OPF$k$NN: Congestion Aware Real-Time Optimal Power Flow With Uncertain Solar Generation

With the growing scale of integration of solar power in the grid, the supply side uncertainties in the grid are increasing. The increase in supply side uncertainties can result in a large number of future states that need to be analyzed for secure and optimal grid operation within a short window of time. The uncertainties in solar power in the future can lead to grid congestion. The grid may force to remain congested until the grid operator solve the optimal power flow (OPF) to decide and implement nonrenewable generation dispatch to avoid congestion in response to the fluctuation in solar power. This requires an early and real-time estimation of congestion and then a real-time optimal power flow (RTOPF) to decide optimal dispatch decisions to avoid congestion. Therefore in this article, we propose a novel two-stage approach called OPF <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$k$</tex-math></inline-formula> nearest neighborhood (OPF<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$k$</tex-math></inline-formula>NN). In the first stage, we propose a machine learning (ML) binary classification problem and optimized ML model to estimate the congested/noncongested state of the grid directly from the generation data. The second stage is a neighborhood-based search approach for RTOPF to decide the optimal generation dispatch of the nonrenewable generation to avoid congestion. The proposed OPF<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$k$</tex-math></inline-formula>NN has been implemented on the IEEE 118 bus system and Polish 2383 bus system to demonstrate the effectiveness of our proposed approach.