One-Day Building Cooling Load Prediction Based on Bidirectional Recurrent Neural Network
Short-term building cooling load prediction is very important for building energy management tasks. Traditional way relies on physical principles. Due to the nonlinearity of the features of the data, it is a challenge for prediction. This work applies the Bidirectional Recurrent Neural Network (BRNNs) in prediction of 24-h ahead building cooling load profiles. The results show that BRNNs have good performance in prediction on building cooling load prediction. The mode can predict the building cooling load profiles effectively.
KeywordsBuilding cooling load Short-term prediction BRNNs
- 6.Schuster, M., Paliwal, K.K.: Bidirectional recurrent neural networks. Signal Process. 45, 2673–2681 (1997)Google Scholar
- 7.Graves, A., Mohamed, A., et al.: Speech Recognition with Deep Recurrent Neural Networks. https://arxiv.org/abs/1303.5778
- 8.Su, Y., Huang, Y., et al.: On Extended Long Short-term Memory and Dependent Bidirectional Recurrent Neural Network. https://arxiv.org/abs/1803.01686