Enhanced pan-European transmission planning methodology

Pache, Camille; Maeght, Jean; Seguinot, Baptiste; Zani, Alessandro; Lumbreras, Sara; Ramos, Andrés; Agapoff, Sergei; Warland, Leif; Rouco, L.; Panciatici, Patrick

doi:10.1109/pesgm.2015.7286270

Cited by 6 publications

(10 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar issues also arise in expansion planning and other configuration problems, where plans are evaluated by solving a massive number of multi-year Monte-Carlo simulations at 15-minute intervals (Pache et al 2015;Deutche-Energue-Agentur 2019). Additionally, the stochasticity introduced by renewable energy sources further increases the number of scenarios to consider.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Predicting AC Optimal Power Flows: Combining Deep Learning and Lagrangian Dual Methods

Fioretto

Mak

Hentenryck

2020

AAAI

126

124

View full text Add to dashboard Cite

The Optimal Power Flow (OPF) problem is a fundamental building block for the optimization of electrical power systems. It is nonlinear and nonconvex and computes the generator setpoints for power and voltage, given a set of load demands. It is often solved repeatedly under various conditions, either in real-time or in large-scale studies. This need is further exacerbated by the increasing stochasticity of power systems due to renewable energy sources in front and behind the meter. To address these challenges, this paper presents a deep learning approach to the OPF. The learning model exploits the information available in the similar states of the system (which is commonly available in practical applications), as well as a dual Lagrangian method to satisfy the physical and engineering constraints present in the OPF. The proposed model is evaluated on a large collection of realistic medium-sized power systems. The experimental results show that its predictions are highly accurate with average errors as low as 0.2%. Additionally, the proposed approach is shown to improve the accuracy of the widely adopted linear DC approximation by at least two orders of magnitude.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Additionally, the stochasticity introduced by renewable energy sources further increases the number of scenarios to consider. Therefore, modern approaches recur to the linear DC-OPF approximation and focus only on the scenarios considered most pertinent (Pache et al 2015) at the expense of the fidelity of the simulations.…”

Section: Introductionmentioning

confidence: 99%

Predicting AC Optimal Power Flows: Combining Deep Learning and Lagrangian Dual Methods

Fioretto

Mak

Hentenryck

2020

AAAI

126

124

View full text Add to dashboard Cite

show abstract

“…Similar issues also arise in expansion planning and other configuration problems, where plans are evaluated by solving a massive number of multi-year Monte-Carlo simulations at 15-minute intervals [26,13]. Additionally, the stochasticity introduced by renewable energy sources further increases the number of scenarios to consider.…”

Section: A Lagrangian Dual-based Approachmentioning

confidence: 99%

“…Additionally, the stochasticity introduced by renewable energy sources further increases the number of scenarios to consider. Therefore, modern approaches recur to the linear DC-OPF approximation and focus only on the scenarios considered most pertinent [26] at the expense of the fidelity of the simulations.…”

Section: A Lagrangian Dual-based Approachmentioning

confidence: 99%

Learning Hard Optimization Problems: A Data Generation Perspective

Kotary,

Fioretto,

Van Hentenryck

2021

Preprint

View full text Add to dashboard Cite

Optimization problems are ubiquitous in our societies and are present in almost every segment of the economy. Most of these optimization problems are NP-hard and computationally demanding, often requiring approximate solutions for largescale instances. Machine learning frameworks that learn to approximate solutions to such hard optimization problems are a potentially promising avenue to address these difficulties, particularly when many closely related problem instances must be solved repeatedly. Supervised learning frameworks can train a model using the outputs of pre-solved instances. However, when the outputs are themselves approximations, when the optimization problem has symmetric solutions, and/or when the solver uses randomization, solutions to closely related instances may exhibit large differences and the learning task can become inherently more difficult. This paper demonstrates this critical challenge, connects the volatility of the training data to the ability of a model to approximate it, and proposes a method for producing (exact or approximate) solutions to optimization problems that are more amenable to supervised learning tasks. The effectiveness of the method is tested on hard non-linear nonconvex and discrete combinatorial problems. IntroductionConstrained optimization (CO) is in daily use in our society, with applications ranging from supply chains and logistics, to electricity grids, organ exchanges, marketing campaigns, and manufacturing to name only a few. Two classes of hard optimization problems of particular interest in many fields are (1) combinatorial optimization problems and (2) nonlinear constrained problems. Combinatorial optimization problems are characterized by discrete search spaces and have solutions that are combinatorial in nature, involving for instance, the selection of subsets or permutations, and the sequencing or scheduling of tasks. Nonlinear constrained problems may have continuous search spaces but are often characterized by highly nonlinear constraints, such as those arising in electrical power systems whose applications must capture physical laws such as Ohm's law and Kirchhoff's law in addition to engineering operational constraints. Such CO problems are often NP-Hard and may be computationally challenging in practice, especially for large-scale instances.While the AI and Operational Research communities have contributed fundamental advances in optimization in the last decades, the complexity of some problems often prevents them from being adopted in contexts where many instances must be solved over a long-term horizon (e.g., multiyear planning studies) or when solutions must be produced under time constraints. Fortunately, in many practical cases, including the scheduling and energy problems motivating this work, one is interested in solving many problem instances sharing similar patterns. Therefore, the application of deep learning methods to aid the solving of computationally challenging constrained optimization problems appears to be a natural approach and has...

show abstract

“…In (Lumbreras & Ramos, 2016) and (Pache, 2015) In Canada, General Electric (GE) evaluated the technical feasibility of integrating large amounts of wind generation into the Canadian electric power system (General Electric Consulting, 2016). The study concluded that 4.6-4.8 GW of additional inter-regional transmission is required to accommodate 35% of the wind penetration in Canada.…”

Section: Wide-area Transmission Studies Around the Worldmentioning

confidence: 99%

Opportunities and benefits for increasing transmission capacity between the US eastern and western interconnections

Figueroa-Acevedo

View full text Add to dashboard Cite

Enhanced pan-European transmission planning methodology

Cited by 6 publications

References 12 publications

Predicting AC Optimal Power Flows: Combining Deep Learning and Lagrangian Dual Methods

Predicting AC Optimal Power Flows: Combining Deep Learning and Lagrangian Dual Methods

Learning Hard Optimization Problems: A Data Generation Perspective

Opportunities and benefits for increasing transmission capacity between the US eastern and western interconnections

Contact Info

Product

Resources

About