Impedance measurement based on binary tree and stack structure

Power electronic based systems are prone to negative impedance instability due to the constant-power nature of the individual components. When designing these systems, impedance measurement is very important because of the close relationship between impedance and stability. The method by injecting a series of sinusoidal perturbation signals to the system is the most widely used one, but there are still some problems. The frequencies of the injection signals are often given by tests, which is not suitable for all situations. The accurate results of some networks can be obtained by measuring only several points but some are not. It is hard to determine what frequencies should be injected into the network to obtain an accurate result because the impedance is unknown for test otherwise you do not need to do the measurement. In this paper, a method based on binary tree and stack structure is proposed to determine the frequencies of injected signals automatically. In this method, impedance measurement is treated as a process of finding a piecewise linear interpolation function by mathematical abstraction and some theories of numerical analysis are used to solve the problem. Compared to the traditional one by one injection with a constant step which has been set by tester before the measurement starts, the proposed method can automatically select a large step to save test time when the impedance changes slow and a small step to make the results be accurate when the impedance changes dramatically, which is more optimized. The simulation results demonstrate the effectiveness of this new method.