Non-Intrusive Load Monitoring (NILM) increases awareness on user energy usage patterns. In this paper, an efficient and highly accurate NILM method is proposed featuring condensed representation, super-state and fusion of two deep learning based models. Condensed representation helps the two models perform more efficiently and preserve longer-term information, while super-state helps the model to learn correlations between appliances. The first model is a deep user model that learns user appliances usage patterns to predict the next appliance usage behavior based on past behaviors by capturing the dynamics of user behaviors history and appliances usage habits. The second model is a deep appliance group model that learns the characteristics of appliances with temporal and electrical information. These two models are then fused to perform NILM. The case study based on REFIT datasets demonstrates that the proposed NILM method outperforms two state-of-the-art benchmark methods.
Accurate short-term electricity load forecasting (STELF) is essential for grid scheduling and low-carbon smart grids. However, load exhibits multi-timescale periodicity and non-stationary fluctuations, making STELF highly challenging for existing models. To address this challenge, an Autoformer–Transformer residual fusion network (ATRFN) is proposed in this paper. A dynamic weighting mechanism is applied to combine the outputs of Autoformer and Transformer through residual connections. In this way, lightweight result-level fusion is enabled without modifications to either architecture. In experimental validations on real-world load datasets, the proposed ATRFN model achieves notable performance gains over single STELF models. For univariate STELF, the ATRFN model reduces forecasting errors by 11.94% in mean squared error (MSE), 10.51% in mean absolute error (MAE), and 7.99% in mean absolute percentage error (MAPE) compared with the best single model. In multivariate experiments, it further decreases errors by at least 5.22% in MSE, 2.77% in MAE, and 2.85% in MAPE, demonstrating consistent improvements in predictive accuracy.