HydraGuard

Inspiration

Gas pipeline operations are at risk from hydrate formation, which can cause blockages and costly disruptions. We were inspired to give operators actionable insights to help reduce the damage done by these occurrences.

What it does

This project analyzes gas pipeline data to predict the likelihood of hydrate formation, which can cause blockages and disrupt operations. It uses a machine learning model to assess patterns such as changes in gas volume and valve behavior. This enables proactive measures to prevent costly issues.

How we built it

We processed raw pipeline data to clean and create relevant features like rate of change and valve effectiveness. Using Python and Scikit-learn, we implemented a Random Forest classifier to predict hydrate formation probabilities for each data entry.

Challenges we ran into

We originally ran into problems with interpreting the data and clearly understanding when the hydrate was going to occur which was necessary to train the model.

Accomplishments that we're proud of

We're proud of getting a working ml model that can run through code and let the user know when a hydrate is about to form.

What we learned

We learned how to know when to use what kind of learning model given a certain type of data, how to use Streamlit, and how to manipulate data using Pandas.