For modern large wind farms, it is interesting to design an efficient diagnosis system oriented to wind turbine generators based on doubly-fed induction machine (DFIM). Intensive research efforts have been focused on the signature analysis to predict or to detect electrical and mechanical faults in induction machines. Different signals can be used such as voltages, currents or stray flux. In the case of wind generators, considering that both the machine signals and the control signals are accessible for diagnostic purposes, there are interesting additional possibilities such as the use of the converter modulating voltages. In this paper, a complete system will be analyzed by suitable simulations and experimentations to deeply study fault influence and to identify the best diagnostic procedure to perform predictive maintenance. Moreover, in order to pass from classical systems to wireless equipments, a virtual platform structure will be presented. Its main features are a real user friendly interface, a low maintenance source code and the possibility of remote monitoring of the DFIM by using a basic Internet connection