SISTEM PEMELIHARAAN PREDIKTIF BERBASIS IoT UNTUK PEMANTAUAN KONDISI MOTOR MENGGUNAKAN ANALISIS GETARAN: IoT-Based Predictive Maintenance System for Motor Condition Monitoring Using Vibration Analysis

Abstrak

The implementation of predictive maintenance systems in industrial environments has become increasingly critical for preventing unexpected equipment failures and optimizing operational efficiency. This study develops an IoT-based predictive maintenance system for motor condition monitoring using vibration analysis and linear regression for trend prediction. The system integrates a Vemsee SN-3001-WZ1-N01-CX vibration sensor with RS-485 communication, Python-based data processing, Firebase Realtime Database, and a mobile application for remote monitoring. The vibration sensor continuously monitors motor condition and transmits data to a Python processing unit, which implements linear regression algorithms to analyze trends and predict threshold levels according to ISO 10816-1 standards. Experimental validation was conducted through systematic monitoring campaigns accumulating 48 hours of operational data across five testing sessions, successfully processing 34,560 vibration data samples. Performance evaluation demonstrated linear regression model with R² coefficient averaging 0.874 and prediction accuracy with MAPE values ranging from 4.1% to 12.8% depending on prediction timeframe. The system achieved 99.2% communication reliability and 98.7% overall uptime while maintaining cost-effectiveness estimated at 60-70% lower than commercial alternatives. These findings indicate that the system provides effective trend prediction for maintenance scheduling, reduces unexpected downtime, and demonstrates the practical application of IoT technology in industrial motor monitoring through real-time data access and mobile visualization.

Kata kunci

• Pemantauan Kondisi Motor