TransfomerFaultPrediction

Fault Prediction in Distribution Transformers using various Machine Learning algorithms.

Distribution Transformer Theory

A distribution transformer is also known as a typical kind of isolation transformer. The main function of this transformer is to alter the high voltage to the normal voltage like 240/120 V to use in electric power distribution. In the distribution system, there are different kinds of transformers available like single phase, 3-phase, underground, pad-mounted, pole-mounted transformer.

• Generally, these transformers are available in different sizes with efficiencies along with insulating oil.
• There are four types of distribution transformer connections available like star-star, delta-delta, star-delta, delta-star and Zig Zag/delta zigzag.

Transformers plays a very important role in the power system. Though they are some of the most reliable component of the electrical grid they are also prone to failure due to many factors both internal or external. There could be many initiators which cause a transformer failure, but those which can potentially lead to catastrophic failure are the following:

1. Mechanical
2. Dieletric Failure
3. Electrical Winding Short-circuit

Dataset Parameters Overview

• OTI- Oil Temperature Indicator
• WTI- Winding Temperature Indicator
• ATI- Ambient Temperature Indicator
• OLI- Oil Level Indicator
• OTIA- Oil Temperature Indicator Alarm OTIT
• Oil Temperature Indicator Trip
• MOG_A- Magnetic oil gauge indicator

Objectives

• Dataset exploration using various types of data visualization.
• Build various Machine Learning models that can predict the Magnetic oil gauge fault indicator.

Steps

1. Import the libraries.
2. Exploratory Data Analysis (EDA).
3. Merge and clean the datasets.
4. Data Visualization.
5. Train-test split and data normalization.
6. Implementing Machine Learning Models.
7. Evaluation and comparison of the models.