Advanced Diagnosis Techniques for Radio Telescopes in Astronomical Applications

The performance of radio telescopes in astronomical applications can be affected by structural variations due to: 1. Misalignment of the feeding structure, resulting in a lateral or axial displacement of the receiver; 2. Wind stress; 3. Gravitational distortion as the antenna is tilted; 4. Thermal distortion with ambient temperature or sunlight. Diagnosis methods are necessary to estimate any deviation of the antenna system from its nominal behavior in order to guarantee the maximum performance. Several approaches have been developed during the years, and among them the electromagnetic diagnosis appears today as the most appealing, because it allows a relatively simple measurement setup and a reduced human intervention. Electromagnetic diagnosis is based on the acquisition of the antenna Far Field Pattern (FFP), with the Antenna Under Test (AUT) working in receiving mode. A natural radio star or a satellite beacon provides the signal source. The acquisition of the FFP typically requires a very large number of field samples to get the complete information about the AUT, and the subsequent measurement process may span over several hours. A prolonged acquisition has significant drawbacks related to the continuous tracking of the source and the inconstancy of the environmental conditions. The purpose of the PhD activity has been focused on an optimized formulation of the diagnosis of radio telescopes aimed at reducing the number of field samples to acquire, and so at minimizing the measurement time. A diagnosis approach has been developed, based on the Aperture Field method for the description of the AUT radiation mechanism. A Principal Component Analysis (PCA) has been employed to restore a linear relationship between the unknowns describing the AUT status and the far field data. An optimal far field sampling grid is selected by optimizing the singular values behavior of the relevant linearized operator. During the activity, a computational tool based on Geometrical Optics (GO) has been developed to improve the diagnosis approach. Indeed, once the Aperture Field is recovered from the inversion of the measured FFP, an additional step is required to assess the AUT status from the phase distribution. Obviously, the computation of the phase distribution should be based on efficient algorithms in order to properly manage electrically large reflectors. The developed GO technique relies on the Fast Marching Method (FMM) for the direct solution of the eikonal equation. A GO approach based on the FMM is appealing because it shows a favorable computational trend. Furthermore, the explicit solution of the eikonal equation opens the possibility to set up an inverse ray tracing scheme, which proves particularly convenient compared to direct ray tracing because it allows to easily select the minimum number of rays to be traced. The FMM is also amenable for parallel execution. In particular, in the present work, the Fast Iterative Method has been implemented on Graphics Processing Units (GPUs). Moreover, the FMM has been accelerated by introducing a tree data structure. The tree allows to manage the mutual interactions between multiple scattering surfaces and the parallelization of the ray tracing step. The method has been numerically tested on simple canonical cases to show its performance in terms of accuracy and speed. Then, it has been applied to the evaluation of the Aperture Field phase required by the reflector diagnosis. During the research activity, the problem of validating the diagnosis algorithms has been also faced. Obviously, a numerical analysis can been carried out to test the model employed to describe the system and to evaluate the performance of the algorithm. To this end, a reliable commercial software exploited to simulate reflector antennas has been exploited. However, to complete the analysis, the experimental validation becomes mandatory, and an experimental outdoor far field test range is required. Accordingly, a test range has been set up thanks to the collaboration with Istituto Nazionale di Astrofisica (INAF) of Naples, Italy. Its realization has involved the full development of the software to drive an Alt-Azimuth positioner and to remotely control the instrumentation. In addition, an upgrade of the internal connections of a Vector Network Analyzer has been performed in order to allow the interferometric acquisition