Work

Load Forecasting via Deep Neural Networks

Year: 2017
Type: conference-paper
Abstract: Nowadays, electricity plays a vital role in national economic and social development. Accurate load forecasting can help power companies to secure electricity supply and scheduling and reduce wastes since electricity is difficult to store. In this paper, we propose a novel Deep Neural Network architecture for short term load forecasting. We integrate multiple types of input features by using appropriate neural network components to process each of them. We use Convolutional Neural Network components to extract rich features from historical load sequence and use Recurrent Components to model the implicit dynamics. In addition, we use Dense layers to transform other types of features. Experimental results on a large data set containing hourly loads of a North China city show the superiority of our method. Moreover, the proposed method is quite flexible and can be applied to other time series prediction tasks.
Author: Wan He
Language: English
Open Access status: diamond
OpenAlex ID: W2774966631
FWCI: 33.76
Cites: 18
Cited by: 316
Related to: 10