An Optimized, Edge-Computed Digital Twin Framework for State Estimation and Thermal Protection of Low-Voltage DC Motors

Abstract

This paper presents a computationally efficient, localized Digital Twin (DT) framework engineered for real-time state tracking and thermal protection of low-voltage DC motors operating under dynamic load conditions. Traditional edge-monitoring systems rely heavily on static threshold alerts, which fail to capture non-linear degradation characteristics. To address this limitation, the proposed architecture deploys an embedded cyber-physical framework using an ESP32 edge-node integrated with an algorithmic Virtual Twin layer running via Node-RED. The Virtual Twin evaluates a dynamic state observer model alongside an enhanced thermal Arrhenius relationship to continuously track the battery State of Charge (SoC) and winding temperature concurrently. Real-time telemetry is transmitted over a lightweight, local messaging topology to ensure seamless bidirectional actuation. Experimental evaluations demonstrate a total fault isolation latency of 34 ms during critical thermal anomalies, proving that the system successfully mitigates cascading power cell destruction within resource-constrained Industrial Internet of Things (IIoT) applications.