B-bem: The Bayesian building energy management Portal
Lead Research Organisation:
University of Cambridge
Department Name: Engineering
Abstract
Energy Management of existing non-domestic buildings is wrought with many challenges, a number of which arguably exist due to the diversity found amongst individual buildings and amongst the humans who occupy them. Buildings are inherently unique systems making it difficult to generalize technology solutions for any individual property. Instead, to make robust investment decisions for the energy-efficient upkeep of a particular building requires some degree of tailored engineering and economic analysis. To understand why this is the case, one need only to consider the chain of questions one would likely need to address for decision-making in an arbitrary building. For instance, we might ask: what is the age of the building and the equipment currently installed in it? Does the heating system need to be replaced? If yes, is the current system a boiler, and if so, how efficiently does it perform? Would the building benefit from a new boiler or an electric heat pump? Would it benefit from replacing the heating distribution pipes? Do the cost / benefits of any of these technologies depend on government tariffs and subsidies? What is the risk faced if any available subsidies are cut in the future? How robust is either technology to the future price of natural gas and electricity? Would that risk be worth taking? Is it too expensive to even start thinking about the options and associated risks? How would a facility manager visualise the options available and possible spreads of benefits and risks for all these aspects?
This project aims to respond to these challenges. Indeed, in order to make sound decisions on future building operation and technology investment, evidence shows that one needs adequate information on a number of engineering, economics, and social science matters pertaining to each individual project. To obtain this information has so-far been viewed as a costly exercise, and has contributed to the general perception that undertaking deep cuts to building energy consumption (achieving more than 15% in energy savings per investment) is an economically risky affair. This proposal is the first to develop and recommend an altogether new approach to performing building audits, energy simulation, uncertainty analysis, data visualization, and finally investment decision-making. It will lead to a marked reduction in the cost of acquiring information for robust retrofit and facility management decisions.
The direct outputs of this project will be a series of software tools for three distinct but related purposes: (i) collecting building data on relevant uncertainty parameters (i.e., "what do we know now?"); (ii) propagating and quantifying uncertainty using building simulation models, measurements obtained from key monitored building sites, and cutting-edge statistical approaches (i.e., Bayesian analysis); and (iii) the display and interpretation of uncertainty.
During the course of the project, workshops will be organised to lay out the current (uncertain) knowledge that has been, until now, largely undocumented in the buildings sector and inaccessible to the energy research community. This includes gaining understanding on the most common faults observed in managing conventional energy systems, and how spatial layouts in building evolve. The graphical presentation of risk information and understanding users' perception of uncertainty and risk will be key elements of these workshops and the research programme. Our software tools, user guidance, and numerical runs of test cases will be made available, as the web-based B-bem portal, via the University of Cambridge web site.
This project aims to respond to these challenges. Indeed, in order to make sound decisions on future building operation and technology investment, evidence shows that one needs adequate information on a number of engineering, economics, and social science matters pertaining to each individual project. To obtain this information has so-far been viewed as a costly exercise, and has contributed to the general perception that undertaking deep cuts to building energy consumption (achieving more than 15% in energy savings per investment) is an economically risky affair. This proposal is the first to develop and recommend an altogether new approach to performing building audits, energy simulation, uncertainty analysis, data visualization, and finally investment decision-making. It will lead to a marked reduction in the cost of acquiring information for robust retrofit and facility management decisions.
The direct outputs of this project will be a series of software tools for three distinct but related purposes: (i) collecting building data on relevant uncertainty parameters (i.e., "what do we know now?"); (ii) propagating and quantifying uncertainty using building simulation models, measurements obtained from key monitored building sites, and cutting-edge statistical approaches (i.e., Bayesian analysis); and (iii) the display and interpretation of uncertainty.
During the course of the project, workshops will be organised to lay out the current (uncertain) knowledge that has been, until now, largely undocumented in the buildings sector and inaccessible to the energy research community. This includes gaining understanding on the most common faults observed in managing conventional energy systems, and how spatial layouts in building evolve. The graphical presentation of risk information and understanding users' perception of uncertainty and risk will be key elements of these workshops and the research programme. Our software tools, user guidance, and numerical runs of test cases will be made available, as the web-based B-bem portal, via the University of Cambridge web site.
Planned Impact
The building services industry has in recent years witnessed a resurgent public interest in building energy retrofits. However, individual building owners - particularly in the non-domestic sector - still face a number of barriers to both efficient building management and future retrofit uptake. Recent academic and government reports have highlighted the persistent incentive and risk gap between the policymakers, who develop macroeconomic schemes for energy demand reduction, and the individual energy consumers who ultimately pay for the success or failure of individual energy-efficiency measures. A related barrier - and the one addressed specifically by this work - is the perceived high cost of acquiring accurate supportive information for individual building management and retrofit decisions. This can be viewed either as an upfront cost of performing accurate building audits and simulation analysis, or a penalty cost resulting from poorly informed decisions. The long-standing issue has contributed, over time, to public perception that energy-efficiency investments in buildings perform unpredictably and should be viewed as economically risky affairs. The ambition of this project is to transform standard energy auditing and simulation processes in order to minimize the transaction costs of building facility management and retrofit decision-making.
Our research will impact those with an interest in managing operational and investment risk, especially when investigating, planning, and undertaking end use energy-demand reduction projects. This audience includes, but is not limited to:
(a) Businesses in facilities and asset management, energy simulation software companies, energy service companies (ESCOs), and the energy bill payers themselves (owners of large building portfolios such as campuses, technology parks, supermarkets) will directly benefit from quantitative risk analysis of investments in energy efficiency of buildings. Furthermore, the Expert Elicitation workshops proposed in WP1 will create objective communication channels leading to improved understanding of user requirements by the research community; improved communication and decision-making tools will come from WP2; and training events and materials designed to help non-specialists interpret uncertain information (WP3). Ultimately, these will serve to enhance the capacity of businesses and organizations to evaluate their investments in energy efficiency based on associated risks.
(b) Government and Agencies responsible for setting benchmarking standards and best practices will find the outcomes of WP1 directly relevant. Indeed, a recent publication from DECC stresses that our scientific understanding of how non-domestic buildings currently operate and the key challenges in making them more efficient is out-of-date. In this sense, the outcomes of WP1: Expert Elicitation will be directly relevant to DCLG and DECC. Another measure of success in research will be in how our work will lead to other new research - especially in the development of new methods for setting variable benchmarks for buildings, with milestones beyond the lifetime of the project.
Our research will impact those with an interest in managing operational and investment risk, especially when investigating, planning, and undertaking end use energy-demand reduction projects. This audience includes, but is not limited to:
(a) Businesses in facilities and asset management, energy simulation software companies, energy service companies (ESCOs), and the energy bill payers themselves (owners of large building portfolios such as campuses, technology parks, supermarkets) will directly benefit from quantitative risk analysis of investments in energy efficiency of buildings. Furthermore, the Expert Elicitation workshops proposed in WP1 will create objective communication channels leading to improved understanding of user requirements by the research community; improved communication and decision-making tools will come from WP2; and training events and materials designed to help non-specialists interpret uncertain information (WP3). Ultimately, these will serve to enhance the capacity of businesses and organizations to evaluate their investments in energy efficiency based on associated risks.
(b) Government and Agencies responsible for setting benchmarking standards and best practices will find the outcomes of WP1 directly relevant. Indeed, a recent publication from DECC stresses that our scientific understanding of how non-domestic buildings currently operate and the key challenges in making them more efficient is out-of-date. In this sense, the outcomes of WP1: Expert Elicitation will be directly relevant to DCLG and DECC. Another measure of success in research will be in how our work will lead to other new research - especially in the development of new methods for setting variable benchmarks for buildings, with milestones beyond the lifetime of the project.
Organisations
- University of Cambridge (Lead Research Organisation, Project Partner)
- University of Tokyo (Collaboration)
- Georgia Institute of Technology (Collaboration)
- LOUGHBOROUGH UNIVERSITY (Collaboration)
- École Normale Supérieure de Cachan (Collaboration)
- UNIVERSITY OF EDINBURGH (Collaboration)
- Delft University of Technology (TU Delft) (Collaboration)
- Aecom (United Kingdom) (Project Partner)
- British Institute of Facilities Management (Project Partner)
- Breathing Buildings Limited (Project Partner)
- BuroHappold (United Kingdom) (Project Partner)
- GI Energy (Project Partner)
Publications
Mouzourides P
(2014)
Searching for the distinctive signature of a city in atmospheric modelling: Could the Multi-Resolution Analysis (MRA) provide the DNA of a city?
in Urban Climate
Rysanek A
(2014)
DELORES - an open-source tool for stochastic prediction of occupant services demand
in Journal of Building Performance Simulation
Ward R
(2014)
A bottom-up energy analysis across a diverse urban building portfolio: retrofits for the buildings at the Royal Botanic Gardens, Kew, UK
in Building and Environment
Wei L
(2015)
Comparative Study on Machine Learning for Urban Building Energy Analysis
in Procedia Engineering
Kiguchi Y
(2015)
Data Driven Model for Rooftop Excess Electricity Generation
Munkhammar J
(2015)
Household electricity use, electric vehicle home-charging and distributed photovoltaic power production in the city of Westminster
in Energy and Buildings
Tian W.
(2015)
High Resolution Energy Simulations at City Scale
Tian W
(2015)
Importance analysis and meta-model construction with correlated variables in evaluation of thermal performance of campus buildings
in Building and Environment
Pacheco-Torres R
(2016)
Efficient energy modelling of heterogeneous building portfolios
in Sustainable Cities and Society
Menberg K
(2016)
Sensitivity analysis methods for building energy models: Comparing computational costs and extractable information
in Energy and Buildings
Mikhaylova O.
(2016)
Benefits and Optimization of District Hybrid Ground Source Heat Pump Systems
Ward R
(2016)
Exploring the impact of different parameterisations of occupant-related internal loads in building energy simulation
in Energy and Buildings
Kumar P
(2016)
Indoor air quality and energy management through real-time sensing in commercial buildings
in Energy and Buildings
Menberg K
(2017)
Exergy analysis of a hybrid ground-source heat pump system
in Applied Energy
Kristensen M
(2017)
Bayesian calibration of building energy models: Comparison of predictive accuracy using metered utility data of different temporal resolution
in Energy Procedia
Kristensen M.H.
(2017)
Bayesian calibration of building clusters using a single geometric building representation
Menberg K.
(2017)
Exergy Analysis Of A Ground- Source Heat Pump System
Barthelmes V
(2017)
Exploration of the Bayesian Network framework for modelling window control behaviour
in Building and Environment
Menberg K.
(2018)
Learning about error terms in energy models by Bayesian calibration
Choi W
(2018)
Bayesian inference for thermal response test parameter estimation and uncertainty assessment
in Applied Energy
Choi W
(2018)
Bayesian inference of structural error in inverse models of thermal response tests
in Applied Energy
Menberg K
(2018)
Influence of error terms in Bayesian calibration of energy system models
in Journal of Building Performance Simulation
Pickering B
(2019)
District energy system optimisation under uncertain demand: Handling data-driven stochastic profiles
in Applied Energy
Ward R
(2019)
A data-centric bottom-up model for generation of stochastic internal load profiles based on space-use type
in Journal of Building Performance Simulation
Menberg K
(2020)
Multi-fidelity approach to Bayesian parameter estimation in subsurface heat and fluid transport models.
in The Science of the total environment
Li Q
(2020)
Effective and scalable modelling of existing non-domestic buildings with radiator system under uncertainty
in Journal of Building Performance Simulation
Ward R
(2021)
A tool for generation of stochastic occupant-based internal loads using a functional data analysis approach to re-define 'activity'
in Journal of Building Performance Simulation
Ward R
(2021)
A study on the transferability of computational models of building electricity load patterns across climatic zones
in Energy and Buildings
| Description | 1. We have developed an improved method for internal load modelling in buildings, which we will publish this year. This new model will lead to improved estimation of electricity demand in existing buildings at high resolution. 2. We have proposed novel metrics for identifying key parameters of buildings that influence energy demand (new developments in sensitivity analysis). These results are now ready for publication. |
| Exploitation Route | During 2016-2017, we will devote some resources to develop an online tool for quantifying uncertainties in key parameters of existing buildings and propagating them as outcomes. |
| Sectors | Construction,Energy,Environment |
| Description | INVITATION FELLOWSHIPS FOR OVERSEAS RESEARCHERS (long-term) |
| Amount | ￥1,676,000 (JPY) |
| Organisation | Japan Society for the Promotion of Science (JSPS) |
| Sector | Public |
| Country | Japan |
| Start | 08/2015 |
| End | 12/2015 |
| Description | Studentship |
| Amount | £38,360 (GBP) |
| Organisation | AECOM Technology Corporation |
| Sector | Private |
| Country | United States |
| Start | 01/2015 |
| End | 02/2020 |
| Description | UK Centre for Digital Build Britain |
| Amount | £81,518 (GBP) |
| Organisation | University of Cambridge |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 10/2018 |
| End | 07/2019 |
| Description | alan turing institute |
| Amount | £364,503 (GBP) |
| Organisation | Alan Turing Institute |
| Sector | Academic/University |
| Country | United Kingdom |
| Start | 10/2018 |
| End | 09/2021 |
| Description | City-scale energy analysis - École Normale Supérieure de Cachan |
| Organisation | École Normale Supérieure de Cachan |
| Country | France |
| Sector | Academic/University |
| PI Contribution | Dr. Choudhary visited the Civil Engineering Department in Summer 2015 for 6 weeks (April-May), and worked with students there on the topic of visualization of city-scale energy data. In September 2017, she was hosted for a month as Invited Professor in the Department. |
| Collaborator Contribution | In 2015 École Normale Supérieure de Cachan provided technical expertise and resources for interactive visualization of data. As a follow up, a student from ENS visited the B-bem group for a 9 month internship period where we mentored him on the topic of quantifying uncertainties in internal loads of buildings at city scale. |
| Impact | A student from École Normale Supérieure de Cachan visited as a research intern for 12 months (October 2015-2016) to work on this topic. Dr. Choudhary was invited as Visiting Professor for 4 weeks in September 2016 and 2018 to work with new students. |
| Start Year | 2015 |
| Description | Collaboration with University of Tokyo |
| Organisation | University of Tokyo |
| Department | Institute of Industrial Science |
| Country | Japan |
| Sector | Academic/University |
| PI Contribution | - Guest Professor at Ooka Lab, Institute of Industrial Science for 4 months (Sept-December 2015) supported by an invitational Fellowship by Japan Society of Promotion of Science. - Interacted with PhD students and staff on the following topics: uncertainty analysis, distributed energy systems, exergy analysis of building energy systems. - Since 2015, we have regular annual visits to each others labs |
| Collaborator Contribution | The Ooka Lab invited Cambridge PhD student Bryn Pickering for 2 week visit in December 2015. We have co-authored 2 peer-reviewed conference articles and 3 journal publications. From the B-bem project, PDRA Kathrin Menberg has been heavily involved in these collaborations. We have worked with University of Tokyo to carry out uncertainty analysis in the estimation of ground thermal properties for geo-energy systems. In turn- University of Tokyo helped us carry out exergy analysis of heat pump systems, which enabled us to have an improved understanding of system efficiencies. |
| Impact | 1. 2018 visiting researcher from U. of Tokyo hosted by Alan Turing Institute 2. Uncertainty Analysis: 2 journal articles in 2018 3. Exergy Analysis: 1 conference publication (2017), and 1 journal paper (2017). 4. Distributed Energy Systems: 1 conference publication in 2016. |
| Start Year | 2015 |
| Description | European Partnership-YH |
| Organisation | Delft University of Technology (TU Delft) |
| Country | Netherlands |
| Sector | Academic/University |
| PI Contribution | Grant application submitted to the Marie Curie Innovative Training Network programme in January 2017 |
| Collaborator Contribution | Grant application submitted to the Marie Curie Innovative Training Network programme in January 2017 |
| Impact | Grant application submitted to the Marie Curie Innovative Training Network programme in January 2017 |
| Start Year | 2016 |
| Description | TEDDINET |
| Organisation | Loughborough University |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Attended workshops, shared research outputs. |
| Collaborator Contribution | Following workshops were organized by TEDDINET 1. TEDDINET Workshop : 27-28/04/2015 2. Joint TEDDINET-Innovate UK event 12-13/10/2015 3. ReCoVER-TEDDINET Workshop 14-15/01/2016 |
| Impact | Networking and help define complementary research with university partners funded within the same scheme. |
| Start Year | 2015 |
| Description | TEDDINET |
| Organisation | University of Edinburgh |
| Country | United Kingdom |
| Sector | Academic/University |
| PI Contribution | Attended workshops, shared research outputs. |
| Collaborator Contribution | Following workshops were organized by TEDDINET 1. TEDDINET Workshop : 27-28/04/2015 2. Joint TEDDINET-Innovate UK event 12-13/10/2015 3. ReCoVER-TEDDINET Workshop 14-15/01/2016 |
| Impact | Networking and help define complementary research with university partners funded within the same scheme. |
| Start Year | 2015 |
| Description | UNCERTAINTY ANALYSIS - Georgia Institute of Technology |
| Organisation | Georgia Institute of Technology |
| Country | United States |
| Sector | Academic/University |
| PI Contribution | co-I (Dr. Yeonsook Heo) and PDRA (Dr. Kathrin Menberg) visited a research team specializing in methods and tools for uncertainty analysis in the context of building energy simulation. Their visit was for 6 weeks, from 18/07/2015 to 31/08/2015. They worked on comparing sensitivity analysis techniques during their time at Georgia Institute of Technology. |
| Collaborator Contribution | The team at Georgia Institute of Technology trained the PDRA (Dr. Kathrin Menberg) in their recently developed toolkit called 'GURA', designed for quantifying parameter uncertainties in dynamic building energy models. |
| Impact | 1. One conference publication and one journal publication (both Menberg et al 2016) resulted from this visit. 2. A PhD student from the team at Georgia Institute of Technology visited Cambridge to work with us from May 2016-March 2017: we are in the process of submitting the journal article resulting from this visit. |
| Start Year | 2015 |
| Description | B-bem: The Bayesian Building Energy Management Portal |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | International |
| Primary Audience | Postgraduate students |
| Results and Impact | Talk at Polytechnic University of Turin, Turin, Italy presenting research outcomes from the B-bem projevt and discussing ongoing and future research collaboration as the next step |
| Year(s) Of Engagement Activity | 2017 |
| Description | Uncertainty Analysis of Energy Systems |
| Form Of Engagement Activity | A talk or presentation |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Industry/Business |
| Results and Impact | Half-day meeting to present B-bem outcomes to researchers in Mitsubishi Electric R&D Centre (Dr Danieal Coakly, Dr Duan Wu) and discuss future opportunities for research collaborations |
| Year(s) Of Engagement Activity | 2017 |
| Description | keynote |
| Form Of Engagement Activity | Participation in an activity, workshop or similar |
| Part Of Official Scheme? | No |
| Geographic Reach | Regional |
| Primary Audience | Postgraduate students |
| Results and Impact | keynote 5th Loughborough-London Student Conference: Energy Demand Pathways: What does the future hold for the built environment?, London, UK June 2018. |
| Year(s) Of Engagement Activity | 2018 |