Improving primary frequency response in networked microgrid operations using multilayer perceptron-driven reinforcement learning

• Published in: IET Smart Grid Vol. 3; no. 4; pp. 500 - 507
• Main Authors: Radhakrishnan, Nikitha, Chakraborty, Indrasis, Xie, Jing, Thekkumparambath Mana, Priya, Tuffner, Francis K, Bhattarai, Bishnu P, Schneider, Kevin P
• Format: Journal Article
• Language: English
• Published: Durham The Institution of Engineering and Technology 01.08.2020; John Wiley & Sons, Inc; Wiley
• Subjects: B8110C Power system control; B8120J Distribution networks; B8120K Distributed power generation; C3110G Frequency control; C3340H Control of electric power systems; C5290 Neural computing techniques; C6170K Knowledge engineering techniques; Controllers; critical end-use loads; Design; Distributed generation; distributed power generation; Dynamic stability; Electrical loads; Electricity distribution; Energy consumption; Energy sources; existing generator controls; fast frequency response; frequency control; Frequency deviation; Frequency response; grid-following inverter-connected renewable energy resources; improving primary frequency response; individual microgrids; invertors; learning (artificial intelligence); multilayer perceptron-driven reinforcement learning; Multilayer perceptrons; networked microgrid operations; networked microgrids; power distribution control; power generation control; power systems; reinforcement-learning-based controller; renewable energy sources; resource sharing; Simulation; Special Issue: Machine Learning in Power Systems; Switching; switching transient scenarios; system collapse; system inertia; System reliability; system topologies; system voltage; Test systems; Topology; power distribution control; system topologies; frequency deviation; multilayer perceptron-driven reinforcement learning; power generation control; existing generator controls; critical end-use loads; distributed power generation; fast frequency response; invertors; individual microgrids; learning (artificial intelligence); system voltage; system collapse; improving primary frequency response; multilayer perceptrons; system inertia; renewable energy sources; reinforcement-learning-based controller; frequency control; grid-following inverter-connected renewable energy resources; resource sharing; power systems; switching transient scenarios; frequency response; networked microgrid operations; networked microgrids
• ISSN: 2515-2947; 2515-2947
• DOI: 10.1049/iet-stg.2019.0261

Individual microgrids can improve the reliability of power systems during extreme events, and networked microgrids can further improve efficiency through resource sharing and increase the resilience of critical end-use loads. However, networked microgrid operations can be subject to large transients due to switching and end-use loads, which can cause dynamic instability and lead to system collapse. These transients are especially prevalent in microgrids with high penetrations of grid-following inverter-connected renewable energy resources, which do not provide the system inertia or fast frequency response needed to mitigate the transients. One potential mitigation is to engage the existing generator controls to reduce system voltage in response to a frequency deviation, thereby reducing load and improving primary frequency response. This study investigates the use of a reinforcement-learning-based controller trained over several switching transient scenarios to modify generator controls during large frequency deviations. Compared to previously used proportional–integral controllers, the proposed controller can improve primary frequency response while adapting to changes in system topologies and conditions.