Distributed load-side frequency control of power systems

Abstract

One trend is arising in modern power systems: increasing penetration of renewable power (e.g., wind and solar). Although it is very attractive, the integration of renewable generation may also lead to notable challenges to power systems due to its uncertainty and intermittency. Consequently, random and rapid fluctuations in power supply may occur which will severely affect the stability of power systems and especially the system frequency. To tackle this problem, the adoption of controllable loads for frequency control has drawn a lot attention. A typical power grid consists of three power network layers, namely transmission, subtransmission and distribution networks. The current load-side control methods for transmission, subtransmission and distribution networks are designed separately and in principle should be coordinated in a granular manner. The goal of this thesis is to design granular control schemes for controllable loads to cooperatively implement frequency regulation in granular and distributed ways. Specifically, it focuses on two concrete aspects. The first focus is to design control schemes for transmission networks with single control area and multiple control areas using load-side controllers in cooperation with generation-side control to regulate the system frequency and the tie-line powers between control areas. A switched consensus-based distributed controller is proposed for each load bus in the single-area system and multi-area system, respectively. Load-side controllers in both of the two cases will start to communicate with neighboring controllers to discover the power imbalance of the corresponding control area once the frequency violates a pre-defined threshold. The load-side controllers will work in a frequency regulation mode (FRM) and adjust power consumption of controllable loads accordingly in the case of the single-area system. Differently, in the case of the multi-area system the control centre in each control area will send the tie-line power information to all the load-side controllers in the area, which together with the power imbalance of the area determines power consumption of controllable loads. After the frequency goes back to a satisfactory range, load- side controllers will switch to a load recovery mode (LRM) and shift their duties to the generators. Different load recovery strategies are designed for the single- area system and multi-area system, respectively. Case studies show that the performance of frequency and tie-line power regulation is improved significantly by the proposed control schemes compared with traditional generation-side control and with minimized impacts on end-users. Based on the results obtained from the first problem, the second focus is to develop granular control schemes to achieve load-side control algorithms across transmission and subtransmission networks. Since loads in transmission networks are usually the aggregation of subtransmission networks and loads, the aforementioned control schemes for aggregate controllable loads in transmission networks need to be granulated down to subtransmission networks. However, not only frequency but also bus voltages will be affected by active power changes in subtransmission networks due to a higher R/X ratio of transmission lines. Thus, electric spring (ES) aggregators consisting of ESs (a new electronic device) associated with noncritical loads are proposed to implement frequency and voltage regulation simultaneously. Various algorithms are adopted to dispense control references from load-side controllers on transmission levels to ES aggregators on subtransmission levels to provide the required active power support and regulate frequency cooperatively. Also, different control methods are used for ES aggregators to regulate bus voltages in the meantime. Power consumption of each ES aggregator is then adjusted accordingly to conduct frequency and voltage regulation simultaneously. The costs for loads participating in frequency and voltage regulation are considered to be minimized using a distributed optimization algorithm. The robustness of the proposed granular control scheme with respect to communication delays is also studied. Simulation results show that ES aggregators can achieve the required active power response and regulate frequency cooperatively, and maintain bus voltages within the acceptable range under the proposed granular control schemes.