Deep Learning for Fault Detection in Transformers Using Vibration Data

Publications - Paper

Deep Learning for Fault Detection in Transformers Using Vibration Data

Description

Related tags

#Electricity Network #Measurement Transformers #Monitoring and Control

Publications - Paper

Deep Learning for Fault Detection in Transformers Using Vibration Data

The aim of this work is to evaluate the virtue of deep neural networks in detecting incipient transformer failures, particularly winding looseness, via vibration data analysis.

The transformer vibration technique is a non-invasive method of monitoring winding looseness. It is based on the analysis of vibration spectra measured by sensors placed on the transformer tank. This work is based on measurements that were carried out in a dedicated laboratory under two different conditions: with and without clamping pressure on the windings. The data analysis, oriented towards fault detection, is performed by feedforward neural networks, which have proven to be effective for reliable prediction according to experimental results. Particular emphasis is given to the robustness of the prediction for sensor loss, and various techniques are performed to evaluate and enforce generalization to out-of-sample data for the classifier obtained.

Download Memory

Projects

Components and materials for safety and resilience (Mat4Res)

Related tags

#Electricity Network #Measurement Transformers #Monitoring and Control

Search in the site by keyword

Publications - Paper

Deep Learning for Fault Detection in Transformers Using Vibration Data

Publications - Paper

Deep Learning for Fault Detection in Transformers Using Vibration Data

Comments