The transition to renewable energy sources requires advanced energy storage solutions to manage their intermittent nature. This thesis explores the feasibility of implementing Battery as a Service (BaaS) for a renewable energy community (RECs) setting, aiming to incorporate this model as part of an existing stacked revenue framework utilized by battery owners across various energy markets. By directly linking battery owners with energy communities, the study shows that renting out battery storage can eliminate intermediary overheads, thus providing financial benefits to both parties. The research addresses two main questions: developing a stacked revenue model for grid-connected batteries including energy communities, and comparing different battery sizing and control methods across various tariff schemes. The findings suggest that the proposed revenue model optimizes energy use and reduces costs for the community. This thesis contributes to the research field by presenting a viable economic model for integrating battery storage into decentralized energy communities.

With the emergence of energy communities, where a number of prosumers (consumers with their own energy generation) invest in shared renewable generation capacity and battery storage, the issue of fair allocation of benefits and costs has become increasingly important. The Shapley value, a solution concept in cooperative game theory initially proposed by Nobel prize-winning economist Lloyd Shapley, has attracted increasing interest for redistribution in energy settings. However, due to its high time complexity, it is intractable beyond communities of a few dozen prosumers. This study proposes a new deterministic method for approximating the Shapley value in realistic community energy settings and compares its performance with existing methods. To provide a benchmark for the comparisons of these methods, we also design a novel method to compute the exact Shapley value for communities of up to several hundred agents by clustering consumers into a smaller number of demand profiles. Experimental analyses with large-scale case studies of a community of up to 200 household consumers in the UK show that the newly proposed method can achieve very close redistribution to the exact Shapley values but at a much lower (and practically feasible) computation cost. Furthermore, it performed similarly to the probabilistic, state-of-the-art approximation method while having smaller time complexity as well as other desirable characteristics for cost redistribution in energy communities.

Peer-to-peer trading and energy communities have garnered much attention over the last few years due to the wider spread of distributed energy resources. Much research has been performed on the mechanisms and methodologies behind their implementation and realisation. However, the efficiency and micro-structure of trading in such markets raise many important challenges. To analyse the efficiency of peer-to-peer energy markets, we consider two different popular approaches to peer-to-peer trading, i.e. centralised and decentralised and explore the economic benefits these models bring given optimal trading schedules computed by a joint schedule optimizer. In both these modes, benefits can be realised mainly due to the diversity in consumption behaviour and renewable energy generation between prosumers in an energy community.
This diversity decreases quickly as more peer-to-peer energy contracts are established and more prosumers join the market, leading to significantly diminishing returns. In this work, we aim to quantify such effects using large-scale real-world data from two trials in the UK, i.e. the Low Carbon London project and the Thames Valley Vision project.
We show that only a small number of peer-to-peer contracts and a fraction of the prosumers are needed to realise the majority of the Gains from Trade.

Renewable energy generation projects are often measured by their peak capacity. A wind farm rated at 25 MW will generate 25 MW of power under the right circumstances. This peak capacity is reached very little in practice. However, these generators are forced to purchase grid operation infrastructure that can handle this peak generation event. The high voltage grid connections are expensive and increasingly difficult to receive permits for. This work presents a solution in which the high voltage grid connection is undersized in comparison to the renewable energy generator. A battery energy storage system is installed in the local grid to solve the issue of excess energy generation (congestion).

A simulation of the local network has been built that models a battery energy storage system (BESS), the network and uses data from a solar park. A case study in which a 19 MW solar park is connected to the high voltage grid with a transformer of only 14 MW as well as a 14 MW | 30 MWh BESS on the network is investigated in the rest of the work. Furthermore a BESS control strategy for price arbitrage on the TenneT imbalance market is presented and encoded such that it can be optimised.

Four heuristics are presented that time and size the congestion issue in a manner the control strategy of the BESS can prepare for and solve congestion when necessary. These heuristics are tested against strategies optimized for revenue maximisation through price arbitrage. While the most aggressive strategies did not solve all the congestion events in these simulations, we found that the heuristic that takes the average generation of the solar park into account performs the best while remaining appropriately conservative.

A basic evolutionary algorithm is presented that optimizes a BESS control strategy for price arbitrage when the BESS is not needed on the local network to solve congestion. Although the strategies earn ~33% less revenue due to the congestion related limitations, the optimisation surrounding congestion does improve revenue by 2.58%.

The results presented in this work suggest that this setup of a local grid can be economically viable and that the BESS can solve the congestion issue when steered with an appropriate control strategy. We hope to inspire parties that battery energy storage systems can earn substantial revenue aside from solving issues on a (local) grid.

Energy communities are not yet fully self sufficient, mostly due to financial factors. Efforts are made to reduce these factors. Communities invest in community-owned assets, which provide more savings compared to individually-owned assets. Prosumers share their loads for better energy distribution, this can provide a significant impact. A good predictor for identifying the financial benefit for a community is the diversity of the consumption behaviors of the prosumers. However, an open question is how the diversity exactly affects the community costs. In this paper, we introduce Two-level K-means, an improvement on K-means, and use it on real consumption data to find energy profiles. We use the energy profiles to model communities, varying in diversity. Finally, we provide an analysis of the affects of diversity on costs. Results from the analysis show that an increased diversity factor can provide financial benefit. This is a result of residual demand being compensated by excess energy generated. However, the added financial benefit depends on the composition of the community.

This paper explores the possibility of using machine learning to improve the profits generated by an energy management system for so called prosumer households. Which are households with their own energy production, consumption, and storage. The commonly used deterministic algorithms only take the current data into account when deciding on how much power to buy or sell. But, if there are flexible energy costs and future energy consumption and production data is available, better choices can be made. For example, you could
store energy at a cheaper time for later usage when the prices are higher. To make these decisions, a new algorithm was created. In this paper different machine learning models are evaluated for generating the future data. The final simulations will use the improved algorithm use the data predicted by these models. For a prosumer with a 10 kWh battery and which produces 70% of its own total energy consumption, the machine learning algorithm decreased the total energy cost by around 7%. While this does show that machine learning
is a viable option to increase the efficiency of an energy management system, there is still much improvement possible to get closer to the theoretical 24% reduction.

In a community energy project, batteries are the asset with the shortest lifespan and are therefore key contributors to cost. Understanding the influence of the battery state of health model on a control algorithm designed for redistribution of benefits in terms of financial gains in a community energy project can help elongate battery lifetime and reduce need for replacement hence minimising costs and reaping environmental benefits. Battery depreciation is predominantly stimulated by cyclic degradation and thus incurred costs are compared by simulating degradation curves for different battery storage systems in terms of chemistry and capacity. Costs are calculated by applying battery models to the control algorithm proposed by Norbu et al. (2021), which factors in cyclic degradation using the rainflow counting algorithm. The experiment explores the influences on cost of different battery chemistry types and capacities. Results demonstrate that lithium-ion batteries, which are the current norm in utility-scale applications, incur the lowest costs. Specifically, lithium manganeseoxide batteries appear to be most effective. Additionally, costs tend to decrease with increasing capacity until a minima corresponding to the optimal battery capacity.

Given the fundamental profit gained by renewable energy assets in climate control, existing control algorithms are urged to be improved to match power supply and demand optimally. This paper explores various designed cases that lead toward an enhanced definition of a control algorithm with optimized behaviour. The core of improvement is exploiting future knowledge, which can be realized by current state-of-the-art forecasting techniques, to effectively store and trade energy.
Based on several thousands of simulations of energy communities in the UK, the proposed smart control algorithm has demonstrated a robust performance and gained notable additional profit in theoretical and practical scenarios using probable data.

The Electric Vehicle Routing Problem (E-VRP) is an extension of the infamous Vehicle Routing Problem which asks which routing decisions an electric vehicle needs to take in order to traverse the network efficiently. Many extensions of this problem have been subject to research in the last decade and now that electric vehicles are starting to pop up in more and more cities, questions are asked about how an electric vehicle should decide which charging station to charge at to minimize their lateness. This problem is of growing significance due to the growth of the amount of electric vehicles and the disruptions which are caused by the longer recharging times of an electrical vehicle when compared to fossil fuel-based vehicles. Since there are countless possibilities and variables to consider in this problem (e.g. the price of electricity or the distance to a charging station), research should be conducted to see which kind of algorithm most satisfies the need of the end-user. To address this problem, this paper proposes several algorithms and compares them to each other based on algorithmic efficiency, average travel time of the vehicles and possible disadvantages when using the algorithms. Through simulations we show that the IARS algorithm as proposed in an earlier paper leads to the overall best performance, but that it lacks efficiency in terms of algorithmical complexity. We also show that when using a shortest path algorithm, the addition of a greedy geometric spanner significantly decreases the time complexity of the algorithm, in some cases reducing the average timespan of the simulation of 1 day by as much as 34%.

Equational reasoning based verification address some of the limitations of classical testing. The Curry-Howard correspondence shows a direct link between type systems and mathematical logic based proofs. Agda is a language with totality and dependent types which makes use of the CH isomorphism to support equational reasoning in its programs. ‘agda2hs’ attempts to bring this formal verification to the Haskell ecosystem, by providing a translation between Haskell and Agda programs. This project will serve to test the viability of this framework by re-writing the Haskell library ‘Data.Map’ in the subset of Agda defined by agda2hs and verifying properties of various functions in the library using Agda and dependent types.

As the popularity of electric vehicles (EVs) increases, congestion at charging becomes a more imminent problem. Congestion at a charging station can lead to long waiting queues and failure of EV owners to charge their vehicles fully before their departure from the station. To combat this issue, this paper explores several candidate scheduling strategies that can be applied for the charging prioritization of EVs at a single station. Through extensive simulations, the efficacy of these strategies is studied under three performance metrics. From the set of strategies studied, we find that earliest deadline first (EDF) and shortest job first (SJF) are the best options in the case that adherence to deadline or a shorter waiting time is most valued, respectively.

With the increasing number of electric vehicles onthe road, the routing problem has become morecomplex. As charging electric vehicles takes longerthan fueling non-electric vehicles, congestion canoccur at charging stations. This might lead to theshortest route not being the fastest route, due tolong waiting times at the stations. By commu-nicating the intentions of each vehicle, they canspread out over multiple stations. This paper in-vestigates the effect of such a routing system on theprofitability of charging stations in comparison toa more simple shortest-path algorithm. In particu-lar, the influence of a charging station’s location onits profitability has been researched for both rout-ing algorithms. In order to do this, a pricing modelhas been developed to extend the routing mod-els used for both the shortest-path algorithm andthe intention-based routing algorithm. Through-out several simulations, it became clear that for theshortest-path algorithm, more centralised stationsobtain a higher profit, whereas for the intention-based routing algorithm there were no significantdifferences in profitability between the more cen-tral stations, and the ones on longer routes.

agda2hs is a project that aims to combine the best parts of Haskell and Agda by providing a common subset between them. It allows programmers to im- plement libraries in Agda, verify their correctness and then translate the result to Haskell so they can be used by Haskell programmers. In this paper, a verified Agda implementation of the Ranged-sets Haskell library is provided, using agda2hs. In or- der to produce a verified implementation of this li- brary, we proved its preconditions, invariants and properties.

Intention Aware Routing System is a route-planning algorithm for electric vehicles that minimizes overall travel time by taking into consideration congestion at charging stations. This paper extends this algorithm to allow choices to be made based on prices at charging stations. The goal of this paper is to find a way to minimize maximum congestion while maximizing overall profit across the stations.

Formal verification works better than testing, since the correctness of a program is proven. It is researched if it is possible and feasible to formally verify the Inductive Graph Library. The library is an abstract class in Haskell and is ported manually to Agda. Agda is a total and dependently typed language and thus can be used as a proof assistant. The functions are first converted to total functions and the preconditions the are proven. Verifying an abstract class is time consuming, since it requires an implemented instance of the abstract class. Stating the properties of the library is possible, but difficult since it requires generalised properties that need to be valid for all instances of the abstract class. The verification process is not completed yet, so no definitive conclusions are made, but the properties that are verified did not produce any issues. agda2hs is used to compile the Agda code to Haskell, this ensures that the verified Agda code is verified Haskell code. This requires the library to fall within the common subset of agda2hs.

This paper introduces an electricity price extension to the intention-aware routing system (IARS) for electric vehicles (EV). The existing intention-aware routing system is used to route electric vehicles who require to charge en-route through a road network. To achieve the objective of minimising the average journey time, the intentions of EVs and waiting times at charging stations are com- municated. Instead of only minimizing travel time, the model extension presented in this paper makes it possible to express a decision trade-off between price and time. A vehicle computes its routing policy such that the combined utility of price and time is as high as possible. In this paper the performance of IARS with a price extension is compared to a greedy maximising algorithm (MAX) in several settings. The increase in utility by using IARS depends on the population of electric vehicles. However, in most experiments conducted in this research IARS achieves a significantly higher average utility.