Generated and Validated Invariants of EPIC Smart Grid

Citation Author(s):
Danish
Hudani
Muhammad
Haseeb
Muhammad
Taufiq
Muhammad Azmi
Umer
Nandha Kumar
Kandasamy
Submitted by:
Muhammad Azmi Umer
Last updated:
Tue, 08/23/2022 - 02:54
DOI:
10.21227/95n2-y902
License:
Creative Commons Attribution
271 Views
Categories:
Machine Learning
Smart Grid
Security
Keywords:
Machine Learning, Anomaly Detection, Cyber-Physical Attack, Cyber-Physical System, Critical Infrastructure, Industrial Control System, Association rule mining, Electrical Power Plant
0
0 ratings - Please login to submit your rating.
ACCESS DATASET CITE SHARE/EMBED

Abstract 

The dataset provided contains the invariants in the form of antecedent and consequent mined using Association Rule Mining with implicit measures such as Confidence, Support, Lift, etc. from an EPIC Plant (Electrical Power and Intelligent Control System).

Instructions: 

The invariants are generated from four different scenarios:

Scenario_01.xlsx: Synchronising two generators without load andangle difference from -180 to 0 and 0 to 180 degrees. The data wasgenerated on Oct 19 2018 from 02:22 PM to 02:54 PM, having 512data points and 292 attributes.

Scenario_02.xlsx: Synchronising two generators with a PV systemand resistive load of 7KW for 20 minutes. The data was generatedon Oct 19 2018 from 03:45 PM to 04:05 PM, having 797 data pointsand 292 attributes.

Scenario_03.xlsx: Synchronising three generators with a resistiveload of 4KW for around 20 minutes. The data was generated on Oct19 2018 from 04:06 PM to 04:28 PM, having 863 data points and 292attributes.
Scenario_04.xlsx: Continuous supply of power to SWaT andWADI testbeds of water treatment and distribution plants respectively. The data was generated on Nov 07 2018 from 02:57 PMto 03:21 PM, having 695 data points and 391 attributes.

The dataset also contains three sample files of the invariants generated from the above four scenarios randomly named Sample_01.xlsx, Sample_02.xlsx, and Sample_03.xlsx. The invariants from these files have been labeled as either global or local. The local invariants are those that might not apply in all scenarios, whereas the global invariants are those that control how the EPIC plant behaves regardless of the scenario in which the plant is functioning. One of the global invariants is given below, which defines that the current value in the component VSD 1 (Variable Speed Driver) will be 1 then the status of Generation Q1 will be True.

VSD1.Current=1=> Generation.Q1 1.STATUS CLOSE=True

Dataset Files

LOGIN TO ACCESS DATASET FILES

 

QUESTIONS?