Method for predicting a rotation fault in the rotor of a vacuum pump, and associated pumping device
The invention pertains to a method for predicting a failure in the rotation of the rotor of a vacuum pump, comprising the following steps: sequences of events related to the change over time of the vacuum pump functional signals are recorded ( 101 ), a match is sought between at least one sequence of events and at least one pre-established association rule precursory pattern of a vacuum pump behavior model within the recorded sequences of events, said pre-established association rule's precursory patterns involving a failure in the rotor rotation ( 102 ), and a time prediction window is deduced during which a failure in the rotor rotation will occur in a vacuum pump ( 103 ). The invention also pertains to a pumping device comprising: a vacuum pump ( 7 ) comprising at least one rotor and one pump body, said rotor having the potential to be driven rotationally within said pump body by a motor of said pump ( 7 ), a functional signal sensor ( 9 ) of said pump ( 7 ), and a means of predicting ( 10 ) a time prediction window during which a failure in the rotor rotation will occur in the vacuum pump ( 7 ), said means for prediction ( 10 ) calculating the predictive time window based on measurements provided by said functional signal sensor ( 9 ).
1. A method for predicting a failure in the rotation of the rotor of a vacuum pump, comprising the following steps performed by a system including a processor:
recording sequences of events related to the change over time of functional signals of the vacuum pump,
seeking a match between at least one sequence of events and at least one precursory pattern of pre-established association rule of a vacuum pump behavior model within the recorded sequences of events, said association rules being established by extracting knowledge, based on data mining, with the restriction of one or more extraction parameters and said precursory patterns of pre-established association rule involving a failure in the rotor rotation, and
deducing a time prediction window during which a failure in the rotor rotation will occur in a vacuum pump, and
performing maintenance on said vacuum pump at a time based on said prediction window,
wherein the functional signals may be transformed into a frequency spectrum, and frequency bands may be selected around the frequencies characteristic of the vacuum pump kinematics within said spectrum,
wherein association rules are established describing the vacuum pump behavior model by extracting knowledge from a learning database comprising a plurality of sequences of events, obtained from a set of vacuum pumps over the vacuum pumps life span running from startup to failure of the rotor rotation, and
wherein from said learning database, a sequence of events is determined, on the one hand by the selection from among said frequency bands of a reference level and intermediate operating levels corresponding to a multiple of the reference level, and characteristics of the vacuum pump operation, and on the other hand, by a duration characteristic of said levels.
2. A method according to claim 1 , wherein said sequences of events relate to the change over time of vibration signals in the vacuum pump.
3. A method according to claim 1 , wherein said sequences of events relate to the change over time of vacuum pump motor current signals.
4. A method according to claim 1 , wherein said association rules are extracted with the restriction of one or more extraction parameters chosen from among: the support, confidence, and maximum duration between each event of the association rule.
5. A method according to claim 1 , wherein one or more association rules are extracted from N−1 sequences of events from the learning database, recorded from N vacuum pumps, and said association rules are verified against the sequence of events which was not used for extraction.