HDSim-RES: A Semantic-Based Binary Function Similarity Detection Framework for Renewable Energy Systems

Renewable energy systems rely on embedded devices with firmware as their operational core. Firmware vulnerabilities can lead to catastrophic consequences, including system downtime, data breaches, or even physical hazards. Function similarity detection provides a proactive approach to identify latent vulnerabilities by comparing firmware with known exploit patterns, ensuring safety in critical infrastructure. This paper proposes a semantic-based binary function similarity detection framework based on text-semantic embedding for renewable energy systems - HDSim-RES, a generalized cross-architecture firmware vulnerability detection framework. The proposed framework employs depth-first traversal for control flow graphs, combined with hardcoding values as feature values, to derive semantic information from binary functions. The experimental results indicate that the proposed framework achieves unprecedented performance in critical dimensions, detects 6 CVE vulnerabilities across real-world firmwares, and bridges the gap between information system security research and renewable energy engineering.