Open Access

In this thesis a beam-based method has been developed to measure the strength and the polarity of corrector magnets (skew quadrupoles and sextupoles) in circular accelerators. The algorithm is based on the harmonic analysis (via FFT) of beam position monitor (BPM) data taken turn by turn from an accelerator in operation. It has been shown that, from the differences of the spectral line amplitudes between two consecutive BPMs, both the strength and the polarity of non-linear elements placed in between can be measured. The method has been successfully tested using existing BPM data from the SPS of CERN, since presently the SIS-18 is not equipped with the necessary hardware. The magnet strength of seven SPS extraction sextupoles was measured with a precision of about 10%. The polarities have been unambiguously measured. This method can be used to detect polarity errors and wrong power supply connections during machine commissioning, as well as for a continuous monitoring of the "nonlinearity budget" in superconducting machines. A second beam-based method has been studied for a fast measurement and correction of betatron coupling driven by skew quadrupole field errors and tilted focusing quadrupoles. Traditional methods usually require a time-consuming scan of the corrector magnets in order to minimize the coupling stop band |C|. In this thesis it has been shown how the same correction can be performed in a single machine cycle from the harmonic analysis of multi-BPM data. The method has been successfully applied to RHIC. It has been shown that the stop band |C| (also known in the American literature as Delta-Qmin) measured in a single machine cycle with the new algorithm is compatible with the value obtained by traditional methods. The measurement of the resonance phase Theta defines automatically the best corrector setting, which was found in agreement with the one obtained with a traditional scan. A third theoretical achievement is a new description of the betatron motion close to the difference resonance in presence of linear coupling. Compared to the matrix formalism the motion is parametrized as a function of the resonance driving term f1001 only (which is proven to be an observable), whereas making use of the matrix approach four parameters need to be measured. Formulae describing the exchange of RMS emittances when approaching the resonances have been already derived in the 70s in the smooth approximation. New formulae have been derived here making use of Lie algebra providing a better description of the emittance behavior. The emittance exchange curves are predicted by new formulae with excellent agreement with multi-particle simulations and the counter-intuitive emittance variation along the ring of the emittance is proven to be related to the variation of f1001. A new way to decouple the equations of motion and explicit expressions for the individual single particle invariants have been found. For the first time emittance exchange studies have been carried out in the SIS-18 of GSI. Transverse RMS emittances have been measured during 2005 from rest gas monitor (RGM) data. Crossing the linear coupling resonance, the transverse emittances exchange completely. It has been observed that this effect is reversible. Applications of this manipulation are: emittance equilibration under consideration for future operations of the SIS-18 as booster for the SIS-100; emittance transfer during multi-turn injection to improve the eficiency and to protect the injection septum in high intensity operations, by shifting part of the horizontal emittance into the vertical plane. The emittance exchange curves obtained experimentally have been compared with analytic formulae providing a fast measurement (in few machine cycles only) of the linear coupling stop band |C|. Technical problems prevented the use of the eight skew quadrupoles installed in the SIS-18 to compensate the linear coupling resonance. It has been observed that the emittance exchange curve is highly sensitive to the beam intensity. Multi-particle simulations with 2D PIC space-charge solver have been run to infer heuristic scaling laws able to quantify the observable stop band, to be used for the resonance compensation. The analysis of BPM and RGM data has been performed making use of new software applications developed for this purpose. The bpm2rdt code for the harmonic analysis of BPM data has been written and tested with real data. The software reads the BPM turn-by-turn data and the Twiss parameters. Then it performs the FFT of these data, finds the peaks of the Fourier spectra and infers the RDT fjklm, the strengths ^hjklm and the local terms lambda-jklm. All these observables are printed out together with the corresponding values of the model, computed from the nominal values of strengths and the Twiss parameters. From the FFT of dual-plane BPM data the linear optics (beta functions and phase advances Delta phi) at the corresponding location is also inferred. From the measurement of f1000, the linear coupling coeffcient C (amplitude and phase) is also computed. The code has been tested by using existing SPS data and new RHIC data. For the on-line analysis of RGM data the rgm2emitt code has been written. The application reads in input the raw data files from the RGM and the beam loss monitor (BLM) respectively, the latter created by the RGM on-line software itself. From the RGM data the transverse beam sizes and emittances are inferred and used together with the BLM data to compute the tune shift during the machine cycle.