University of Twente Student Theses
Digital Twin-Based Planning Support System for Urban Heat Island Mitigation
Afzalinezhad, Amir (2024) Digital Twin-Based Planning Support System for Urban Heat Island Mitigation.
|
PDF
8MB |
| Abstract: | Urban growth is accompanied by increased impervious surfaces and urban building densities, contributing to the (UHI) phenomenon. This phenomenon refers to areas in cities with higher temperatures than rural areas. This phenomenon has a negative impact on human health and quality of life by increasing heat stress and the urban environment. Due to UHI's adverse effects, multiple studies have been conducted to address this challenge and mitigate UHI in urban areas. This study introduces an innovative Digital Twin- Based Planning Support System (DT-PSS) that can help urban planners mitigate the UHI formation and its effect in cities. This methodological research explored the development of a DT-PSS within the Unreal Engine (UE) platform for Wuppertal City, Germany. The research integrated real-time temperature data, a Machine Learning (ML) model, and remote sensing data to predict the temperature and assess the impact of urban planning scenarios on UHI ahead of implementation. This research is divided into three parts: data analysis, ML model training, and DT-PSS creation. Nine variables are selected as predictor variables of LST from the literature: Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Normalized Difference built-up Index (NDBI), Patch Density (PD), Edge Density (ED), Aggregation Index (AI), population density, Land use, and Digital Elevation Model (DEM). Among the selected variables, NDVI, NDBI, and NDWI showed the highest correlation with LST. The study used four different regression models to train them based on the predictor variables: Random Forest (RF), Support Vector Machine (SVM), Artificial Neural Networks (ANNs), and Polynomial regression mode to predict LST and UHI. The RF model demonstrated high accuracy in predicting LST, with an R-squared value of 0.86 and a Mean Absolute Error (MAE) of 1.09, outperforming other ML models. To create DT-PSS, a 3D city model of Wuppertal is developed in City Engine and imported into the UE to reduce the computational workload for the UE platform. The trained RF model and real-time temperature data from sensors in Wuppertal City are imported into UE to include prediction capability to the DT-PSS. Integrating the RF model and sensor data into the UE allowed for real-time updates and scenario assessment. The DT-PSS prototype is developed for a small neighbourhood as a test area. The DT- PSS performance is evaluated by holding a workshop with Wuppertal municipality officials. Feedback is received regarding tool usability and the potential to enhance the performance of the tool for future research. The research also found that involving stakeholders from the beginning step is one of the primary and fundamental requirements in the process of creating and designing a DT-PSS. User requirements and needs should be met since they are the final users of the tool. The DT-PSS suggested in this study provides urban planners with an innovative tool to evaluate the effects of different planning/mitigation scenarios on urban temperature and UHI. These research findings proved that satellite imagery provides valuable insights regarding the historical and current data and the state of land cover and temperature. This research also identified several limitations, such as data availability, GIS software and UE interoperability, and computation constraints. However, the findings of this research highlight the potential of the proposed DT-PSS tool for broader application. Further research should focus on enhancing model accuracy and data integration and including more predictor variables to expand the functionality of the tool for more complex and detailed urban scenarios such as green roofs or redesigning streets with vegetation. |
| Item Type: | Essay (Master) |
| Faculty: | ITC: Faculty of Geo-information Science and Earth Observation |
| Subject: | 38 earth sciences, 74 (human) geography, cartography, town and country planning, demography |
| Programme: | Geoinformation Science and Earth Observation MSc (75014) |
| Link to this item: | https://purl.utwente.nl/essays/104873 |
| Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Show download statistics for this publication
Show download statistics for this publicationRepository Staff Only: item control page