The impact of a reduced training subspace on the prediction accuracy of neural networks for hygrothermal predictions
- 5 November 2020
- journal article
- research article
- Published by Taylor & Francis in Journal of Building Performance Simulation
- Vol. 14 (1) , 20-37
Abstract
Performing a probabilistic assessment of a building component can easily become computationally inhibitive. To solve this issue, the hygrothermal model can be replaced by a metamodel, which mimics the original model with a strongly reduced calculation time. In this paper, convolutional neural networks are used to predict hygrothermal performance. Because neural networks do not extrapolate well outside their training subspace, it is important to select the training data wisely so that the network can be used to predict for a wide variety of cases, while keeping training time as low as possible. The impact of a reduced training subspace is investigated by training a network on a limited number of wall types or exterior climates and evaluate its prediction accuracy for different wall geometries or other climates. The results showed that is indeed possible to train on a well-considered reduced subspace, while maintaining high accuracy, though it does not necessarily save training time.Keywords
Funding Information
- European Union’s Horizon 2020 Research and Innovation Programme (637268)
This publication has 22 references indexed in Scilit:
- Internal insulation applied in heritage multi-storey buildings with wooden beams embedded in solid masonry brick façadesBuilding and Environment, 2016
- Reliability of Interior Thermal Insulation as a Retrofit Measure in Historic Wooden Apartment Buildings in Cold ClimateEnergy Procedia, 2015
- Interior insulation for wall retrofitting – A probabilistic analysis of energy savings and hygrothermal risksEnergy and Buildings, 2015
- Probabilistic design and analysis of building performances: Methodology and application exampleEnergy and Buildings, 2014
- Multi-objective optimization of a building envelope for thermal performance using genetic algorithms and artificial neural networkEnergy and Buildings, 2013
- Hygric performance of a massive masonry wall: How do the mortar joints influence the moisture flux?Construction and Building Materials, 2013
- Monte-Carlo based uncertainty analysis: Sampling efficiency and sampling convergenceReliability Engineering & System Safety, 2013
- Towards modelling of decay risk of wooden materialsEuropean Journal of Wood and Wood Products, 2010
- Extrapolation limitations of multilayer feedforward neural networks2003
- Spatial and temporal distribution of driving rain on a low-rise buildingWind and Structures, 2002