Scenario generation is an important step in the operation and planning of power systems with high renewable penetrations. In this work, we proposed a data-driven approach for scenario using generative adversarial networks, which based on two interconnected deep neural networks. Compared existing methods probabilistic models that are often hard to scale or sample from, our method data-driven, captures energy production patterns both temporal spatial dimensions large number correlated...

Control of complex systems involves both system identification and controller design. Deep neural networks have proven to be successful in many tasks, however, from model-based control perspective, these are difficult work with because they typically nonlinear nonconvex. Therefore still identified controlled based on simple linear models despite their poor representation capability. In this paper we bridge the gap between model accuracy tractability faced by networks, explicitly constructing...

Advances in Machine Learning (ML) have led to its adoption as an integral component many applications, including banking, medical diagnosis, and driverless cars. To further broaden the use of ML models, cloud-based services offered by Microsoft, Amazon, Google, others developed ML-as-a-service tools black-box systems. However, classifiers are vulnerable adversarial examples: inputs that maliciously modified can cause classifier provide adversary-desired outputs. Moreover, it is known...

Load forecasting plays a critical role in the operation and planning of power systems. By using input features such as historical loads weather forecasts, system operators utilities build forecast models to guide decision making commitment dispatch. As techniques becomes more sophisticated, however, they also become vulnerable cybersecurity threats. In this paper, we study vulnerability class load algorithms analyze potential impact on operations, shedding increased dispatch costs....

Recent advances in Machine Learning (ML) have led to its broad adoption a series of power system applications, ranging from meter data analytics, renewable/load/price forecasting grid security assessment. Although these data-driven methods yield state-of-the-art performances many tasks, the robustness and applying such algorithms modern grids not been discussed. In this paper, we attempt address issues regarding ML applications systems. We first show that most current proposed systems are...

In this paper, we propose a novel scenario forecasts approach which can be applied to broad range of power system operations (e.g., wind, solar, load) over various horizons and prediction intervals. This is model-free data-driven, producing set scenarios that represent possible future behaviors based only on historical observations point forecasts. It first applies newly-developed unsupervised deep learning framework, the generative adversarial networks, learn intrinsic patterns in renewable...

Abstract The spatiotemporal distribution of intangible cultural heritage in Fujian Province, China, and the factors that influence it were explored using multiple spatial scales. samples include five batches Chinese national-level six provincial-level items, totaling 554. involve city-scale, county-scale, traditional dwellings, analysis uses various methods, such as gravity migration theory GeoDetector model. results show that, terms historical timescales, moving trajectory center Province...

Neural solvers based on the divide-and-conquer approach for Vehicle Routing Problems (VRPs) in general, and capacitated VRP (CVRP) particular, integrates global partition of an instance with local constructions each subproblem to enhance generalization. However, during phase, misclusterings within subgraphs have a tendency progressively compound throughout multi-step decoding process learning-based policy. This suboptimal behavior turn, may lead dramatic deterioration performance overall...

We present a method to generate renewable scenarios using Bayesian probabilities by implementing the generative adversarial network (Bayesian GAN), which is variant of networks based on two interconnected deep neural networks. By formulation, generators can be constructed and trained produce that capture different salient modes in data, allowing for better diversity more accurate representation underlying physical process. Compared conventional statistical models are often hard scale or...

Recent proliferation of electric vehicle (EV) charging load has imposed vital stress on power grid. The stochasticity and volatility EV behaviors render it challenging to manipulate the uncertain demand for grid operations management. Charging scenario generation can serve future integration by modeling uncertainties simulating various realistic sessions. To this end, we propose a denoising Diffusion-based model coined DiffCharge, which is capable yielding both battery-level station-level...

Modern buildings encompass complex dynamics of multiple electrical, mechanical, and control systems. One the biggest hurdles in applying conventional model-based optimization methods to building energy management is huge cost effort capturing diverse temporally correlated dynamics. Here we propose an alternative approach which model-free data-driven. By utilizing high volume data coming from advanced sensors, train a deep Recurrent Neural Networks (RNN) could accurately represent operation's...

Distributed energy resources (DERs) can serve as non-wire alternatives (NWAs) to capacity expansion by managing peak load avoid or delay traditional projects. However, the value stream derived from using DERs NWAs is usually not explicitly included in DER planning problems. In this paper, we study a problem that co-optimizes investment and operation of timing expansion. By including decision variable, naturally incorporate NWA-related problem. Furthermore, show even though resulting...

Scenario generation is an important step in the operation and planning of power systems with high renewable penetrations. In this work, we proposed a data-driven approach for scenario using generative adversarial networks, which based on two interconnected deep neural networks. Compared existing methods probabilistic models that are often hard to scale or sample from, our method data- driven, captures energy production patterns both temporal spatial dimensions large number correlated...

Many decision-making problems in engineering applications such as transportation, power system and operations research require repeatedly solving large-scale linear programming with a large number of different inputs. For example, energy systems high levels uncertain renewable resources, tens thousands scenarios may need to be solved every few minutes. Standard iterative algorithms for network flow problems, even though highly efficient, becomes bottleneck these applications. In this work,...

The novel coronavirus (COVID-19) pandemic has posed unprecedented challenges for the utilities and grid operators around world. In this work, we focus on problem of load forecasting. With strict social distancing restrictions, power consumption profiles world have shifted both in magnitude daily patterns. These changes caused significant difficulties short-term Typically algorithms use weather, timing information previous levels as input variables, yet they cannot capture large sudden...

The dc optimal power flow (DCOPF) problem is a fundamental in systems operations and planning. With high penetration of uncertain renewable resources systems, DCOPF needs to be solved repeatedly for large amount scenarios, which can computationally challenging. As an alternative iterative solvers, neural networks are often trained used solve DCOPF. These approaches offer orders magnitude reduction computational time, but they cannot guarantee generalization, small training error does not...

The prosperity of smart mobile devices has made crowdsensing (MCS) a promising paradigm for completing complex sensing and computation tasks. In the past, great efforts have been on design incentive mechanisms task allocation strategies from MCS platform's perspective to motivate users' participation. However, in practice, participants face many uncertainties coming their environment as well other participants' strategies, how do they interact with each make decisions is not understood. this...