Authors : Walaa H. Elashmawi, Nesma A. Kotp and Ghada El Tawel
Abstract: Monitoring, analyzing and diagnostic faults of industrial gas turbine are not an easy way by using conventional methods to complexity of faults. Artificial neural network is deem an active tool to analysis and diagnose faults. Here, we suggest an efficient neural network model due to monitor, analysis and diagnose faults of the gas turbine engine for on-line treatment with a twofold advantage. First, the model is able to diagnose the fault in case of uncertainty or corrupted data by using semi-intelligent artificial neural network. Second, it can predict the extent of the deterioration of the performance efficiency of the turbine engine by using intelligent artificial neural network through a simple GUI. The experiment has been done on five faulty conditions and the proposed neural network model tested with new dataset. The results have proven that, the proposed model produced satisfactory results with 10-10 Mean Square Error (MSE) that considered optimal results when compared with training data sets.
Walaa H. Elashmawi, Nesma A. Kotp and Ghada El Tawel, 2018. A Novel Proposed Neural Network MAD (Monitoring, Analysis and Diagnose) Model for Industrial Gas Turbine. International Journal of Soft Computing, 13: 92-101.