AI-Driven Time-Series Forecasting for Predicting Machine Failure Risk in Industrial Manufacturing Environments Using Real-Time Multisensory Data

Abstract

The combination of Artificial Intelligence (AI) and the Industrial Internet of Things (IIoT) has changed manufacturing into a data-driven environment. However, many industries still depend on reactive maintenance, which sends alerts only after issues occur. This results in downtime, financial losses, and safety concerns. This study suggests a mixed AI framework for predicting machine failure risks by using real-time data from multiple sensors. It combines Random Forest (RF) for classifying maintenance needs and Long Short-Term Memory (LSTM) networks for forecasting sequential time-series data. By merging these methods, the system can detect anomalies instantly and predict future trends, providing early warning alerts before operation limits are exceeded. The research uses the Smart Manufacturing IoT-Cloud Monitoring Dataset, which contains 100,000 sensor records from 50 industrial machines. Selected features include temperature, vibration, pressure, humidity, and energy use due to their links with mechanical wear. The results show that the RF model achieved 89% accuracy but had low recall (~45%) for failure events. In contrast, LSTM reached 80% accuracy with significantly higher recall (~78%). This highlights the value of recall and clarity in safety-critical situations. The hybrid model addresses the usual limitations of predictive maintenance by creating a two-tier intelligence system: RF gives immediate classification, while LSTM constantly forecasts sensor behavior. This setup enables real-time alerts based on risk priorities. The system also supports edge computing deployment, reducing latency in detecting anomalies and enabling proactive measures. Overall, AI-driven predictive analytics can greatly cut unplanned downtime, prolong the life of machinery, and improve workplace safety. This redefines predictive maintenance as a strategic intelligence layer in modern manufacturing.