Chi Zhang

Mr. Chi Zhang


Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
14 October 2026 in Machines
Verified
Journal Article
Open access
PDF file
CZChi Zhang
Chi Zhang
ZSZhengyi Sun
Zhengyi Sun
XWXin Wang
Xin Wang
HYHaitao Yan
Haitao Yan
CZCong Zhu
Cong Zhu
RNRisha Na
Risha Na
Fast and selective faulted-section identification is challenging in multiterminal medium-voltage DC (MVDC) ring networks because the initial fault transient is affected by converter dynamics, fault resistance, line parameters, and multiple feeding paths. This paper proposes a local identification method for pole-to-pole faults in a four-terminal ±10 kV modular multilevel converter (MMC)-based MVDC ring network, using hybrid DC circuit-breaker (HCB) commutation characteristics. The breaker-current magnitude is used to initiate hybrid DC circuit breaker operation. The identification feature is the magnitude of the mechanical-switch (MS) current slope estimated during the passive intervals of the clamped commutation stage. This feature is used to classify faults as in-zone or out-of-zone. An in-zone fault is confirmed when three consecutive valid slope estimates satisfy the threshold criterion. Because only local HCB measurements and control signals are required for online identification, the method does not rely on interterminal communication or time synchronization. Offline threshold setting, however, requires the network topology and component parameters. After classification, the HCBs bounding the faulted line continue the interruption process, whereas triggered out-of-zone HCBs maintain semiconductor conduction and subsequently reclose their MSs. PSCAD/EMTDC simulations on a four-terminal ±10 kV MVDC ring network demonstrate correct faulted-section identification at the end of a 200 μs decision interval measured from local HCB triggering, using a baseline protection sampling interval of 1 μs, over the tested fault locations and fault resistance. Additional sampling-interval tests show no observed identification failures at 1 and 2 μs for the evaluated cases, whereas failures occur at 5 μs. Additional bounded measurement-noise tests show correct discrimination for the benchmark fault at the tested nominal noise levels up to 1% for the selected noise realizations. These results demonstrate the feasibility for the studied configuration; broader parameter variations and practical measurement and execution constraints require further validation.