Modern distribution networks (DNs) suffer from unstable conditions, such as fast voltage fluctuations, due to the intermittent nature of integrated renewable energy sources. An emerging type of power electronics-based controller called the electric spring (ES) is believed to be a promising solution for alleviating this instability by performing voltage/var control (VVC).Existing literature on the ES mostly study circuit topologies and associated control methodologies in the device-level, while some other works demonstrate the utility of the ES in system-level planning. However, decision-making problem of the network operation involving the ES technology has never been formulated.

In this thesis, our first contribution involves proposing AC power flow models for the ES-equipped bus and t...[ Read more ]

Modern distribution networks (DNs) suffer from unstable conditions, such as fast voltage fluctuations, due to the intermittent nature of integrated renewable energy sources. An emerging type of power electronics-based controller called the electric spring (ES) is believed to be a promising solution for alleviating this instability by performing voltage/var control (VVC).Existing literature on the ES mostly study circuit topologies and associated control methodologies in the device-level, while some other works demonstrate the utility of the ES in system-level planning. However, decision-making problem of the network operation involving the ES technology has never been formulated.

In this thesis, our first contribution involves proposing AC power flow models for the ES-equipped bus and the ES-equipped DN. Based on the models, we formulate a centralized problem to solve VVC decisions for the ES-equipped DN. To solve the problem, we propose a second-order cone program relaxation method with exactness guaranteed.

Our second contribution is motivated by the privacy and scalability concerns since the ES is designed to become a household device in future power grids. We distinguish the distribution system operator and the ES-equipped bus agents and formulate distributed optimization problems based on the respective concerns of the operator and the bus agents. The alternating direction method of multipliers (ADMM) algorithm is adopted to guarantee the convergence and computational efficiency.

Numerical tests are demonstrated through a 34-bus distribution network, validating that in both ways, the fast VVC provided by ESs can effectively improve the network properties.