Neural network based short-term electric load forecasting: EMS-integrated and PC-based stand-alone systems
- ABB Systems Control, Santa Clara, CA (United States)
This paper presents the application of a neural network (NN) based short-term electric load forecast model that is being used for the energy control center of the electric utilities. This NN based short-term load forecast program has been developed with a version integrated with the electric utility`s Energy management System (EMS), as well as a PC-based stand-alone version. The model forecasts the hourly electrical load for the current day and up to seven days. A multi-layer neural network is used to provide a non-linear mapping between weather parameters and electric load. Using historical weather parameters and electric load. Using historical weather parameters (such as dry bulb temperature, relative humidity, wind velocity and light intensity), and historical hourly loads, a neural network is trained for each day type and each weather-defined season. The forecast of weather parameters can be obtained by a weather station for the forecast period. The program is capable to generate hourly weather forecast if the forecast form the weather service is partial, such as if only a few hours per day are available, or even if the maximum or minimum daily values of the temperature forecast is available. A separate NN model has also been developed for identifying seasons based on the historical weather data. This paper will discuss features of the system, the neural network models and algorithm, and a sample result of the program performance.
- OSTI ID:
- 376140
- Report Number(s):
- CONF-960426--
- Country of Publication:
- United States
- Language:
- English
Similar Records
Hourly load forecasting by neural networks
An artificial neutral network hourly temperature forecaster with applications in load forecasting