Electrifying a large amount of heat demand is expected to impact future network infrastructure due to the scale and seasonal nature of heat demand. This project aims to develop and apply methods to explore optimal decarbonisation pathways to determine likely future heating technology mixes against a backdrop of policy, cost and demand uncertainties.
Objectives
Package 1:
Phase 1 - Geospatial Analysis:
Generate high spatial resolution heat energy demand estimates based on actual energy consumption data and investigate how these heat demands relate to type of heating system, dwelling characteristics and social demographics.
Characterise households by their behaviour and attitudes towards alternative heating technology uptake.
Phase 2 – SAFE-HD Model:
Develop the SAFE-HD agent based model and explore the spatial distributions of future electric heat demand under uncertainty.
Phase 3 – Network Impact Assessments:
Conduct network impact assessments with the aim of identifying least regret network investment options required.
Recommend how the tools and methods developed for the SAFE-HD project can be used by all DNO’s.
Package 2:
Extrapolate individual LCTs effect on ADMD.
Assess the impact of heat pumps and electric water heating in off gas grid areas.
Learnings
Outcomes
The uptake of electrified heat is a complex matter and will be subject to developments in policy, technical and economic factors varying across different demographics. The high spacial resolution of the agent based model allows the intricacies of future energy scenarios to be addressed on a street level. The project will enable sensitivities within projected uptake to be modelled and assessed by network planners when considering the implications that the electrification of heat will have on future demand growth.
The SAFE-HD model will notably be used as part of work package 3 for the CREDS Flex Fund research programme that is broadly concerned with assessing post-COVID impacts on demand, flexibility and infrastructure. As part of the planned CREDS project, as well as other potential academic activities at Strathclyde, the major SAFE-HD model updates will hopefully include:
o a revision of model inputs and the re-calibration of agent investment behaviour in line with state-of-the art research on post-COVID demand and energy systems changes for Net Zero
o an increased consideration of stakeholder agency beyond owner occupied households
o additional heating technologies
o additional energy tariffs available to households
Lessons Learnt
Learnings gained from each phase are outlined below. Phase 2 has been split into model development and model simulation.
Phase 1 - Geospatial Analysis
1. Heat demand typically becomes lower as deprivation increases.
2. Households with electric storage heaters have the lowest normalised heat demands.
3. Emissions abatement costs are highly sensitive to heat demand and technology assumptions.
4. There is potential for a demand rebound when implementing low carbon heating.
5. A national heat strategy must consider within-sector diversity and broader benefits.
Phase 2(a) – SAFE-HD Model Development
1. Big data techniques are required to develop a spatially explicit ABM that has both localised and GB-wide coverage using only standard desktop computing and considering heating technology investment decisions on a individual household basis.
2. Obtaining this level of spatial detail for the entire of GB necessitates some simplifications through the use of informed assumptions. This is mostly because of both the computational power afforded by a standard desktop PC and the model developmental resource constrains imposed by a PhD. However, this is also a result of some data limitations faced when only using freely and publicly available datasets.
3. Incorporating consumer choice in energy systems planning and modelling activities should; consist of calibration and validation activities; be thoroughly grounded by the use of a behavioural theoretical model(s); be developed with a deep understanding of the research context; consider how the ambiguities of heating technology characteristics are actually perceived by households (particulalry with regards to heating regime change and installation inconveniences).
Phase 2(b) – SAFE-HD Model Simulations
1. The large set of heterogeneous agents modelled for SAFE-HD are found to behave in a manner that provides a degree of confidence in the novel ABM framework developed.
2. Consumer investment behaviour is found to be highly sensitive to policy interventions, techno-economic developments and other dynamic factors including (but not limited to) societal attitudes, perceived market share of heating technologies and how heating technology characteristics are actually perceived by households.
3. In general, most households are not willing to adopt heat pumps unless they are subject to regulation (such as no new fuel oil-fired heating permitted from a given date', or are subject to a gas conversion scheme to hydrogen).
4. Agent-based modelling is highly flexible for incorporating future developments as well as accounting for complex and non-linear phenomena, particularly for those that are attributable to human behaviour.
Phase 3 - Synthesis of Learning
There is a great deal of overlap between SAFE-HD and National Grid's GB Spatial Clean Heat Model. Both projects broadly seek to develop spatially resolved methods to better explore heat decarbonisation pathways by accounting for consumer choice and localised factors all while having GB-wide coverage. Despite this, it is apparent from available project reporting that different methods have been used. Based on the interim learning from the SAFE-HD project, particularly surrounding the sensitivities in technology uptake rates observed across the full range of consumer types modelled, but also for different scenario parameters and assumptions, it would be particularly useful to coherently compare approaches and findings.
