Approximating Trajectory Constraints With Machine Learning – Microgrid Islanding With Frequency Constraints

Zhang, Yichen; Chen, Chen; Liu, Guodong; Hong, Tianqi; Qiu, Feng

doi:10.1109/tpwrs.2020.3015913

Cited by 41 publications

(18 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…NN embedding has only been applied within the power systems research community in few applications so far. Authors in [70] approximate the nonlinear function between microgrid operating point and frequency nadir in order to encode frequency constraints in a scheduling problem. Surrogate frequency constraints are then encoded into the linearized Unit Commitment problem in [71].…”

Section: E Embedding Neural Network In Optimization Problemsmentioning

confidence: 99%

Closing the Loop: A Framework for Trustworthy Machine Learning in Power Systems

Stiasny¹,

Chevalier²,

Nellikkath³

et al. 2022

Preprint

View full text Add to dashboard Cite

Deep decarbonization of the energy sector will require massive penetration of stochastic renewable energy resources and an enormous amount of grid asset coordination; this represents a challenging paradigm for the power system operators who are tasked with maintaining grid stability and security in the face of such changes. With its ability to learn from complex datasets and provide predictive solutions on fast timescales, machine learning (ML) is well-posed to help overcome these challenges as power systems transform in the coming decades. In this work, we outline five key challenges (dataset generation, data pre-processing, model training, model assessment, and model embedding) associated with building trustworthy ML models which learn from physics-based simulation data. We then demonstrate how linking together individual modules, each of which overcomes a respective challenge, at sequential stages in the machine learning pipeline can help enhance the overall performance of the training process. In particular, we implement methods that connect different elements of the learning pipeline through feedback, thus "closing the loop" between model training, performance assessments, and re-training. We demonstrate the effectiveness of this framework, its constituent modules, and its feedback connections by learning the N-1 small-signal stability margin associated with a detailed model of a proposed North Sea Wind Power Hub system.

show abstract

Section: E Embedding Neural Network In Optimization Problemsmentioning

confidence: 99%

Closing the Loop: A Framework for Trustworthy Machine Learning in Power Systems

Stiasny¹,

Chevalier²,

Nellikkath³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Because of that, the size of the training dataset should be moderately small, especially for deeper tree structures. A dynamic model is presented in [17] to generate the training data. The generated data is trained by the deep neural network.…”

Section: Introductionmentioning

confidence: 99%

Robust frequency constrained uc using data driven logistic regression for island power systems

Rajabdorri

Lobato

Sigrist

2022

IET Generation Trans & Dist

View full text Add to dashboard Cite

In the current practice of short-term power scheduling, online power reserves are used to address generation mismatches and contingencies. Neither online inertia nor the speed of the committed units is considered in the scheduling process. With the increasing injection of uncertain renewable energy sources, this practice is starting to fall short, especially in island power systems, where the primary frequency response is already scarce, and any contingency leads to potentially poor frequency response. This paper introduces a data-driven linear constraint to improve the post-fault frequency quality in island power systems. A coherent initial data-set is obtained by simulating the system frequency response of single outages. Then logistic regression is employed as a predictive analytic procedure to differentiate the acceptable and unacceptable incidents. To compare the conventional methods with the proposed approach and also to handle the uncertain nature of renewable energy generation, an adaptive robust unit commitment formulation is utilized. Results for the island power system of La Palma show that depending on the chosen cut-point on the logistic regression estimation the proposed method can improve the frequency response quality of the system while reducing the operation costs.

show abstract

“…An iterative algorithm is developed based on the cutting plan approach to incorporate the post-contingency frequency constraints, which increases its implementing complexity. Deep learning is applied in [15] to approximate the nonlinear nadir constraint using a neural network such that an MILP-based microgrid scheduling problem can be formulated. Nevertheless, the detailed SI modeling from IBGs is not considered and the uncertainty associated with the noncritical load shedding due to the forecasting errors and the relays has not been discussed.…”

Section: Introductionmentioning

confidence: 99%

Frequency-Constrained Resilient Scheduling of Microgrid: A Distributionally Robust Approach

Chu

Zhang

Teng

2021

Preprint

View full text Add to dashboard Cite

In order to prevent the potential frequency instability due to the high Power Electronics (PE) penetration under an unintentional islanding event, this paper presents a novel microgrid scheduling model which explicitly models the system frequency dynamics as well as the long/short term uncertainty associated with renewable energy resources and load. Synthetic Inertia (SI) control is applied to regulating the active power output of the Inverter-Based Generators (IBGs) to support the postislanding frequency evaluation. The uncertainty associated with the noncritical load shedding is explicitly modeled based on the distributionally robust formulation to ensure resilient operation during islanding events. The resulted frequency constraints are derived analytically and reformulated into Second-Order Cone (SOC) form, which are further incorporated into the microgrid scheduling model, enabling optimal SI provision of Renewable Energy Sources (RESs) from the micorgrid perspective. With the SOC relaxation of the AC power flow constraints, the overall problem is constructed as a mixed-integer SOC Programming (MISOCP). The effectiveness of the proposed model is demonstrated based on modified IEEE 14-bus system.

show abstract

Approximating Trajectory Constraints With Machine Learning – Microgrid Islanding With Frequency Constraints

Cited by 41 publications

References 38 publications

Closing the Loop: A Framework for Trustworthy Machine Learning in Power Systems

Closing the Loop: A Framework for Trustworthy Machine Learning in Power Systems

Robust frequency constrained uc using data driven logistic regression for island power systems

Frequency-Constrained Resilient Scheduling of Microgrid: A Distributionally Robust Approach

Contact Info

Product

Resources

About