Search

Practical application of deep learning based non-intrusive load monitoring (NILM) system requires the deep neural network model to generalize on new unseen data. Existing NILM solutions are not suitable for real-world application due to their poor disaggregation accuracy on new unseen data. In order to address this problem, this paper presents a NILM algorithm that uses data augmentation to generate synthetic data for training deep convolutional neural network models for each target appliance. Proposed data augmentation technique works by combining on and off-durations of a target appliance from various datasets, and forms a unified and comprehensive synthetic aggregate and sub-meter profiles. Apart from proposed algorithm, this paper also proposes an evaluation approach that relies on total predicted energy and ground-truth energy of an appliance to provide detailed insights about total overlapping energy, missing energy and extra energy predicted by the algorithm. Comparison results on our proposed evaluation approach showed that proposed disaggregation algorithm was able to predict energy that was 60% overlapping with ground truth energy and 36% energy was extra. Overall results showed that overlapping energy was 2.5 times more, and extra-predicted energy was 60% less than state-of-the-art algorithms in unseen test cases. [ABSTRACT FROM AUTHOR]

Considering the complementary characteristics of wind and solar energy, expanding an existing wind farm with photovoltaic panels can significantly smooth fluctuation of output power and improve operation economy. This paper proposes a two-stage approach to optimize the wind farm expansion. Based on the National Aeronautics and Space Administration data, modified meteorological models are developed considering the correlation between wind speed and solar irradiation. Taking into account fluctuation of output power, utilization of electrical equipment, and losses of renewable energy, a multi-objective optimization model is established. Two scenarios with different transformer ratings are analyzed to determine whether to expand electrical equipment. The Monte Carlo simulation is utilized to generate meteorological data in the first stage. The Pareto optimal solution set is searched by the multi-objective particle swarm optimization algorithm to determine the final solution in the second stage. A case study was conducted to validate the proposed approach. [ABSTRACT FROM PUBLISHER]

Superconducting magnetic energy storage (SMES) is expected to be utilized in the power grid for dynamic power compensation. However, a SMES status such as magnet current (including parameters of power conditioning system) and temperature of superconducting (SC) magnet will restrict the output capability of SMES. This paper proposes a new method to evaluate SMES status, which will guarantee the dynamic thermal stability of SC magnet applied in power grid. First, it is generally analyzed that the output capability (P\max) of a SMES is restricted by the initial (standby) current (I_0) in the SC magnet for a specific application of SMES. To prove that P_{\max} also depends on the temperature (T_0)$ of the SC magnet, the thermal characteristics of the SC magnet is simulated with a 5 MJ SMES model. Based on these restricted conditions and the simulation results, a SMES status evaluation method (SSEM)is constructed and implemented. Then, with combination of test data gotten from open-loop experiments of a real 150-kJ/100-kW conduction-cooled SMES, the presented SSEM is realized by simulation in MATLAB/Simulink, the simulation results preliminarily verify the feasibility and effectiveness of SSEM. Furthermore, SSEM is applied to dynamic experiments, whereas the 150-kJ/100-kW SMES is set for damping power oscillation after the short-circuit fault in a physical simulated power grid. The proposed SSEM presents a potential way to optimize SMES operation in the power system. [ABSTRACT FROM PUBLISHER]