Smart Sense Ai

Unexpected machine failures cause production losses, maintenance expense, and safety risks, especially in small and medium-sized manufacturing facilities that depend on periodic inspection. This paper presents SmartSense AI, a low-cost, retrofit-oriented condition-monitoring prototype that combines vibration, temperature, ambient humidity, and motor-current sensing with an ESP32-based Internet of Things (IoT) node. Sensor readings are transmitted over Wi-Fi to a cloud data layer, where time-window features are evaluated by a Random Forest classifier and a statistical threshold rule. A web/mobile dashboard presents machine status through green, amber, and red alerts. The source project report describes laboratory testing on a motor/pump setup under healthy and induced-fault conditions, including unbalance and overheating. It reports classification accuracy of approximately 92–95%, dashboard refresh latency below 2 s, and alerts 24–48 h before a simulated failure, with an estimated hardware cost of INR 2,000–3,000 per node. These figures are reported project results; the source does not provide the dataset size, class-wise metrics, train/test protocol, or detailed uncertainty analysis, so independent replication and field generalisation remain necessary. The design offers a practical starting point for affordable predictive maintenance, subject to broader validation.