MILP-Based Fault Diagnosis Model in Active Power Distribution Networks

With a rapid growth in the deployment of renewable and distributed energy resources (DERs) in active power distribution networks, especially the inverter-interfaced distributed generators (IIDGs), existing fault section estimation methods are facing additional challenges. This paper proposes an optimization model for fault diagnoses in active power distribution systems, which aims at simultaneous identifications of suspected fault line sections and potential false alarms sent by remote fault indicators (RFIs). First, mathematical expressions of expected and false alarms of oriented RFIs (ORFIs) and non-oriented RFIs (NRFIs) are formulated to construct a basic fault diagnosis model. Minimizing the number of false alarms is chosen as the objective function to find the most credible one from all possible fault hypotheses which represent different combinations of elements in the outage area. Next, considering IIDGs, a modified fault diagnosis model is established by eliminating misleading components from the objective function which are produced by incompetent RFIs and have failed to detect weak currents. The monitoring data provided by distribution level phasor measurement units (D-PMUs) for weak IIDG fault currents are leveraged to construct the equivalent RFI current, which is compared with the trigger threshold to find incompetent sensors. The proposed problem is formulated as a mixed-integer linear programming (MILP) model, which is solved by off-the-shelf commercial solvers. The effectiveness of the proposed fault diagnosis model is validated by test cases representing several DERs and false alarms applied to a utility feeder.

Automated summary

This paper proposes an optimization model for fault diagnoses in active power distribution systems, aiming to simultaneously identify suspected fault line sections and potential false alarms sent by remote fault indicators. By analyzing expected and false alarms, minimizing the number of false alarms is chosen as the objective function to find the most credible fault hypothesis. A modified fault diagnosis model is established, leveraging monitoring data to construct equivalent RFI current and identify incompetent sensors.