Modification of building energy simulation tool TRNSYS for modelling nonlinear heat and moisture transfer phenomena by TRNSYS/MATLAB integration

Software for numerical simulation of various types of energy used in buildings, i.e. building energy simulation (BES), have become an essential tool for recent research pertaining to building physics. TRNSYS is a well-known BES used in both academia and the construction industry for a wide range of simulations, such as the design and performance evaluation of buildings and related facilities for heating, cooling, and ventilation. TRNSYS has a modular structure comprising various components, and each component is interconnected and compiled through a common interface using a FORTRAN compiler. Its modular structure enables interactions with various external numerical simulation tools, such as MATLAB, Python, and ESP-r. For ordinary simulations of building energy load using TRNSYS, the generic module Type 56 is usually recommended, which provides detailed physics modelling of building thermal behaviours based on unsteady energy conservation equations and Fourier's law for each building material. However, Type 56 explicitly depends on the transfer function method to discretise the original differential equations; therefore, it cannot model nonlinear phenomena, such as latent heat and moisture transfer between a building surface and ambient air. In other words, the current TRNSYS cannot be used to estimate the effectiveness of evaporation during cooling, which is a typical passive design method. Hence, the authors developed a MATLAB/TRNSYS integration scheme, in which TRNSYS was modified to model simultaneous heat and moisture transfer from the wet roof surface of a building. This scheme enabled TRNSYS to calculate the rate of evaporative heat and moisture transfer dynamically from the roof surface, assuming a control volume approximation of the roof surface. Finally, the effect of evaporative cooling on the thermal performance of an Indian building was estimated using the modified model.