Securing Fault Diagnosis in IoT-Enabled Industrial Systems Using Homomorphic Encryption

Securing fault diagnosis in IoT-enabled industrial systems is critical to preventing data breaches and ensuring reliable operation. This study proposes a novel approach that integrates homomorphic encryption (HE) into the fault diagnosis pipeline, enabling secure processing of sensitive sensor data. Using the CWRU Bearing Dataset, we demonstrate that the HE-based system achieves high accuracy (97.92%) while preserving data confidentiality. Despite the computational overhead introducing latency (1100 ms), the trade-off remains acceptable for non-real-time applications. The system’s ability to protect privacy without significantly compromising performance makes it well-suited for sensitive industries such as healthcare and manufacturing. Future work includes optimizing HE algorithms and exploring hybrid encryption schemes to enhance scalability and efficiency. This research highlights HE as a promising solution for secure and effective fault diagnosis in IoT systems.