Assessment of power system equipment for equipment maintenance and/or risk mitigation
View Patent ↗Among other things, one or more techniques and/or systems are provided for assessing power system equipment of a power system. Historical sensor data and/or historical field test data collected from the power system equipment may be utilized to develop a health profile of the power system equipment. The health profile is indicative of a predicted health (e.g., or probability of failure) of the power system equipment. In one embodiment, the health profile further comprises a health index score which evaluates that health of the power system equipment in terms of the importance of the power system equipment to the power system. Using the health profiles developed for a plurality of power system equipment, a maintenance strategy for at least a portion of the power system may be developed.
1. A method for assessing power system equipment to predict a health of the power system equipment, comprising:
acquiring historical sensor data yielded from one or more sensors operably coupled to the power system equipment, the historical sensor data indicative of one or more sensor measurements;
acquiring historical field test data yielded from one or more field tests performed on the power system equipment, the historical field test data indicative of one or more field test measurements;
merging the historical sensor data for the power system equipment and the historical field test data for the power system equipment to form a merged data set, the merging comprising:
resolving a conflict between the historical sensor data for the power system equipment and the historical field test data for the power system equipment based upon a conflict resolution parameter that gives preference to first data for the power system equipment representing a measurement of a characteristic that more closely resembles temporally proximate measurements of that characteristic than second data for the power system equipment;
using the merged data set to generate a health profile of the power system equipment, the health profile indicative of a predicted health of the power system equipment; and
replacing at least part of the power system equipment based upon the predicted health indicating a likelihood of failure of the power system equipment.
2. The method of claim 1 , the power system equipment comprising a power transformer.
3. The method of claim 1 , comprising:
evaluating the historical sensor data and the historical field test data to identify an anomaly corresponding a first measurement that deviates from a set of measurements beyond a specified threshold; and
excluding the anomaly from the merged data set.
4. The method of claim 1 , the merging comprising:
resolving a second conflict between the historical sensor data and the historical field test data in favor of the historical field test data.
5. The method of claim 1 , the merging comprising:
temporally correlating the historical sensor data with the historical field test data.
6. The method of claim 1 , the merging comprising:
identifying historical sensor data to exclude from the merged data set prior to the merging.
7. The method of claim 1 , comprising:
comparing the health profile to new sensor data and new field test data to determine if a current state of the power system equipment deviates from the health profile by more than a specified threshold.
8. The method of claim 7 , comprising:
responsive to determining that the current state of the power system equipment deviates from the health profile by more than the specified threshold, issuing a maintenance request to perform maintenance associated with the power system equipment.
9. The method of claim 1 , wherein the health profile comprises a health index score for the power system equipment, the health index score derived based upon an importance of the power system equipment in a power system and a risk of failure of the power system equipment.
10. The method of claim 9 , comprising:
developing a maintenance strategy for maintaining the power system, wherein the maintenance strategy ranks the power system equipment relative to second power system equipment according to the health index score of the power system equipment and a second health index score of the second power system equipment.
11. The method of claim 1 , the using the merged data set comprising:
developing an algorithm for generating the profile based upon the historical sensor data and the historical field test data.
12. The method of claim 1 , the using the merged data set comprising:
analyzing the historical sensor data and the historical field test data using:
clustering analysis;
classification analysis;
association analysis;
prediction analysis; and
anomaly detection analysis.
13. A system for predicting a health of power system equipment, comprising:
a historian data repository configured to receive, over time, sensor data yielded from one or more sensors operably coupled to the power system equipment to create historical sensor data and configured to receive, over time, field test data yielded from one or more field tests performed on the power system equipment to create historical field test data;
a machine learning component configured to:
generate a health profile of the power system equipment using the historical sensor data and the historical field test data, the health profile indicative of a predicted health of the power system equipment;
determine an importance of the power system equipment in a power system based upon at least one of a number of consumers that are supplied electricity from the power system equipment or a level of redundancy of the power system equipment in the power system; and
assign a health index score to the power system equipment based upon the importance of the power system equipment in the power system in relation to a risk of failure of the power system equipment; and
a maintenance component configured to replace at least part of the power system equipment based upon the health index score.
14. The system of claim 13 , the machine learning component configured to:
compare the health profile to new sensor data and new field test data to determine if a current state of the power system equipment deviates from the health profile by more than a specified threshold.
15. The system of claim 13 , the machine learning component configured to determine the importance of the power system equipment based upon the number of consumers that are supplied electricity from the power system equipment.
16. The system of claim 13 , the maintenance component configured to develop a maintenance strategy for maintaining the power system, wherein the maintenance strategy ranks the power system equipment relative to second power system equipment according to the health index score of the power system equipment and a second health index score of the second power system equipment.
17. The system of claim 13 , comprising a data pre-processing component configured to form a merged data set by temporally correlating the historical sensor data and the historical field test data.
18. The system of claim 17 , the data pre-processing component configured to identify an anomaly and to exclude the anomaly from the merged data set.
19. The system of claim 13 , the machine learning component configured to determine the importance of the power system equipment based upon the level of redundancy of the power system equipment in the power system.
20. A non-transitory computer readable medium comprising processor executable instructions that when executed perform a method for assessing power system equipment to predict a health of the power system equipment, the method comprising:
acquiring historical sensor data yielded from one or more sensors operably coupled to the power system equipment, the historical sensor data indicative of one or more sensor measurements;
acquiring historical field test data yielded from one or more field tests performed on the power system equipment, the historical field test data indicative of one or more field test measurements;
merging the historical sensor data for the power system equipment and the historical field test data for the power system equipment to form a merged data set, the merging comprising:
resolving a conflict between the historical sensor data for the power system equipment and the historical field test data for the power system equipment based upon a conflict resolution parameter that gives preference to first data for the power system equipment representing a measurement of a characteristic that more closely resembles temporally proximate measurements of that characteristic than second data for the power system equipment;
determining an importance of the power system equipment in a power system based upon at least one of a number of consumers that are supplied electricity from the power system equipment or a level of redundancy of the power system equipment in the power system;
assigning a health index score to the power system equipment based upon the importance of the power system equipment in the power system in relation to a risk of failure of the power system equipment;
using the merged data set and the health index score to generate a health profile of the power system equipment, the health profile indicative of a predicted health of the power system equipment; and
replacing at least part of the power system equipment based upon the predicted health indicating a likelihood of failure of the power system equipment.