Detect faults in gas turbine installation, using adaptive network with fuzzy inference

N. Hadroug and A. Hafaifa, University of Djelfa, Algeria; and M. Guemana, University of Médéa, Algeria Automatic fault detection becomes more essential as the modernization of industrial facilities grows more complex, to the detriment of the human operator. This article presents a fault diagnosis system based on artificial intelligence techniques (fuzzy and neural networks systems) applied to a gas turbine installed at a gas compression station in Algeria. This technique, with its generalization capabilities and memory, provides an effective diagnostic tool, giving the operator more information on the behavior of the gas turbine in different operating modes. Fault detection challenges. Fa