Thesis:

Three-phase induction motor troubleshooting using adaptive neural network technique

As one of the key and widely used components in various industries, including petrochemical, power plants, aerospace, and heavy industries, three-phase induction motors play a fundamental role in converting electrical energy to mechanical energy. These motors are an ideal choice for industrial applications due to their simplicity of design, strength and high reliability. However, problems such as aging and erosion can lead to various defects and breakdowns, which will have negative effects on the performance and productivity of industrial systems. Since even a small fault in an electric motor may cause total system failure and reduced production, fast and accurate fault detection is of particular importance in order to prevent unexpected shutdowns and reduce maintenance costs. In this research, ANFIS method is used to identify and diagnose electrical and mechanical defects of induction motors. By using the input data from the engine and analyzing them with ANFIS, the defects have been detected and the detection accuracy has been calculated by comparing the frequencies available in the Fourier analysis (FFT) with a certain threshold. The research results show that the proposed ANFIS method has been able to provide high accuracy in the identification and diagnosis of defects and effectively help to reduce costs and times related to repairs and maintenance.

Keywords: Three-phase induction motor, adaptive neural network (ANFIS), Fourier analysis (FFT), fault diagnosi