Search

Innovative and efficient ways to perform numerical simulations are worth investigations to reduce the computational complexity of building models and permit to solve complex problems. This paper proposes a Proper Generalised Decomposition (PGD) model reduction technique for heat and moisture multizone modelling. First, the transient multizone problem is written using PGD formalism and solved in a non-incremental way to assess hygrothermal fields. The computational complexity and accuracy of results of the reduced-order model and the large original model are compared, highlighting that the fields are precisely represented with the PGD resolution. The complexity of the reduced-order model is lower, offering interesting numerical savings, particularly when the number of the building zones increases. The second advantage of the method is for solving parametric problems. Two illustrations are proposed. The first considers a climate and ventilation flow rates parametric multizone problem. The PGD model analyse the behaviour of building zones as a function of these two extra-parameters. For the second, an initial condition and source-term PGD parametric solution is computed. It is then coupled with a wall model to perform whole-building hygrothermal simulation. For both studies, the model represents accurately the hygrothermal fields in the different zones with important computational savings. [ABSTRACT FROM AUTHOR]

Innovative and efficient ways to carry out numerical simulations are worth of investigation to reduce the computational complexity of building models and make it possible to solve complex problems. This paper presents a reduced order model, based on Proper Generalised Decomposition (PGD), to assess 2-dimensional heat and moisture transfer in walls. This model is associated with the multizone model Domus using an indirect coupling method. Both models are co-simulated to perform whole-building hygrothermal simulation, considering 2D transfer in walls. The whole-building model is first validated with data from the IEA Annex 41. Then, a case study is considered taking into account a 2-zones building with an intermediary shared wall modelled in 2 dimensions to illustrate the importance of the technique to analyse the hygrothermal behaviour of the wall. It has been highlighted that the whole model enables to perform more precisely analyses such as mould growth on the internal surface. In addition, important theoretical numerical savings (90%) are observed when compared to the large original model. However, the effective numerical savings are not so important (40%) due to the limitations of the co-simulation method. [ABSTRACT FROM AUTHOR]