A convolutional neural network for the hygrothermal assessment of timber frame walls

Currently, no general guidelines exist to determine damage-free timber frame wall compositions, as this requires a comprehensive study, taking into account the variability in boundary conditions. Such a probabilistic assessment is typically carried out through a Monte-Carlo approach, which easily becomes computationally inhibitive. This paper thus makes use of a convolutional neural network as metamodel to replace the hygrothermal model, previously demonstrated to be accurate when predicting the hygrothermal response of massive masonry walls. It is shown that the network can accurately predict the hygrothermal response, and that it can be employed with confidence to estimate the moisture damage risks.