Implementation of a Massively Parallel Dynamic Security Assessment Platform for Large-Scale Grids

Konstantelos, Ioannis; Jamgotchian, Geoffroy; Tindemans, Simon H.; Duchesne, Philippe; Cole, Stijn; Merckx, Christian; Štrbac, Goran; Panciatici, Partick

doi:10.1109/pesgm.2018.8586141

Cited by 18 publications

(35 citation statements)

References 9 publications

(9 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The problem increases in complexity for larger and more realistic power systems. A large number of samples is required to learn accurate classifiers [3]. However, we estimate the increase in the complexity of the optimization problem will be only moderately higher than the increase in an equivalent OPF problem.…”

Section: F Computational Feasibilitymentioning

confidence: 96%

“…In general, the aim of such work-flows is to provide a scalable way of managing uncertainty and system complexity within the tight constraints of real-time operation in an effort to improve the TSO's and DSO's situational awareness. As such, most research efforts until now have focused on studying the computational performance of such platforms [3] as well exploring the statistical model for generating operating points in a multivariate setting (e.g. [4], [5]) and machine learning approaches for building useful security rules [1].…”

Section: A Existing Approachesmentioning

confidence: 99%

“…The population of pre-fault operating points X and X t can, for example, be obtained by sampling an underlying statistical model fitted to historical data, while the population of labels Y and Y t is obtained via simulation Y : X → Y (see [3], [20] and [21]…”

Section: A Security Rules For Classificationmentioning

confidence: 99%

“…This is evidenced by the fact that terminal node n3 is not pure, but contains mixed class labels. Furthermore, wrong predictions may occur when the DT has been trained on an insufficiently large number of samples, or if the training and testing populations differ [3], e.g. due to the respective underlying model to generate the samples, such as in a topological change in the power system (e.g., as studied in [13]).…”

Section: A Security Rules For Classificationmentioning

confidence: 99%

See 3 more Smart Citations

Data-Driven Power System Operation: Exploring the Balance Between Cost and Risk

Cremer

Konstantelos

Tindemans

et al. 2019

IEEE Trans. Power Syst.

Self Cite

View full text Add to dashboard Cite

Supervised machine learning has been successfully used in the past to infer a system's security boundary by training classifiers (also referred to as security rules) on a large number of simulated operating conditions. Although significant research has been carried out on using classifiers for the detection of critical operating points, using classifiers for the subsequent identification of suitable preventive/corrective control actions remains underdeveloped. This paper focuses on addressing the challenges that arise when utilizing security rules for control purposes. Illustrative examples and case studies are used to show how even very accurate security rules can lead to prohibitively high risk exposure when used to identify optimal control actions. Subsequently, the inherent tradeoff between operating cost and security risk is explored in detail. To optimally navigate this tradeoff, a novel approach is proposed that uses an ensemble learning method (AdaBoost) to infer a probabilistic description of a system's security boundary. Bias in predictions is compensated by the Platt Calibration method. Subsequently, a general-purpose framework for building probabilistic and disjunctive security rules of a system's secure operating domain is developed that can be embedded within classic operation formulations. Through case studies on the IEEE 39-bus system, it is showcased how security rules derived from supervised learning can be efficiently utilized to optimally operate the system under multiple uncertainties while respecting a user-defined balance between cost and risk. This is a fundamental step toward embedding data-driven models within classic optimisation approaches.

show abstract

Section: F Computational Feasibilitymentioning

confidence: 96%

Section: A Existing Approachesmentioning

confidence: 99%

Section: A Security Rules For Classificationmentioning

confidence: 99%

Section: A Security Rules For Classificationmentioning

confidence: 99%

See 2 more Smart Citations

Data-Driven Power System Operation: Exploring the Balance Between Cost and Risk

Cremer

Konstantelos

Tindemans

et al. 2019

IEEE Trans. Power Syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the European project iTesla, decision trees are used for online security assessment. The platform developed within this project for online static and dynamic security assessment is presented in [95], [96].…”

Section: Learning a Modelmentioning

confidence: 99%

Recent Developments in Machine Learning for Energy Systems Reliability Management

2020

View full text Add to dashboard Cite

This paper reviews recent works applying machine learning techniques in the context of energy systems reliability assessment and control. We showcase both the progress achieved to date as well as the important future directions for further research, while providing an adequate background in the fields of reliability management and of machine learning. The objective is to foster the synergy between these two fields and speed up the practical adoption of machine learning techniques for energy systems reliability management. We focus on bulk electric power systems and use them as an example, but we argue that the methods, tools, etc. can be extended to other similar systems, such as distribution systems, micro-grids, and multi-energy systems.

show abstract

Decomposing convexified security‐constrained AC optimal power flow problem with automatic generation control reformulation

Waseem

Manshadi

2021

Int Trans Electr Energ Syst

View full text Add to dashboard Cite

This paper presents a reformulation for the automatic generation control (AGC) in a decomposed convex relaxation algorithm. It finds an optimal solution to the AC optimal power flow (ACOPF) problem that is secure against a large set of contingencies. The original ACOPF problem which represents the system without contingency constraints, is convexified by applying the second-order cone relaxation method. The contingencies are filtered to distinguish those that will be treated with preventive actions from those that will be left for corrective actions. The selected contingencies for preventive action are included in the set of security constraints. Benders decomposition is employed to decompose the convexified Security-Constrained ACOPF problem into a master problem and several security check subproblems. Subproblems are evaluated in a parallel computing process with enhanced computational efficiency. AGC within each subproblem is modeled by a set of proposed valid constraints, so the procured solution is the physical response of each generation unit during a contingency. Benders optimality cuts are generated for the subproblems having mismatches and the cuts are passed to the master problem to encounter the security-constraints. The accuracy of the relaxation results is verified using the presented tightness measure. The effectiveness of the presented valid AGC constraints and scalability of the proposed algorithm is demonstrated in several case studies.

show abstract

Implementation of a Massively Parallel Dynamic Security Assessment Platform for Large-Scale Grids

Cited by 18 publications

References 9 publications

Data-Driven Power System Operation: Exploring the Balance Between Cost and Risk

Data-Driven Power System Operation: Exploring the Balance Between Cost and Risk

Recent Developments in Machine Learning for Energy Systems Reliability Management

Decomposing convexified security‐constrained AC optimal power flow problem with automatic generation control reformulation

Contact Info

Product

Resources

About