Federated Database System for Multi-Organization Sharing

The electric power industry, particularly large utilities such as PLN, increasingly requires cross-regional
data analysis to support planning, monitoring, and decision-making. However, data sharing between
organizational units is constrained by privacy concerns, data ownership policies, and the risk of sensitive
operational information leakage. Centralized SCADA data analytics therefore become difficult to implement. This
paper proposes a federated SCADA data analytics approach based on a Federated Database System that enables
multi-organization collaboration without transferring raw data. Each organization retains its local database, while
only privacy-protected statistics and model parameters are shared. Secure Multi-Party Computation (SMC)
principles and Differential Privacy (DP) mechanisms are applied to ensure confidentiality and prevent data
reconstruction. In addition, a lightweight artificial intelligence model based on local linear regression is integrated
to demonstrate federated AI capability. Experiments using substation capacity data from two regions, West Java
and East Java, show that the proposed system can successfully generate global statistics such as total capacity,
average capacity, maximum capacity, and number of substations with minimal accuracy degradation. The results
confirm that federated analytics is a secure and practical solution for SCADA-based data collaboration in the
energy sector.
Keywords: Federated Database, SCADA System, Secure Multi-Party Computation, Differential Privacy.