Optimal Power Flow in Distribution Scheme Using Load Forecast

As load forecasting in distribution becomes increasingly important considering the new push for microgrids, distributed generation and system security studies, new methods for handling distribution’s unique challenges are required. Data availability, data quality, and system dynamics make distribution an especially difficult environment to analyze. Neural networks provide a powerful tool to analyze how load is impacted by time and weather. The objective of this study is the application of neural networks to learn a feeder’s load behavior and forecast a 24-hour load profile based on historical weather and SCADA data. Further, this load forecasting is used to perform an optimal power flow calculation for minimizing violations and deviations from normal on a distribution scheme that can be served from one of several sources.