Use Case Oil & Gaz
Anomaly Detection
Our renewable energy customer would like to detect anomalies in their insulators in their power generation facilities in order to reduce outages or unplanned outages, improve reliability and optimize the cost of operations.
Challenges
Detecting anomalies in power generation facilities requires advanced technology to capture, process, and analyze multispectral data efficiently. The goal is to enhance reliability, minimize outages, and reduce maintenance costs.
Multispectral Imaging
Generate an anomaly detection solution for critical assets based on multispectral image analysis: (real image, ultrasound and infrared) by capturing information from drones
Predictive Maintenance
Use artificial intelligence to enable proactive and targeted asset maintenance.
Data Processing
Handling large volumes of high-resolution image data and ensuring efficient processing for real-time or near-real-time anomaly detection.
Solution
The approach is based on a solution combining Computer Vision and predictive analytics.
Using drones to capture information about isolators (real image, with ultrasound and infrared).
Development of a multispectral image analysis model based on Deep Learning to analyze and classify information/images.
Prediction of multiple categories with high accuracy (e.g., good, broken, contaminated, chipped, flashed) to generate a maintenance order.
Results
Improved maintenance by increasing visibility into the condition of equipment and associated maintenance if necessary, and reducing unnecessary manual inspection activities.
Advanced image analysis capability to provide accurate information on anomalies and potential asset failures
-70%
operating and maintenance costs
