NAPS 2026 Papers
We’re pleased to share that two recent papers from PowerWorld contributors have been accepted for presentation at the North American Power Symposium (NAPS) 2026. These works highlight ongoing advances in power flow modeling, including stochastic external input integration and robust forecast application methods.
An Extensible Approach for Modeling the Impact of Deterministic and Stochastic External Inputs in the Power Flow
This paper presents an approach for modeling the impact of weather and other external inputs (EXIs) in the power flow or optimal power flow (OPF) with consideration of both deterministic and stochastic models. The need for this work is due to growing dependence of large-scale grids on the weather and the potential need to model the impact of many other EXIs. The work presents a relatively easy to implement and extensible modeling approach for directly including EXIs in the power flow or OPF by extending the capability of contingency analysis to do Monte Carlo simulations. Results are demonstrated using earthquake scenarios on two synthetic grids with 37 and 993 buses, respectively.
A Robust Method for Applying Forecast States to AC Power Flow Cases
Power system studies frequently require the application of forecast data, such as load or renewable generation projections, to an existing solved power flow case. However, direct application of forecast values may result in convergence failure due to the nonlinear nature of the AC power flow equations. This paper presents a simple and robust algorithm for applying arbitrary forecast data while maintaining solvability. The method parameterizes forecast application as a scalar interpolation between a reference state and a target forecast state and iteratively searches for solvable intermediate operating points using adaptive step reduction. When a solvable intermediate state is identified, it is used as a new reference, allowing the algorithm to progressively advance toward the full forecast condition. The approach requires no modification to the underlying power flow solver and operates as a wrapper around existing tools. The method is evaluated on a 10,000-bus synthetic system across 168 hourly scenarios with varying solar availability and load forecasts. Results demonstrate that the algorithm reliably recovers solvable cases that fail under direct forecast application, with modest computational overhead. The impact of step size parameters on robustness and performance is also analyzed.