Sensitivity analysis of impedance measurement algorithms used in distance protection

Abstract

This paper presents the methodology for Global Sensitivity Analysis (GSA) of impedance measurement algorithms that are used in distance protection of transmission lines. The analysis is using Quasi-Monte Carlo technique in calculating variance of the measurement error as well as parts of this variance that are contributed by uncertainty of some parameters. Specifically, fault resistance is the parameter that cannot be measured and it is treated as uncertain. Such parameters are called factors and they are represented with specific distribution function within specified domain of variation. The proposed methodology is implemented using automation of DIgSILENT and SIMLAB programs. The program in DPL scripting language of DIgSILENT is developed to automate simulation of transmission line faults when uncertain factors are varied. Each variation represents a sample in factor space. They are generated using Sobol quasirandom sequence in SIMLAB. This program also calculates variance-based sensitivity indices based on the factor space samples and corresponding simulation results obtained in DIgSILENT. To illustrate this methodology, the paper presents results of the GSA of SEL-412 impedance measurement algorithm implemented as a model in DIgSILENT.