[Paper Review] Outage Detection in Power Distribution Networks.

Raffi Sevlian, Yue Zhao+3 more|arXiv (Cornell University)|Mar 27, 2015

Optimal Power Flow Distribution14 references3 citations

TL;DR

This paper proposes a tree-structured power distribution network outage detection framework that integrates real-time power flow measurements and load forecasts to optimize sensor placement. By minimizing missed detection probability, the method reduces required sensor density by 60% for a 10% drop in network-wide detection reliability, demonstrating significant efficiency gains in feeder-level monitoring.

ABSTRACT

An outage detection framework for power distribution networks is proposed. Given the tree structure of the distribution system, a framework is developed combining the use of real-time power flow measurements on edges of the tree with load forecasts at the nodes of the tree. Components of the network are modeled. A framework for the optimality of the detection problem is proposed relying on the maximum missed detection probability. An algorithm is proposed to solve the sensor placement problem for the general tree case. Finally, a set of case studies is considered using feeder data from the Pacific Northwest National Laboratories and Pacific Gas and Electric. We show that a 10 \% loss in mean detection reliability network wide reduces the required sensor density by $60 \%$ for a typical feeder if efficient use of measurements is performed.

Motivation & Objective

• To develop a reliable outage detection framework for radial distribution networks with a tree topology.
• To minimize the probability of missing outages through optimal sensor placement using real-time measurements and load forecasts.
• To reduce sensor deployment costs by optimizing sensor density while maintaining acceptable detection reliability.
• To validate the framework using real feeder data from Pacific Northwest National Laboratories and Pacific Gas and Electric.

Proposed method

• The framework models the distribution network as a tree, with power flow measurements collected on edges and load forecasts at nodes.
• It formulates the outage detection problem using a maximum missed detection probability criterion to ensure reliability.
• An optimization algorithm is developed to solve the sensor placement problem for general tree-structured networks.
• The method leverages real-time measurements and forecasted loads to improve detection accuracy and reduce false negatives.
• The approach is evaluated using actual feeder data from two major utilities to assess performance under realistic conditions.
• A trade-off between detection reliability and sensor density is quantified using a probabilistic detection model.

Experimental results

Research questions

• RQ1 How can outage detection reliability be maximized in radial distribution networks using minimal sensor deployment?
• RQ2 What is the optimal sensor placement strategy for tree-structured distribution systems to minimize missed detections?
• RQ3 To what extent can sensor density be reduced while maintaining acceptable outage detection performance?
• RQ4 How do real-time measurements and load forecasts jointly improve detection accuracy?
• RQ5 What is the trade-off between detection reliability and sensor count in practical distribution feeders?

Key findings

• A 10% reduction in mean network-wide detection reliability results in a 60% reduction in required sensor density for a typical distribution feeder.
• The proposed framework enables efficient use of real-time measurements and load forecasts to significantly reduce sensor deployment needs.
• The sensor placement algorithm is effective for general tree-structured distribution networks.
• Case studies using real data from Pacific Northwest National Laboratories and Pacific Gas and Electric confirm the method’s practical viability and performance gains.
• The framework achieves high detection reliability with substantially lower sensor density compared to conventional approaches.
• The optimization framework successfully balances detection reliability and sensor cost, offering a scalable solution for modern distribution systems.