The safety of utilisation is one of the chief current preoccupations. The user must be able to know what impact the failure of one ore more network elements will have on the continuity of supply, as well as the countermeasures to be taken against degradation of the state of the system. It is therefore necessary to isolate critical instances of faults, to analyse their consequences taking into account the various local automatic mechanisms (adjustment and protection systems), and to search for methods to be implemented for maintaining the network in an assured state of use. A hierarchical approach is presented for solving this problem. Critical instances of faults are selected with the aid of an operations research algorithm specially adapted to the peculiarities of electrical networks. The consequences, both static and dynamic, are analysed by models of increasing complexity