Hydrate Alert

config.toml
main
main
data_page
data_page
data_page
data_page

Hydrate Alert

Inspiration

Our inspiration for the Hydrate Alert project came from the need to detect potential hydrate formations in gas injection systems. Hydrate formation can lead to severe operational issues, including pipeline blockages and equipment failure. By creating a system that can monitor and alert users of significant changes in gas volume, we aimed to provide a proactive solution to this problem.

What it does

The Hydrate Alert web application is designed to analyze and monitor data from gas injection systems. The application reads real-time CSV files containing gas injection data, including metrics like the instantaneous gas meter volume and valve percentage. It identifies significant drops in the gas injection volume, which may indicate potential hydrate formation. If such a decline is detected, the app triggers an alert to inform operators. It also includes visualizations to help users better understand trends and potential issues.

How we built it

Streamlit for the frontend, allowing for easy file uploads, real-time visualizations, and simple navigation between pages.
Python powering the backend, utilizing libraries like Pandas for data handling and manipulation, and Matplotlib for plotting graphs. The core logic involves analyzing gas injection metrics, detecting significant volume drops, and issuing alerts when the system detects potential hydrate risks. We also implemented data cleaning and interpolation to handle missing values.

Challenges we ran into

Handling Large CSV Files: Managing large datasets and ensuring smooth real-time plotting of data while avoiding performance issues was difficult.
Real-time Data Updates: Simulating live data and continuously updating the graph proved tricky due to the need to refresh the plot dynamically while maintaining app performance.
Data Cleaning: Dealing with incomplete data (such as missing values) required careful preprocessing and interpolation techniques to ensure the accuracy of our alerts.

Accomplishments that we're proud of

Real-time Data Processing: The ability to process and visualize data line-by-line and update the graph live is a key feature we’re happy with.
User Alerts: Successfully implementing an alert system that notifies users when a potential hydrate risk is detected.
Streamlined User Interface: The simple and intuitive interface built with Streamlit allows users to easily upload data, view trends, and get real-time alerts.

What we learned

How to handle real-time data and update visualizations efficiently using libraries like Matplotlib and Streamlit.
How to preprocess data for accurate analysis, especially when dealing with missing or inconsistent data.
We learned how to design a system that automatically detects and alerts users about potential system issues based on thresholds.

What's next for Hydrate Alert

In the future, we plan to:

Enhance the Alert System: Introduce more sophisticated alerting mechanisms, such as email or SMS notifications.
Optimize Performance: Work on optimizing the app to handle larger datasets more efficiently, possibly integrating a database for more scalable storage and retrieval.
Add More Data Sources: Support additional data inputs from other sensors or systems to expand the app’s monitoring capabilities.
Integrate Machine Learning: Investigate machine learning models to predict hydrate risks based on historical data trends rather than just threshold-based detection.