A detailed study on random resonances was conducted to facilitate further beam power ramp-up in the 3-GeV rapid cycling synchrotron of the Japan Proton Accelerator Research Complex. Systematic experiments employing a low-intensity beam revealed considerable excitation of the 2nd-order random resonances, namely horizontal half-integer and linear sum resonances, located just above the current operating point. The half-integer resonance was well compensated by using a trim quadrupole magnet, without simultaneously exciting the other higher-order resonances. Similarly, the linear sum resonance was found to be compensated by making local bumps in the location of sextupole fields. We identified the source of lattice imperfection driving the random resonances as the leakage field from extraction magnets by implementing a theoretical procedure based on resonance driving terms. These resonance compensation schemes have been confirmed to mitigate beam loss even for high-intensity beams and effectively improve the operating point tunability.

In high-intensity ion accelerators where beam density is high and beam velocity is low, the beam receives a significant impulse from the self field. This results in a sudden emittance growth over a short distance. This emittance growth manifests at the local level, necessitating the employment of analytical and suppression methods that differ from conventional approaches. Conventional methods presuppose periodic structures, whereas the present study employs methods that do not rely on such assumptions. In this presentation, we propose a series of research results, including an in-depth analysis of the conditions under which emittance growth occurs in regions where the influence of the self field is strong. We also put forward a novel suppression method and examine its efficacy.

In the 3-GeV RCS of J-PARC, we have already achieved user operation at the designed 1 MW beam power. The beam loss and the corresponding machine activation have been sufficiently minimized to obtain a stable operation. To cope with user demands, the beam power of the RCS has to be increased far beyond the designed 1 MW. For that purpose, we have performed numerical simulation and beam studies for 1.5 MW beam power by increasing both peak current and injection pulse length of the linac beam injected into the RCS. We have demonstrated an extremely low beam loss rate of ~1E-4 corresponding to a loss power of even less than 0.1 kW against the collimator capacity of 4 kW. The beam loss occurs only at the injection energy localizing well at the collimator section. Recent beam test and simulation results at 1.5 MW beam power are presented.