Application of generalized linear models to evaluate nuclear EMP tests

High Altitude Nuclear Electromagnetic Pulses (HEMP) are among most extreme external signals that can interfere with an electric device. To ensure the survival of vital components even in case of a HEMP event, several procedures have been designed to test DUTs in a lab environment. As a consequence, field tests with HEMP simulators are mandatory. A valid estimation of the level which will cause a failure of the DUT due to the external HEMP (either generated artificially or by a real life HEMP event) will allow engineers to fix better field levels for HEMP susceptibility tests and to avoid overtesting. This paper will apply binary regression models such as as logit or probit on the results of HEMP tests from the literature and estimate an more applicable test level as a first step. Concluding, additional possibilities of the regression models are discussed.