INTEGRATED, PREDICTIVE VIBRATION ANALYSIS OF ROTATIONAL MACHINE WITHIN ELECTRONICS RACK

• US 20150292982A1
• Filed: 11/11/2014
• Published: 10/15/2015
• Est. Priority Date: 04/15/2014
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: CRUZ, Ethan E., NOTOHARDJONO, Budy D., ZHANG, Jing, CHERNEGA, Gary J., JOURDAN, Udo, TSFASMAN, Arkadiy O.
• Granted Patent: US 9,915,584 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G01H 1/003 of rotating machines G01H1/10 takes precedence
• CPC Class Codes: G01M 13/045 Acoustic or vibration analysis
• CPC Class Codes: G06F 11/0754 by exceeding limits

Abstract

Predictive monitoring of a rotational machine of a cooling apparatus is provided by integrating a predictive vibration analyzer into an electronics rack being cooled by the cooling apparatus. The integrating includes associating at least one sensor of the predictive vibration analyzer with the rotational machine of the cooling apparatus. The analyzer includes a predetermined harmonics table of one or more base operational frequencies and associated rotational speed harmonics for the rotational machine indicative of one o r more rotational faults, and the predictive vibration analyzer automatically evaluates vibration of the rotational machine during operation of the machine by analyzing vibration data therefor, and automatically ascertains, based on the vibration data, and the predetermined harmonics table, whether the rotational machine is predicted to possess a rotational fault of the one or more rotational faults.

Claims

1. A method comprising:

integrating a predictive vibration analyzer into an electronics rack, the integrating comprising associating at least one sensor of the predictive vibration analyzer with a rotational machine of a cooling apparatus to cool one or more components of the electronics rack, to generate vibration data therefor, the predictive vibration analyzer comprising;

a predetermined harmonics table of one or more base frequencies of vibration and associated rotational harmonics for the rotational machine indicative of one or more rotational faults therein; and

evaluating, by the predictive vibration analyzer, vibration of the rotational machine during operation thereof by analyzing the vibration data, and automatically ascertaining based thereon, and based on the predetermined harmonics table, whether the rotational machine is predicted to possess a rotational fault of the one or more rotational faults.

2. The method of claim 1, wherein the at least one sensor of the predictive vibration analyzer comprises at least one triaxial accelerometer, and the integrating comprises coupling the at least one triaxial accelerometer to the rotational machine to facilitate generating the vibration data during operation of the rotational machine.

3. The method of claim 1, wherein the predictive vibration analyzer automatically transforms the vibration data to a plurality of sample frequency signals with associated signal amplitudes, and wherein the automatically ascertaining comprises generating, from the plurality of sample frequency signals, a peak amplitude table of sample frequency signals of the plurality of sample frequency signals with peak amplitude values in at least one of an x, y, or z direction.

4. The method of claim 3, wherein generating the peak amplitude table of sample frequency signals comprises, for a selected frequency signal of the plurality of sample frequency signals, determining whether the selected frequency signal has a peak signal amplitude by comparing the associated signal amplitude of the selected frequency signal to an average signal amplitude for a window of sample frequency signals, of the plurality of sample frequency signals, about the selected frequency signal, and determining whether the associated signal amplitude of the selected frequency signal exceeds the average signal amplitude for the window of sample frequency signals by a set percentage.

5. The method of claim 4, further comprising predefining the set percentage for the rotational machine.

6. The method of claim 4, wherein generating the peak amplitude table further comprises automatically determining, by the predictive vibration analyzer, whether to add the selected frequency signal to the peak amplitude table when its associated signal amplitude exceeds the average signal amplitude for the window by the defined percentage, the automatically determining comprising ascertaining whether the peak amplitude table already contains, within a preset frequency range of the selected frequency signal, an existing peak amplitude value, and if so, substituting for the existing peak amplitude value the selected frequency signal'"'"'s signal amplitude within the peak amplitude table when that signal amplitude is greater than the existing peak amplitude value.

7. The method of claim 6, further comprising repeating the determining whether the selected frequency signal possesses a peak amplitude signal to evaluate the selected frequency signal in each of the x, y, and z directions.

8. The method of claim 3, further comprising referencing, by the predictive vibration analyzer, the predetermined harmonics table and the peak amplitude table in determining for a base frequency of vibration in the predetermined harmonics table a count of a number of harmonic frequencies of the base frequency that are present in the peak amplitude table, and if three or more harmonic frequencies are present for the base frequency within a set frequency range, automatically predicting, by the predictive vibration analyzer, that the rotational machine possesses the rotational fault of the one or more rotational faults.

9. The method of claim 8, further comprising determining for the base frequency of vibration a number of directions that associated harmonic frequencies of the base frequency are present in the peak amplitude table, and automatically predicting, by the predictive vibration analyzer, that the rotational machine possesses the rotational fault where associated harmonic frequencies are present in multiple directions.

10. The method of claim 8, further comprising repeating the determining the count of the number of harmonic frequencies in the peak amplitude table that are present for each base frequency in the harmonics table, and based thereon, predicting whether the rotational machine will possess the rotational fault.

11. The method of claim 8, further comprising, where three or more harmonic frequencies are identified in the peak amplitude table associated with the base frequency, determining whether the base frequency is also present in the peak amplitude table in each of the x, y, and z directions, and if so, automatically indicating by the predictive vibration analyzer that the rotational machine will possess the rotational fault.