Vis enkel innførsel

dc.contributor.advisorHjelmstad, Jens F.nb_NO
dc.contributor.advisorCella, Tommaso
dc.contributor.authorChase, Briannb_NO
dc.date.accessioned2014-12-19T13:49:00Z
dc.date.accessioned2015-12-22T11:48:53Z
dc.date.available2014-12-19T13:49:00Z
dc.date.available2015-12-22T11:48:53Z
dc.date.created2013-12-13nb_NO
dc.date.issued2013nb_NO
dc.identifier679231nb_NO
dc.identifierntnudaim:9477
dc.identifier.urihttp://hdl.handle.net/11250/2370912
dc.description.abstractGround Surveillance Radar (GSR) requires a phased array, conformal cylindricalantenna solution that has specific features (electronically scan the azimuth andelevation, with angular width of ca. 6° in both directions, and dimensions of ca. 30 cm inradius and height), working between 10 GHz and 10.5 GHz. The task is to prepare therequirements and specifications for the antenna in cooperation with the system user,then conduct a study and simulate the resulting performance where analysis are carriedout in CST Microwave Studio, and MATLAB. Then the results are evaluated in relation tothe required specification. Efforts seek to define the optimal parameters for theantenna.With that in mind, the suitable antenna structure was chosen to be a microstrippatch. These antennas have the advantages of light weight, high gain, conformability tovarious surfaces, low cost of fabrication, and versatility in terms of input impedance.Their ability to integrate with microwave integrated circuits, plus their low profilesmakes them an unmatched rival for other antenna types.As the first step, a single element, single layer rectangular microstrip patch antennawas designed and simulated. Next, a planar array consisting of these single elementswas optimized and synthesized. Antenna simulations in this work have been mainlycarried out in CST Microwave Studio, and some of them in MATLAB.The studies began by calculating the necessary dimensions for the single elementpatch antenna, and then parameterizing the CST with the obtained dimension values,and simulation in CST. As in any practical system, there is always a deviation betweenthe theoretical values and the practical results. Thus some changes had to be done in theantenna dimensions, so that it fitted the system requirements and needs; requirementssuch as return loss value (S11) and the resonance frequency (10.25 GHz) of the antenna.The simulation values based on the theoretical calculations, were S11=‐9.59 dB at 9.65GHz. By adjusting and optimizing the dimensions of the patch, a reduction of return lossvalue was observed from ‐9.59dB to ‐47.32dB at the desired resonant frequency whichwas 10.25 GHz, indicating a great match at that frequency. Furthermore, some antennaparameters such as the substrate height (h), patch width (W) and length (L), and themetallic layer thickness (Mt) were also changed to see how each of these factorscontributed to the overall antenna performance.As the next step in the design, an arbitrary 24×12 planar array was constructed, andelements were excited with different input excitation signals, and it was observed thatthe binomial signal gave the lowest SLL value of all excitation signals, but the 3dB angular width of the beam (both in θ and φ directions) was too wide to be used in thearray. The lowest values of angular width were obtained by the uniform excitations,which was ideal, but the SLL value was too high (just ‐13.2 dB), which again was outsidethe specification boundary of at least ‐20dB. All of the input excitation signals exhibitedroughly the same amount of directivity. Thus, based on the results, Chebyshevexcitation was chosen that had an SLL value of ‐30 dB, and an angular width of 6.1° and8.2°. On top of all these requirements, the appearance of grating lobes on the radiationpattern had to minimized as much as possible, as they represented the unwantedradiation in directions other than the direction of interest.In the final stage, after choosing the suitable input signal waveform, theoptimization of the spacing?s between the antenna elements was carried out, to havethe lowest value of SLL (obtained: ‐22.5dB) and lowest possible angular width(obtained: 6.1° and 6.5° in the θ and φ directions respectively). Both these values werewithin the boundary of the specifications and thus were satisfying.Furthermore the effect of parasitic mutual coupling in the simulated planar arraywas studied, and it was shown that it degraded the array performance by changing thesingle element antenna input impedance (Zin), which resulted in a mismatch betweenthe feed and the antenna itself, which ultimately caused a different antennaperformance and behavior, e.g. the S11 value changed dramatically.One of the requirements of this antenna solution was its shape and ability to bephased steered. It was supposed to be a conformal, cylindrical, phased array antenna.Unfortunately due to some problems with the CST program, and not having thesufficient and required information to be able to connect and log in to its website (to usethe online help, tutorials and instructions, provided by the CST producers on how tobend, and phase steer the planar structures), the writer, was not able to bend the planararray structure to a cylindrical form. Therefore an octagonal (rather than cylindrical)multi‐beam antenna (MBA) was constructed. Just to have a comparison, a linear arrayin CST is rather easier to be bent, because it extends in only one direction, and on top ofthat, one does not necessarily need to design the feed network for the antenna elements,as they can be fed by the already‐defined discrete ports in CST from the feed entrance ofthe antenna elements, with desired amplitude distribution and phase to the elements toform a linear phased array antenna. Thus bending the structure becomes much easierthan in the planar structure case. A planar array however, extends in both x and ydirections, and therefore it needs special knowledge of the CST program to feed theelements (with the suitable amplitude and phase distribution) and bend the structure.However, despite all these difficulties, one can rely on the works done in the previousstudies [5,6] that basically have bent the microstrip patch antenna and concluded thatthe curvature of the antenna resulted in a significant influence on the fringing fields ofthe antenna, and consequently, an impact on the εreff, which ultimately led to a change in antenna performance. For a given cylindrical antenna with radius R, it was shown in[5,6] that fresonant changed very little, so it could be considered untouched compared tothe planar single element antenna. Furthermore decreasing R resulted in a reduction inresonant resistance and Q factor, while the antenna efficiency and its bandwidthincreased and the radiation pattern widened.nb_NO
dc.languageengnb_NO
dc.publisherInstitutt for elektronikk og telekommunikasjonnb_NO
dc.titleDesign of a Conformal Ground Surveillance Radar (GSR) Antennanb_NO
dc.typeMaster thesisnb_NO
dc.source.pagenumber101nb_NO
dc.contributor.departmentNorges teknisk-naturvitenskapelige universitet, Fakultet for informasjonsteknologi, matematikk og elektroteknikk, Institutt for elektronikk og telekommunikasjonnb_NO


Tilhørende fil(er)

Thumbnail
Thumbnail
Thumbnail

Denne innførselen finnes i følgende samling(er)

Vis enkel innførsel