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.
- Research Organization:
- Pacific Northwest National Lab. (PNNL), Richland, WA (United States)
- Sponsoring Organization:
- USDOE
- DOE Contract Number:
- AC05-76RL01830
- OSTI ID:
- 1227075
- Report Number(s):
- PNNL-SA-92471
- Resource Relation:
- Conference: Power and Energy Society General Meeting (PES 2013), July 21-25, 2013, Vancouver, BC, Canada, 1-5
- Country of Publication:
- United States
- Language:
- English
Similar Records
Implications of Regional Transmission Organization Design for Renewable Energy Technologies
Combining multi-objective optimization and bayesian model averaging to calibrate forecast ensembles of soil hydraulic models