Niranjan, S and Shrivas , A (2014) To model breakdown voltage using artificial neural networks of solid insulating materials. BTech thesis.
During manufacture, insulating materials may have voids which are source to electrical trees. Due to partial discharge, the insulating material degrades and breakdown occurs. The factors contributing to the breakdown are difficult to determine. As the equation describing the function is unknown, function estimation, which has some of its own useful properties, a major field of Artificial neural networks, is used. In this project using Artificial Neural Network, we develop models which intakes four different possible inputs that effect the breakdown which are the insulating sample thickness (t), void thickness (t1), void diameter(d) and the materials¡¦ permittivity (ƒÕr) predicts the breakdown voltage as a function of these four inputs. The Neural Network needs to be trained to be able to predict the Breakdown Voltage as close as possible. For the purpose of training , experimental data using a cylinder plane electrode system is used. The different dimensions used will be used to create the voids artificially. The parameters are selected after detail studying of the models as to which would generate best results. After the training is completed, the breakdown voltage as a function of the four input parameters is predicted. The results are very convincing as the error with which it is predicted is very less. Hence, this again proves the capability and effectiveness of using simulation models. MATLAB 2010 is used for doing the simulation process.
|Item Type:||Thesis (BTech)|
|Uncontrolled Keywords:||Partial Discharge , Artificial neural network, Breakdown voltage, permittivity, void diameter,void thickness.|
|Subjects:||Engineering and Technology > Electrical Engineering > Image Processing|
|Divisions:||Engineering and Technology > Department of Electrical Engineering|
|Deposited By:||Hemanta Biswal|
|Deposited On:||11 Aug 2014 09:23|
|Last Modified:||11 Aug 2014 09:23|
Repository Staff Only: item control page