Description:PLEASE SEE THE ATTACHED SPECIFICATION , Claims:PLEASE SEE THE ATTACHED SPECIFICATION

We claim:
1. A method for detecting faults in a bearing of industrial equipment, the method comprising:
analysing vibration data of the bearing to determine a plurality of vibration
characteristics at a set of bearing frequencies;
calculating a likelihood of fault based on a comparison of the plurality of vibration
characteristics with a set of predefined fault vibration characteristics at the set of bearing
frequencies;
determining a Cumulative SUM (CUSUM) parameter for the bearing based at least on
the likelihood of fault; and
detecting a fault in the bearing based on the CUSUM parameter.
2. The method of claim 1, wherein determining the plurality of vibration characteristics
comprises:
determining the set of bearing frequencies corresponding to the bearing based on a fault
frequency and a resonant fault frequency of the bearing;
estimating frequency characteristics of the vibration data for the set of bearing
frequencies based on a frequency analysis of the vibration data; and
determining the plurality of vibration characteristics based on the frequency
characteristics.

3. The method of claim 1, wherein the set of predefined fault vibration characteristics are
determined based on one or more characteristics of a faulty bearing.
4. The method of claim 1, wherein determining the CUSUM parameter is based on a
maximum value of a sum of a previous CUSUM parameter and the likelihood of fault, and
a zero value, wherein the previous CUSUM parameter indicates a CUSUM parameter of
the vibration data obtained in a preceding time period, and wherein the vibration data is
obtained periodically.
5. The method of claim 1, wherein detecting the fault comprising:
determining a fault threshold based on a rate of incorrect detections of the fault in the
bearing; and detecting the fault in the bearing upon determining that the CUSUM parameter exceeds
the fault threshold.
6. A system for detecting faults in a bearing of industrial equipment, the system comprising:
5 a memory; and
a processor, communicatively coupled with the memory, the processor is configured
to:
analyse vibration data of the bearing to determine a plurality of vibration
characteristics at a set of bearing frequencies;
10 calculate a likelihood of fault based on a comparison of the plurality of vibration
characteristics with a set of predefined fault vibration characteristics at the set of bearing
frequencies;
determine a Cumulative SUM (CUSUM) parameter for the bearing based at least
on the likelihood of fault; and
15 detect a fault in the bearing based on the CUSUM parameter.
7. The system of claim 6, wherein to determine the plurality of vibration characteristics, the
processor is configured to:
determine the set of bearing frequencies corresponding to the bearing based on a fault
20 frequency and a resonant fault frequency of the bearing;
estimate frequency characteristics of the vibration data for the set of bearing frequencies
based on a frequency analysis of the vibration data; and
determine the plurality of vibration characteristics based on the frequency
characteristics.
25
8. The system of claim 6, wherein the processor is configured to determine the set of
predefined fault vibration characteristics based on one or more characteristics of a faulty
bearing.
30 9. The system of claim 6, wherein the processor is configured to determine the CUSUM
parameter based on a maximum value of a sum of a previous CUSUM parameter and the
likelihood of fault, and a zero value, wherein the previous CUSUM parameter indicates a
CUSUM parameter of the vibration data obtained in a preceding time period, and wherein
the vibration data is obtained periodically

10. The system of claim 6, wherein to detect the fault, the processor is configured to:
determine a fault threshold based on a rate of incorrect detections of the fault in the
bearing; and
5 detect the fault in the bearing upon determining that the CUSUM parameter exceeds
the fault threshold.