In her paper entitled “Classification of Importance Factors based on their Influence on Extreme Values in HPEM Studies”, Mae AlMansoori, Senior Mechanical Researcher at Directed Energy Research Centre (DERC), at Technology Innovation Institute (TII), has proposed a correlation metric in order to evaluate the influence of a set of given random variables on both the central tendencies and the extreme values of physical quantities related to high-power electromagnetic sources. The test case of this study is a surrogate model of a Vircator, known to be low efficiency High Power Microwave (narrow-band) sources. In such system, mechanical uncertainties as well as the plasma expansion phenomena and shot-to-shot variabilities have a significant influence on their performances. Statistical methods have been applied to central tendencies and extreme values in order to analyze the influence of random variables and particularly to correctly capture the shape of the tails of the probability distributions. Combining the extreme value theory, specifically the peak over a very high threshold approach, and classical descriptive statistics for central tendencies, the peak output power variability has been evaluated. Based on the implemented framework, dominant factors have been ranked depending on their influence on the extreme value statistics as well as central tendencies leading to an efficient management of uncertainties through a correlation estimator. An optimization technique has then been introduced in order to study the possibility of increasing the peak output power leveraging the optimized input parameters. Using a 3D simulation code, the results of the surrogate model have been confirmed. The next steps of this research will be the design and manufacture of the optimized Vircator.
Shown below the 3D model of a Vircator (Figure1.(a)) and the schematic diagram of the system under study showing the considered uncertain parameters (Figure1.(b)).
Figure 1. The test case of this study. (a) Vircator 3D model (b) Scheme of the complete system of a Vircator model with a non-ideal pulsed power source.
This work is part of the PhD project of Mae Al Mansoori related to uncertainties management applied to High Power Electromagnetic sources with a focus on efficiency optimization and in the framework of an international research collaboration composed of Technology Innovation Institute, University of Clermont Auvergne (France) and Helmut Schmidt University (Germany).