When people ask what I got out of my PhD, I usually start with the research questions. Can traditional building design hold up as the climate changes? Do Passivhaus principles actually work in the tropics? But underneath those questions, the four studies that made up my thesis gave me something more concrete: a real technical toolkit, built from building performance simulation, field instrumentation, and data analysis.
Simulating Airflow and Heat
Across all four studies, I used IES Virtual Environment as my main tool for modelling how buildings perform over a full year — ApacheSim for thermal simulation, MacroFlo for natural ventilation. For the roof study, I went further and used ANSYS Fluent to run computational fluid dynamics (CFD), modelling steady-state 3D turbulent airflow with the standard k–ε turbulence model. That meant getting the fundamentals right: building a clean, unstructured mesh, setting up realistic atmospheric boundary layers for suburban terrain, and sizing the simulation domain according to standard guidelines (3H upstream, 15H downstream).
For the Passivhaus study, I compared PHPP’s simpler steady-state calculations against IESVE’s more detailed dynamic simulations across nineteen cities worldwide, to see where the two methods disagreed for naturally ventilated buildings in hot climates.
Understanding the Physics, Not Just the Software
To run these simulations properly, I had to actually understand the equations behind them — continuity, momentum, energy, turbulence — rather than just treating the software as a black box. I used the Discrete Ordinates model to simulate solar heat gain, and calculated wet-bulb and heat-index temperatures from raw temperature and humidity data (using Stull’s formula and the Rothfusz regression). This shifted the analysis away from simple air temperature and toward the numbers that actually predict heat stress and health risk.
Building My Own Weather Data
Good weather data for Myanmar barely existed when I started my PhD — the free databases we rely on today didn’t exist yet. I bought historical weather files from ASHRAE (about $75 each) and built future-climate weather files myself, applying IPCC and WWF temperature projections to historical data using an established “shift” method, then filling gaps with my own sensor readings. Every result in the thesis that involves future climate scenarios rests on this groundwork, even though it’s the part nobody sees.
Monitoring Real Buildings
For the vulnerability study, I ran a full year of continuous monitoring using IoT weather stations (Netatmo), cross-checked against independent data loggers (Tinytag) to make sure the readings were accurate. Designing that kind of study — where to place sensors, how to pair indoor and outdoor readings, what resolution to use, how to validate the data — is a completely different skill from running simulations. Using real-world data to test and correct my simulation assumptions was, honestly, the most rigorous piece of work across all four papers.
Designing Experiments That Actually Isolate Variables
The Passivhaus study tested ten building envelope designs against thirty shading options and four different weather files. That’s a large experimental matrix, and the point of building it that way was to isolate each variable — thermal mass, insulation, shading geometry — while holding everything else steady. Keeping track of which variable is driving which result at that scale is a skill I sharpened across all four studies.
Turning Data Into Findings
Each study produced tens of thousands of hourly data points. Making sense of that meant real statistical work: box plots, threshold analysis (what percentage of the year a space overheats, for example), and regression trends, all checked against recognised standards like CIBSE Guide A, ASHRAE 55, and NOAA’s heat-index classifications.
Research Is a Team Sport
The roof study was done with Conrad, who brought CFD expertise that sharpened the airflow analysis. The Passivhaus study was developed with Renata, who shaped the shading and envelope methodology. Even a PhD that can feel like a solo project is stronger with collaborators — they test your assumptions, share the technical load, and you end up with a professional network that outlasts the paper itself.
What I’m Grateful For
None of these skills came from nowhere. Each one showed up because a specific research question demanded it, and each one carried into the next study instead of starting from scratch. Three or four years of applied building research doesn’t hand you one narrow expertise — it builds a working fluency across simulation, instrumentation, and data analysis. And that fluency travels: to new climates, new building types, new questions entirely.
Leave a Reply