Voltage control in power distribution grids equipped with large-scale renewable energy sources

Abstract

High integration of renewable energy sources (RESs) and battery energy storage systems (BESS) into power distribution networks brings new challenges to grid control schemes. The old assumption of unidirectional power flow (from transmission system to loads) is no longer valid when almost every load bus can become temporarily or permanently a generation bus due to increasing power from renewable sources such as photovoltaic (PV) generation. In fact, a power flow direction can alter at any grid location many times a day, impacting existing voltage regulation schemes. In particular, due to the lack of coordination between different voltage control apparatus, a dynamic change in PV generation can cause the grid voltage magnitude to go beyond acceptable limits. Hence, daily power fluctuations, e.g. caused by moving clouds, can lead to over- or under-voltage situations with a regulation time that takes minutes before the voltage is recovered and the grid is ready for a next, unexpected disturbance. This prolonged and inaccurate traditional regulation process affects electrical devices connected to grids with medium- and large-PV generation. Even more importantly, insufficient voltage regulation can prevent new developments of RESs plants because power utility companies might not allow for more RESs generation if a grid voltage is prone to fluctuate.This research is focused on the behaviour of power distribution systems exposed to large-scale PV generation and BESS facilities. Specifically, the study aims to address questions of how those modern power technologies influence voltages and their control in a distribution network as well as how to improve cooperation between traditional and modern grid regulation devices by means of advanced control strategies. To answer the research questions, the study investigates the real-time performance of a distribution grid voltage control in the presence of dynamically changing large-scale RESs and BESS generation and provides a novel real-time coordinated control to address voltage fluctuations caused by uncertain PV generation. Furthermore, the developed voltage control method is based on a model-driven approach with the real-time controller predicting hundreds of possible regulation trajectories and choosing the one that fulfils optimisation criteria the most. In addition, the multidimensional, nonlinear, mixed-integer optimisation is executed based on parallel-computed time-series grid simulations, which reproduce grid behaviour, incorporate mutual interactions of voltage controllers and account for PV variability. Uniquely, the algorithm also considers autonomous responses from upstream voltage regulation devices and includes them into the solution.Next, the investigation was conducted on real-time models of a semi-rural grid located in South-East Queensland, Australia, with a large-scale PV plant (3.15MWp of power) and BESS facility as large as 0.6MW/0.76MWh. As a result of implementing the proposed control coordination to the studied grid, voltage fluctuations caused by dynamic PV variability are regulated quicker and more accurately in comparison with existing control schemes. This leads to a substantial reduction in voltage variations both in magnitude and time, eliminates unnecessary control operations and prevents voltage hunting conditions. Then, the developed control method has been successfully applied to incorporate BESS devices into the voltage control scheme. In addition, control effectiveness has been analysed with and without BESS to demonstrate advantages of using BESS voltage support in a grid with high R/X ratio. Hence, it has been proven that battery energy storage can play a vital role in a distribution grid control with minimised impact on primary BESS operations (e.g. load peak shaving and energy market contribution). In summary, due to the proposed coordinated control method, it became possible to effectively counteract voltage fluctuations. The solution can be used to enhance PV generation in weak MV feeders, which are prone to suffer from PV power variations.Finally, the thesis also reports on a new real-time co-simulation platform, which combines real-time power system simulations in a RTDS device with advantages of algorithms prototyping in a MATLAB environment. In addition, the RTDS communicates with MATLAB throughout the TCP/IP-based protocol to exchange grid data and control signals. The platform performance tests have been completed with the implemented communication channel to determine platform limits of applicability. As a result of the RTDS-MATLAB combination, advanced controllers, which require computation power offered by modern processors can be validated in real-time even at early stages of their development. The proposed solution offers great simulation and prototyping flexibility and can be broadly used in many voltage control studies – both for wide-area distributed control problems and for individual controller algorithms.