PhD Thesis

PhD thesis repository

Author: Sina Rahimi

Program: Ph.D. in Building Science, Toronto Metropolitan University

Year: 2026

Supervisory team: Prof. Russell Richman, Dr. Patrick Kastner, and Prof. Umberto Berardi

Overview

Cities are increasingly exposed to heat stress because dense urban form, impervious surfaces, reduced vegetation, and anthropogenic activity modify local temperature, humidity, wind, and radiation conditions. These local changes are often described through the Urban Heat Island (UHI) effect and directly influence outdoor thermal comfort, building energy demand, and climate resilience.

This repository documents the thesis Integrated Microclimate Simulation: A Novel Approach Towards Simulating Urban Micro-Climate. The thesis develops and validates a full-physics, neighborhood-scale urban microclimate workflow by coupling momentum, heat, moisture, radiation, vegetation, terrain, and soil heat-flux processes in OpenFOAM-based solvers and connecting the results to design-oriented comfort mapping.

The work focuses on bridging the gap between research-grade CFD and practical design workflows. It extends and validates two complementary solver frameworks:

1. buoyantHumidityPimpleFoam (BHPF): a humidity-coupled OpenFOAM solver developed from buoyantPimpleFoam for local/building-block scale simulations.

2. urbanMicroclimateFoam (UMCF): a multi-region, coupled microclimate solver used for district/campus-scale simulations and integrated with the Eddy3D/Outdoor+ workflow.

Research Problem

Many design-stage microclimate tools simplify important processes such as humidity transport, vegetation-air interaction, long-wave/short-wave radiation exchange, and realistic time-varying boundary conditions. These simplifications can lead to inaccurate prediction of heat stress, reduced confidence in mitigation strategies, and limited usability for architects, engineers, and planners.

This thesis addresses the need for a validated workflow that can:

• represent coupled heat, air, moisture, radiation, vegetation, terrain, and soil processes;

• use realistic meteorological forcing and field measurements;

• produce decision-ready thermal-comfort maps;

• remain computationally feasible for neighborhood-scale design studies; and

• operate within accessible design workflows such as Rhino/Grasshopper, Eddy3D, and Outdoor+.

Research Questions

|—|—|—|

Main Contributions

• Developed and validated a humidity-coupled OpenFOAM solver, BHPF, by extending buoyantPimpleFoam to solve a transient humidity transport equation.

• Implemented time-varying inlet humidity using table-driven boundary conditions based on measured atmospheric data.

• Quantified the impact of humidity coupling on nocturnal urban microclimate predictions and WBGT-based comfort analysis.

• Configured, extended, and validated UMCF for campus-scale simulations with coupled aerodynamics, heat, moisture, radiation, vegetation, terrain, and soil heat flux.

• Ported UMCF to Windows/BlueCFD, reducing the barrier for use in design offices that rely on Windows-based Rhino/Grasshopper workflows.

• Integrated the simulation workflow with Eddy3D/Outdoor+ using QC-checked case templates, tuned OpenFOAM dictionaries, and post-processing automation.

• Developed a validation workflow that uses one station for forcing and independent stations for model evaluation.

• Translated CFD outputs into decision-ready Wet Bulb Globe Temperature (WBGT) maps for heat-stress and mitigation analysis.

• Benchmarked UMCF against ENVI-met, showing comparable or improved accuracy with shorter runtime for the tested Georgia Tech case.

Tools and Frameworks

|—|—|

Case Studies

1. Toronto Metropolitan University (TMU) Campus

The TMU case study tested the customized BHPF solver in a dense downtown campus environment. The simulation used a six-hour nocturnal period from midnight to 06:00 on 28 July 2021, when stable stratification and humidity effects are important.

Key features:

• dense urban geometry around the TMU campus;

• time-varying inlet temperature and relative humidity from rooftop measurements;

• validation against the rooftop weather station on the Architecture building;

• comparison between humidity-enabled BHPF and the baseline temperature-only solver;

• WBGT and apparent-temperature post-processing for comfort analysis.

TMU validation results:

|—|—:|

2. Georgia Tech Campus

The Georgia Tech case study tested UMCF at the campus scale under a hot, humid summer day. The site includes mixed building heights, open plazas, vegetated areas, shaded routes, and multiple measurement stations.

Key features:

• campus-scale domain with buildings, vegetation, terrain, and meteorological forcing;

• validation using a dense network of temperature and relative-humidity stations;

• hourly simulation over a full 24-hour period;

• spatial mapping of temperature, humidity, wind velocity, and WBGT at pedestrian height;

• comparison with ENVI-met under harmonized boundary and material settings.

Georgia Tech validation results:

|—|—:|

UMCF versus ENVI-met

A like-for-like comparison was conducted between UMCF and ENVI-met for the Georgia Tech campus. Inputs were harmonized across building material properties, soil settings, vegetation representation, and weather forcing.

Key findings:

• UMCF reproduced daytime warming and evening cooling more smoothly for the tested stations.

• UMCF relative-humidity predictions followed early-morning and late-evening measurements more closely.

• ENVI-met showed sharper midday temperature spikes and humidity dips in the tested configuration.

• Under comparable multi-core runs, UMCF completed the tested case faster than ENVI-met.

• The comparison supports the use of coupled CFD when geometry-sensitive heat, moisture, shading, and vegetation effects are important.

Methodology

The thesis workflow follows a consistent sequence for both solvers:

1. Geometry and domain setup

   Prepare buildings, ground, vegetation, and terrain for CFD simulation.

2. Mesh generation

   Use blockMesh, snappyHexMesh, and region-specific meshing strategies to resolve near-building and pedestrian-level gradients.

3. Boundary-condition preparation

   Apply time-varying temperature, humidity, wind profiles, radiation forcing, and surface/material properties.

4. Physics coupling

   Couple airflow, heat transfer, humidity transport, radiation exchange, vegetation drag/transpiration, and soil/terrain heat exchange.

5. Solver execution

   Run BHPF for local humidity-sensitive simulations or UMCF for fully coupled district/campus-scale simulations.

6. Validation

   Compare model outputs with on-site observations at matching heights, timestamps, and receptor locations.

7. Thermal comfort post-processing

   Convert CFD fields to WBGT and other comfort-relevant outputs using Python and ParaView workflows.

8. Design interpretation

   Identify hot spots, cool routes, shaded areas, ventilation corridors, and potential mitigation opportunities.

Steps to Reproduce the Workflow

This repository is intended to document the thesis workflow. Some raw measurement data, case files, or third-party tools may not be included publicly because of size, licensing, or data-sharing restrictions.

1. Install the required simulation environment:

   • OpenFOAM or BlueCFD;

   • ParaView;

   • Python 3;

   • Rhino/Grasshopper with Eddy3D and Outdoor+ where applicable;

   • ENVI-met if reproducing the benchmark comparison.

2. Prepare the study geometry:

   • import building and terrain geometry;

   • simplify or clean surfaces for CFD meshing;

   • define vegetation zones and Leaf Area Density (LAD) parameters;

   • export surfaces for OpenFOAM meshing.

3. Generate the computational mesh:

   • run blockMesh;

   • run snappyHexMesh for buildings and vegetation;

   • check mesh quality with checkMesh;

   • refine local regions where pedestrian-level gradients are important.

4. Prepare meteorological boundary conditions:

   • process measured or reference station data;

   • build time-varying temperature and humidity tables;

   • define ABL wind profiles and turbulence quantities;

   • configure radiation and surface/material properties.

5. Run the solver:

   • use BHPF for local humidity-enabled simulations;

   • use UMCF for coupled air-solid-radiation-vegetation simulations;

   • run in parallel for large campus-scale domains.

6. Validate the results:

   • sample model outputs at weather-station locations;

   • align timestamps and measurement heights;

   • compute RMSE, bias, regression plots, and difference series.

7. Generate comfort maps:

   • compute WBGT from CFD outputs;

   • visualize 2 m pedestrian-level fields;

   • identify critical hot spots and mitigation opportunities.

Extracted Tables

The following key thesis tables are included in assets/tables/:

|—|—|

Key Outputs

The workflow produces:

• temperature fields at pedestrian height;

• relative-humidity fields;

• wind-speed and wind-vector fields;

• WBGT heat-stress maps;

• station-based time-series comparisons;

• RMSE and bias metrics;

• UMCF versus ENVI-met comparison plots;

• design-oriented maps for heat mitigation and comfort assessment.

Discussion

The results show that humidity is not a secondary variable in urban heat-stress analysis. At night and during humid periods, moisture affects buoyancy, density, and comfort indices. The TMU case demonstrated that BHPF could reproduce measured nocturnal temperature and humidity cycles with low error.

The Georgia Tech case showed that coupled vegetation, radiation, and heat-moisture transport are necessary to resolve campus-scale microclimate differences. Vegetated areas produced local cooling and higher humidity, while exposed plazas and roofs showed higher WBGT during critical midday hours. These findings are directly relevant to shade planning, tree placement, material selection, and heat-risk mapping.

Limitations and Future Work

Several limitations remain:

• the Georgia Tech analysis focused on a single clear summer day;

• vegetation was modeled using simplified porous-media and LAD assumptions rather than species-specific physiological parameters;

• pavement and soil processes were simplified and should be expanded in future versions;

• radiation modeling should be refined for dense districts, glazing, reflections, and nighttime cooling;

• wind validation was limited by the availability of site-specific wind measurements;

• more benchmarking is needed across additional cities, climates, geometries, and tools.

Future work should extend the workflow to multi-day and seasonal simulations, improve vegetation and soil physics, add richer radiation validation, and develop reduced-order or surrogate models for faster design iteration.

Citation

If this repository supports your work, please cite the thesis:

@phdthesis{Rahimi2026IntegratedMicroclimate,

  author  = {Rahimi, Sina},

  title   = {Integrated Microclimate Simulation: A Novel Approach Towards Simulating Urban Micro-Climate},

  school  = {Toronto Metropolitan University},

  year    = {2026},

  address = {Toronto, Ontario, Canada},

  type    = {PhD thesis}

}

Author

Sina Rahimi

Ph.D., Building Science

Toronto Metropolitan University

Acknowledgement

This work was developed with the support and guidance of Prof. Russell Richman, Dr. Patrick Kastner, Prof. Umberto Berardi, Dr. Miljana Horvat, and Dr. Helen Stopps. The thesis also acknowledges collaboration and discussions with Perkins&Will, the Outdoor+ development team, and Georgia Tech collaborators who supported the validation dataset and applied microclimate workflow.