Safe Bayesian Optimization for Data-Driven Power Electronics Control Design in Microgrids: From Simulations to Real-World Experiments

Weber, Daniel; Heid, Stefan; Bode, Henrik; Lange, Jarren Hilton; Hüllermeier, Eyke; Wallscheid, Oliver

doi:10.1109/access.2021.3062144

Cited by 8 publications

(7 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the corresponding article, the RL agent lacks behind concerning the steady-state error compared to a classical PI controller but outperforms it in terms of total demand distortion. The measurement-based validation of the simulation setting used for the power grid application can be found in [22]. As shown in [23], RL is already applied in literature to solve power grid stability and power quality problems, but in a discrete state and action space without the usage of deep neural networks and without focusing on stationary accuracy of the reference tracking problem.…”

Section: A State Of the Artmentioning

confidence: 99%

“…It is assumed to be a part of the environment such that the control agent never needs to handle any signals in the abc frame. Utilization of this transformation is a well-established method, as no further system information is necessary for applying this simplification [22], [28]. A dead time of one simulation step is assumed within the plant system to account for the controller calculation delay that is common for digital control systems.…”

Section: A Electrical Power Grid Applicationmentioning

confidence: 99%

“…A dead time of one simulation step is assumed within the plant system to account for the controller calculation delay that is common for digital control systems. The same simulation configuration as described in [22] is used in the following. The output ŷk consists of the measurement signals depicted in Fig.…”

Section: A Electrical Power Grid Applicationmentioning

confidence: 99%

“…IV-A is investigated. The PI controller, implemented as a cascade of voltage and current controllers, is parameterized using safe Bayesian optimization [22] over 100 iterations. More information about the parameterization can be found in [37].…”

Section: B Performance Analysismentioning

confidence: 99%

See 3 more Smart Citations

Steady-State Error Compensation in Reference Tracking and Disturbance Rejection Problems for Reinforcement Learning-Based Control

Weber¹,

Schenke²,

Wallscheid³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Reinforcement learning (RL) is a promising, upcoming topic in automatic control applications. Where classical control approaches require a priori system knowledge, data-driven control approaches like RL allow a modelfree controller design procedure, rendering them emergent techniques for systems with changing plant structures and varying parameters. While it was already shown in various applications that the transient control behavior for complex systems can be sufficiently handled by RL, the challenge of non-vanishing steady-state control errors remains, which arises from the usage of control policy approximations and finite training times. To overcome this issue, an integral action state augmentation (IASA) for actor-critic-based RL controllers is introduced that mimics an integrating feedback, which is inspired by the delta-input formulation within model predictive control. This augmentation does not require any expert knowledge, leaving the approach model free. As a result, the RL controller learns how to suppress steady-state control deviations much more effectively. Two exemplary applications from the domain of electrical energy engineering validate the benefit of the developed method both for reference tracking and disturbance rejection. In comparison to a standard deep deterministic policy gradient (DDPG) setup, the suggested IASA extension allows to reduce the steady-state error by up to 52 % within the considered validation scenarios.

show abstract

Section: A State Of the Artmentioning

confidence: 99%

Section: A Electrical Power Grid Applicationmentioning

confidence: 99%

Section: A Electrical Power Grid Applicationmentioning

confidence: 99%

Section: B Performance Analysismentioning

confidence: 99%

See 2 more Smart Citations

Steady-State Error Compensation in Reference Tracking and Disturbance Rejection Problems for Reinforcement Learning-Based Control

Weber¹,

Schenke²,

Wallscheid³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…These systems collect surplus electricity from sustainable sources and provide electricity during periods of high demand or low generation. The main objectives of utilizing these storage systems are to enhance energy efficiency, reduce reliance on fossil fuels, and optimize storage and usage costs [7], [8].…”

Section: Introductionmentioning

confidence: 99%

Cost optimization of electricity in energy storage system by dynamic programming

Mishra,

Routray,

Pattanayak

et al. 2024

IJAPE

View full text Add to dashboard Cite

This paper presents a dynamic programming solution for the cost optimization of an electric storage system. The objective is to minimize the total cost of meeting electricity demand over a specified time interval, considering energy constraints and costs. The proposed algorithm efficiently determines the optimal energy discharge and charge strategies for the storage system, resulting in reduced overall costs. The effectiveness and efficiency of the algorithm are demonstrated through various test cases, highlighting its potential for real-world applications in energy storage systems and electric grid management. It also provides an overview of different types of electrical storage systems, review recent research on optimization techniques for energy storage, and examines recent studies on the optimization of electrical storage systems for specific applications, such as peak load shaving and grid stability. Through this comprehensive analysis, we hope to shed light on the current state of the field and identify areas for further research and improvement.

show abstract

Steady-State Error Compensation for Reinforcement Learning-Based Control of Power Electronic Systems

2023

View full text Add to dashboard Cite

Safe Bayesian Optimization for Data-Driven Power Electronics Control Design in Microgrids: From Simulations to Real-World Experiments

Cited by 8 publications

References 32 publications

Steady-State Error Compensation in Reference Tracking and Disturbance Rejection Problems for Reinforcement Learning-Based Control

Steady-State Error Compensation in Reference Tracking and Disturbance Rejection Problems for Reinforcement Learning-Based Control

Cost optimization of electricity in energy storage system by dynamic programming

Steady-State Error Compensation for Reinforcement Learning-Based Control of Power Electronic Systems

Contact Info

Product

Resources

About