**1. Introduction**

The term 'passive buildings' has become common nowadays. However, when this term was coined, it originally related to a building that included passive solar systems [1] and was designed to maximize solar gains [2]. Although the passive solar energy gains in buildings are predictable and controllable, protecting buildings against overheating is often neglected during the design process, leading to the construction of buildings with high cooling loads even in heating-dominated climates [3] and in mild ones [4]. This occurs because the measures to protect buildings against the risk of overheating are commonly poorly understood given the dynamicity of solar heat gains.

The lack of thermal inertia in many modern buildings exacerbates the risk of reaching too high indoor temperatures. In the case of lightweight buildings, there is no opportunity to accumulate the available exterior (solar) energy gains for later use efficiently [5]. This is why phase change materials (PCM) that store large amounts of energy thanks to the latent heat needed for a phase change process have been proposed over the last two decades [6].

Specific PCM algorithms have been added to various kinds of simulation tools, as the dynamic control of PCM and of their phase change process requires to be controlled dynamically. One of the largely adopted tools for PCM simulations is EnergyPlus™, which will be considered in this study, and it is briefly presented in the next section.

#### **2. Simulation Method of Phase Changing Materials in EnergyPlus**™

The first EnergyPlus™ version, including a PCM simulation algorithm, was released over a decade ago. Pedersen et al. developed an implicit finite difference thermal model of building surfaces that simulates the performance of PCM using their enthalpy law [7]. This model has been incorporated into EnergyPlus™ and combined with the general transformation based algorithm.

In recent years, many studies have been performed on the validation and verification of the results of the EnergyPlus™ calculations using different procedures: correspondence between the measured and simulated wall surface temperatures [8], heat flux density [9], and internal air temperatures [10], just to cite a few. These works did not lead to modification of the source code of the program, but only to the comparison and assessment of the reliability of the results.

The subjects of the simulation were mainly wall components with PCM. The impact of these components on thermal conditions in the simulated rooms and on the thermal properties of the entire building has been examined in previous research using fiber insulation by Kosny et al. [11] and concrete tiles by Naraid et al. [12]. In most cases, the authors reported a relatively strong agreemen<sup>t</sup> between simulation results and test data [13–15]. In many of these studies, the authors benchmarked EnergyPlus against controlled field data [14] and confirmed the value of this software.

Numerous simplifications are often used during numerical modeling. For example, the complex geometry of the products containing PCM is replaced with a homogeneous layer to which the enthalpy data is assigned. It is not surprising that the results of some experimental investigations performed did not find a strong agreemen<sup>t</sup> between the test results and the simulation ones [16]. For example, Castell et al. [16] evaluated software simulations in EnergyPlus™ with measurements done inside 3 × 3 × 3 m test cells in Spain. The authors suggested that the reason for the discrepancy of the results could be the inconsistency of the weather data used for calculations and the actual variable thermal conditions. Recent studies indicating strong agreemen<sup>t</sup> between simulation and research results in a moderate United States of America climate [17] and in the sub-tropical one of Hong Kong [18] confirmed that the right weather data and material data are key factors for reducing the discrepancy between simulation and test results. Lee et al. showed that the differences between experimental and predicted total heat transfer values were under 5% [19].

The results of calculations in the EnergyPlus™ program were also compared with the results from the HEATING program [20]. Errors in the routing of EnergyPlus™ were, hence, corrected over the years [21]. For example, in version 8, code modifications allowed an acceleration of the calculation process and the inclusion of variable values for the thermal conductivity coefficient.

In EnergyPlus™, the calculation of the energy balance of building surface constructions is based on the conduction transfer function (CTF), but in the case of more advanced constructions, such as PCMs or variable thermal conductivity, a more flexible approach in the form of a conduction finite difference algorithm is used [22]. Because of the implicit solution of the equation set, it is more efficient to set a time step shorter than those used for the CTF solution algorithm [20]. Tabares-Velasco recommended a simulation time step of fewer than three minutes for a more accurate prediction of the behavior of PCM [21]. This approach is critical to allow to fully control the dynamic changes that occur in a PCM during a phase transition. However, one of the critical elements that still remain poorly modeled is the hysteresis of PCM, an aspect that is described in the next section.
