Calculate&Plot the evaluated S/N ratios ofTaguchi Method L18

Calculation of optimum inputs using the Taguchi Method

Squirrel-cage induction motors are the most widely used type of electric motors in modern industry. The investigation of motor faults and their dynamic behavior requires an accurate simulation of a healthy motor. For dynamic simulation, appropriate input and output parameters must be defined. In this study, stator resistance, stator inductance, rotor bar resistance, rotor bar inductance, rotor end-ring resistance, and rotor end-ring inductance were selected as input parameters. These parameters represent measured values rather than optimized values. The selected output parameters were rotational speed (rpm), torque, and stator current, which are also provided in the nameplate data of the actual motor.

Using MATLAB/Simulink, the simulated motor outputs were adjusted to closely match the nameplate data of the laboratory motor. This enabled comparison between simulated motors under healthy and faulty conditions, facilitating analysis in both the time and frequency domains, identification of vibration root causes, and investigation of electromagnetic characteristics of faulty motors.

The primary objective of this study was to determine the optimal input parameters that minimize the difference between the simulated outputs and the nameplate values of the actual motor. The Taguchi method was employed for parameter optimization. The input parameters were successfully optimized to achieve minimal deviation from the target values. This approach enabled accurate estimation of output parameters without relying on commercial software, supported by a novel analytical equation developed during the study. All MATLAB implementations are presented in the paper.

The optimized simulation results showed strong agreement with the nameplate data, confirming that the proposed methodology can accurately model a healthy motor. Consequently, the developed model can be used as a reliable basis for dynamic simulations aimed at studying various motor faults.

To determine the optimal values of the six input parameters, the following Taguchi-based procedure was applied:

Step 1: Distribution into Orthogonal Arrays

The parameters Rs,Rb,Re,Ls,Lb,LeR_s, R_b, R_e, L_s, L_b, L_eRs,Rb,Re,Ls,Lb,Le stored in the orthogonal array file (orth18.txt) were expanded into three columns each, corresponding to three experimental levels. For example, RsR_sRs was divided into Rs1,Rs2,Rs3R_{s1}, R_{s2}, R_{s3}Rs1,Rs2,Rs3, where levels 1, 2, and 3 were assigned to the respective columns. The same process was applied to all parameters.

Step 2: Replacement with S/N Ratios

The level indicators in each column were replaced with the corresponding signal-to-noise (S/N) ratios derived from rotational speed, torque, and stator current measurements for each experimental run. The evaluated S/N ratios were calculated using Equation (1) in the referenced work.

Step 3: Reconstruction of the Evaluation Table

The evaluated S/N ratios were reorganized into a table consisting of six rows (parameters) and three columns (levels).

Step 4: Taguchi Graph Analysis

Taguchi response graphs were generated using the evaluated S/N ratios to identify the optimal level of each parameter.

This work builds upon the methodology presented in:

A. Purvee, “Determination of Optimal Parameters of Simulated Motors Based on the Taguchi Method,” IEEE PEDES: Power Electronics, Drives, and Energy Systems.