Abstract
The rapid integration of battery energy storage systems (BESS) into modern power grids necessitates robust safety architectures to prevent catastrophic failures such as thermal runaway. Current diagnostic frameworks are often reliant on isolated, single-parameter thresholds and lack direct linkage to hardware-level execution. This work proposes a comprehensive, multi-parameter early warning and hierarchical protection methodology that bridges the gap between software-based fault detection and practical electromechanical response. The proposed algorithm calculates a dynamic, weighted safety score in real-time by structurally integrating internal battery management system (BMS) metrics, such as cell temperature and voltage deviation, with enclosure-level environmental indicators, particularly early gas emissions. To ensure ultra-low latency, the architecture utilizes discrete analog signal conditioning to bypass computational delays found in complex software models. The assessed score drives a hierarchical state machine branching to a preventive pre-alarm phase initiating HVAC support and load reduction, and a critical protection phase executing closed-loop electromechanical isolation. Transient state simulations performed via LTspice verify the deterministic performance of the system. These simulations demonstrate the system's ability to identify leading anomalies, secure critical response time, and guarantee absolute hardware isolation before critical thermal safety limits are exceeded.