Business Challenge and Objectives

• High dust deposition on insulators/transformers leads to a rise in insulator temperature, causing grid failure
• The grid failure in-turn results in revenue losses for the organization
• The client currently uses state-of-the-art automation systems to capture the images of insulator/transformers at the sub stations to monitor the dust deposition on them
• The client wanted to fetch and manage video feed data from 0.3 million towers
• The client also required an analytics ecosystem to trigger preventive cleaning alerts on-time to avoid grid failure

The Solution

The TekLink Team identified the challenges in storage, access, and analysis of the data created. The TekLink provided a holistic solution considering the complexity of the factors involved in the grid functioning and other business needs.

• Pre-processing (image scaling, background noise removal and others) the real-time data and images, and offline data (manual, image feed, video streams).
• Migration of the pre-processed data to the Data Warehouse
• Using predictive models for data training
• Building deep learning algorithms like convolution learning (CNN), using TensorFLow to categorize the clean and dirty images of the insulators
• Implementing a smart predictive system to provide regular updates and critical warnings for preventive maintenance

Key Benefits

The client can now:

• Efficiently process and categorize the insulator image data feed
• Minimize the consequential losses due to power supply downtime
• Improve preventive maintenance with the clearing alerts generated from analytical models
• Leverage AI in data processing to reduce error probabilities eliminating human intervention