Search

• Identification of vulnerable NTL system areas based on concepts of measurement's innovation and n-tuple of critical measurements. • Synthetic measurement model for GE analysis enhancement. • Framework for NTL estimation in unbalanced distribution networks. This paper presents a hybrid data-driven physics model-based framework for distribution networks nontechnical power loss estimation. Nontechnical power loss is defined as energy delivered to the consumers but not billed by the utility. These losses, unlike technical losses, are not inherent to the process of transportation of electricity. State-of-the-art solutions for nontechnical power loss estimation are either data-driven or physics model-based. However, due to the evolving nature of nontechnical power losses, data-driven solutions by themselves are not sufficient. Physics model-based analytical solutions, otherwise, which consider a quasi-static system model, rely solely on physics phenomena observation, however it is virtually impossible to model all grid dynamics. In this case, the nexus of data-driven physics model-based analytic models enable the solution of the problem. The hybrid framework is composed of three interdependent processes. First, an unbalanced load flow analysis is performed to obtain an initial estimate of the operating system state. Second, a data-driven method for consumer classification is applied. Third, synthetic measurements are created considering the measurement's innovation and n -tuple of critical measurements aiming to improve gross error analysis. Solution validation is made considering the IEEE 4-bus, 13-bus and 123-bus unbalance test feeders. Comparative test results highlight decreased nontechnical power loss estimation errors. Simplicity of implementation, with easy-to-obtain parameters, built on the classical weighted least squares state estimator, indicate potential aspects for real-life applications. [ABSTRACT FROM AUTHOR]

• Comprehensive physics-based analytical model for HIF location. • Statistical behavior of estimates is analyzed to identify the faulted section. • One-ended data based frequency domain measurement model. This paper presents a physics-based analytical model for high impedance fault location in power distribution systems. The presented analytical solution is composed by two interdependent processes. First, fault distance is estimated through a Weighted Least Square approach applied to solve an overdetermined and nonlinear algebraic system of equations. Second, an analytical method based on the statistical behavior of the estimates obtained from the frequency domain system model solution is presented to address the multiple estimates problem, identifying the faulted section. The presented fault location physics-based analytical model considers the capacitive effect of distribution lines. Furthermore, voltages and currents measured from only one terminal are used. The formulation is evaluated using the IEEE 13-bus and a modified version of the IEEE 34-bus test feeders. Comparative analysis with state-of-the-art methods is presented. Average errors of 1.95% are obtained. Easy to implement analytical model, without hard to derive parameters, built on the weighted least squares state estimator, highlight potential aspects for real-life applications. [ABSTRACT FROM AUTHOR]