Local flexibility markets for distribution network operation

Abstract

The continuing replacement of fossil fuel-based power generators with renewable intermittent sources like wind and solar has increased the power system’s need for flexibility. Previous research has shown that some of these additional flexibility demands can be met by demand response of residential customers. However, managing residential power consumption for system purposes, i.e. balancing consumption and generation, synchronizes loads and can cause congestion in distribution networks. At the same time, newly installed distributed energy resources, such as electric vehicles, photovoltaic panels and electrical heating systems, are increasing residential power demand. Recent advances in information and communications technology allow distribution system operators (DSOs) to better observe these developments, and could enable them to apply alternative strategies to the classical approach of network expansions, which is becoming increasingly costly in the light of artificial load synchronization. To this end a congestion management mechanism is needed to operate distribution networks cost efficiently in the future. Local flexibility markets have the potential to unlock DSOs access to flexibility resources. Flexibility services could then be used to mitigate potential congestion and increase overall network efficiency and reliability. This thesis addresses issues related to the introduction of local flexibility markets in three parts. The first part examines how reactive power control of inverter-based distributed energy resources can be implemented under current grid codes such that it is compatible with local flexibility markets. Based on this examination, a method for designs of unit-specific local reactive power controllers is proposed, which is able to take distribution network topology and individual customer characteristics into account while requiring no additional information and communication infrastructure. The approach proves robust against overvoltages in a distribution network with a high share of solar generation. Compared to similar approaches, the method is easily scalable, and can thus be used in large distribution networks and may be applied to other inverter-based equipment such as electric vehicles or heat pumps. The second part of the thesis discusses tools and methods which DSOs need to participate in local flexibility markets. Two essential functions are developed which allow DSOs to quantify the benefit of flexibility services. First, a method is proposed to calculate the expected gains based on a probabilistic distribution network analysis. The method is tested in a real implementation of a local flexibility market with an aggregation of 800 thermostatically controlled loads in the context of the EcoGrid 2.0 project. Results show that flexible consumption can be shaped to deliver reliable flexibility services to DSOs, despite the presence of large uncertainties. Services can act as an insurance policy for DSOs against network congestion and power outages, thus preventing such rare events and related social cost. Second, a method is proposed to create flexibility models for aggregators and DSOs, with a particular focus on aggregations of thermostatically controlled loads. DSOs require such models to evaluate the financial benefit and reliability of flexibility options. The proposed method is based solely on aggregated behind-the-meter power consumption data which is generally available to DSOs, and thus requires no further privacy-intrusive customer data. The impact of aggregation size on baseline uncertainty and flexibility model accuracy is analyzed in detail. The flexibility model is applied to predict the response of an aggregation of thermostatically controlled loads in a series of verification experiments, which reveal that it is able to predict the aggregated behaviour with good accuracy. Finally, in the third part of the thesis, design questions pertaining to local flexibility markets are addressed. Flexibility services of local flexibility markets need to be specifically designed for DSO needs such as voltage control, substation congestion management, loss minimization or load reductions during planned maintenance. Flexibility services are commonly defined by two different approaches: Baseline services define an expected behaviour or schedule, a baseline, and quantify power deviations from that baseline. By contrast, capacity limitation services set temporary power consumption caps, and penalize exceeding the cap. Based on experience from the large scale demand response field trial in the EcoGrid 2.0 project and a review of existing proposals, a list of essential requirements for flexibility services is developed. Flexibility services need to fulfil these requirements to represent an effective and reliable option for DSOss. However, baseline flexibility services are shown to fail certain requirements. Thus, local flexibility markets are recommended to employ capacity limitation services as a primary product. To handle clearing of flexibility markets with these services, a market clearing method built upon an adapted Vickrey-Clarke-Groves-based auction is proposed. The mechanism provides a balanced trade-off among desirable economic properties, including budget-balancedness, incentive-compatibility and stability. Results of a case study of a medium voltage feeder show that aggregators and the DSO benefit from trading capacity limitation services under the proposed setup and accordingly would want to participate. Implementing such a market would allow the DSO to defer infrastructure investment while ensuring secure operation. Overall, this thesis presents a coherent setup for local flexibility markets. A functional framework for DSOs is developed, allowing them to include flexibility options in their distribution network operational strategy. The framework was tested and demonstrated in a series of large-scale demand response experiments on the Danish island Bornholm. Experimental results show that flexibility services represent a reliable option and can be delivered and verified with limited information and simple information and communication infrastructure. The real-life demonstration of a local flexibility market raised questions regarding the nature of flexibility services, and it was found that capacity limitation services are best suited to address issues in the distribution network. Finally, an effective market mechanism for capacity limitation services with desirable properties is proposed. The thesis provides insight into the operation of local flexibility markets and proposes solutions to encountered barriers, thereby supporting the real-world establishment of such markets.