Master's Thesis Lukas Schmitt


Thermographic detection of energetic and comfort-relevant parameters for controlling indoor climate

Thermographic reconstruction of a measurement scene Copyright: EBC Thermographic reconstruction of a measurement scene

Common HVAC-Systems typically use the indoor air temperature as regulating variable for maximizing the indoor thermal comfort. However, well-established comfort models prove that solely using this parameter is not always sufficient in this context. This suggests that control strategies can be improved through a model providing additional comfort-relevant parameters.

The presented thesis proposes a method for thermographically detecting and assigning indoor surface temperatures in a real-time application. A geometrical calibration enables the integration visualization and assessment of measurement data of multiple infrared sensors in a joint model. In order to assign the measured temperatures, a ray-tracing algorithm is used. The above-mentioned processes result in an adaptive thermographic reconstruction of the measurement scene providing the mean radiant temperature, radiation asymmetry information and the vertical air temperature stratification as comfort-relevant parameters. The latter is determined based on an energy balance at the wall surface in various heights. Radiative, convective and conductive heat exchange is calculated using ray tracing, empirical correlations and the finite-difference method.

A field test validates the proposed method. The ray-tracing algorithm requires high model resolutions and a precise geometrical calibration to detect temperatures of low-emissivity surfaces properly. Otherwise distorted and reflected temperatures can cause high measurement errors. The mean radiant temperature and the radiation asymmetry can be detected appropriately, independent of the model resolution. Calculating the vertical temperature stratification exposes a high sensitivity of the model for detecting surface temperatures and the wall temperature distribution properly, because small temperature differences within the measurement uncertainty must be distinguished. Thermal comfort is finally assessed using the PMV-model with regard to the detected parameters.