Title: Design of Predictive Lifter Chain based on HMI and IOT Dashboard
Authors: Budi Setiyono, Yosua Alvin, Hendrianto Arief Hidayat
Volume: 8
Issue: 11
Pages: 33-43
Publication Date: 2024/11/28
Abstract:
Maintenance is very crucial in the world of manufacturing. PT Toyota Motor Manufacturing Indonesia (TMMIN) is one of the global automotive factories that is aggressively improving predictive maintenance methods on its production machines. This method aims to reduce the potential for repair downtime when production is in progress. In predictive maintenance, various IoT sensors and programming algorithms are used to predict machine conditions. In this study, the LSTM (Long Short Term Memory) method was used which functions to predict the elongation of the lifter chain. In this research, the Human Machine Interface (HMI) interface and IoT Thingsboard Dashboard were also developed as real-time monitoring. Sending data from the OPC Client to the Thngsboard server uses HTTP Request and MQTT communication. In data delivery testing, MQTT has better quality than HTTP requests with a throughput value of 8918627.32 bps, delay0.23 s, jitter 0.03 s, and packet loss 1.017%. Meanwhile, the LSTM algorithm testing carried out obtained prediction results with RMSE values of 2.24, MSE 52.15, MAPE 0.33, MAE 5.78, and R^2 0.996 for the 3-Stacked LSTM model with 1000 epochs. Apart from that, the HMI interface system and Thingsboard IoT Dashboard also succeeded in displaying sensor values for real-time monitoring.