Automatic estimation process and device for a flight parameter vector in an aircraft, as well as detection methods and assemblies for a failure affecting such a vector

• US 9,116,520 B2
• Filed: 06/08/2012
• Issued: 08/25/2015
• Est. Priority Date: 07/26/2011
• Status: Active Grant

• Current Assignee: Airbus Operations (European Aeronautic Defence and Space Company EADS NV)
• Original Assignee: Airbus Operations (European Aeronautic Defence and Space Company EADS NV)
• Inventors: Goupil, Philippe, Cazes, Florian, Dayre, Rmy, Le Berre, Herv, Mailhes, Corinne, Chabert, Marie, Michel, Patrice
• Primary Examiner(s): Sample, Jonathan L
• Application Number:
  US13/492,120
• Publication Number:
  US 20130030610A1
• Time in Patent Office: 1,173 Days
• US Class Current: 1/1
• CPC Class Codes: G05D 1/0077 using redundant signals or controls

Abstract

Embodiments of the disclosure include methods and devices in which a flight parameter vector for an aircraft is automatically estimated and then compared to a threshold to determine whether the flight parameter vector is accurate. The detection device estimates a coefficient vector in an observation window, calculates an estimated flight parameter value, and minimizes an error associated with the observation window when calculating the estimated flight parameter value of the flight parameter vector. Comparisons may be performed by a failure detector or a flight control computer to determine failures impacting the flight parameter vector.

Claims

1. A method of automatically estimating a flight parameter vector when in flight used by an electrical flying control system of an aircraft, the method comprising:

selecting, by an assembly of information sources, a plurality of observed explanatory values from a plurality of explanatory values, wherein each explanatory value represents a flight parameter of the aircraft;

estimating, by an estimated value calculator, a current coefficient vector in a current observation window and a previous coefficient vector in a previous observation window, wherein the current observation window includes a sampling of the observed explanatory values for a most recent period of time and the previous observation window includes a sampling of the observed explanatory values for a previous period of time;

calculating a first estimation of an estimated flight parameter value from the plurality of observed explanatory values observed during the current observation window and the previous coefficient vector, and a second estimation of the estimated flight parameter value from the plurality of observed explanatory values observed during the current observation window and the current coefficient vector, wherein the first and second estimations of the estimated flight parameter value are calculated by implementing a linear relationship between the respective estimated flight parameter value and the respective plurality of observed explanatory values by a Partial Least Squares PLS regression;

comparing the first and second estimations of the estimated flight parameter value and an observed flight parameter vector to detect at least one failure impacting the estimation of the flight parameter vector; and

estimating the flight parameter vector when in flight by using the current coefficient vector in a calculation of the flight parameter vector to minimize error in the flight parameter vector.

2. The method according to claim 1, wherein a flight parameter vector includes at least one estimated flight parameter value that is associated with at least one flight parameter.

3. The method according to claim 1, wherein the observation window includes a plurality of successive samples.

4. The method according to claim 1, further comprising:

accommodating a plurality of non-centered signals by implementing an adjusting input variable.

5. The method according to claim 1, wherein the comparing of the first and the second estimations and the observed flight parameter vector further comprises:

calculating a decision value from the first estimation, the second estimation, and the plurality of observed explanatory values;

comparing the decision value to a threshold; and

detecting a failure when the decision value is higher than the threshold.

6. The method according to claim 5, wherein the threshold is determined based on detection and false alarm probabilities.

7. The method according to claim 5, wherein the detecting of a failure further comprises:

detecting when the decision value is higher than the threshold during a confirmation time.

8. The method according to claim 1, further comprising:

detecting a malfunction based upon a development change in the current coefficient vector.

9. The method according to claim 8, wherein detecting a malfunction further comprises:

calculating a criterion that is representative of an intra-vectorial development of the current coefficient vector; and

comparing the criterion to a predetermined value so that the malfunction is detected when the criterion is higher than the predetermined value during a confirmation duration.

10. The method according to claim 8, wherein detecting a malfunction further comprises:

calculating a criterion that is representative of statistics associated with the current coefficient vector; and

comparing the criterion to a predetermined value so that the malfunction is detected when the criterion is lower than the predetermined value during a confirmation duration.

11. An automatic estimation device for a flight parameter vector when in flight used by an electrical flying order system of an aircraft, the device comprising:

an assembly of information sources that select a plurality of observed explanatory values from a plurality of explanatory values and transmit the plurality of observed explanatory values to an estimated value calculator, wherein each explanatory value represents a flight parameter of the aircraft; and

the estimated value calculator, which is connected to the assembly of information sources by a link and which performs the following functions;

estimate a current coefficient vector in a current observation window and a previous coefficient vector in a previous observation window, wherein the current observation window includes a sampling of the observed explanatory values for a most recent period of time and the previous observation window includes a sampling of the observed explanatory values for a previous period of time,

calculate a first estimation of an estimated flight parameter value from the plurality of observed explanatory values observed during the current observation window and the previous coefficient vector, and a second estimation of the estimated flight parameter valude from the plurality of observed explanatory values observed during the current observation window and the current coefficient vector, wherein the first and second estimations of the estimated flight parameter value are calculated by implementing a linear relationship between the respective estimated flight parameter value and the respective plurality of observed explanatory values by a Partial Least Squares PLS regression;

compare the first and second estimations of the estimated flight parameter value and an observed flight parameter vector to detect at least one failure impacting the estimation of the flight parameter vector; and

estimate the flight parameter vector when in flight by using the current coefficient vector in a calculation of the flight parameter vector to minimize error in the flight parameter vector.

12. The device according to claim 11, further comprising:

a measurement device that measures on the aircraft the plurality of explanatory values from which the plurality of observed explanatory values is selected, the measurement device being connected to the estimated value calculator by a link.

13. The device according to claim 11, further comprising:

a detector configured to detect a malfunction is based upon a development change in the current coefficient vector.