Nonintrusive load monitoring (NILM) identifies the status and power consumption of various household appliances by disaggregating the total power usage signal of an entire house. Efficient and accurate load monitoring facilitates user profile establishment, intelligent household energy management, and peak load shifting. This is beneficial for both the end users and utilities by improving the overall efficiency of a power distribution network. Existing approaches mainly focus on developing an individual model for each appliance. Those approaches typically rely on a large amount of household-labeled data that are hard to collect. In this article, we propose a multi-appliance-task framework with a training-efficient sample augmentation (SA) scheme that boosts the disaggregation performance with limited labeled data. For each appliance, we develop a shared-hierarchical split structure for its regression and classification tasks. In addition, we also propose a 2-D attention mechanism in order to capture spatio-temporal correlations among all appliances. With only one-day training data and limited appliance operation profiles, the proposed SA algorithm can achieve comparable test performance to the case of training with the full dataset. Finally, simulation results show that our proposed approach features a significantly improved performance over many baseline models. The relative errors can be reduced by more than 50% on average.

The increasing penetration of grid-forming inverters has significantly complicated the operating characteristics of modern power grids, which calls for reduced-order models that preserves the system's main responses for large-scale analysis. This paper presents a model reduction approach to study the dynamic responses of inverter-dominated networked microgrids (MGs) under disturbance. Coherency-based aggregation technique is used to identify the study area and external area. The study area is modeled in detail while the external area is successively simplified. Specifically, the reduced-order model of the external area is first simplified by developing reduced-order models of inverter and network; it is then further simplified using linear truncation. To replicate the interactions between the two areas, the external area is aggregated into a controllable voltage source whose operating states are collaboratively determined by the states from both areas. The developed reduced-order model represents a closed-loop simulation that combines both linear model reduction for reduced system dimensions and detailed nonlinear model for better accuracy. At last, the performance of the developed model reduction approach is validated using time-domain simulation.

Photovoltaic (PV) systems play a critical role in renewable energy resource grid integration, and levelized cost of energy (LCOE) is commonly used to evaluate PV system feasibility in modern power grids. In this work, a revised PV system LCOE calculation model is derived to quantify the potential of LCOE reduction. Particularly, the grid support functions are valuated to offset the investment and operation costs of PV systems, which thereby reduces the LCOE. Meanwhile, PV system service lifetime is also estimated with the derived PV inverter reliability model, considering the critical components (i.e., semiconductor devices and capacitors). The case studies with field datasets are conducted to validate the effectiveness of the developed LCOE calculation model.