Online Impedance Identification and Transfer Function Modeling in PHIL Applications

Power Hardware-in-the-Loop (PHIL) provides a safe and reproducible environment for testing real hardware components coupled to simulated networks via power amplifiers (PAs). Ensuring stable and safe operation requires considering the entire control loop behavior, including the transmission characteristics of the hardware under test (HuT). In particular, knowledge of the HuT impedance is essential for implementing a digital twin or time delay compensation in the control loop. This paper presents a robust method for identifying the HuT's frequency behavior online using harmonic injection and discrete rational transfer function modeling. The frequency response is calculated from measured voltage and current signals and fitted using a differential evolution (DE) optimization algorithm, which is resilient to measurement noise, incomplete data, and suitable for non-linear models in the z-domain. Furthermore, a real-time capable approach for estimating the linearized impedance at the mains frequency is demonstrated using a grid-feeding solar inverter. Experimental results confirm the method's effectiveness under both passive and active operating conditions.