Method and system for estimating the wear of a rotating machine including a bearing

1. A method for estimating the wear of a rotating machine including at least one bearing, the method comprising:

measuring vibrations of the rotating machine during its operation, using a vibration sensor;

automatically extracting, using an electronic detection device, from a signal measured by the vibration sensor, a first signal representative of components of a first frequency range of the measured signal, and a second signal representative of a second frequency range of the measured signal, wherein the first frequency range is the range [10 Hz, 1 kHz] and the second frequency range is the range [10 Hz, 10 kHz];

calculating, from the first signal, a first data set belonging to a time domain of the first signal;

extracting first calculation elements from the first data set;

further calculating, from the second signal, a second data set belonging to a frequency domain of the second signal;

further extracting second calculation elements from the second data set; and

determining a health index of the bearing from the first and second calculation elements.

2. The method according to claim 1 , wherein the measured signal is an acceleration signal representative of measured vibrations, wherein the first data set comprises a velocity of the measured vibrations, and wherein the second data set comprises an envelope of a spectrum of the acceleration of the measured signal.

3. The method according to claim 1 , wherein the first calculation elements comprise at least one element chosen from the following elements: a root mean square of the first data set, and a root mean square of the measured signal, and wherein the second calculation elements comprise at least one root mean square of harmonics of the second data set associated with a predefined failure frequency of the bearing.

4. The method according to claim 1 , wherein said determining a health index of the bearing comprises:

extrapolating the evolution of the health index over time, comprising calculating at least one future value of the health index from at least some previous calculated values of the health index; and

estimating a remaining lifetime of the bearing by comparing the at least one future value of the health index to at least one pre-defined limit.

5. The method according to claim 4 , further comprising emitting an alert via the electronic detection device if the health index exceeds one of the pre-defined limit values.

6. The method according to claim 1 , wherein said determining a health index of the bearing comprises calculating a statistical distance relative to a reference data set, for each of the extracted calculating elements.

7. The method according to claim 6 , wherein the statistical distance is a distance separating the calculation elements extracted from the reference data set.

8. The method according to claim 6 , wherein the statistical distance is a Mahalanobis distance.

9. The method according to claim 1 , the method being implemented periodically with a pre-defined periodicity.

10. A mechanical system, comprising:

a rotating machine, comprising:

a shaft; and

at least one bearing mounted on said shaft;

a vibration sensor to measure vibrations of said rotating machine; and

an electronic detection device for detecting a defect of said at least one bearing, the electronic detection device being programmed to:

measure vibrations of said rotating machine during its operation, using said vibration sensor;

automatically extract, from a signal measured by said vibration sensor, a first signal representative of components of a first frequency range of the measured signal, and a second signal representative of a second frequency range of the measured signal, wherein the first frequency range is the range [10 Hz, 1 kHz] and the second frequency range is the range [10 Hz, 10 kHz];

calculate, from the first signal, a first data set belonging to a time domain of the first signal;

extract first calculation elements from the first data set;

further calculate, from the second signal, a second data set belonging to a frequency domain of the second signal;

further extract second calculation elements from the second data set; and

determine a health index of said at least one bearing from each of the first and second calculation elements.