Implementation of Neuro- Fuzzy and Multi-Layer Perceptron System Intelligent Techniques for Main Fault Diagnosis of Rotating Machinery

Nowadays, Fault detection of rotating machinery by diagnosing sings of starting point and growth of defect using intelligent techniques, discovering the defected parts and the reason behind them and prediction of remaining working life of the machine play an important role in preserve the machine from severe defects and the high price of repairing it. The goal of this paper is using the Adaptive Neural - Fuzzy Inference Systems (ANFIS) and Multi-Layer Perceptron (MLP)for detecting the original defects in rotating machinesincluding unbalancing, Bearing defects, Looseness and misalignment. So,in this study addition to the creation of this mechanism for automatic fault diagnosis, improve accuracy and speed of the network was also performed.Therefore, using the Principal Component Analysis (PCA), the input matrix was reduced to acceptable amontand the effectiveness of the ANFIS and MLP networks in detection of defects were compared with each other.To achieve this goal, mentioned networks were trained using feature vectors extracted from the spectrum frequency and waves.The obtained results showed that for 84 final measurements, the ANFIS and MLP networks have 91 and 78 averages percent successful in detecting the defects, respectively.