Monitoring of rotary machine by mean of self adaptive growing neural network

For the improvement of reliability, safety and efficiency advanced methods of supervision, fault detection and fault diagnosis become increasingly important for many technical processes. To determine the condition of an inaccessible fault in an operating mechanical system, the vibration signal of the machine is continuously monitored by placing sensors close to the source of the vibrations. These signals are further processed to extract the features and identify the status of the machine. This paper deals with a new scheme for the detection and diagnosis of localized defects in machine rotating components based on Self Adaptive Growing Neural Network (SAGNN) classifier which is a novel approach to pattern classification propounded for tackling fault diagnosis tasks. Feature vector is extracted from signal in time domain and SAGNN is trained and used as a diagnosis classifier. In order to evaluate SAGNN, the data sets obtained from vibration signals are classified with both SAGNN and Radial Basis Function (RBF) classifiers. The proposed algorithm is applied to the fault diagnosis of mechanical rotary machine, and the test results show that classification by mean of SAGNN identify the faults of rotary elements (Gears, bearing and belts) more accurately and has better diagnosis performance compared to the RBF.