Spectral Decomposition-Based Mitigation of Radiative Side-Channel Noise Intrusion in a Cascaded SST

Abstract

Solid-state transformers (SSTs) are an emerging class of converters that have shown potential to replace traditional low-frequency transformers (LFTs). With cascading capabilities similar to modular multilevel converters, SSTs can efficiently handle higher voltage and power levels. A cascaded SST setup usually uses a decentralized hierarchical control architecture, with primary controllers overseeing operations at the module level and secondary controllers keeping an eye on operations at the aggregated phase level. Despite their potential to replace LFTs, reliability concerns hinder their widespread adoption, given their vulnerability to cyberattacks. Nonintrusive attacks, such as side-channel noise intrusion (SNI), threaten the SSTs' stability by corrupting the measurement feedback signals fed to the controllers. This letter proposes a novel technique for rapid intrusion detection and mitigation of radiative SNI threats to provide resilience to the secondary layer of the cascaded SSTs. The technique entails spectral decomposition of the noise-tampered signal, followed by noise-free signal reconstruction of the targeted frequency component. Experimental results validated the solution deployed in a cascaded SST experimental prototype's secondary control layer to deal with radiative noise on global input voltage sensor feedback.