Inverter based renewable generation (RG), especially at the distribution level, is supposed to trip offline during an islanding situation. However, islanding detection is done by comparing the voltage and frequency measurements at the point of common coupling (PCC), with limits defined in the form of ride-through curves. Current practice is to use the same limit throughout the year independent of the operating conditions. This could result in the tripping of RG at times when the system is already weak, thereby posing a threat to voltage security by heavily limiting the load margin (LM). Conversely, heavily relaxing these limits would result in scenarios where the generation does not go offline even during an islanding situation. The proposed methodology focuses on optimizing low-voltage ride-through (LVRT) settings at selective RGs as a preventive control for maintaining a desired steady-state voltage stability margin while not sacrificing dependability during islanding. The proposed process is a multi-stage approach, in which at each stage, a subset of estimated poor-quality solutions is screened out based on various sensitivities. A full continuation power flow (CPFLOW) is only run at the beginning and in the last stage on a handful of remaining candidate solutions, thereby cutting down heavily on the computation time. The effectiveness of the approach is demonstrated on the IEEE 9-bus system.