Algorithmic approach to predictive maintenance

The rapid adoption of IoT-based smart-building technologies is enabling predictive maintenance that outperforms traditional reactive or purely scheduled strategies. This work presents a practical end-to-end workflow for predictive AHU maintenance by combining real-time sensor data, machine learning, and a BIM-based digital twin. In a case study, an industrial AHU system was created with IoT sensors (18 zone temperatures) and simulated over 11 days under both normal conditions and injected faults (e.g. valve malfunctions, stuck valves, filter blockages, fan stress), yielding labeled datasets of healthy vs. faulty behavior. After data cleaning and feature engineering to capture temporal patterns, two complementary models were trained: a Random Forest classifier for real-time fault detection and diagnosis, and a Long Short-Term Memory (LSTM) network for sequence-based early fault prediction. With Random Forest model we have these three alarm level, fault probability, and fault code for each timestamp as a label. This makes it possible to find problems (like overheating or a blocked filter) right away with more than 90% accuracy. The LSTM model learns how the sensors change over time, reaching an accuracy of 90% by sending out early warning signals then unusual trends continue across multiple readings. then the outputs put into a live digital twin (Building Information Model), which keeps each virtual sensor update with the most recent values and shows AI-driven alerts. The twin uses color-coded heatmaps and dashboards to show facility managers the current temperature, Air pressure , air flow and alarm status in real live time. by Combining AI and digital twin technology it creates a powerful predictive maintenance tool that can find, diagnose, and even predict AHU problems, which reduces unplanned downtime and makes maintenance scheduling more efficient. The framework is modular and can be expanded to include more types of sensors also, like humidity, and vibration, for more thorough monitoring. This will make it easier to use smart predictive maintenance in more industrial settings.