Search

To enhance the quality of energy management tasks, accurately representing the thermal dynamics of buildings is crucial. Traditional methods aim to improve the building model in regards to an arbitrary statistical metric, before feeding the trained model to the optimization-based energy management process. In this paper, we advocate for a more integrated approach, consisting of incorporating the downstream optimization directly into the training pipeline. The goal is to improve the building model in strategic operating zones, where the greatest impact on decision-making will be achieved. To that end, we first formulate the thermal dynamics as ordinary differential equations (ODEs) using neural networks. The model parameters are then updated through an end-to-end gradient-based training strategy wherein the downstream optimization is used as the loss function. To increase the robustness of the approach, the proposed loss is combined with traditional physics-informed accuracy-oriented training, employing a novel coordinated gradient descent algorithm. Simulation results show the effectiveness of the proposed modeling method, regarding both the optimality of decisions and their physical interpretability.

Roadmaps toward a low-carbon renewable energy industry demand substantial bulk energy storages to account for non-dispatchability of renewables. Liquid Air Energy Storage (LAES) has gained recognition as one of few bulk-scale energy storage facilities not limited by geographical requirements, unlike pumped hydro and compressed air energy storage systems. However, the comparatively low efficiency of freestanding LAES facilities hinders their widespread stationing in power and energy systems. There has been a recent uptick of interest in the potential efficiency gains of coupling LAES with the Liquified Natural Gas (LNG) regasification process. Thus, for the first time, this article presents a day-ahead dispatch model for a LAES coupled with an LNG regasification process (hereafter, LAES-LNG), interacting with electricity and LNG markets as involved energy carriers. Through opportune sequence-aware LNG and electricity procuring, the coupled facility can benefit not only from the released cold energy during the LNG regasification process but also from co-firing regasified air and natural gas to produce and sell electricity. Focusing on a realistic market environment, the proposed dispatch model is derived as a two-stage stochastic setup. The cost-effectiveness of the proposed LAES-LNG facility is validated through a short-term daily test and a probabilistic economic feasibility study.

Day-ahead wind power forecast (DWPF) is an indispensable component for accommodating wind energy in power systems. Traditional DWPF models focus on providing information, merely, around the average wind power for day-ahead periods. This article goes beyond the traditional approach and formulates a DWPF problem that renders information on intra-period wind variability with high resolution (second-wise), by predicting the intra-period temporal distribution of wind power for day-ahead horizons. This additional information could be of particular interest to decision-makers when high-resolution wind variability is crucial for optimal decisions, e.g., wind power reserve scheduling. First, a parametric and several entropy-based losses are tailored to this problem to acquire candidate solutions using classical approaches. Then, a differentiable loss, based on the Wasserstein distance (WD), is dedicatedly developed to overcome the inherent limitations of the tailored losses. The superiority of the proposed WD-based loss is, first, verified by comparing its predictions with those of classical losses, using on-site wind measurements. Furthermore, the superiority of the predictions obtained by the proposed WD loss, in the context of day-ahead wind power scheduling, is further verified using operational electricity market data, with profit erosion as an error measure.

In this paper, the reliability of cyber-physical energy hubs (CPEHs) is evaluated based on the Monte Carlo simulation (MCS) method. Previous works in the related literature are generally focused on the physical layer of energy hubs (EHs). However, the security of EHs heavily relies on the reliable and safe performance of the cyber system. Therefore, this study considers the failure of cyber-physical components, data packet errors, and information transmission delay to improve the reliability assessment accuracy. The physical layer of the proposed CPEH is divided into two sections, namely heat and power sections, and reliability indices are calculated for both sections. Various scenarios are suggested to precisely analyze the effects of the cyber layer consideration on the CPEH reliability results. Sensitivity analysis on the failure rate of the components is also conducted to understand better the performance of the CPEH in situations with a higher risk of heat components’ failure. The notable impacts of the cyber layer disturbances on the CPEH reliability evaluation are demonstrated through the simulation results.

Uncertain distributed energy resources and uneven load allocation cause the three-phase unbalance in distribution networks (DNs), which may harm the health of power equipment and increase the operational cost. There are emerging opportunities to balance three-phase DNs with a number of power electronic devices installed in the system. In this paper, we propose a novel two-level coordination strategy to improve the network balancing performance, where soft open points (SOPs) and phase switch devices (PSDs) are hierarchically coordinated in the network. At the upper level, a new type of SOPs with the function of phase switching is designed to explore the cross-phase power transfer ability; at the lower level, PSDs are utilized to flexibly allocate individual loads to specific phases. The two-level coordination strategy is typically formulated as a mixed-integer nonlinear programming (MINLP) problem. To solve the model accurately and efficiently, we develop a successive linearization algorithm to approximate it to a mixed-integer linear programming (MILP) problem at each iteration. On this basis, we propose a heuristic time-independent fixing algorithm to further ease the computational burden by eliminating a large number of integer variables in the MILP problem. Numerical simulations are conducted to validate the effectiveness, accuracy, and efficiency of the proposed method.

The increasing share of renewables in the electricity generation mix comes along with an increasing uncertainty in power supply. In the recent years, distributionally robust optimization has gained significant interest due to its ability to make informed decisions under uncertainty, which are robust to misrepresentations of the distributional information (e.g., from probabilistic forecasts). This is achieved by introducing an ambiguity set that describes the potential deviations from an empirical distribution of all uncertain parameters. However, this set typically overlooks the inherent dependencies of uncertainty, e.g., spatial dependencies of weather-dependent energy sources. This paper goes beyond the state-of-the-art models by embedding such dependencies within the definition of ambiguity set. In particular, we propose a new copula-based ambiguity set which is tailored to capture any type of dependencies. The resulting problem is reformulated as a conic program which is kept generic such that it can be applied to any decision-making problem under uncertainty in power systems. Given the Optimal Power Flow (OPF) problem as one of the main potential applications, we illustrate the performance of our proposed distributionally robust model applied to i) a DC-OPF problem for a meshed transmission system and ii) an AC-OPF problem using LinDistFlow approximation for a radial distribution system.

Real-time electricity prices are economic signals incentivizing market players to support real-time system balancing. These price signals typically switch between low- and high-price regimes depending on whether the power system is in surplus or shortage of generation, which is hard to capture. In this context, we propose a new Transformer-based model to assist the short-term trading strategies of market players. The proposed model offers high-performance probabilistic forecasts of real-time prices while providing insights into its inner decision-making process. Transformers rely on attention mechanisms solely computed via feed-forward networks to explicitly learn temporal patterns, which allows them to capture complex dependencies such as regime switching. Here, we augment Transformers with subnetworks dedicated to endogenously quantify the relative importance of each input feature. Hence, the resulting forecaster intrinsically provides the temporal attribution of each input feature, which foster trust and transparency for subsequent decision makers. Our case study on real-world market data of the Belgian power system demonstrates the ability of the proposed model to outperform state-of-the-art forecasting techniques, while shedding light on its most important drivers.

By sharing common assets such as the power grid, prosumers are closely interrelated by their actions and interests. Game theory provides powerful tools for increased coordination among the prosumers to optimize the energy resources. However, depending on the prosumer profiles and the market rules, the individual bills may notably differ and prove to be unfair. In this work, we analyze the outcomes of three relevant game-theoretical billing methods, which are innovatively transposed to the day-ahead scheduling of energy exchange within a liberalized residential community dominated by distributed energy resources. The first two approaches rely on a (static) daily billing scheme, while the third considers a multi-temporal (continuous) billing. The Nash equilibria are computed using distributed algorithms, hence ensuring individual decision-making and avoiding third-party dependencies. The cost distributions are assessed using both a qualitative and a quantitative comparison based on various prosumer profiles in a modern smart grid. It is shown that, depending on the billing option, either the contribution towards the entity (i.e., the ability to improve the global solution) or the individual empowerment (i.e., the ability to bargain) can be preferentially incentivized.