Control of DC microgrids for distributed generation including energy storage

Abstract

In recent years DC microgrids have attracted significant interest in research, with various literature published covering areas such as system design, control systems, and energy management. This interest can be attributed to the increase in use of DC energy generation systems, such as photovoltaics (PVs), and energy storage systems, such as battery banks. The aims of this research were: (a) to build a laboratory-based DC microgrid system on which control algorithms can be developed, applied, and tested, (b) to implement and improve the primary control system, and (c) to develop a battery management system. A lab-based 48V DC microgrid consisting of two 2.5kW Buck converters, a 1.5kW Bidirectional converter connected to a 24V battery bank, and a resistive load bank was built. Detailed modelling of the Buck and Boost converters was performed, deriving the small signal models and the transfer functions for both converters, which were needed to design the control systems. The two Buck converters and the Bidirectional converter were designed, built, and tested. The converters were connected in parallel and shared a common resistive load using droop control. Three droop control methods were implemented and tested: V-I droop, I-V droop, and a newly proposed method called Combined Voltage and Droop (CVD). A battery management system (BMS) was developed to provide high-level control to the Bidirectional converter. The Buck and Bidirectional converters operated successfully both as standalone units and within a DC microgrid configuration. V-I droop control provided correct current sharing capability with good results, however its load sharing response was slower than CVD control. I-V droop control resulted to be unstable during practical implementation due to the high gain and bandwidth of the voltage control loop, which interacted with the bandwidth of the anti-aliasing filter in the voltage feedback path. The proposed CVD method solved the instability issues experienced by using I-V droop, making the control system work in a stable way by providing a means to adjust the bandwidth of the voltage control loop. Successful operation was also attained from the DC microgrid setup, which was operated in two scenarios: (1) with the two Buck converters and the Bidirectional converter (operated as a Boost converter) all sharing the resistive load among them, and (2) with the Bidirectional converter (operated as a Buck converter) charging the battery bank and the two Buck converters supplying the load current and input current of the Bidirectional converter. The BMS was successfully tested with simulations, utilizing the load current and state of charge (SOC) of the battery bank to select the mode of operation of the Bidirectional converter among battery charging, load sharing/supplying, and idle modes. Through this project, an experimental lab-based DC microgrid was built, serving as a valuable setup for further research on control algorithms of renewable energy conversion systems. By using the lab-based DC microgrid, the new CVD droop control method was developed, which offers advantages over the other droop methods. The BMS developed set up the basis for further development in the area. From this thesis there are two main contributions: (1) the novel droop control method (Combined Voltage and Droop method) which offers an alternative to the standard I-V droop control system, and (2) an algorithm for a BMS which provides simple but effective control of the Bidirectional converter and its storage system.