An Empirical Method for Estimating Thermal System Parameters Based on Operating Data in Smart Grids

An experimental methodology was developed for online system identification of a thermal system or heated space. In this setting, the intelligent controller detects system parameters during normal operation and adapts its performance accordingly. The ultimate goal is to demonstrate that load leveling with demand side management can be used to reduce peak power consumption while maintaining residential room temperatures at a comfortable level. A prototype enclosure was built and equipped with a heater and thermal measuring equipment. Data was collected during a 17 hour temperature regulation experiment using a bang-bang controller similar to those commonly used for residential heating control. First and second order mathematical models were developed for thermal system identification. The mathematical models utilized the collected temperature data to estimate the net thermal resistance and capacitance using system identification techniques. Results showed the second order model to match the real system characteristics reasonably well. It was found that even for a small prototype enclosure, the estimated thermal parameters showed quite large values of thermal capacitance which can be a great asset for demand side management and control applications in a smart grid. The system identification method developed here is an important step toward the development of intelligent controllers.