Electric spring applications in smart grids : from modeling to control

Abstract

In the future smart grids, renewable sources, dominantly photovoltaic (PV) and wind generations, will be largely deployed to reduce the carbon emission and improve the energy availability. Such a dramatic change will bring new technical challenges due to intermittent nature of renewable energy. The traditional control paradigm of generation following demand has been proved to be insufficient in tackling the new challenges. A new control concept of demand following generation must be explored, as the penetration of wind and solar keeps increasing. In such a view of demand-side management, power system researchers and engineers have responded to the new challenge with various demand response methods. Based on power electronics, electric springs (ES) have been proposed as a fast demand response technology to realize the demand-follow-generation control paradigm. An ES is typically connected with a noncritical load to form a smart load. Such a smart load is capable of adapting its own power consumption to enhance the stability of grid voltage or frequency. With various structures and control methods reported, ES technology has received increasing attention. However, some gaps once neglected in the rapid development of ES require to be addressed immediately. Specifically, a versatile but simple dynamic model of ES is imperative for complex analyses, as the previous ES model only suits for original ES with limited functions. Also, the power flow of ES and the smart load should be closely examined, as the existing works often ignore the energy limitation on the ES when utilizing its real power exchanging ability. In this thesis, an in-depth investigation is conducted on these issues. Chapter 3 presents a modular ES model incorporating the dynamics of ES, controller design, and noncritical load characteristic. A second-order model of ES is simplified based on theoretical derivation and parametric estimation. The modular approach further allows the controller and the load modules to be designed independently but combined with the ES dynamic model in the d-q frame. Various demonstrations verified the accuracy and usability of the proposed model. In Chapter 4, a smart load with a novel configuration integrating ES and PV systems is proposed. To reduce supply-demand power imbalance, the possibility of adjusting the power consumption of this smart load while delivering maximum PV power is investigated, while battery storage is not a necessity in the proposed system. Validated by simulated comparisons and experimental test, the proposed system can be a solution better than energy storage and other ES-based ones. In Chapter 5, a coordinated battery management system (BMS) is proposed and tested on a Shunt ES. A hybrid battery model is extracted for state-of-charge (SOC) control. Verified both in simulation and experiment, this proposed BMS successfully coordinates the real power for frequency stabilization and SOC control by using variable weights. Such a BMS fills the vacancy of SOC control in existing ES technologies. It can be concluded from the results in this thesis that the ES is further developed as a demand response technology for the emerging smart grids.