Thermal models of buildings are helpful to forecast their energy use and to enhance the control of their mechanical systems. However, these models are building-specific and require a tedious, error-prone and time-consuming development effort relying on skilled building energy modelers. Compared to white-box and gray-box models, data-driven (black-box) models require less development time and a minimal amount of information about the building characteristics. In this paper, autoregressive neural network models are compared to gray-box and black-box linear models to simulate indoor temperatures. These models are trained, validated and compared to actual experimental data obtained for an existing commercial building in Montreal (QC, Canada) equipped with roof top units for air conditioning. Results show that neural networks mimic more accurately the thermal behavior of the building when limited information is available, compared to gray-box and black-box linear models. The gray-box model does not perform adequately due to its under-parameterized nature, while the linear models cannot capture non-linear phenomena such as radiative heat transfer and occupancy. Therefore, the neural network models outperform the alternative models in the presented application, reaching a coefficient of determination R² up to 0.824 and a root mean square error down to 1.11 °C, including the error propagation over time for a 1-week period with a 5-minute time-step. When considering a 50-hour time horizon, the best neural networks reach a much lower root mean square error of around 0.6 °C, which is suitable for applications such as model predictive control.

Modeling phase change materials (PCMs) thermal behavior requires solving a system of non-linear equations to account for temperature-dependent thermal capacity and thermal conductivity. These properties depend on the PCM temperature and state (solid, liquid or mushy). Most models rely on enthalpy-temperature or specific heat-temperature curves to consider the variable thermal capacity during heating and cooling processes. These curves are generally obtained through experimental methods such as a Differential Scanning Calorimetry (DSC) test or the T-history method. Significant differences can be observed between the results of these methods, due to different experimental conditions. In order to clarify the influence of experimental conditions, experimentations on a bio-based PCM are performed with varying heat transfer rates and different configurations (PCM samples and PCM-equipped walls). Enthalpy-temperature or specific heat-temperature curves are computed for each case using an inverse method. A comparison between the results obtained with different methods and different heat transfer rates shows significant differences. The phase change temperature range obtained with the inverse method applied to the PCM samples is larger than the range obtained with the DSC test. The tests on the PCM-equipped walls show that varying heat transfer rates has a significant impact on the phase change temperature range and the hysteresis between heating and cooling curves. Higher rates increase the hysteresis and shift the phase change temperature range towards colder temperatures. Given the observed differences between properties obtained from different experimental conditions, it is recommended to carefully select the method used to define PCM enthalpy-temperature curves, taking into account the modeling application (PCM configuration and expected heating / cooling rates).