A gap-filling method for room temperature data based on autoencoder neural networks

Building Simulation Conference proceedings

Abstract

This study explores the applicability of a deep learning-based approach for reconstructing missing room temperature data from different domains where relatively few training samples are available. For that purpose, the existing convolutional, long short-term memory (LSTM) and feed-forward autoencoders were combined with a suitable domain adaptation procedure. Eventually, the developed models were evaluated on data collected in four buildings with significant differences in thermal mass, design and location. The findings pointed out that the domain adaptation can be conducted efficiently by using a small data sample from the target domain. Additionally, the results showed that the proposed model can reconstruct up to 80 % of the missing daily room temperature inputs with RMSE accuracy of 0.6 °C.

Authors 5

1. Romana Markovic

   Karlsruhe Institute of Technology

   Affiliation as printed

   Building Science Group, Karlsruhe Institute of Technology, Karlsruhe, Germany

2. Jérôme Frisch Aachen E3D - Institute of Energy Efficiency and Sustainable Building

   RWTH Aachen University

   Affiliation as printed

   E3D -Institute of Energy Efficiency and Sustainable Building, RWTH Aachen University, Aachen, Germany

3. Andreas Wagner

   Karlsruhe Institute of Technology

   Affiliation as printed

   Building Science Group, Karlsruhe Institute of Technology, Karlsruhe, Germany

4. Francesco Causone

   Politecnico di Milano

   Affiliation as printed

   Department of Energy, Politecnico di Milano, Milano, Italy

5. Christoph van Treeck Aachen E3D - Institute of Energy Efficiency and Sustainable Building

   RWTH Aachen University

   Affiliation as printed

   E3D -Institute of Energy Efficiency and Sustainable Building, RWTH Aachen University, Aachen, Germany