Share Your Requirements
Wellsite
LANCERProject Description
A leading energy technology provider, Wellsite focused on optimizing drilling operations, approached Oodles to develop a predictive analysis proof of concept. The client aimed to reduce downtime by identifying drilling motor stalls and forecasting potential failures. A deep learning-based system was designed to analyze directional drilling data, detect anomalies, and deliver actionable insights in real time.
Scope Of Work
The client sought Oodles to tackle the challenge of unplanned motor stalls that disrupted drilling operations and increased costs. They required a solution to preprocess drilling data, build a stall detection model, and expand it into predictive analytics. Areas of work included data cleaning, feature engineering, model training, evaluation, backend services, and a simple frontend dashboard for visualization.
Our Solution
To address the client’s requirements, a predictive analytics framework was developed combining advanced machine learning techniques, scalable architecture, and intuitive interfaces.
Key Implementations:
- Data Preparation: Preprocessed raw EDR files, structured datasets, and labeled stall events for accurate training.
- Exploratory Data Analysis: Conducted detailed data exploration to uncover patterns in pressure, torque, and event timelines.
- Model Development: Used sci-kit learn and scipy to train machine learning models capable of detecting and predicting stalls.
- Backend Services: Built APIs using Flask for data handling, model serving, and stall predictions.
- Frontend Visualization: Created a lightweight UI with Flask templates to display predictive results and stall event graphs.