Due to its reduced communication overhead and robustness to failures, distributed energy management is of paramount importance in smart grids, especially in microgrids, which feature distributed generation (DG) and distributed storage (DS). Distributed economic dispatch for a microgrid with high renewable energy penetration and demand-side management operating in grid-connected mode is considered in this paper. To address the intrinsically stochastic availability of renewable energy sources (RES), a novel power scheduling approach is introduced. The approach involves the actual renewable energy as well as the energy traded with the main grid, so that the supply-demand balance is maintained. The optimal scheduling strategy minimizes the microgrid net cost, which includes DG and DS costs, utility of dispatchable loads, and worst-case transaction cost stemming from the uncertainty in RES. Leveraging the dual decomposition, the optimization problem formulated is solved in a distributed fashion by the local controllers of DG, DS, and dispatchable loads. Numerical results are reported to corroborate the effectiveness of the novel approach.

This paper develops a novel graph convolutional network (GCN) framework for fault location in power distribution networks. The proposed approach integrates multiple measurements at different buses while taking system topology into account. The effectiveness of the GCN model is corroborated by the IEEE 123 bus benchmark system. Simulation results show that the GCN model significantly outperforms other widely-used machine learning schemes with very high fault location accuracy. In addition, the proposed approach is robust to measurement noise and data loss errors. Data visualization results of two competing neural networks are presented to explore the mechanism of GCNs superior performance. A data augmentation procedure is proposed to increase the robustness of the model under various levels of noise and data loss errors. Further experiments show that the model can adapt to topology changes of distribution networks and perform well with a limited number of measured buses.

Non-intrusive load monitoring (NILM) addresses the challenging task of decomposing the aggregate signal of a household's electricity consumption into appliance-level data without installing dedicated meters. By detecting load malfunction and recommending energy reduction programs, cost-effective NILM provides intelligent demand-side management for utilities and end users. In this paper, we boost the accuracy of energy disaggregation with a novel neural network structure named scale- and context-aware network (SCANet), which exploits multi-scale features and contextual information. Specifically, we develop a multi-branch architecture with multiple receptive field sizes and branch-wise gates that connect the branches in the sub-networks. We build a self-attention module to facilitate the integration of global context. We incorporate an adversarial loss and on-state augmentation to further improve the model's performance. Extensive simulation results tested on open datasets corroborate the merits of the proposed approach, which significantly outperforms state-of-the-art methods.

A large number of geo-distributed data centers begin to surge in the era of data deluge and information explosion. To meet the growing demand in massive data processing, the infrastructure of future data centers must be energy-efficient and sustainable. Facing this challenge, a systematic framework is put forth in this paper to integrate renewable energy sources (RES), distributed storage units, cooling facilities, as well as dynamic pricing into the workload and energy management tasks of a data center network. To cope with RES uncertainty, the resource allocation task is formulated as a robust optimization problem minimizing the worst-case net cost. Compared with existing stochastic optimization methods, the proposed approach entails a deterministic uncertainty set where generated RES reside, thus can be readily obtained in practice. It is further shown that the problem can be cast as a convex program, and then solved in a distributed fashion using the dual decomposition method. By exploiting the spatio-temporal diversity of local temperature, workload demand, energy prices, and renewable availability, the proposed approach outperforms existing alternatives, as corroborated by extensive numerical tests performed using real data.

Integrating renewable energy into the modern power grid requires risk-cognizant dispatch of resources to account for the stochastic availability of renewables. Toward this goal, day-ahead stochastic market clearing with high-penetration wind energy is pursued in this paper based on the DC optimal power flow (OPF). The objective is to minimize the social cost which consists of conventional generation costs, end-user disutility, as well as a risk measure of the system re-dispatching cost. Capitalizing on the conditional value-at-risk (CVaR), the novel model is able to mitigate the potentially high risk of the recourse actions to compensate wind forecast errors. The resulting convex optimization task is tackled via a distribution-free sample average based approximation to bypass the prohibitively complex high-dimensional integration. Furthermore, to cope with possibly large-scale dispatchable loads, a fast distributed solver is developed with guaranteed convergence using the alternating direction method of multipliers (ADMM). Numerical results tested on a modified benchmark system are reported to corroborate the merits of the novel framework and proposed approaches.

Through spatial multiplexing and diversity, multi-input multi-output (MIMO) cognitive radio (CR) networks can markedly increase transmission rates and reliability, while controlling the interference inflicted to peer nodes and primary users (PUs) via beamforming. The present paper optimizes the design of transmit- and receive-beamformers for ad hoc CR networks when CR-to-CR channels are known, but CR-to-PU channels cannot be estimated accurately. Capitalizing on a norm-bounded channel uncertainty model, the optimal beamforming design is formulated to minimize the overall mean-square error (MSE) from all data streams, while enforcing protection of the PU system when the CR-to-PU channels are uncertain. Even though the resultant optimization problem is non-convex, algorithms with provable convergence to stationary points are developed by resorting to block coordinate ascent iterations, along with suitable convex approximation techniques. Enticingly, the novel schemes also lend themselves naturally to distributed implementations. Numerical tests are reported to corroborate the analytical findings.