Control in stand-alone microgrids

Abstract

Power systems are shifting from a centralized to a distributed generation paradigm. As more and more distributed generators, such as renewable energy sources and microturbines, are installed, the new power generation paradigm has led to the introduction of microgrids. Stand-alone microgrids generally refer to small-scale localized medium or low voltage power networks temporarily disconnected or permanently isolated from the utility grid. Their unique characteristic and the uncertainties of renewables introduce unprecedented challenges for the stable operation of these microgrids. This thesis proposes different control schemes to overcome these new challenges.

Firstly, model predictive control with demand response is employed to control stand-alone microgrids with synchronous generators. Since synchronous generators and conventional equipment such as load-tap changers may not be capable of regulating the microgrids under the fast-fluctuating power outputs of renewables, electric springs are implemented to engage non-critical loads in providing grid-ancillary services. Model predictive control is used to coordinate synchronous generators and electric springs in voltage and frequency regulation while minimizing the impact of demand response on non-critical loads.

Secondly, a distributed consensus-based dispatchable energy source powered inverter (DESI) control method is proposed for fully inverter-based stand-alone microgrids. In view of issues such as the conventional reactive power-voltage magnitude droop control method failing to accurately share reactive power among DESIs due to a mismatch in output impedance, and inaccurate reactive power sharing may result in inverter overloading and circulating current among them, a consensus-based DESI control method is proposed to improve the reactive power sharing accuracy.

Thirdly, a distributed feedback optimization-based control framework for grid-forming inverters (GFMIs) in fully inverter-based stand-alone microgrids is proposed. Historically, GFMIs adopt a hierarchical control strategy with the optimization process implemented in the slowest control layer. This control strategy may not drive the microgrids fast enough to follow the optimal operating point, which may change quickly due to the fast-fluctuating power outputs from renewables. To solve this issue, feedback optimization is employed to optimally control GFMIs in a faster time scale and drive microgrids to stay closer to their optimal operating points.

Fourthly, a unified control method for DESIs and renewable energy source powered inverters (RESIs) is proposed under the framework of distributed feedback optimization. DESIs and RESIs are generally controlled separately as GFMIs and grid-following inverters (GFLIs). The active power outputs of GFLIs are usually controlled to follow the maximum power point tracking profiles, with the reactive power outputs driven to zero. This control strategy neglects their reactive power support capability. To coordinate DESIs and RESIs and unleash RESIs’ reactive power support potential, the microgrids’ control problem is formulated as a single optimization problem, with the decision variables being the control input of the microgrids. Feedback optimization is used to solve the optimization problem in a distributed manner.

In summary, this thesis proposes different types of control methods, such as model predictive control, consensus-based control, and feedback optimization-based control, to achieve different control targets for stand-alone microgrids.