2018 · 62 citations · 20 references
ComputingComputer ScienceEnergy ManagementEnergy SystemsMachine LearningBuilding AutomationBuilding Energy SystemsDeep LearningDemand-side ManagementEnergy ConservationEnergy UseDisaggregationEnd-use Energy DisaggregationEnergy Consumption MonitoringEnergy MonitoringLoad MonitoringMean Absolute ErrorReal-time MonitoringRecurrent Neural Networks
Residential and commercial buildings consume more electricity than any other sector. Thus, energy saving through smart electrification in those buildings is the best way to reduce overall energy demand. Smart electrification includes end-use appliance energy consumption monitoring in real-time, which can be achieved through Non-Intrusive Load Monitoring (NILM). Recently, deep learning algorithms have been introduced for energy disaggregation but their effectiveness in real-time appliance load monitoring is questionable. In this paper, we have presented two deep recurrent neural networks models: LSTM and GRU. We have introduced regularization to improve our proposed model's performance and have tested them on unseen buildings during training. We have achieved promising results with proposed Regularized LSTM model in terms of accuracy, f1 score and mean absolute error. These results suggest using this model in real-time energy disaggregation of end-use appliances.
20
Sepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 98.1K citations
Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
Jun‐Young Chung, Çağlar Gülçehre, Kyunghyun Cho et al. · arXiv (Cornell University) · 2014 · 10.8K citations · Full text
Klaus Greff, Rupesh K. Srivastava, Jan Koutník et al. · IEEE Transactions on Neural Networks and Learning Systems · 2016 · 6.8K citations · Full text