PMU data can go missing due to several factors - faulty PMU or PDC or communication channel, cyber attack, etc.
Bad data can be also caused due to some of the above reasons or maybe due to some valid reasons such as faults occurring in the power system.
But be the reason valid or invalid, the signal processing algorithm will consider the data to be missing irrespective of the cause and could treat it as a fault and can opt for control strategies accordingly.
My question is directed at gaining insights about how to identify whether the missing/bad data occurred due to a valid (fault, etc) or invalid reason (cyberattack, device failure, etc) by monitoring any other data from any other device or using any algorithm.
If you can share any references, it will also be quite useful.
Thanks a lot in advance.
Interesting Q
Below my thoughts ..not tested ..
1. Try to define" bad data" such as unaccepatble limit transgressions etc
2.Gather and store data of past events to obtain tolerances at specific nodes
3. Examine inconsistencies in Data with near networks physical data from voltmeters/ ammeters/ wattmeters/ phase meters.
4. Use some sort of code from information theory to determine outliers
5. Bayesian networks
6. ANN
Good luck
Cheers
It may be by suitably extrapolating from exact neighboring element data (bus or line depending upon the system) at that very point of time. It is some sort of estimation to be applied.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Analysis
- Data Analysis
- Dataset
- Big Data