Grid-friendly and autonomous operation of distributed generators in power grids

Abstract

Power grids are evolving towards distributed and islanded small-scale grids. A small-scale grid can be cast as a network of distributed generators (DGs) that are integrated through power electronics converters. This brings many converter control-related issues as well as opportunities for power grids. The current mainstream studies are primarily focused on resolving the issues that power grids suffer from through the development of grid-friendly services of DGs, but in stark on (i) what issues of grid-friendly control imposing on DG themselves and (ii) what opportunities of DGs providing for mitigation of potential issues.

In this thesis, we identify the issues associated with DGs applied in power grids (whether they are grid connected or disconnected), unravel the cause-effect relationships and seek out mitigation solutions for the potential issues. Our contributions are outlined as follows.

First, we conduct an overview on the grid-friendly control methods of DGs and evaluate the limitations of frequency support from DGs when certain control method engages. These results are valuable to prevent a deteriorated frequency response from DGs, as well as to expand the range of stable operation for DGs and power grids.

Second, we establish basic concepts of electromechanical oscillations (EMOs) that grid-friendly DGs are involved in, and provide countermeasures to mitigating such EMOs. We develop a linearized DG model for wind turbines as an example, which preserves the dynamics of rotor speed and output frequency. We explain the negative damping of the electromechanical time-scale control in DGs imposed on their rotor-speed and output-frequency dynamics, followed by the development of EMO damping strategies. Analysis results and the effectiveness of damping control are validated by simulations in single-machine and multi-machine systems.

Third, we optimize the autonomous operation control (AOC) of multiple DGs-based power grids. Recalling the requirements on DGs’ autonomous operation in islanded power grids, we design a novel AOC (NAOC) on local active power-frequency (P−f)/reactive power-voltage magnitude derivative (Q–dot(V), where dot(V) is the time derivate of V) droop mechanism. The proposed NAOC is capable to maintain accurate reactive power sharing amongst DGs, compared to conventional AOC that is built on P−f/Q−V droop concept. To address the inferior performance of NAOC under loosely interconnected grids and large load perturbations, we develop an optimization-based AOC (OAOC) that is implemented according to P−f/Q−dot(V) mechanism and neighboring communication. The performance and advantages of NAOC and OAOC are verified by simulations in two islanded systems under a wide range of scenarios.

Fourth, when the connected power grid goes black, we propose a black-start control and sequence for DGs to serve in a “build-up” grid restoration, and a superseding islanded operation control to stepwise pick up local loads. We develop a modified PLL control to form a synchronous reference frame for DG control and a field control to smoothly build DGs terminal voltage. Also, we design a strategy for the self-sustaining DGs-based grid to resynchronize with a utility grid that has been restored. Using simulations, the proposed control strategies demonstrate the capability and effectiveness of DGs to perform local black-starts.