Automatic data processing of traction motor measured data and vibration analysis of test bench

Abstract: One of the goals of ABB AB is to develop highly efficient electric motors for traction application. The demand for traction motors is increasing due to the rise in electric vehicles sale and railway locomotive engines. Highly efficient traction motors will assist in reducing the pollution caused by fossil fuels and help make the earth a better place to live by leveraging sustainable energy. The electrical and mechanical characteristics of electric motors are measured and analyzed in the lab. The measured data of the electric motor in the lab are analyzed using the conventional way. One of the significant challenges in a conventional way is to isolate the system with various limitations, and it offers very few choices for measurement. The data management of measured observation readings is affected severely due to this, and it is then risky to determine and analyze the characteristics of electric motors. The first aim is to develop an automatic data processing algorithm for the measurement data collected from the specific setup of the electrical machine. The data processing is done using the MATLAB tool. Statistical methods such as mean, median, moving mean, moving median, Gaussian model for handling missing data, outliers, and data smoothing methods have been implemented to get accurately measured datasets as a part of this thesis. In addition, a study of vibrational analysis of the test bench assembly was performed for the traction motor. The natural frequency of test bench assembly is computed on the Finite Element Method (FEM) tool. All the natural frequencies of the test bench assembly with the traction motor are analyzed, and some of them were closed to the excitation frequency of the traction motor.  This study found that the resonance frequency of the test bench assembly has to be prevented while operating the traction motor during lab to strengthen the life of the test bench.