Application Of Machine Learning In Failure Time Prediction Of Computer Cooling Fan

Abstract

Machines are essential in both life and production. For life, a machine that performs efficiently and smoothly contributes to the replacement of many labors, decreasing workload and raising productivity. For production, machinery will help firms create quicker, with higher product quality, and at a lower cost. Furthermore, the adoption of technology provides firms with the required competitive advantages for long-term success. A machine failure will affect a specific step of manufacturing, affecting the entire process. This results in inefficient manufacturing lines, expensive repairs, project delays, and decreased output. As a result, industrial operations demand solutions that reduce maintenance costs while increasing equipment uptime and availability. Predictive maintenance is a more efficient and cost-effective approach for maintenance. Maintenance should be predicted using situation-appropriate machine learning approaches. This thesis proposes the Artificial Neural Network (ANN), Random Forest (RF), Support Vector Machine (SVM) and Extreme Gradient Boosting (XGBoost) to predict the motor failure time.