An anonymization framework for IEC 61850 substation communications: Field-level and topology-aware privacy

Substation datasets, like those using the IEC61850 standard, hold sensitive information about power flows, equipment statuses, and network configurations. This data could expose vulnerabilities to knowledge-based cyberattacks, making utility providers hesitant to share it publicly for research. While encryption enhances security, it often diminishes the dataset’s utility for research purposes. To address the trade-off between security and utility, we introduce an anonymization technique specifically for the IEC61850 standard, demonstrated on the GOOSE protocol. Our method involves two main approaches: anonymizing sensitive and quasi-identifying fields within packets to preserve data utility, and injecting dummy packets using one of our proposed algorithms to effectively obscure network topology. Using the first method, we publish an anonymized dataset derived from substation communications captured in our testbed to support ongoing research. We evaluated the framework’s effectiveness through a comprehensive communication pattern analysis, including time, flow, statistical, and entropy analyses, and field anonymization testing. Our study highlights the critical importance of maintaining privacy in substation data sharing while ensuring data remains useful for research, setting the foundation for extending this framework across multiple substation protocols in future studies.