taylorfyq/opendss-faultlocation

Rodrigo Rianelly's graduation project about "Distributed photovoltaic generation impact on fault location in radial distribution networks".

Hello everyone, my name is Rodrigo Rianelly and this is the code of my graduation project!

Here's the abstract:

This work aims to establish the impact of the distributed photovoltaic generation on fault location in radial distribution networks. A case study is proposed using EPRI ckt24 to do so. Photovoltaic systems are dimensioned for this circuit with respect to the penetration limit. Fault simulations are performed on OpenDSS in order to acquire voltage and current values at the substation. These results are recorded in a database and utilized to train a neural network of the multilayer perceptron class, which is capable of determining with 90% average success rate the fault location between 6 possible geographic regions in the circuit in presence or not of distributed photovoltaic generation.

And here are the keywords: Radial distribution networks. Distributed photovoltaic generation. Photovoltaic systems. Penetration limit. OpenDSS. EPRI ckt24. Fault location. Supervised learning. Neural network. Multilayer perceptron.

You can look into Fault_location_ckt24.pptx to see more details. The full document will be published within a few weeks, in portuguese language, by University of Brasília.

In order to run the codes, you should have Python 3, OpenDSS, a .db reader and some python packages such as win32com, numpy, matplotlib and sqlite3.

Since it takes a long time to run the fault simulations that build up the database, I have uploaded the files to Google Drive: https://drive.google.com/open?id=1jzBDv_J_BDSlXoOs8bTm0COY-eiDCQmk