When hybridizing GPS with IRS, the integrity of GPS signals has to be checked so that slowly growing errors on GPS measurements don't affect inertial calibration. Several solutions exist to perform fault detection and exclusion of GPS signal. Some of these depend only on GNSS information but the capacity of RAIM (Receiver Autonomous Integrity Monitoring) is limited. This algorithm greatly depends on the constellation geometry (at least 5 SV are needed for detection) and doesn't have sufficient availability for stringent phases of flight. AAIM (Aircraft Autonomous Integrity Monitoring) algorithms are taking into account additional onboard information such as IRS. The Solution Separation method can be implemented for RAIM or AAIM. The aim of this paper is to describe the technique inspired from Solution Separation used to compute isolation radius in addition to the classical detection protection level, and to present some simulation results we obtain by implementing this algorithm in a tightly coupled Kalman. After reviewing civil aviation requirements and defining the tightly coupled Kalman filter that was used, the detection and isolation satellite failure method is described. This method is inspired from the Solution Separation method which is generalised to N subsets of N-1 Kalman filters used to define the isolation radius as presented in [2]. Finally simulated results of the algorithm for NPA are shown.