Markov chain optimisation for energy systems (Ext.)
Lead Research Organisation:
Queen Mary University of London
Department Name: Sch of Mathematical Sciences
Abstract
This is an extension of the Fellowship: 'Optimal Prediction in Local Electricity Markets'. In this project we will develop novel approaches to the optimisation of energy systems under uncertainty. Our approach, based on methods of computationally intensive statistics, offers significant advances on multiple fronts relative to the state of the art. Firstly more detailed and appropriate representations of random variations will be made possible, to address the increasingly important question of the integration of renewable power generation. Secondly we will apply cutting-edge approaches in computationally intensive statistics to reduce the computational time required for the optimisation of energy systems under detailed models of uncertainty, and to develop methods capable of scaling up to large power systems. We will work together with both established and start-up energy companies in the UK to maximise the potential impact of our work. The developed methods will be general in their applicability across energy systems and this research will also support the technical development in the UK of heat networks, a potentially efficient method of delivering water and space heating to multiple buildings. Our research therefore offers multiple contributions to the 'Energy trilemma' of delivering affordable, clean and reliable energy and to the COP21 agenda.
Planned Impact
Electrical power systems are said to be the largest industrial systems created by humankind. The necessary equilibrium between generation and demand is maintained by a complicated control system, whose modelling and optimisation is a high-dimensional problem, and the UK spends approximately GBP 14 billion per year on power production. While the scale of potential economic impacts is therefore clear, there are also potential environmental impacts. Current industrial approaches to power system optimisation, based for example on mixed integer programming, are appropriate for traditional highly predictable thermal generators and predictable, passive demand. When uncertainty is significant due for example to economic feedback or the complex spatial and temporal properties of wind generation forecast errors, however, deterministic formulations can lead to solutions having both higher real-world operational cost and a greater requirement for capital-intensive and potentially highly polluting balancing reserve such as diesel farms, when compared to stochastic formulations. Such uncertainty is set to increase as UK market incentives sharpen and storage technologies evolve and as power systems incorporate further renewable generation following the recent COP21 Paris Agreement governing carbon dioxide reduction measures from 2020. More powerful approaches to the stochastic optimisation of power systems, based for example on modern computationally intensive statistics, are therefore increasingly important.
The extent of potential impacts from research in mathematical optimisation under uncertainty has been highlighted by a 2011 report of the Department of Energy (DOE, see Alexander (2013), Case for Support). As I understand it this is due to be re-asserted in the coming 2016 report "Analytical Research Foundations for the Next-Generation Electric Grid" of the US National Academies. The DOE report highlights the following research priorities:
* the modelling of non-Gaussian noise processes,
* the development of scalable algorithms to address the interaction of uncertainty with integrality constraints,
* development of methods capable of exploiting problem structures such as networks, and constraints of multiple types.
All of the above priorities are addressed in the present proposal.
There are a large number of particular power systems optimisation problems, with a range of timescales (near real-time security, short-term efficiency, mid-term maintenance and natural resource planning and long-term capital planning) and objective functions (average cost, reliability, emissions). We will provide a methodology based on MCMC potentially applicable in all the above contexts. In this way our work will be directly applicable to questions including the value of flexibility, adequacy of generation reserve levels, and suitability of the generation mix, in power systems and in heat networks.
A key part of delivering these impacts will relate to implementation and software. We will therefore take care to perform algorithm development using standard statistical libraries for MCMC as far as possible, and otherwise to make code available openly on our web pages. Further there are a number of initiatives worldwide promoting both open energy data and modelling (including OpenEI, Open Power System Data, and openmod), and we will pursue opportunities to contribute to these forums. In industrial collaborations we will put in place appropriate symmetric agreements relating to intellectual property (IP), ensuring that outputs from our own research are owned by our institutions while enabling partners to retain their proprietary IP. In this way we will ensure that our own research outputs remain available to be used widely in order to maximise their impact.
The extent of potential impacts from research in mathematical optimisation under uncertainty has been highlighted by a 2011 report of the Department of Energy (DOE, see Alexander (2013), Case for Support). As I understand it this is due to be re-asserted in the coming 2016 report "Analytical Research Foundations for the Next-Generation Electric Grid" of the US National Academies. The DOE report highlights the following research priorities:
* the modelling of non-Gaussian noise processes,
* the development of scalable algorithms to address the interaction of uncertainty with integrality constraints,
* development of methods capable of exploiting problem structures such as networks, and constraints of multiple types.
All of the above priorities are addressed in the present proposal.
There are a large number of particular power systems optimisation problems, with a range of timescales (near real-time security, short-term efficiency, mid-term maintenance and natural resource planning and long-term capital planning) and objective functions (average cost, reliability, emissions). We will provide a methodology based on MCMC potentially applicable in all the above contexts. In this way our work will be directly applicable to questions including the value of flexibility, adequacy of generation reserve levels, and suitability of the generation mix, in power systems and in heat networks.
A key part of delivering these impacts will relate to implementation and software. We will therefore take care to perform algorithm development using standard statistical libraries for MCMC as far as possible, and otherwise to make code available openly on our web pages. Further there are a number of initiatives worldwide promoting both open energy data and modelling (including OpenEI, Open Power System Data, and openmod), and we will pursue opportunities to contribute to these forums. In industrial collaborations we will put in place appropriate symmetric agreements relating to intellectual property (IP), ensuring that outputs from our own research are owned by our institutions while enabling partners to retain their proprietary IP. In this way we will ensure that our own research outputs remain available to be used widely in order to maximise their impact.
Organisations
- Queen Mary University of London (Fellow, Lead Research Organisation)
- OGTC (Collaboration)
- Energy Systems Catapult Ltd (Collaboration)
- FUTURE DECISIONS LTD (Collaboration)
- National Grid UK (Collaboration)
- Origami Energy Limited (Project Partner)
- COHEAT Ltd (Project Partner)
- Uniper Technologies Ltd. (Project Partner)
People |
ORCID iD |
John Moriarty (Principal Investigator / Fellow) |
Publications
Martyr R
(2018)
Optimal control of a commercial building's thermostatic load for off-peak demand response
in Journal of Building Performance Simulation
Moriarty J
(2018)
Frequency violations from random disturbances: an MCMC approach
Mijatovic A
(2019)
Asymptotic variance for random walk Metropolis chains in high dimensions: logarithmic growth via the Poisson equation
in Advances in Applied Probability
Moriarty J
(2019)
A Metropolis-class sampler for targets with non-convex support
Hamadène S
(2019)
A probabilistic verification theorem for the finite horizon two-player zero-sum optimal switching game in continuous time
in Advances in Applied Probability
Goodridge M
(2020)
Distributions of cascade sizes in power system emergency response
Nesti T
(2020)
Large Fluctuations in Locational Marginal Prices
Goodridge M
(2020)
Distributions of cascade sizes in power system emergency response
Kosmala T
(2020)
Markov risk mappings and risk-sensitive optimal prediction
Maldon Patrice Goodridge
(2021)
Supplementary Material from A rare-event study of frequency regulation and contingency services from grid-scale batteries
Arvizu JM
(2021)
Reinforcing the role of competition platforms.
in Patterns (New York, N.Y.)
Moriarty J
(2021)
A Metropolis-class sampler for targets with non-convex support
in Statistics and Computing
Patrice Goodridge M
(2021)
A rare-event study of frequency regulation and contingency services from grid-scale batteries
in Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences
Cucuringu Mihai
(2021)
AN MBO SCHEME FOR CLUSTERING AND SEMI-SUPERVISED CLUSTERING OF SIGNED NETWORKS
in COMMUNICATIONS IN MATHEMATICAL SCIENCES
Martyr R
(2021)
Nonzero-Sum Games of Optimal Stopping and Generalized Nash Equilibrium Problems
in SIAM Journal on Control and Optimization
Nesti T
(2021)
Large fluctuations in locational marginal prices.
in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences
Wu Q
(2021)
Equity2Vec
Cucuringu M
(2021)
An MBO scheme for clustering and semi-supervised clustering of signed networks
in Communications in Mathematical Sciences
Goodridge M
(2021)
Hopping between distant basins
Kosmala T
(2022)
Markov risk mappings and risk-sensitive optimal prediction
in Mathematical Methods of Operations Research
Goodridge M
(2022)
Hopping between distant basins
in Journal of Global Optimization
Martyr R
(2022)
Discrete-time risk-aware optimal switching with non-adapted costs
in Advances in Applied Probability
Goodridge M
A rare-event study of frequency regulation and contingency services from grid-scale batteries
in Philosophical Transactions of the Royal Society A
Nesti T
Large fluctuations in locational marginal prices
in Philosophical Transactions of the Royal Society A
Description | Simulating rare events is a challenging task because they arise infrequently. However, to maintain reliability in electrical power systems it is highly important to understand the causes or characteristics of rare events. Our power system is now in a period of transition, as it decarbonises and more randomly fluctuating renewable generation is connected. Historical information on rare power system events is therefore becoming less applicable, and simulation methods offer an alternative. Motivated by this need, we have developed Markov Chain Monte Carlo methods designed for rare event simulation. These methods have resulted in the publication of a number of research articles, and have stimulated further research through a novel application to global optimisation. |
Exploitation Route | Our rare event sampling method is general, making essentially no assumptions on the system which it tests. It can be used by the operators of critical infrastructure to simulate rare events causing system stress. For example, power system operators may simulate large fluctuations in generation which could cause an unacceptably large rate of change in the AC (alternating current) frequency, or large price fluctuations. Another example is in the design of robust physical structures or financial trading strategies, or in sampling potential failure modes of an artificial intelligence algorithm. |
Sectors | Aerospace, Defence and Marine,Digital/Communication/Information Technologies (including Software),Energy,Financial Services, and Management Consultancy,Transport |
URL | https://sites.google.com/site/jmoriartygroup/home/research |
Description | This project developed novel approaches to the optimisation of energy systems under uncertainty based on methods of computationally intensive statistics. The main methodological advance was the Skipping Sampler, a Markov Chain Monte Carlo method designed to sample rare events more efficiently. In addition to publishing this methodology, we applied the sampler to the study of cascading failures in power systems. This work also led to the publication of a novel method in the domain of global optimisation, namely the Basin Hopping with Skipping algorithm. Beyond academia, this work on computational approaches to optimisation in power systems led us to develop RangL, a novel machine learning competition platform. In collaboration with the Alan Turing Institute and the Net Zero Technology Centre we ran a global machine learning challenge titled 'Pathways to Net Zero', based on a computational planning problem for the transition to a Net Zero UK offshore energy industry. The RangL code base was subsequently used in a collaboration with the Energy Systems Catapult to develop a demonstration digital twin for the UK energy system, a project commissioned by the UK government department for Business, Energy and Industrial Strategy. |
First Year Of Impact | 2022 |
Sector | Energy,Government, Democracy and Justice |
Impact Types | Policy & public services |
Description | Research England Policy Impact Funding |
Amount | £19,228 (GBP) |
Organisation | United Kingdom Research and Innovation |
Sector | Public |
Country | United Kingdom |
Start | 01/2022 |
End | 06/2022 |
Description | Research fellowship for data-centric engineering programme |
Amount | £276,297 (GBP) |
Funding ID | R-LRF-JM1 |
Organisation | Alan Turing Institute |
Sector | Academic/University |
Country | United Kingdom |
Start | 04/2020 |
End | 04/2022 |
Description | The Mathematics of Energy Systems |
Amount | £185,000 (GBP) |
Funding ID | MES |
Organisation | Isaac Newton Institute for Mathematical Sciences |
Sector | Academic/University |
Country | United Kingdom |
Start | 01/2019 |
End | 05/2019 |
Description | FD |
Organisation | Future Decisions Ltd |
Country | United Kingdom |
Sector | Private |
PI Contribution | Development of an optimal control algorithm for participation in demand response using a commercial building's thermostatic load |
Collaborator Contribution | Problem formulation and feedback during algorithm development |
Impact | The developed algorithm has provided Future Decisions with a new capability |
Start Year | 2016 |
Description | National Grid control actions project |
Organisation | National Grid UK |
Country | United Kingdom |
Sector | Private |
PI Contribution | Through a Knowledge Transfer Network (KTN) study group, we produced two algorithms to derive acceptable sets of 'Bid-Offer Acceptances' (BOAs) for use in National Grid's Balancing Mechanism. BOAs are instructions send to electricity generators, in an agreed format, in order to keep supply and demand balanced. |
Collaborator Contribution | National Grid framed the problem and provided constructive feedback during the study group. |
Impact | 'Algorithms to Devise Optimal Power System Control Actions'. Technical report, KTN Energy Systems Study Group with Industry (2018). |
Start Year | 2018 |
Description | Pathways To Net Zero Reinforcement Learning Competition for the Net Zero Technology Centre |
Organisation | OGTC |
Country | United Kingdom |
Sector | Public |
PI Contribution | My research team developed the reinforcement learning challenge and managed the competition aspects |
Collaborator Contribution | The Net Zero Technology Centre (formerly the OGTC) and the Offshore Renewables Catapult collaborated with us in developing their Integrated Energy Vision model into a reinforcement learning challenge |
Impact | The output was a reinforcement learning competition, which gave the global AI community access to a specially developed version of the NZTC's Integrated Energy Vision model. The competition was to find the optimal deployment of zero-carbon technologies in order to reach Net Zero in 2050, and to share methodological best practice. The winners are due to disseminate their results at an industry webinar hosted by the NZTC in March 2022, and in an academic seminar at the California Institute of Technology later in 2022. A journal paper is also planned based on the challenge. |
Start Year | 2021 |
Description | Secondment to Energy Systems Catapult Ltd to assist with Digital Twin Demonstrator project for BEIS |
Organisation | Energy Systems Catapult Ltd |
Country | United Kingdom |
Sector | Private |
PI Contribution | I had the role of Product Owner, responsible for the development of a Visual Demonstrator for a potential future Digital Twin of the UK energy system for the UK Government Department for Business, Energy and Industrial Strategy (BEIS). One of my team was responsible for developing a reinforcement learning environment, which translated modelling output from BEIS into a form suitable for the Visual Demonstrator. |
Collaborator Contribution | The Energy Systems Catapult were responsible for scoping and managing the overall project, and for development of a Technical Demonstrator to illustrate the potential role of new data feeds and modelling approaches. |
Impact | The Energy System Digital Twin Demonstrator was delivered to the UK government department for Business, Energy and Industrial Strategy (BEIS) and showcased to BEIS in October 2022. Follow-on work is ongoing at the Energy Systems Catapult. |
Start Year | 2021 |
Company Name | Future Decisions Ltd. |
Description | Future Decisions Ltd. is a provider of software and secure networking services for building management systems. Their products optimise both internal air quality and energy consumption from heating, ventilation and cooling units. |
Year Established | 2015 |
Impact | Future Decisions has developed an automated control service for buildings which improves indoor air quality by reducing the measured levels of pollutants; increases the energy efficiency of the building's heating, ventilation and cooling system; and also provides demand response to support the power grid. In 2017 the company completed a six-figure contract installing its technology in a new building in central London and in 2020 it was granted a UK patent on its technology. |
Website | http://www.futuredecisions.net |
Description | CPD course |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Professional Practitioners |
Results and Impact | I developed and co-organised a continuous professional development (CPD) workshop, titled 'Air quality in urban areas: Harnessing data to breathe easy'. The workshop raised awareness among relevant professionals on indoor air quality standards, highlighting both challenges around implementation, and data-driven solutions to improving indoor air quality which also unlock electricity demand response. It was attended by 34 relevant professionals and, according to the feedback collected, its attendees were likely to take steps to improve indoor air quality reaching over 4,000 people. |
Year(s) Of Engagement Activity | 2019 |
URL | https://www.qmul.ac.uk/maths/news-and-events/events-/air-quality-in-urban-areas-harnessing-data-to-b... |
Description | Industry webinar |
Form Of Engagement Activity | A broadcast e.g. TV/radio/film/podcast (other than news/press) |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Industry/Business |
Results and Impact | The Net Zero Technology Centre created a challenge, in conjunction with The Alan Turing Institute, Oxquant and ORE Catapult to discover, through the power of artificial intelligence, the optimal pathway to net zero that maximised jobs and economic value for the UK. Building on the model used to produce the Reimagining a Net Zero North Sea: An Integrated Energy Vision for 2050 report, the RangL challenge sought to gamify these modelled parameters into a time step environment where AI could be deployed to calculate the most rewarding steps in the pathways to net zero by 2050. The main objective of this industry collaboration was to crowdsource the global AI community to showcase reinforcement learning and other optimisation algorithms in finding optimal deployments of blue hydrogen, green hydrogen, and offshore wind to generate the most rewarding economic and environmental benefits based on CAPEX investment, OPEX investment, decommissioning costs, jobs, revenue, and emissions. |
Year(s) Of Engagement Activity | 2022 |
URL | https://www.youtube.com/watch?v=Azuzkd0WjyU&list=PLv3aNE_KExFOT0EAMn0epeDMbNKvfTM_P&index=3&t=41s |
Description | Keynote talk at UK Energy Storage Conference 2019 |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Industry/Business |
Results and Impact | I gave the closing keynote talk 'Storage in the Digital World' at the 2019 UK Energy Storage conference, to an audience of approximately 70 from both industry and academia, leading to questions and discussion afterwards. |
Year(s) Of Engagement Activity | 2019 |
URL | https://conferences.ncl.ac.uk/ukes2019/ |
Description | MES |
Form Of Engagement Activity | Participation in an activity, workshop or similar |
Part Of Official Scheme? | No |
Geographic Reach | International |
Primary Audience | Professional Practitioners |
Results and Impact | I was principal organiser for The Mathematics of Energy Systems, a four month international visitor research programme at the University of Cambridge. The programme was highly interdisciplinary, involving 100 visiting researchers across mathematics, economics and power systems engineering, from the Americas, Europe, Asia and Australasia, with 3 international workshops and 2 industry outreach days attracting a further 250 people, and sponsorship from National Grid and Google DeepMind. The programme has given rise in particular to a theme issue of Philosophical Transactions of the Royal Society A based upon it. |
Year(s) Of Engagement Activity | 2019 |
URL | http://www.newton.ac.uk/event/mes |
Description | Opening talk at the Financial Times Digital Energy Summit |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Industry/Business |
Results and Impact | I gave an opening talk at the 2018 Financial Times Digital Energy Summit, to 179 delegates including 70 senior level attendees. This sparked discussion afterwards, leading to an ongoing research relationship between the Alan Turing Institute and the Oil and Gas Technology Centre. |
Year(s) Of Engagement Activity | 2018 |
URL | https://live.ft.com/Events/2018/FT-Digital-Energy-Summit |