Abstract: A method for automatically diagnosing a part of a rotating machine based on a time signal generated by the rotating machine, includes constructing a diagram from the signal, including the following procedures: splitting the signal into a plurality of sub-signals; for each sub-signal, calculating the Fourier transform of the sub-signal to obtain a vibratory energy per frequency; constructing the diagram, the diagram being a matrix having rows each corresponding to a speed of rotation of the rotating machine, and columns each corresponding to a frequency of the Fourier transform divided by a speed of rotation of the rotating machine, the matrix including, for each row and each column, the corresponding vibratory energy; supervised training of an artificial neural network to provide, from a diagram, an operating class included in a set of operating classes including at least one nominal operating class and one defective operating class, using the trained network.

Abstract: A method for monitoring an epicyclic gear train of an aircraft includes the following steps: acquiring, at a predetermined sampling frequency, first values (5(ti)) of a signal formed by a progressive mechanical wave generated in the epicyclic gear train; measuring, at a plurality of successive instants, values (Vmes _r(tj)) of a speed of rotation of at least one of the toothed wheels of the gear train; calculating values (Vc(tj)) of a speed of rotation of a point of contact between two toothed wheels of the epicyclic gear train; determining second values (S(ç½)) of the signal formed by a progressive mechanical wave generated in the epicyclic gear train, the second values being sampled depending on a phase of the point of contact and forming secondary mechanical wave data; and using the secondary mechanical wave data (S(ç½)) to detect an anomaly related to the operation of the epicyclic gear train.

Abstract: A method for monitoring an epicyclic gear train of an aircraft includes the following steps: acquiring, at a predetermined sampling frequency, first values (5(ti)) of a signal formed by a progressive mechanical wave generated in the epicyclic gear train; measuring, at a plurality of successive instants, values (Vmes_r(tj)) of a speed of rotation of at least one of the toothed wheels of the gear train; calculating values (Vc(tj)) of a speed of rotation of a point of contact between two toothed wheels of the epicyclic gear train; determining second values (S(ç½)) of the signal formed by a progressive mechanical wave generated in the epicyclic gear train, the second values being sampled depending on a phase of the point of contact and forming secondary mechanical wave data; and using the secondary mechanical wave data (S(ç½)) to detect an anomaly related to the operation of the epicyclic gear train.

Abstract: A method of analysis of the state of operation of a machine including a learning step supplementing a reference database with one or more thresholds for one or more indicators calculated on the basis of signals delivered by a sensor associated with the machine, the learning step including the following operations implemented by a computer processing unit; an acquisition of signals characteristic of normal operation and of abnormal operation of the machine; of each of the signals characteristic of normal operation, formation of at least one so-called deviation signal by implementing a mathematical operation having as attributes the signal characteristic of normal operation and one of the signals characteristic of normal or abnormal operation other than the signal characteristic of the normal operation; for each of the deviation signals, calculation of an indicator; determination of an indicator threshold representative of a limit between normal operation and abnormal operation of the machine.

Abstract: A method of analysis of the state of operation of a machine including a learning step supplementing a reference database with one or more thresholds for one or more indicators calculated on the basis of signals delivered by a sensor associated with the machine, the learning step including the following operations implemented by a computer processing unit; an acquisition of signals characteristic of normal operation and of abnormal operation of the machine: of each of the signals characteristic of normal operation, formation of at least one so-called deviation signal by implementing a mathematical operation having as attributes the signal characteristic of normal operation and one of the signals characteristic of normal or abnormal operation other than the signal characteristic of the normal operation; for each of the deviation signals, calculation of an indicator; determination of an indicator threshold representative of a limit between normal operation and abnormal operation of the machine.

Abstract: A method of monitoring twisting vibration of a rotary shaft of a turbine engine includes obtaining a vibratory acceleration signal from a sensor located on a stationary component of the turbine engine. This vibratory signal is characterized by a carrier frequency. The method also includes evaluating a frequency spectrum of the vibratory signal and searching for a pair of spectral lines of amplitudes greater than at least a first threshold. The lines are distributed in the spectrum on either side of the carrier frequency of the vibratory signal, and spaced apart therefrom by a twisting frequency of the shaft. Where appropriate, the method includes issuing a warning message.

Abstract: A method for detecting resonance in a rotor shaft of a turbine engine, in which a resonance frequency (Fr) of the rotor shaft belonging to a resonance band is centered on an estimated resonance frequency (Frest), is disclosed. The method including: measuring a parameter (P) for regulating the fuel of the turbine engine is measured to obtain an excitation signal; forming a spectrum of the excitation signal; comparing an amplitude of the excitation signal is compared with a resonance threshold (S) on the resonance band of the spectrum; and emitting a resonance detection alarm if the resonance threshold (S) is exceeded.

Abstract: A method for forming a global spectrum (Sg) of an analogue signal (A) to be digitized, in which: the analogue signal (A) is sampled with a first analogue-digital converter (21, 22) of determined Shannon frequency (Fs) so as to obtain an aliasing spectrum (Sr), the analogue signal (A) not having been previously filtered by an anti-aliasing filter; a base spectrum (Sb) is subtracted from the aliasing spectrum (Sr) so as to obtain an aliased spectrum (Sre), the base spectrum (Sb) corresponding to an aliasing-free spectrum of the said analogue signal; a dealiased spectrum (Sd) is computed on the basis of the aliased spectrum (Sre) as a function of the Shannon frequency (Fs) of the converter (21, 22); the base spectrum (Sb) is concatenated with the dealiased spectrum (Sde) so as to form the global spectrum (Sg) of the analogue signal (A).

Abstract: A method of monitoring twisting vibration of a rotary shaft of a turbine engine includes obtaining a vibratory acceleration signal from a sensor located on a stationary component of the turbine engine. This vibratory signal is characterized by a carrier frequency. The method also includes evaluating a frequency spectrum of the vibratory signal and searching for a pair of spectral lines of amplitudes greater than at least a first threshold. The lines are distributed in the spectrum on either side of the carrier frequency of the vibratory signal, and spaced apart therefrom by a twisting frequency of the shaft. Where appropriate, the method includes issuing a warning message.

Abstract: A method for detecting resonance in a rotor shaft of a turbine engine, the resonance frequency (Fr) of the said rotor shaft belonging to a resonance band centred on an estimated resonance frequency (Frest), a method in which: a parameter (P) for regulating the fuel of the turbine engine is measured to obtain an excitation signal; a spectrum of the excitation signal is formed; the amplitude of the excitation signal is compared with a resonance threshold (S) on the resonance band of the said spectrum; and a resonance detection alarm is emitted if the resonance threshold (S) is exceeded.

Abstract: A method for forming a global spectrum (Sg) of an analogue signal (A) to be digitized, in which: the analogue signal (A) is sampled with a first analogue-digital converter (21, 22) of determined Shannon frequency (Fs) so as to obtain an aliasing spectrum (Sr), the analogue signal (A) not having been previously filtered by an anti-aliasing filter; a base spectrum (Sb) is subtracted from the aliasing spectrum (Sr) so as to obtain an aliased spectrum (Sre), the base spectrum (Sb) corresponding to an aliasing-free spectrum of the said analogue signal; a dealiased spectrum (Sd) is computed on the basis of the aliased spectrum (Sre) as a function of the Shannon frequency (Fs) of the converter (21, 22); the base spectrum (Sb) is concatenated with the dealiased spectrum (Sde) so as to form the global spectrum (Sg) of the analogue signal (A).