Online Parameter Estimation for Induction Heating Systems with Quasi-Resonant Inverter

Induction heating has become one of the most dominant heating technology in many industrial and domestic applications. Accurate modeling and control of such heating systems require an estimation of the load and the coil parameters. For the cases without a current sensor, the parameter estimation is usually done at so-called load-detection stage before starting the heating. This process needs, however, a certain execution time, and it is usually implemented by applying computationally-complex algorithms. This paper proposes an online estimation technique to estimate the load and the coil parameters of a quasi-resonant-inverter-topology-based induction heating system. The estimation is performed during the normal operation of the inverter without any pre-operation steps. The proposed method is implemented by fitting the sampled voltage to an exponentially-damped sinusoidal signal. This curve-fitting process is based on an iterative numerical algorithm which does not require any non-linear optimization or complex signal processing steps, thus making the method implementable in the existing digital control platform. An experimental prototype emulating the quasi-resonant-based induction heating system is used to verify the efficiency of the proposed estimation technique.