708: {of fuel cell reactants}
- H01M 8/04753: {of fuel cell reactants}
- H01M 8/04835: {of fuel cell reactants}
- H01M 8/04992: characterised by the implementation of mathematical or computational algorithms, e.g. feedback control loops, fuzzy logic, neural networks or artificial intelligence
- Y02E 60/50: Fuel cells
- H01M: PROCESSES OR MEANS, e.g. BATTERIES, FOR THE DIRECT CONVERSION OF CHEMICAL ENERGY INTO ELECTRICAL ENERGY
- Y02E: REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION

## Abstract

The aim of the invention is to allow a precise and quick regulation of the influencing variables in the operation of a fuel cell. A regulator is used for the regulation, said regulator being designed on the basis of a model of the gas conditioning unit in the form of a coupled nonlinear multivariable system in that the coupled nonlinear multivariable system is decoupled and linearized using Lie derivatives, and the regulator is designed for the decoupled linear multivariable system. The regulator calculates the control variables (uG, uS, uN, Q) for at least three present actuators of the influencing variables of the gas conditioning unit at each sampling time of the regulation using specified target variables (yj,dmd) and at least the three actuators of the gas conditioning unit adjust the calculated control variables (uG, uS, uN, Q at each sampling time of the regulation.

## Claims

### Claim 1

A gas conditioning system for a reaction gas of a fuel cell, comprising:
a gas source configured and arranged to provide the for a reaction gas to the gas conditioning system;
a humidifier unit configured and arranged to humidify the reaction gas and thereby achieve a desired a relative humidity of the reaction gas;
a temperature control unit configured and arranged to maintain the temperature of the reaction gas at a set temperature;
a mass flow control unit configured and arranged to control mass flow rate of the reaction gas; and
a pressure control unit configured and arranged to control the pressure of the reaction gas,
a control unit including a controller configured and arranged to control the gas conditioning system by computing manipulated variables for at least three actuators from the group of the humidifier unit, the temperature-control unit, the mass flow control unit, and the pressure control unit, to thereby set a predefined time curve of at least three associated influencing variables on the operation of the fuel cell from the group of the relative humidity, the temperature, the mass flow, and the pressure of the reaction gas in a reaction gas line, and wherein
the controller is designed on the basis of a model of the gas conditioning unit in the form of a coupled, nonlinear multivariable system of the form
x
.
=
f

(
x
)
+
∑
i
=
1
m


g
i

(
x
)

u
i
y
=
h

(
x
)
,
wherein the multivariable system is decoupled and linearized by means of Lie derivatives for controller design.

### Claim 2

A method for controlling a gas conditioning unit for a fuel cell comprising:
designing a controller using a coupled, nonlinear multivariable
x
.
=
f

(
x
)
+
∑
i
=
1
m


g
i

(
x
)

u
i
system of the form y=h(x) as model of the gas conditioning unit, in that the coupled, nonlinear multivariable system is decoupled and linearized by means of Lie derivatives and the controller is designed for the decoupled, linear multivariable system, wherein
computing with the controller, for every sampling time of the control, from predefined setpoint variables (yj,dmd), the manipulated variables {dot over (Q)} for at least three provided actuators of the gas conditioning unit from the group of a humidifier unit for humidifying the reaction gas to set a relative humidity of the reaction gas, a temperature control unit to set a temperature of the reaction gas, a mass flow control unit to control the mass flow rate of the reaction gas, and a pressure control unit to control the pressure of the reaction gas; and
operating the at least three actuators to set the manipulated variables at every sampling time of the control.

### Claim 3 (depends on 2)

The method according to claim 2, characterized in that each output variable (yj) of the coupled, nonlinear multivariable system is derived with respect to time according to its relative degree (δi) in the form
y
j
(
δ
j
)
=
L
f
δ
j

h
j
+
∑
i
=
1
m


(
L
g
i

L
f
δ
j
-
1

h
j
)

u
i
,
wherein Lfδi denotes the δj-th Lie derivative, and that the δj-th derivatives with respect to time of the output variables (yj) are equated with new input variables from which a linear, non-coupled state space model of the form ż=Acz+Bcv can be derived, having defined matrices Ac and Bc and a new state vector
zT(h1(x)Lfh1(x) . . . Lfδ1-1h1(x)h2(x)Lfh2(x) . . . Lfδ2-1h2(x) . . . Lfδm-1hm(x))T.

### Claim 4 (depends on 3)

The method according to claim 3, characterized in that the controller is designed for the linear, non-coupled state space model of the form
ż=Acz+Bcv.

### Claim 5 (depends on 4)

The method according to claim 4, characterized in that a controller having two degrees of freedom having a feedforward controller (FW) and a feedback controller is designed, wherein the outputs of the feedforward controller (FW) and the feedback controller are added to the new input variables (v) and the feedforward part results from the δj-th derivatives with respect to time of the setpoint variables (yj,dmd), and the feedback part corrects the error (e) between the present actual values of the output variables (yj) and the values of the setpoint variables (yj,dmd).

### Claim 6 (depends on 5)

The method according to claim 5, characterized in that error components (ej,δj) are defined according to ej,0=∫0tej(τ)dτ, ej,1=ej, ej.2=ėj, . . . , ej,δj=ej(δj-1), from which ė(t)=(Ac−BcK)e(t) results, having the defined matrices Ac and Bc and a matrix K having parameters of the feedback controller, which are defined to achieve a desired control behavior of the feedback controller.

### Claim 7 (depends on 2)

The method according to claim 2,
characterized in that, using the property of the differential flatness of the nonlinear, coupled multivariable system, the time curve of the state variables (x) are calculated according to
y=F(x)
x=Φ(y,{dot over (y)}, . . . ,y(n-1))
u=Ψ(y,{dot over (y)}, . . . ,y(n))
beforehand from the time curve of the setpoint variables (yj,dmd).
