Master's thesis · Lund University · 2026

In collaboration with Ana Sofia Compean Becerra (joint thesis) · Zerotude AB (industry supervision)

Energy renovation of existing office buildings

This project develops a decision-support framework for the energy renovation of existing office buildings in Sweden. Two district-heated office buildings in Stockholm were selected as case studies and their energy models calibrated using measured data. Different renovation measures were applied through parametric simulations, and the results evaluated with techno-economic analyses to identify renovation strategies that remain effective and robust under uncertainty.

Cutaway model of an office building with air handling, lighting and heating distribution

37 %

energy saving by AHU improvement

40 %

average ROI with converting from CAV to VAV system

3–6

years payback for the ventilation measures in both buildings

Awarded the Miljöfond scholarship of Sveriges Ingenjörer (the Swedish Association of Graduate Engineers) for its contribution to the sustainable development of the built environment.

Published at lup.lub.lu.se/student-papers/record/9230933

The 2024 recast of the EU buildings directive introduces binding minimum standards for existing non-residential buildings. The worst-performing 16 % of the stock must reach a threshold by 2030, rising to 26 % by 2033; transposition into Swedish law is due by May 2026. This creates substantial renovation pressure in a short time.

The difficulty lies in the data: in existing buildings, simulated and measured consumption regularly diverge, and many inputs cannot be verified. The thesis therefore develops a procedure that does not eliminate this uncertainty but carries it explicitly.

Aim of the project

To develop a framework that accounts for model uncertainty and identifies cost- and energy-effective renovation pathways, with potential for scalability.

1

Fitted energy model

Building models were calibrated using measured heating data and validated according to ASHRAE Guideline 14 criteria.

2

Uncertainty & parametric simulation

Three energy models were retained for each building to account for uncertainty. The most influential parameters affecting heating demand were identified and evaluated.

3

Techno-economic evaluation

Renovation packages were assessed based on energy savings, payback period, and return on investment across all retained models.

01

Case studies

Two district-heated Stockholm offices — one with near-absent, one with good heat recovery — span the realistic baseline range.

Building A

Building B

Function

Office (100 %)

Office (84 %) · Retail (16 %)

Year built · heated area

1890 · 1,360 m²

1898 · 3,088 m²

Ventilation

CAV · heat recovery 12 %

CAV · heat recovery 60–75 %

02

Model fitting

Three calibrated models per building keep the residual uncertainty visible instead of hiding it.

1

Sensitivity screening

Non-observable inputs went through a sensitivity screening: anything moving heating demand by more than 5 % became a fitting variable. Ventilation parameters ranked clearly first in both buildings.

Average effect on heating demand

2

Model fitting

All combinations of the governing parameters were simulated and scored against measured heating data under ASHRAE Guideline 14.

  • CV(RMSE) ≤ 15 %
  • NMBE within ±5 %
  • Monthly filter in the heating season
3

Representative models

Rather than a single best fit, three energy models per building were retained as representative cases.

Building A
3 envelope variants

Building B
3 envelope & ventilation variants

03

Measures and packages

Six measures from ventilation, envelope and auxiliary systems — assessed individually and in every combination.

Ventilation

AHU replacement

CAV → VAV

Envelope

Roof insulation

Window replacement

Auxiliary

LED lighting

Pump replacement

Every on/off combination was simulated as a package — 64 packages for both buildings, for each fitted model.

  • Each measure was analysed individually by energy savings and simple payback.
  • All combinations were then assessed as packages to rank their profitability, prioritise measures and derive an investment-based renovation pathway for each building.

04

Results for individual measures

Ventilation dominates: replacing the AHU and converting CAV→VAV deliver by far the largest savings.

Annual energy saving in kWh/m²·a — averages across the three fitted models per building

  • Ventilation measures deliver the highest savings.
  • The ranking of measures is consistent across all fitted models.
  • Variation in savings reflects baseline uncertainty.

05

Renovation pathway

The same three-step pathway holds in both buildings: ventilation and lighting first, envelope last.

ROI

LED lighting + AHU replacement

> 30 %

+

+ CAV → VAV

> 25 %

+

+ Roof insulation

≈ 20 %

+

+ Window replacement

≈ 10 %

Prepared as a basis for decisions by building owners and investors — conceptually comparable to Germany's individual renovation roadmap (iSFP).

ROI — return on investment: average annual cost saving relative to the invested cost

06

Key findings

The ranking is robust

The same renovation sequence emerges in both buildings, across all fitted models and baseline assumptions.

Uncertainty does not always affect the outcome

Savings shifted between the fitted models — the order of measures did not.

A path toward scalability

If impactful parameters can be reliably documented, this framework can scale.

Software

  • Rhino
  • Grasshopper/Honeybee
  • EnergyPlus
  • Revit
  • Excel

MSc Energy-efficient and Environmental Building Design, Lund University · 2026