Integrationof Data Mining and Design of Experiment to Diagnosis Machine Efficiency

Abstract

The study focused on three phase induction motor because it is playing a vital role in production development. The motor may get the problem in some parts may effect in the wholeefficiency. Biannually maintenance increases the life of the machine, but many companies, especially in the third nation, do the maintenance as soon as the machine gets fail. The studyapplied two strategies to diagnosis and improve the efficiency of the machine. Design of experiment conducted to extract the main effect and interactions between variables. Three variables current, voltage, and power factor applied to understand the main effect in the exploited power (P). Power factor recorded the most important factor impacts the exploited power. Then, data mining utilized with two algorithms; random forests and linear regression. These algorithms used to predict the power factor — six variables collected for data mining, current, power losses, voltage, apparent power, resistance, and exploited power. These variables used to predict the power factor. From both strategies, we found that there is a strong relationship between exploited power and power factor, and those variables have a positive impact on the efficiency of the machine. Also, power losses and current have a negative impact on the power factor. Voltage did not give significant important whether in the design of experiment or data mining. Therefore, power losses and current must be controlled to keep the efficiency, and this can be done by regular maintenance by professional workers.