Methods and apparatus for detection of process parameter anomalies

1. A system for identifying anomalies in actual time series data from multiple sources, comprising: a processing system, implemented via one or more computer systems, configured to receive the actual time series data, wherein the processing system comprises a neural network generating a nominal predictive time series dataset according to the actual time series data, and wherein the processing system: compares the actual time series data to the nominal predictive time series dataset; and identifies an outlier in the actual time series data compared to the nominal predictive time series dataset; and an output responsive to the processing system and providing information relating to the identified outlier.

2. A system for identifying anomalies according to claim 1 , wherein the neural network comprises a recurrent auto-encoding neural network.

3. A system for identifying anomalies according to claim 1 , wherein:

the neural network groups actual time series data from a plurality of sources to generate a multivariate nominal predictive time series dataset; and

the processing system groups actual time series data from the plurality of sources to create a multivariate actual time series dataset, compares the multivariate actual time series dataset to the multivariate nominal predictive time series dataset, and identifies the outlier in the multivariate actual time series dataset compared to the multivariate nominal predictive time series dataset.

4. A system for identifying anomalies according to claim 3 , wherein the processing system identifies primary clusters of correlated multivariate actual time series datasets.

5. A system for identifying anomalies according to claim 4 , wherein the processing system identifies secondary clusters comprising primary clusters correlating at least one of temporally and spatially.

6. A system for identifying anomalies according to claim 1 , wherein the processing system selects the plurality of sources according to a similarity matrix populated with historical actual time series data.

7. A system for identifying anomalies according to claim 1 , wherein an architecture of the neural network is dynamically alterable at a training time according to at least one of a size and a shape of a set of training data.

8. A system for identifying anomalies according to claim 1 , wherein the processing system identifies the outlier according to an outlier detection parameter, and wherein the processing system modifies the outlier detection parameter according to a user feedback input.

9. A system for identifying anomalies according to claim 1 , wherein the processing system further comprises a supervised learning machine model, and wherein the supervised learning machine model assigns a validity probability score to the outlier.

10. A system for identifying anomalies according to claim 1 , wherein the processing system groups multiple outliers using a similarity matrix.

11. A system for identifying anomalies according to claim 1 , wherein the processing system identifies the outlier in conjunction with a sensitivity, and wherein the processing system adjusts the sensitivity in response to an invalidly identified outlier.

12. A system for identifying anomalies according to claim 1 , wherein the processing system identifies the outlier in conjunction with a sensitivity, and wherein the processing system automatically increases the sensitivity over a time period.

13. A system for identifying anomalies in multiple streams of actual time series data from multiple power generation sources, comprising: a recurrent auto-encoding neural network configured to receive the actual time series data, wherein the recurrent auto-encoding neural network generates a multivariate nominal predictive dataset based on a group of the streams of the actual time series data from the multiple power generation sources; a processing system, implemented via one or more computer systems, responsive to the recurrent auto-encoding neural network, wherein the processing system: compares the multivariate nominal predictive dataset data sot to an actual multivariate dataset based on the group of the streams of the actual time series data from the multiple power generation sources to generate a delta dataset; and identifies an outlier in the delta dataset; and an output responsive to the processing system and providing information relating to the identified outlier.

14. A system for identifying anomalies according to claim 13 , wherein the processing system identifies primary clusters of the actual multivariate datasets.

15. A system for identifying anomalies according to claim 14 , wherein the processing system identifies secondary clusters comprising primary clusters correlating at least one of temporally and spatially.

16. A system for identifying anomalies according to claim 13 , wherein the processing system selects the multiple power generation sources according to a similarity matrix populated with historical actual time series data.

17. A system for identifying anomalies according to claim 13 , wherein an architecture of the recurrent auto-encoding neural network is dynamically alterable at a training time according to at least one of a size and a shape of a set of training data.

18. A system for identifying anomalies according to claim 13 , wherein the processing system identifies the outlier according to an outlier detection parameter, and wherein the processing system modifies the outlier detection parameter according to a user feedback input.

19. A system for identifying anomalies according to claim 13 , wherein the processing system further comprises a supervised learning machine model, and wherein the supervised learning machine model assigns a validity probability score to the outlier.

20. A system for identifying anomalies according to claim 13 , wherein the processing system groups multiple outliers using a similarity matrix.

21. A system for identifying anomalies according to claim 13 , wherein the processing system identifies the outlier in conjunction with a sensitivity, and wherein the processing system adjusts the sensitivity in response to an invalidly identified outlier.

22. A system for identifying anomalies according to claim 13 , wherein the processing system identifies the outlier in conjunction with a sensitivity, and wherein the processing system automatically increases the sensitivity over a time period.

23. A method for identifying anomalies in multiple streams of actual time series data from multiple power generation sources, comprising:

generating a multivariate nominal predictive dataset based on a group of the streams of the actual time series data from multiple power generation sources;

generating a delta dataset according to a comparison of the multivariate nominal predictive dataset to an actual multivariate dataset based on the group of the streams of the actual time series data from the multiple power generation sources;

identifying an outlier in the delta dataset; and

providing information relating to the identified outlier at an output.

24. A method for identifying anomalies according to claim 23 , wherein generating the multivariate nominal predictive dataset comprises using a recurrent auto-encoding neural network to generate the multivariate nominal predictive dataset.

25. A method for identifying anomalies according to claim 23 , further comprising identifying primary clusters of the actual multivariate datasets.

26. A method for identifying anomalies according to claim 25 , further comprising identifying secondary clusters comprising primary clusters correlating at least one of temporally and spatially.

27. A method for identifying anomalies according to claim 23 , further comprising selecting the multiple power generation sources according to a similarity matrix populated with historical actual time series data.

28. A method for identifying anomalies according to claim 23 , wherein generating the multivariate nominal predictive dataset comprises using a neural network, and further comprising dynamically altering an architecture of the neural network at a training time according to at least one of a size and a shape of a set of training data.

29. A method for identifying anomalies according to claim 23 , further comprising identifying the outlier according to an outlier detection parameter; and modifying the outlier detection parameter according to a user feedback input.

30. A method for identifying anomalies according to claim 23 , further comprising assigning a validity probability score to the outlier.

31. A method for identifying anomalies according to claim 23 , further comprising grouping multiple outliers using a similarity matrix.

32. A method for identifying anomalies according to claim 23 , further comprising identifying the outlier in conjunction with a sensitivity; and

adjusting the sensitivity in response to an invalidly identified outlier.

33. A method for identifying anomalies according to claim 23 , further comprising identifying the outlier in conjunction with a sensitivity; and

automatically increasing the sensitivity over a time period.