Simulated Annealing Algorithm Of Inversing Atmospheric Duct From Sea Clutter

Atmospheric duct near sea surface is an anomalous atmosphere structure which can guide radio waves to distance far beyond the horizon with less attenuation. The atmospheric duct phenomenon greatly affects the capability of Radar, Communication, Electronic Support Measures (ESM) and some other electromagnetic-radiation based systems. Therefore, forecasting the real-time refractivity parameters in radar working area is both theoretically and practically significant to the studies on radar, communication, electronic support measures system, etc.This dissertation refers the formation and classification of the atmospheric duct as well as its impact on the EM-wave propagation. The forecasting technology of the atmospheric duct is introduced. With the atmospheric boundary layer theory, the modified refractivity profiles of the different atmosphere layers in evaporation duct are discussed. The Georgia Institute of Technology (GIT) model and the permittivity of seawater's empirical formulas are introduced. The relationship between the permittivity and the temperature, the permittivity, the salinity of seawater are respectively discussed along with their influence on electromagnetic wave propagation.By analyzing the Helmholtz equation, the parabolic equation (PE) is obtained as the atmospheric refractivity distribution is inhomogeneous. The distribution of electromagnetic field is induced with the application of Split Fast Fourier transform (Split-FFT) and Discrete Mixed Fourier Transform (DMFT). The dissertation also introduces AREPS software modules which are applicable to calculating electromagnetic field transportation loss. With the DMFT and AREPS the electromagnetic field transportation loss in various atmospheric ducts are calculated.According to simulated annealing algorithm, two types of atmospheric ducts are inversed based on radar clutter simulation data. The inversion processes, inversion error and the rate of convergence were analyzed. By comparing with the cure of real measurement profile, the inversion profile fits well.