Distributed Generation Capacity Assessment of Active Distribution Systems
Date
2023
Authors
Mahmoodi, Masoume
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Significant decarbonisation (reducing the intensity of carbon emissions) in different economic sectors, such as electricity, is required to mitigate the effects of climate change. Driven by this, our electricity systems are undergoing a massive transition from centralised fossil fuel-based generation to renewable energy sources. However, increasing uptake of renewable-based distributed generation (DG) has caused technical issues such as over-voltage and line/transformer overloading in electricity distribution systems. These issues, in turn, deteriorate the ability of the grid to accommodate more DG. This motivates a DG capacity assessment which determines the amount of DG installations that maximises the overall DG generation without causing any technical issues, which is the main focus of this thesis.
We first introduce the DG capacity of an unbalanced distribution system as a range rather than a single value. We develop two optimisation-based models to obtain the lower and upper bounds of the network's DG capacity. We also evaluate the impact of the fairness element and active network management (ANM) schemes on improving the network's capacity. Finally, we investigate the effectiveness of different standard local ANM schemes and compare them with a central approach that assumes a live, two-way communication as an upper limit benchmark for the DG capacity.
Next, we deal with the data uncertainty (e.g. load and DG output uncertainties) in the DG capacity assessment problem. To this end, we build our DG capacity assessment problem on the paradigm of robust optimisation (RO) methodology. In RO, the range of possible values for uncertain parameters is defined as an uncertainty set. Then the decisions are determined to guarantee the results under all realisations in the uncertainty set. We develop polyhedral uncertainty sets that account for strong spatial correlation between renewable energy sources within a given geographical area, as well as the coincidence factor for the electrical demands in the network. The coincidence factor represents how likely the individual demands peak simultaneously. We then use the adjustable robust counterpart (ARC) methodology to extend our robust DG capacity assessment accounting for the capability of fast-acting devices, e.g. inverters, to respond to demand and DG output uncertainties.
However, since the RO-based approaches do not utilise the probability distributions of the uncertain parameters, they opt for an over-conservative solution. They, therefore, demonstrate a poor out-of-sample performance when the uncertainty realises away from the worst-case scenario. Thus we explore the use of an alternative approach, namely distributionally robust optimisation (DRO), to tackle this issue. In DRO approaches, the actual but unknown distribution of the uncertain parameters is assumed to belong to an ambiguity set. Then, the decisions are determined to guarantee the results under the worst-case distribution in the ambiguity set. We develop a data-driven adjustable distributionally robust DG capacity assessment model considering ANM techniques in three-phase distribution systems.
Finally, we account for the consumers' privacy to decide about their operation/investment decisions while guaranteeing the network's security through a dynamic operating envelope (DOE) received from the distribution system operators. A DOE is a convex set that defines acceptable real and reactive powers at a consumer connection point without violating network constraints. We propose a novel methodology to calculate a DOE by applying the right-hand side decomposition technique through which we allocate the network's capacity to mutually independent consumers. The consumers are then free to make their decisions as long as they comply with their DOEs. We then build a hierarchical DER capacity assessment within DOEs that considers the uncertainty of solar power and demand when making investment decisions.
Description
Keywords
Citation
Collections
Source
Type
Thesis (PhD)
Book Title
Entity type
Access Statement
License Rights
Restricted until
Downloads
File
Description
Thesis Material