Deep neural network framework simplifies AC power flow contingency analysis

Researchers have developed a novel framework for AC power flow contingency analysis that utilizes a single deep neural network (DNN) to predict post-contingency operating states. This approach aims to reduce the computational burden associated with traditional methods, which often require extensive outage-specific training data. The proposed method formulates state prediction as a fixed-point iteration and includes semidefinite programming (SDP) formulations to certify the convergence conditions for the DNN. Tests on the IEEE 118-bus system showed that the SDP formulations are effective and the method accurately estimates post-contingency states with few iterations. AI

IMPACT This research could lead to more efficient and accurate grid security assessments by reducing the computational cost of contingency analysis.

RANK_REASON Academic paper detailing a new method for power grid analysis. [lever_c_demoted from research: ic=1 ai=0.7]

Read on arXiv cs.LG →

• AC power flow
• deep neural network
• IEEE 118-bus system
• Md Obaidur Rahman
• semidefinite programming

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →