Large-scale building age classification for urban energy demand estimation

Large-scale building age classification for urban energy demand estimation: Geo and satellite data for building age identification

Garbasevschi, Oana (TU Delft Technology, Policy and Management; TU Delft Policy Analysis)

Korthals Altes, W.K. (mentor)
Verma, T. (mentor)
Lefter, I. (graduation committee)
Wurm, Michael (mentor)
Estevam Schmiedt, Jacob (mentor)
Schiricke, Björn (graduation committee)

Delft University of Technology

Engineering and Policy Analysis

2020-03-27

Urban areas are the biggest consumers of electricity and energy consumption is only likely to increase with rapid urbanization. Out of the urban building stock residential buildings require continuous supply of energy for space heating and appliances. To answer to this demand in a sustainable way policy maker need to design energy efficiency strategies that must rely on accurate and traceable models. These models estimate energy demand based on a series of building features, out of which building age is of prime importance because it predicts the insulation properties of the building. To support the energy modelling process, we propose a method of automatically identifying building age from spatial data at a large scale. We identify features of buildings that are significant for age prediction and determine which set of features has best prediction power at national scale, in Germany. It is expected that the accuracy of classification will be strongly related to sampling design and data availability. The final results will be used to identify the impact of misclassification errors on estimating energy use in urban energy models, providing in this manner a measure of the reliability of such models.

spatial data
building energy model
supervised learning
energy efficiency

http://resolver.tudelft.nl/uuid:ba4511dd-4ba0-43eb-b6c8-ce3896cd5dc3

Student theses

master thesis

© 2020 Oana Garbasevschi