Method and system for predicting residual useful life of an air filter
A method for predicting the residual useful life of an air filter arrangement is described. The method includes the following steps: providing a plurality of predetermined reference degradation curves; measuring a degradation parameter of the filter arrangement; and estimating the residual useful life of the filter arrangement by comparing the predetermined reference degradation curves and an actual degradation curve defined by measured values of the degradation parameter.
1. A method for predicting the residual useful life of an air filter arrangement, the method comprising:
measuring a degradation parameter of the filter arrangement; and
estimating the residual useful life of the filter arrangement by comparing a plurality of predetermined reference degradation curves and an actual degradation curve defined by measured values of the degradation parameter.
2. The method of claim 1 , wherein the degradation parameter is the pressure loss across the filter arrangement.
3. The method of claim 1 , wherein comparing the plurality of predetermined reference degradation curves and the actual degradation curve further comprises applying a similarity-based prediction algorithm.
4. The method of claim 1 , further comprising:
calculating a similarity measure between at least some of the reference degradation curves (Cj) and the actual degradation curve; and
calculating the residual useful life of the filter arrangement as a weighted sum of the residual useful life values of the reference degradation curves, the sum being weighted by the similarity measures.
5. The method of claim 4 , wherein the similarity measure is calculated for a time frame.
6. The method of claim 4 , wherein each similarity measure is calculated based on an Euclidean distance between the actual degradation curve of the filter arrangement and the respective reference degradation curve, and wherein the Euclidean distance is transformed into the corresponding similarity measure by means of a Gaussian kernel.
7. The method of claim 6 , wherein the Gaussian kernel has a bandwidth between 0.1 and 5.
8. The method of claim 1 , wherein when the measured degradation parameter is above a limit value, a physics-based estimated value of the residual useful life is combined with the residual useful life value estimated on the basis of the reference degradation curves.
9. The method of claim 8 , wherein the physics-based estimated value of the residual life of the filter arrangement is calculated using a regression applied to the actual degradation curve of the filter arrangement.
10. The method of claim 9 , wherein the regression is selected from the group consisting of: a linear regression, a quadratic regression, or a combination thereof.
11. The method of claim 8 , wherein the physics-based estimated residual useful life of the filter arrangement and the residual useful life of the filter arrangement estimated on the basis of the reference degradation curves are combined to one another through a weighing parameter, such that the weight of the physics-based estimated value becomes predominant over the residual useful life estimated on the basis of the reference degradation curves as the degradation parameter increases from the limit value to a threshold value indicative of the end of life of the filter arrangement.
12. The method of claim 8 , wherein the physics-based estimated residual useful life of the filter arrangement and the residual useful life of the filter arrangement estimated on the basis of the reference degradation curves are combined to one another through the formula wherein
f
=
median
[
0
;
Δ
p_alarm
-
x
Δ
p_lim
;
1
]
(
6
)
and wherein
RUL SE is the residual useful life estimated on the basis of the reference degradation curves
RUL PE is physics-based estimated residual useful life,
Δp_lim is a limit value,
x is the measured degradation parameter.
13. A gas turbine engine system comprising
a gas turbine engine and
a turbine inlet air filter system, comprised of at least one filter arrangement; wherein a filter degradation measuring system is combined with the filter arrangement and interfaced with a residual useful life estimation system, configured to:
measure a degradation parameter of the filter arrangement; and
estimate the residual useful life of the filter arrangement by comparing a plurality of predetermined reference degradation curves and an actual degradation curve defined by measured values of the degradation parameter.
14. The gas turbine engine system of claim 13 , wherein the degradation parameter is the pressure loss across the filter arrangement.
15. The gas turbine engine system of claim 13 , wherein the system is further configured to apply a similarity-based prediction algorithm.
16. The gas turbine engine system of claim 13 , wherein the system is further configured to
calculate a similarity measure between at least some of the reference degradation curves (Cj) and the actual degradation curve; and
calculate the residual useful life of the filter arrangement as a weighted sum of the residual useful life values of the reference degradation curves, the sum being weighted by the similarity measures.
17. The gas turbine engine system of claim 16 , wherein the similarity measure is calculated for a time frame.
18. The gas turbine engine system of claim 16 , wherein each similarity measure is calculated based on an Euclidean distance between the actual degradation curve of the filter arrangement and the respective reference degradation curve, and wherein the Euclidean distance is transformed into the corresponding similarity measure by means of a Gaussian kernel.
19. The gas turbine engine system of claim 13 , wherein when the measured degradation parameter is above a limit value, a physics-based estimated value of the residual useful life is combined with the residual useful life value estimated on the basis of the reference degradation curves.
20. The gas turbine engine system of claim 19 , wherein the physics-based estimated value of the residual life of the filter arrangement is calculated using a regression applied to the actual degradation curve of the filter arrangement.
21. A method for predicting the residual useful life of an air filter arrangement, the method comprising:
measuring a degradation parameter of the filter arrangement; and
estimating the residual useful life of the filter arrangement by comparing a plurality of predetermined reference degradation curves and an actual degradation curve defined by measured values of the degradation parameter,
wherein when the measured degradation parameter is above a limit value, a physics-based estimated value of the residual useful life is combined with the residual useful life value estimated on the basis of the reference degradation curves, and
wherein the physics-based estimated residual useful life of the filter arrangement and the residual useful life of the filter arrangement estimated on the basis of the reference degradation curves are combined to one another through the formula
RUL ( Ti )= f·RUL SS ( Ti )+(1− f )· RUL PB ( Ti ) (5)
wherein
f
=
median
[
0
;
Δ
p_alarm
-
x
Δ
p_lim
;
1
]
(
6
)
and wherein
RUL SE is the residual useful life estimated on the basis of the reference degradation curves
RUL PE is physics-based estimated residual useful life,
Δp_lim is a limit value,
x is the measured degradation parameter.