# US 2015/0177030 A1 · Transient multivariable sensor evaluation

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

- Number: US 2015/0177030 A1
- Document ID: US20150177030A1
- Status: Published application
- Kind: A1
- Application number: 14/134,790
- Filed: 2013-12-19
- Published: 2015-06-25
- Earliest claimed: 2013-12-19
- Assignees: Uchicago Argonne, LLC
- Applicants: Uchicago Argonne, LLC
- Inventors: Richard B. Vilim; Alexander Heifetz
- Web page: https://patentgrep.com/patent/US20150177030A1

## CPC

- G01D 3/08: with provision for safeguarding the apparatus, e.g. against abnormal operation, against breakdown
- G06F 2218/20: by applying autoregressive analysis
- G01D: MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A method and system for performing transient multivariable sensor evaluation. The method and system includes a computer system for identifying a model form, providing training measurement data, generating a basis vector, monitoring system data from sensor, loading the system data in a non-transient memory, performing an estimation to provide desired data and comparing the system data to the desired data and outputting an alarm for a defective sensor.

## Claims

### Claim 1

A method of performing transient multivariable sensor evaluation in a system, comprising the steps of,
identifying a model form;
providing training measurement data;
generating a basis vector;
monitoring system data from a sensor;
loading the system data;
performing an estimation to provide desired data; and
comparing the system data to the desired data and outputting an alarm for a defective sensor.

### Claim 2 (depends on 1)

The method as defined in claim 1 wherein the step of identifying the model form comprises determining whether the system conforms to an assumed model form.

### Claim 3 (depends on 2)

The method as defined in claim 2 wherein the model form is in accord with,

x

(
t
)

t
=
f
a

(
x

(
t
)
,
u

(
t
)
)
and y(t)=g(x(t),u(t)) where,
x(t)=n*p×1 state vector
u(t)=r×1 input vector
y(t)=p×1 output vector

### Claim 4 (depends on 3)

The method as defined in claim 3 further including the step of transforming the model form into a time-discretizing form such that input and output variables for the system data are sampled at periodic time points wherein the vector u(t) can change only at discrete times tk where k=0, 1, 2 . . . and tk+1−tk=Δ and y(t) is sampled only at times tk.

### Claim 5 (depends on 4)

The method as defined in claim 4 further including the step of integrating across the sampled time points comprise processing,
δ


x
k
+
1
=
θ
k

δ


x
k
+
θ
k

δ


u
k


and
δ


y
k
=
C
k

δ


x
k
+
D
k

D
k

δ


u
k
where
θ
k

(
t
,
j
)
=
δ


g

(
i
)
δ


x

(
j
)


k


θ
k

(
i
,
j
)
=
δ


f
d

(
i
)
δ


u

(
j
)


k


C
k

(
i
,
j
)
=
δ


g

(
i
)
δ


x

(
j
)


k
k


D
k

(
i
,
j
)
=
δ


g

(
i
)
δ


u

(
j
)


k
k


### Claim 6 (depends on 5)

The method as defined in claim 5 further including transforming to an input-output version of the model form.

### Claim 7 (depends on 6)

The method as defined in claim 6 wherein the model form comprises an auto-regressive moving average (ARMA) form.

### Claim 8 (depends on 7)

The method as defined in claim 7 wherein the ARMA form comprises,
δ


y
k
+
A
1

δ


y
k
-
1
+
…
+
A
n

δ


y
k
-
n
=
C
o

δ


u
k
+
…
+
C
n

δ


u
k
-
n
E


δ


Z
k
=
0
E
=
[
I
A
1
A
2
…


A
n
-
C
0
-
C
1
…
-
C
n
]
.

δ


Z
k
=
[
δ


y
k
δ


y
k
-
1
⋮
δ


y
k
-
n
δ


u
k
δ


u
k
-
1
⋮
δ


u
k
-
n
]

### Claim 9 (depends on 8)

The method as defined in claim 8 wherein the ARMA form for causality operation comprises,
[
A
-
C
]

[
δ


Y
δ


U
]
=
0
where
A
=
[
I
A
1
A
2
…


A
n
]


C
=
[
C
0
C
1
…


C
n
]
.

and
δ


Y
=
[
δ


y
k
δ


y
k
-
1
⋮
δ


y
k
-
n
]


δ


U
=
[
δ


u
k
δ


u
k
-
1
⋮
δ


u
k
-
n
]


### Claim 10 (depends on 9)

The method as defined in claim 9 wherein the ARMA form for input controllable operation comprises
[δY1δY2 . . . δYK−n]T[CL−1A]T=[δU1δU2 . . . δUK−n]T

### Claim 11 (depends on 9)

The method as defined in claim 9 wherein the ARMA form for output observable operation comprises,
[δY1δY2 . . . δYK]=AL−1C[δU1δU2 . . . δUK−n]
[δU1δU2 . . . δUK−n]T[AL−1C]T=[δY1δY2 . . . δYK−n]T

### Claim 12 (depends on 9)

The method as defined in claim 9 wherein the ARMA form is re-cast into a reduced form based on the following two mathematical properties of stable physical systems,
hij[n]=0, n<0 and
Σn=−∞∞|hij[n]|=Σn=0∞|hij[n]|<∞,
where h represents a physical component and h[n]=x[n]=Σn=∞∞h[k]x[n−k]
and in a discrete time domain becomes,

(
y
1

[
n
]
⋮
y
l

[
n
]
)
=
(
h
11

[
n
]
…
h
1


m

[
n
]
⋮
⋱
⋮
h
l


1

[
n
]
…
h
l


m

[
n
]
)
*
(
x
1

[
n
]
⋮
x
m

[
n
]
)
(
h
11

[
0
]
…
h
11

[
K
]
…
h
1


m

[
0
]
…
h
1


m

[
K
]
-
1
…
0
⋮
h
l


1

[
0
]
…
h
l


1

[
K
]
…
h
l


m

[
0
]
…
h
l


m

[
K
]
0
…
-
1
)
·
(
x
1

[
n
]
⋮
x
1

[
n
-
K
]
⋮
x
m

[
n
]
⋮
x
m

[
n
-
K
]
y
1

[
n
]
⋮
y
l

[
n
]
)
=
0

### Claim 13 (depends on 12)

The method as defined in claim 12 wherein the step of generating the basis vectors comprises specifying normal sensor measurement accuracy, initializing number of time lags to zero, generating a training matrix from measurements of the system data by applying
A=(X1 . . . Xr),
computing a singular value decomposition of the training data matrix by applying,
Xnεnull(C); and
testing whether σ points are less than the normal sensor measurement accuracy, and if is less than, then form the matrix of the basis vectors, and if not less than the normal sensor measurement accuracy, increment by K and return to composing the training data matrix step and then recycling until the σ points are less than the normal sensor measurement accuracy.

### Claim 14

A system for performing transient multivariable sensor evaluation, comprising:
providing a computer system having a non-transient memory with a computer software program stored therein for executing instructions to execute the steps of identifying a model form;
the non-transient memory storing training measurement data;
the system generating a basis vector;
the system monitoring system data;
the non-transient memory providing the system data;
the computer system performing an estimation to provide desired data; and
the computer system comparing the system data to the desired data and outputting an alarm for defective sensors.

### Claim 15 (depends on 14)

The system as defined in claim 14 wherein the computer system identifying the model form comprises determining whether the system conforms to an assumed model form.

### Claim 16 (depends on 14)

The system as defined in claim 14 wherein the model form is in accord with,

x

(
t
)

t
=
f
a

(
x

(
t
)
,
u

(
t
)
)
and y(t)=g(x(t),u(t)) where,
x(t)=n*p×1 state vector
u(t)=r×1 input vector
y(t)=p×1 output vector

### Claim 17 (depends on 14)

The system as defined in claim 14 further including the computer system operating to transform the model form into a time-discretizing form with that input and output variables for the system data are sampled at periodic time points wherein the vector u(t) can change only at discrete times tk where k=0, 1, 2 . . . and tk+1−tk=Δ and y(t) is sampled only at times tk.

### Claim 18 (depends on 14)

The system as defined in claim 14 further including the computer system operating to integrate across the sampled time points applying,
δ


x
k
+
1
=
θ
k

δ


x
k
+
θ
k

δ


u
k


and
δ


y
k
=
C
k

δ


x
k
+
D
k

D
k

δ


u
k
where
θ
k

(
t
,
j
)
=
δ


g

(
i
)
δ


x

(
j
)


k


θ
k

(
i
,
j
)
=
δ


f
d

(
i
)
δ


u

(
j
)


k


C
k

(
i
,
j
)
=
δ


g

(
i
)
δ


x

(
j
)


k
k


D
k

(
i
,
j
)
=
δ


g

(
i
)
δ


u

(
j
)


k

### Claim 19 (depends on 14)

The system as defined in claim 14 further including the computer operating to transform an input-output version of the model form.

### Claim 20 (depends on 14)

The system as defined in claim 14 wherein the model form comprises an auto-regressive moving average (ARMA) form.
