The need for fail-safe operation of permanent magnet synchronous machines in applications requiring high reliability has prompted the development of techniques for detecting incipient stator winding faults, rotor eccentricities and demagnetization of permanent magnets. It is desired to detect the presence of a fault during normal operation to avoid catastrophic failures. Fault mitigation may allow post-fault operation; however, deploying the proper mitigation scheme requires categorization of the fault type and severity. In this talk, I will describe how Flux2D aided the development of a fault detection and separation algorithm.