A fast method for simulating single room dynamic thermal loads using model order reduction

This paper presents a fast computational model for simulating indoor dynamic temperature and thermal load using model order reduction (MOR) algorithms. By analyzing the building's thermal behaviour, thermal balance equations for the building envelope and indoor air are formulated. The state-space method is employed to discretize these equations in both time and space domains, leading to the original system (OS). A typical model is solved numerically using the direct numerical solution (DS) method, and the results are found to be consistent with the ASHRAE 140 thermal load simulation data. The DS method consumes significant computational resources. To enhance efficiency and reduce resource consumption, the Krylov Subspace (KS) and Laguerre orthogonal polynomial (LP) MOR algorithms are introduced. These result in relative errors below 10⁻⁵ and reduce simulation time by approximately 92.2% for yearly simulation.