skip to main content

Title: An Ensemble Approach for Forecasting Net Interchange Schedule

The net interchange schedule (NIS) is the sum of the transactions (MW) between an ISO/RTO and its neighbors. Effective forecasting of the submitted NIS can improve grid operation efficiency. This paper applies a Bayesian model averaging (BMA) technique to forecast submitted NIS. As an ensemble approach, the BMA method aggregates different forecasting models in order to improve forecasting accuracy and consistency. In this study, the BMA method is compared to two alternative approaches: a stepwise regression method and an artificial neural network (ANN) trained for NIS forecasting. In our comparative analysis, we use field measurement data from the Pennsylvania, New Jersey, and Maryland (PJM) Regional Transmission Organization (RTO) to train and test each method. Our preliminary results indicate that ensemble-based methods can provide more accurate and consistent NIS forecasts in comparison to non-ensemble alternate methods.
; ; ; ; ;
Publication Date:
OSTI Identifier:
Report Number(s):
DOE Contract Number:
Resource Type:
Resource Relation:
Conference: Power and Energy Society General Meeting (PES 2013), July 21-25, 2013, Vancouver, BC, Canada, 1-5
IEEE, Piscataway, NJ, United States(US).
Research Org:
Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
Sponsoring Org:
Country of Publication:
United States