Smart, Sustainable and Grid-Friendly Buildings: Optimal Integration in LV Grids

To mitigate rising CO2 emissions, the EU has put in place an unprecedented emphasis on the modernization and sustainable development of the building sector. The most recent EU directive imposes that all new buildings should be nearly zero-energy (nZE), i.e., achieve a yearly balance between consumption and on-site renewable energy production. To ensure the profitability of sustainable investments, buildings should progressively become smarter, optimally steering their behaviour from an economic perspective. As integral parts of the so-called smart grid, their sustainable development should also be examined through the lens of "optimal grid integration", so as to study the developing building-grid interaction. In that light, the all-encompassing entity that the scientific community should be working towards is the smart sustainable building (SSB). With low voltage (LV) networks hosting the vast majority of buildings, they make the perfect staging ground for tackling the above issues. However, the literature scarcity precludes the community from solidly positioning on how to best implement the desired large-scale SSB integration. The major limitation is the nigh-exclusive focus on either the building or network side. However, the SSB-network organism should be examined as a whole. Aside from enriching the state-of-the-art, the main goal of this thesis is to propose a holistic SSB integration approach for smart LV networks benefiting all involved parties.

The main scope of this dissertation is to investigate the optimal integration of SSBs in smart LV networks, so as to maximize their innate flexibility potential. Within a framework driven by DSO-end-user collaboration, the concept of "optimal integration" is tied to the allocation of "authority" and responsibilities between involved parties. Thus, this work investigates the pros and cons of each SSB integration approach, and attempts to gauge which one brings about the best outcome for all involved participants. Building on this, it subsequently addresses a fundamental concern: "how can smart LV networks optimally utilize SSBs to mitigate operational issues and to develop into superior entities in support of the wider smart distribution system"?

The first part of the thesis focuses on bottom-up (BU) integration, the driving force being building optimization, with the end-user at the epicenter. It examines various smart grid-based building archetypes (smart, sustainable, grid-friendly) and explores their interaction with the grid in conjunction with the end-user's own objectives. Novel concepts and methodologies are developed for the uncertainty-aware operational planning and real-time management of buildings, as well as for the "fairer" pricing of their flexibility services. Emphasis is placed on the "owner-first" mentality, i.e., the notion that end-users should prioritize their own objectives when adapting to the needs of the grid. 

The second part shifts its focus to top-down (TD) integration, the driving force being network optimization, with the DSO at the epicenter. With multi-period optimal power flow (MP-OPF) formulations lying at its core, comprehensive network and device models are developed in service of an exhaustive analysis of the various flexibility options, thus setting the benchmark for all comparisons. A wide variety of different techniques are proposed for addressing scalability, including mathematical approximations, technical simplifications, heuristic algorithms and local control approaches; both deterministic and uncertainty-aware optimization settings are examined. Aside from examining the flexibility potential of smart LV networks, emphasis is placed on identifying the extend of the BU-TD "gap", and on evaluating the necessary steps towards bridging it. 

Finally, the third part bridges the customer-grid divide by developing a hybrid integration approach that balances out the positives of the BU and TD integration approaches with practical levels of "authority" and responsibility for each involved party. Particular emphasis is placed on addressing highly practical issues, including uncertainty, scalability/granularity, failures or issues in communication and data availability, end-user "diversity" and the potential of providing higher-level ancillary services in an organized manner. 

This dissertation shows that the existing DSO-end-user collaboration paradigm is becoming outdated as i) the needs of LV networks constantly change and ii) end-users become increasingly complex entities. In driving the massive integration of SSBs in smart LV networks, one should explore the issue from both viewpoints and, to serve all desired objectives, define realistic ways to bridge the conceptual "gap". A hybrid approach is an excellent compromise between BU and TD integration and can address a wide variety of practical aspects like uncertainty, scalability communication/data issues, end-user "diversity" and sophisticated ancillary services provision. The most valuable contribution of this work is perhaps the further establishment of the term "smart sustainable building" in the scientific discussion (far more limited use at the onset of gENESiS), and its proper support through advanced methodologies and numerical results. This is a major improvement over the better-known "smart building" concept, which is no longer compatible with the European sustainability strategy.