Table 9 Techniques for identifying l.a.a.s (synthesized from existing literature).

⁹⁵

Network-based long short-term memory

Long short-term memory (LSTM) network-based approach for detecting malicious code in smart meters via CPU or MCU power consumption side channel. Smart meters were tested in real life and showed 92% accuracy.

⁹⁵

Algorithm of lightweight index

A simple algorithm that detects hidden and dangerous tap changes might assist. The approach relies on logic. Bad orders and inaccurate data only affect part of the variables you measure and estimate. This technique relies on branch currents and tap-modifying transformer end-node voltages.

⁹⁶

Bad data detection

Detection can identify cyberattack reactions’ unusual phase changes. The detection method included four branch ratios and terminal injection current indicators. Additionally, discrete index evaluation counted reference values during phase shift selection.