Statistical response of EM-driven cables inside an overmoded enclosure

This paper deals with probabilistic modeling of the electromagnetic (EM) response of cables inside a complex cavity subjected to well overmoded EM penetration. Theoretical studies indicate that the field amplitudes and cable currents squared both should have a /spl chi/-square distribution with two degrees of freedom, but our observations indicate that a log normal fit is empirically better unless the data, if experimentally obtained, is first passed through a carefully tailored trend-removing filter. If a cable model is driven by statistically simulated enclosure fields, similar extreme care must be taken with the numerical generation of these driving fields. The major innovation reported here is the development of an algorithm that models cable-drive fields simultaneously having a /spl chi/-square power-flux distribution and the physically mandated local autocorrelation at a spatial point as the frequency is swept or at a fixed frequency as the power flux sensor is moved around to map the cavity response. Nature is quite adept at creating a cable drive with these simultaneous attributes, but computer emulation had proved very exasperating. Our algorithm, as an unplanned bonus, also has the capability to transform random numbers from one distribution to another. For instance, one can input normally distributed power-flux values and obtain as the output /spl chi/-square or log normally distributed power-flux values. The reverse transformations are also allowed.