Induction motors, often referred to as the workhorse of the industry, play an important role in production today. Therefore, it's important to diagnose any faulty induction motor prior to failure to avoid costly and time consuming production stops.This thesis compares methods for diagnosing induction motors in operation, MCSA, Park's vector approach and the standard method used by Swedish industries today. By simulating broken rotor bars and short-circuit in one of the windings, each method is tested and the results are compared using matlab.The results need further verification due to uncertainties of the condition of the test motor, even though the experiment gives a hint of the different advantages of each method. A combination of the methods should be able to tell if the motor needs further examination.