Periodic Reporting for period 1 - SCIENCES (Smart City Innovations and Experiments using New Climate and Energy Simulations)
Periodo di rendicontazione: 2022-10-01 al 2025-02-28
The SCIENCES project contributes to the development of climate-neutral and smart cities. To achieve this mission, the main focus is innovating methods to simulate interactions between buildings and their outdoor environment. The first innovation is developing methods that use 3D city models, thermal images, and weather data as inputs to predict building energy consumption and outdoor conditions. Then, it is aimed to determine how simplified physics-based models and/or machine learning can improve simulations of interactions between buildings and their outdoor environment across different climates and scales. Finally, the intention is to incorporate carbon exchanges between buildings and their outdoor environments into simulation methods, in addition to thermal interactions. The final objective is fundamental to understanding how cities can potentially act as carbon sinks, and thus, to finding more effective strategies for achieving climate neutrality.
Activities related to data collection:
- A field experiment was conducted at Carnegie Mellon campus during the summer 2024 to collect weather data at the street level and thermal images of an urban canyon. The data are aimed to validate simulation methods of thermal interactions between buildings and their outdoor environment in a humid cold climate during an extreme heat event. Data collected during the field experiment are publicly available in the 4TU.ResearchData.
- Collaboration with the National University of Singapore (NUS) to collect similar data, which could be used to validate simulation methods in the tropics.
Activities related to modelling:
- A method was developed to automatically generate detailed physics-based building energy models from CityJSON, a JSON-based format for 3D city modelling. CityJSON can be used to express 3D city models up to a level of detail of 2.2. The method was tested on a 3D city model of NUS campus.
- A data-driven urban canopy model was developed to study the impact of interactions between buildings and their outdoor environment on the calibration of an urban building energy model.
- Data-driven building energy models were developed to perform simulations of interactions between buildings and their outdoor environment at the neighborhood or city scales.
Activities related to training:
- I audited a course of introduction to machine learning at Carnegie Mellon University.
- I co-instructed a course on smart and sustainable buildings at Carnegie Mellon University.
- A field experiment was conducted at Carnegie Mellon campus during the summer 2024 to collect weather data at the street level and thermal images of an urban canyon. The data are aimed to validate simulation methods of thermal interactions between buildings and their outdoor environment in a humid cold climate during an extreme heat event. Data collected during the field experiment are publicly available in the 4TU.ResearchData.
- Collaboration with the National University of Singapore (NUS) to collect similar data, which could be used to validate simulation methods in the tropics.
Activities related to modelling:
- A method was developed to automatically generate detailed physics-based building energy models from CityJSON, a JSON-based format for 3D city modelling. CityJSON can be used to express 3D city models up to a level of detail of 2.2. The method was tested on a 3D city model of NUS campus.
- A data-driven urban canopy model was developed to study the impact of interactions between buildings and their outdoor environment on the calibration of an urban building energy model.
- Data-driven building energy models were developed to perform simulations of interactions between buildings and their outdoor environment at the neighborhood or city scales.
Activities related to training:
- I audited a course of introduction to machine learning at Carnegie Mellon University.
- I co-instructed a course on smart and sustainable buildings at Carnegie Mellon University.
- I created a unique dataset that can be used to study urban overheating during an extreme heat event in a humid cold climate, like the one experienced in Pittsburgh.
- I developed one of the first data-driven urban canopy model that rely on simplified physics and genetic optimization. The model is capable of assessing outdoor conditions with little computational efforts while considering urban morphology with a high level of detail.
- I was one of the first in determining the statistical significance in considering interactions between buildings and their outdoor environment during the calibration of an urban building energy model.
- I developed simplified building energy models that can easily be coupled with the data-driven urban canopy model for neighborhood- or city-scale simulations.
- I developed one of the first data-driven urban canopy model that rely on simplified physics and genetic optimization. The model is capable of assessing outdoor conditions with little computational efforts while considering urban morphology with a high level of detail.
- I was one of the first in determining the statistical significance in considering interactions between buildings and their outdoor environment during the calibration of an urban building energy model.
- I developed simplified building energy models that can easily be coupled with the data-driven urban canopy model for neighborhood- or city-scale simulations.