Strategies for robust renovation of residential buildings in Switzerland

The study addresses how to renovate existing residential buildings in Switzerland in a way that is robust to major future uncertainties (climate, energy grid mix, costs, user behavior) while minimizing lifecycle greenhouse gas (GHG) emissions and costs. Switzerland provides a representative context for Northern Europe, with ~64% of buildings heated by oil and gas and ~70% of the stock built before stringent efficiency standards. Conventional decarbonization focuses on energy systems and efficiency, but high renovation rates risk a lifecycle “carbon spike” if embodied emissions of materials are neglected. The research question is to identify optimal, robust renovation strategies—combining envelope and heating system measures—that minimize lifecycle GHG emissions and life cycle costs under uncertainty, and to compare conventional versus bio-based (non-conventional) insulation solutions. The purpose is to inform building-scale and stock-scale decisions that align with climate goals while avoiding unintended increases in embodied carbon and addressing social equity through operational energy cost impacts.

Prior work has examined renovation strategies at single-building and stock scales and used multi-objective optimization for costs, emissions, and comfort. However, most studies do not comprehensively include uncertainty; only a few consider limited uncertain parameters such as electricity mix evolution, climate change, replacement timing, or material properties. The authors’ previous work combined multiple uncertainty sources with robust multi-objective optimization but focused on a single case. Conventional thermal retrofits often rely on fossil-based insulation (e.g., polystyrene, aerogel) with high embodied GHG and hygrothermal drawbacks. Literature on bio-based alternatives (straw, hemp) has matured, demonstrating technical feasibility (e.g., straw bales, hemp blocks/hemp-lime) and potential to store biogenic carbon, with increasing industrial adoption. Gaps remain in integrating comprehensive uncertainty quantification with multi-objective optimization across representative buildings and directly comparing conventional versus non-conventional materials within a unified LCA-LCCA framework.

• Case studies: Six representative Swiss multi-family residential buildings (covering construction periods before 1919, 1919–1945, 1945–1960, 1960–1970, 1971–1980, and 1981–1990) were selected from the eRen project database (193 buildings) to reflect the stock currently needing renovation. Some pre-existing roof insulation was retained as initial state.
• Renovation options: Heating system replacement (gas boiler, wood pellets boiler, heat pump), thermal envelope insulation (external walls, roof, ground floor) with conventional materials (EPS, glass wool, rock wool, cellulose) and non-conventional fast-growing bio-based materials (straw bales, hemp mats, hempcrete), and window replacement (double/triple glazing; frames: aluminum, PVC, wood). Renewable onsite energy (e.g., PV) was excluded.
• LCA-LCCA scope: EN 15978 stages included: production (A1–A3), replacement (B4), operational energy for heating (B6), end-of-life (C3–C4). Only climate change was assessed, using GWP (kgCO2eq). Embodied carbon data for A1–A3 and C3–C4 from KBOB 2016. Operational energy demand for heating was modeled with monthly quasi-steady-state per Swiss SIA 380/1; useful heating demand was converted to final energy via efficiency and then multiplied by carrier-specific GWP from KBOB.
• LCCA: Investment, operational, and replacement costs were included; annual repairs modeled as a percentage of investment. Reference study period: 60 years (Swiss standards). Functional unit: complete 60-year lifecycle of the renovated building.
• Biogenic carbon: For non-conventional materials, dynamic carbon storage was modeled with time-dependent characterization factors and KBOB-based release parameters.
• Uncertainties: Design parameters (e.g., insulation material/type, heating system, insulation thickness) and exogenous parameters (e.g., future climate change, service life of components, future electricity mix and energy prices, occupant behavior, inflation). Parameter ranges and distributions are provided in SI.
• Robust optimization: Multi-objective robust optimization minimizing LCA (GWP) and LCCA was performed using NSGA-II, suitable for mixed discrete-continuous variables. To reduce computational burden under uncertainty, adaptive surrogate modeling was used, coupling NSGA-II with Kriging (Gaussian process regression). Probabilistic comparisons used 5th–95th percentiles; Pareto fronts were assessed; median solutions per heating type were compared against non-renovated and conventional baselines.
• Generalization and stock scaling: Representativeness of the six buildings was validated by scaling their heating demands by energy reference area per construction period and comparing to measured Swiss stock heating consumption (difference ~4%). Potential annual GHG savings from applying optimal scenarios to the Swiss stock were then estimated.

• Heating system replacement is the most influential, robust action across all periods, driving the largest reductions in lifecycle GHG emissions.
• Non-conventional bio-based insulation (straw/hemp) consistently outperforms conventional insulation in GWP and heating demand: optimized non-conventional solutions select higher insulation thicknesses, yielding lower energy consumption and lower GWP; conventional solutions tend to minimal insulation.
• Robustness: Solutions with wood pellet boilers and heat pumps are more robust than gas boiler options in LCA performance.
• GWP reduction magnitude: For wood boiler cases, GWP values with non-conventional materials are on average three times lower than with conventional materials; uncertainty ranges do not overlap, indicating non-conventional materials perform better under all circumstances examined.
• LCCA: Non-conventional materials show lower average life cycle costs across buildings, but cost uncertainty leads to overlapping ranges; a clear cost advantage emerges when initial heating demand exceeds ~150 kWh/m²·a, where operational savings offset higher upfront investments.
• Cost structure: Non-conventional renovation increases investment costs but yields larger operational savings; the operational savings share typically exceeds the increased investment share.
• Windows: Optimal solutions rarely prescribe window replacement (3 of 18 scenarios for conventional materials; 1 of 18 for non-conventional), due to high embodied carbon and investment cost versus limited performance gains.
• Insulation thickness: Conventional materials’ optimal external wall insulation commonly lies between 0–10 cm; whole-envelope insulation rarely exceeds 20 cm. Non-conventional cases frequently select the maximum modeled thickness, notably 70 cm straw bale on external walls and roofs in most cases.
• Stock-scale potential: Applying optimized scenarios with heat pumps or wood pellets plus non-conventional insulation to the Swiss residential stock can save up to 87% of annual GHG emissions.
• Social equity: Conventional minimum-insulation strategies (even with renewable heating) can leave occupants with higher energy bills; deep conventional retrofits reduce bills but cause a significant upfront “carbon spike.” Thick bio-based insulation with renewable heating simultaneously achieves low GHG and low energy bills.

The results directly answer the research question by identifying renovation strategies that remain effective under wide-ranging uncertainties: prioritize replacing fossil-based heating systems and, where feasible, deploy thick bio-based insulation. These strategies minimize lifecycle GHG emissions and can reduce operational costs, achieving robustness to future climate, energy mix, and cost uncertainties. The strong, non-overlapping GWP advantages of non-conventional materials underscore the importance of embodied carbon in a decarbonizing grid context. Policy implications are substantial: conventional deep renovation without fossil heating replacement complies with energy codes but increases embodied impact and does not reduce lifecycle GHG, suggesting standards should pivot to lifecycle GHG minimization. Scaling analyses show the six cases are representative (4% deviation in heating demand from stock data), enabling projection of up to 87% annual GHG savings at stock scale with combined renewable heating and bio-based insulation. Broader relevance extends to European contexts as grids decarbonize to Swiss-like intensities, which elevates the role of embodied emissions and further favors bio-based strategies. Resource and implementation considerations include straw availability (EU-wide potential with limited land-use impact), limited sustainable wood energy potential in Switzerland (necessitating a mix of wood-based systems and heat pumps), and particulate emissions from individual wood boilers (mitigated by modern boilers and district heating filtration). The findings also highlight social justice dimensions: bio-based strategies deliver both low emissions and lower energy bills, while conventional strategies can either spike embodied carbon (deep retrofit) or yield higher occupant energy costs (minimal insulation).

This work integrates LCA and LCCA with comprehensive uncertainty treatment and robust multi-objective optimization to identify renovation strategies for six representative Swiss residential buildings. The main contributions are: (1) demonstrating that replacing fossil heating systems is the primary lever for lifecycle GHG reductions; (2) showing that maximizing fast-growing bio-based insulation thickness, combined with renewable heating, delivers robust, low-carbon, and lower operational cost outcomes; (3) revealing that with conventional insulation, optimal robust solutions favor only limited insulation thicknesses rather than deep retrofits; and (4) establishing stock-scale representativeness and large potential annual GHG savings. The study argues for updating building standards toward lifecycle GHG reduction, especially as grids decarbonize and embodied emissions dominate. Future research should extend the scope to include onsite renewables (e.g., PV), assess district heating heat pumps and practical deployment constraints, update models with evolving energy prices and discount rates, and employ dynamic simulations for overheating and potential cooling needs.

• Scope limitations: Renovation measures were limited to envelope insulation, window replacement, and heating system replacement; onsite renewables such as PV were excluded. Including PV could change optimal solutions (e.g., combined window-façade strategies).
• Constructability/urban constraints: Thick bio-based insulation may be infeasible in some urban contexts due to space constraints, despite recent projects demonstrating thick straw bale walls.
• Heating technology scope: Only individual residential heat pumps were modeled; practical constraints (space, aesthetics) may necessitate district heating heat pumps in some cases.
• Economic inputs: Energy prices and discount rates have changed since the 2022 energy crisis; future updates could alter outcomes.
• LCI staging: While LCI data covered production and end-of-life stages, the study did not disaggregate their individual impacts in results, which may influence findings.
• Energy modeling: Operational performance used monthly quasi-steady-state analysis (Swiss standard). Although adequate for heating, dynamic simulation could better capture overheating and cooling needs.
• Figures excluding replacements: Some presented deterministic figures excluded replacement-stage impacts/costs, which may understate lifecycle totals in those visuals.