Efficient data analysis for accurate fault localization in medium-voltage grids


The project will involve researching and developing methods for localizing ground faults. The goal is to speed up the search for actual fault positions and potentially prevent cascading faults that can occur with conventional localization methods, thus leading to more stable and safer distributed grid operation.

Project description

Fault localization in energy grids can be efficiently and accurately carried out by analyzing high-resolution digital data. GridData developed a method that can localize so-called ground faults in medium-voltage grids through a comparison with fault signatures. Creation of the required signatures and the comparison at runtime is time-consuming however. At the moment, the former also requires expert knowledge and manual intervention.

The goal of the EDaF project is to develop automated and scalable algorithms validated on real grids, as well as an integrated architecture to reduce data volumes and the effort required for the signature-based fault localization.

Research contribution

The EDaF project is structured around three engineering work packages:

  1. Real grid scenarios and solutions concepts
  2. Automated, scalable methods and IT infrastructures for fault localization
  3. Integration, evaluation and field testing.

The project will involve examining software architectures that satisfy the requirements. The corresponding prototype architecture concept will be evaluated with the help of distributed grid operators in real operating environments. fortiss is primarily responsible for using sensitivity analyses to determine the position of ground faults on the basis of real and simulated signature recordings of electric current. For these analyses fortiss will conduct research into approaches from the field of artificial intelligence and machine learning.


Project duration

01.05.2019 - 31.03.2022

Dr. Markus Duchon

Your contact

Dr. Markus Duchon

+49 89 3603522 30

More information

Project partner