Nonlinear identification modeling and compensating of electromagnetic pulse measurement system

A method of Hammerstein model direct identification under step-like signal excitation is proposed to build a nonlinear model of the pulse voltage measurement system in continuous domain. With the dynamic calibration data in time-domain, the systems Hammerstein model with a four-order polynomial nonlinear element and a three-order linear element in continuous domain is established. The measured results verify the validity of the model. Then a compensation element to improve the sensors dynamic performance is designed, which includes a dynamic linear compensator and a static nonlinear corrector. After compensation, the lower cut-off frequency of the system widens from dozens of kHz to DC, and the response time of the system shortens form s level to ns level. The feasibility of the compensation design is also verified by the reconstruction result of measured pulse.