Synthesis and Characterization of Multi-level Pseudo-Random Sequences as Excitation Signals for System Identification

Identifying the frequency characteristics of systems prone to noise and non-linear distortions, such as grid-connected components, requires careful consideration of the excitation signal. For instance, to obtain the linear behaviour of such a system, the effects of the non-linear distortions should be reduced in the system identification procedure. This paper discusses the synthesis and suitability of multi-level pseudo-random sequences for system identification in the presence of harmonic distortion. Three different methods of synthesising multi-level pseudo-random sequences are presented and the properties of the resulting signals are discussed. It is shown that the design of these sequences can be optimised such that certain harmonics are suppressed. This is beneficial for minimizing, detecting or separating the effects of nonlinear distortions in an identification experiment. A method that uses a field programmable gate array is developed to generate controllable multi-level pseudo-random sequences. A frequency response estimation of an equivalent circuit model of a lithium-ion battery is conducted to demonstrate the use of multi-level signals in system identification. It is shown that the estimation corresponds well with the analytical response.