Description
AI-based Heat Comfort Prediction. See attached pdf for details.
Tasks
In the project Climasens: Interactive Microclimate modelling for decision support (https://climasens.at/en/about-climasens), we
develop a microclimate model to support decision making. Our heat simulator based on PALM-4U solves the heat equations at
urban scales on a GPU.
We are looking for a motivated student with a good knowledge of python programming (Pytorch or Tensorflow) to design and
validate an AI-based prediction model of urban heat comfort. The idea is to learn and predict heat comfort in a complete spatial
domain based on just a few measured points and some other input parameters. In addition, the optimal number and location of
such measurement points could be learned. If successful, the results can be validated with real world experiments in our case
study in Seestadt.
Requirements
- Knowledge of Python (Pytorch or Tensorflow)
- Interest in designing and validating an AI-based prediction model
- More knowledge is always appreciated
Environment
If successful the work will be integrated into the scenarify (http://visdom.at) framework, which is developed by VRVis. The diploma thesis will be supervised by Dr. Viktor Birschitzky, VRVis Vienna.
Contact
Additional Images and Files
| Attachment | Size |
|---|---|
| DA Heat AI Simulation | 1.18 MB |