IMPLEMENTASI METODE ANFIS DALAM MEMPREDIKSI MAINTENANCE PADA MESIN E-FILL

Abstract

This study applies the Adaptive Neuro-Fuzzy Inference System (ANFIS) to predict maintenance times for the E-FILL machine. Rapid technological advancements necessitate effective predictive maintenance systems to optimize machine performance and ensure smooth production processes. This research designs and tests a predictive model using ANFIS, evaluating its accuracy and efficiency based on RMSE values. The results show that the ANFIS model can make accurate predictions, with an RMSE of 0.000003 for training and 0.0001 for testing. The performance of the machine monitored using ANFIS indicated an Overall Equipment Effectiveness (OEE) of 85.31% at 2000 RPM, 67.06% at 2500 RPM, and 65.81% at 3000 RPM. These values reflect stable performance under various operating conditions. These findings suggest that ANFIS can significantly contribute to developing predictive maintenance systems, ensuring timely maintenance and minimizing downtime. This research provides valuable insights into the application of ANFIS in the industrial sector to enhance operational efficiency and productivity.