Determining the load characteristics of domestic induction heaters

The most challenging problems which induction heating has faced with are variety in pans and end-users' preferences. Inductor-pan system is modelled with a series RL circuit composed of one resistor and one inductor. Here, the resistor and inductor's specifications are determined according to the material and diameter of pan, frequency and voltage level of source. In this thesis, five different type of pans are selected according to their materials. Analyses are performed based on measurements collected via LCR meter and an application circuit. The study showed that the resistance changes in accordance with pan type and the frequency. Usually, the equivalent inductance increases when inductor-pan coverage ratio decreases. However, it is not valid for all pan types. Moreover, the properties of supply voltage are another key factor for determining the load characteristics, since the increment or decrement in voltage level changes load characteristics depending on pan type. Awareness of load characteristics is a significant advantage for optimum and reliable design. For this purpose, using the data that are acquired from tests a back propagation algorithm is devised. The load characteristics are determined by using an artificial neural network which is modelled by the devised algorithm. The algorithm obtained from artificial neural network is used in accordance with microcontroller of the induction heater. Comparisons are made according to measurement and artificial neural networks for different pans that are not included to the experiments. According to results, by using artificial neural networks, load characteristics can be determined with high accuracy.