A techno-economic analysis of domestic thermal energy storage systems from a suppliers’ standpoint and their value for the UK electricity market
Student: Hisham Abunassar
The United Kingdom is committed to decarbonise the heating sector to reduce greenhouse gas emissions by 80-100% by the year 2050. This requires the immediate elimination of natural gas, and the swift shift towards electricity-based heating. The project aims to analyse the techno-economic performance of domestic thermal energy storage technologies by developing a business model that allows energy suppliers to control these assets at a broader system level to ensure domestic space and water heating self-sufficiency, capture benefits over the entire electricity supply chain, and generate profits in the process.
Supervisors:
- Professor Graham Hughes, Department of Civil and Environmental Engineering
- Johan Du Plessis, Tepeo
- Chris Carver, Tepeo
Economic potential and development optimisation of electrolysis for interseasonal energy storage
Student: Antoine Crepel
The aim of this project is to model the economic potential of electrolysers operating on arbitrage in order to help in prioritising their technical development. We have designed a model to assess the dispatch and revenues of inter-seasonal storage, based on historic power prices in the UK. We are using this to provide insights on the value to investors of increased efficiency, current density, cell voltage, durability and storage capacity, relative to further efforts to reduce capital cost. This could help stakeholders of the hydrogen sector prioritise the technical development of electrolysis and foster clean hydrogen scale-up.
Supervisors:
- Dr Iain Staffell, Centre for Environmental Policy
A Reinforcement Learning framework for battery energy trading under real-time pricing
Student: Yann Delclos
Battery costs have been rapidly reducing making them an increasingly interesting investment and source of value on the grid. At the residential level, customers facing real-time pricing can use energy storage to trade electricity and generate profit while homogenising the system load. However, to secure reliable revenues, the trader must make optimal decisions about when to charge/discharge the battery under the uncertainty of future energy prices. The project presents an algorithm to maximise home battery revenues by optimising the timing of its charge/discharge cycles. The problem falls within the purview of optimal decision-making over time and is solved via Reinforcement Learning.
Supervisors:
- Dr Fei Teng, Department of Electrical and Electronic Engineering
- Cris Lowery, Baringa Partners
- Remy Nguyen, Baringa Partners
Net Imbalance Forecasting with Recurrent Neural Networks
Student: John Long
This project aims to utilise this relevant data and advances in deep neural networks to predict the net imbalance volume of the UK national gird. The implications of this are to also be discussed with the aim of providing greater clarity for policymakers, market participants and system operators.
Supervisors:
- Andy Hadland, Arenko Group
- Jose Ortiz de Lanzagorta, Evergreen Innovations
- Dr Sam Cooper, Dyson School of Design Engineering
Development of an optimal economic dispatch tool for Battery Energy Storage Systems
Student: Carlos Rene Millan Medina
Looking into the future energy grid, a considerable amount of effort is being focused on making it reliable, flexible and efficient. With the integration of intermittent renewable technologies, Energy Storage Systems (EES) have taken the spotlight in hopes of facilitating the transition. However, prices for EES and batteries, in particular, have stalled the deployment of the assets. To increase the interest of investors, developers and consumers this project aims to develop a tool for Battery Energy Storage Systems to maximise the value of the operations and transactions made between the system and the energy markets.
Supervisors:
- Dr Fei Teng, Department of Electrical and Electronic Engineering
- Cris Lowery, Baringa Partners
- Remy Nguyen, Baring Partners
Recurrent Neural Networks for energy price forecasting in energy storage applications
Student: Ioannis Skoupras Giannikis
The project is concerned with the development of sophisticated deep learning algorithms for the purpose of predicting elements of the UK energy market and making more educated trading decisions. The interest is focused around the Balancing Mechanism, a complicated market structure that is fundamental for flexible asset operation and the matching of supply and demand in real time. Machine learning has become an essential tool for every market participant, since it ensures the optimal utilisation of assets and, ultimately, more pronounced economic benefits. The project uses Recurrent Neural Networks due to their suitability for time-series data with long-term dependencies and seasonal nature.
Supervisors:
- Dr Samuel Cooper, Dyson School of Design Engineering
- Andy Hadland, Arenko Group
- Jose Ortiz de Lanzagorta, Evergreen Innovations
Flexibility and Demand Side Response in London's Energy Market: A Stakeholder Analysis with KTAB
Student: Alexander Zuend
Flexibility is vital to the operation of electricity systems as it keeps supply and demand in balance. Until recently this was a fortunate by-product of thermal power generation using fossil fuels as a convenient storage medium. The rapid uptake of renewables poses a major challenge to this model. Currently, we see multiple approaches to Flexibility and DSR. Several pilot projects are funded by national and local governments; and new business models are currently developed by DNOs, Retailers and Innovators. The question of this study is thus: 'Who is going to reward people and businesses for domestic /small-scale DSR in London?'
Supervisors:
- Dr Mark Workman, Energy Futures Lab
- Dr Brian Efird, KAPSARC