Efficient testing of fault direction algorithms based on design of experiments methods

Distribution networks are structurally very diverse and continuously evolving due to the transition towards decentralized energy systems. As earth faults are an inevitable aspect of power systems, a reliable detection of earth faults is a key factor to ensure security of supply within power systems. Given the distinctive structures of distribution networks and the influencing fault parameters, each fault can be considered as unique. This diversity necessitates extensive testing of detection devices and algorithms, which is crucial for developing new algorithms capable of addressing a wide range of fault scenarios. To address this issue, this paper proposes a simulation-based test environment for fault direction algorithms that incorporates a wide range of input parameters, enabling the generation of diverse fault scenarios. To manage the large number of scenarios, Design of Experiments (DoE) methods are applied to optimize test case selection. Using the statistical software JMP, test scenarios are created and integrated into MATLAB Simulink for automated simulation. The results demonstrate that DoE methods significantly reduce the number of test cases while maintaining diversity. Among the evaluated methods, the D-optimal approach proves to be most effective, providing a foundation for developing fault direction algorithms and enhancing network reliability.