Browsing by Author "Farmer, Warren"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    A novel approach to identify frequency transient (un)stable sub-networks in low-inertia power systems with high renewable energy penetration
    (Stellenbosch : Stellenbosch University, 2021-12) Farmer, Warren; Rix, Arnold J.; Stellenbosch University. Faculty of Engineering. Dept. of Electrical and Electronic Engineering.
    ENGLISH ABSTRACT: Due to global environmental concerns, power systems are transitioning from conventional to renewable energy sources (RES). The integration of large-scale inverter-based RES introduces numerous challenges to power system stability, specifically frequency transient stability. The power system becomes more complex and will continue evolving in this direction. Thus, the conventional assumptions and approaches to evaluate the transient stability of the power system frequency become obsolete. The focus of this dissertation is on the impact of network topology on frequency transient stability. This research considers the network's spatial element and the influence of spatially distributed variable inverter-based RES. The power system network is the medium connecting all the generation and load units. Unlike active mitigating measures like virtual inertia control, the network topology can serve as a passive means to maximise frequency transient stability, i.e., the network can form an embedded noise filter. The approach involves Graph Theory and state-space domain representation to model and analyse the power system topology's impact on frequency transient stability. The analytical results undergo validation through simulations of practical and realistic power system networks (IEEE systems) using the DigSilent Power- Factory simulation package. Modelling the power system network as a graph made it possible to make the following contributions to the research field: 1) Give insight into the short time scale (below one-minute temporal res- olution) stochastic wind speed signal as input disturbance (noise signal) to the network. Wind speed increments fit a Gaussian function, and a Markov chain can be used to describe the temporal characteristics. The Markov chain shows that the fluctuating perturbations are biased to continue in the same direction as the previous increment, which is a characteristic of the short time scale temporal nature of wind fluctuations. 2) Understanding the spatial-temporal perturbations from the power system network's perspective due to inverter-based RES utilities. Using high temporal resolution wind speed data shows that the spatial cross-correlations of wind speed fluctuations are low, and due to the transmission line reactance, the impacts of the fluctuating perturbations are mainly within the local network. 3) Identify and understand the topological measures contributing to local- area frequency robustness. An analytical investigation into the graph Lapla- cian facilitated the derivation of the network metrics relevant to transient stability. These network metrics are verified through simulation and shows that the weighted node degree, hierarchical position, and effective reactance to the perturbation location are relevant. The impact of an imbalance disturbance propagates with mechanical wave characteristics, and the derivation shows that the propagation velocity is a function of inertia and transmission capacity. The topological measures influence the impact propagation, and they contribute to how the network dissipates the disturbance impact energy. 4) The correlation between stochastic generation utilities, like wind or solar photovoltaic, is illustrated and shown to be a crucial limiting factor consider- ing large-scale inverter-based RES integration. The simulation results verify that the rate-of-change-of-frequency (RoCoF) is inversely proportional to the effective reactance to the perturbation location. It is thus worth considering the effective reactance between stochastic generation and inertia sources in low-inertia power systems when planning the placement of inverter-based RES to minimise the excitation of the RoCoF responses. 5) A new approach to evaluate power system inertia for frequency transient stability, based on the previously mentioned ndings, is proposed to identify and group nodes/buses to form local inertia areas. Spectral Graph Theory and clustering serve to be valuable in identifying local inertia areas based on network regions that maximise the within-cluster mesh and group nodes with coherent dynamics. These inertia areas provide a spatial awareness of a power system network's frequency transient stability. The proposed approach makes it possible to determine where inertia or other equivalent ancillary service support is of greater importance and thus optimise the inertia distribution to maximise system frequency robustness in low-inertia power systems. 6) Contributing to understanding the concept of optimising a power sys- tem network in terms of topology, real and virtual inertia placement, and the distribution of stochastic generation utilities to maximise system frequency transient stability. These research contributions ultimately point to converting distributed noise input from spatially distributed and stochastic RES generation to a robust and stable frequency signal through the topology of the power sys- tem network to help network planners design optimal, RES rich, and stable networks.