Challenges in Energies and Utilities
Transmission towers and substations form core infrastructure elements that ensure efficient supply of power across the country. Power lines span several thousands of miles, delivering energy to several substations before reaching to their consumers. A typical transmission tower has a variety of components like conducting wires, insulators, bird guards, marker balls, ground bonds, dampers, cotter pins and many more. All these components are subject to natural wear-tear and are constantly pounded by weather (rain, cold, heat) resulting in deterioration over a prolonged periods of time.
Power generating companies periodically fly past these towers in helicopters to manually inspect assets for health and maintenance. Surveying with helicopters over several thousand miles of power lines is expensive, slow and risky to their workers.
Adopting AI to transform monitoring and maintaining assets
Unmanned Arial vehicles (Drones) are slowly being adopted by industries to reduce risk to their workers, survey assets more frequently and literally avoid risk to their workers. In addition to adopting efficient ways to capture images/videos, there is a need to engage technologies around computer vision that analyze data to identify defective assets. Defective assets could be corroded conductors, broken insulators, misplaced marker balls, missing cotter pins or even birds settling in with nests. Accelerating the rate of capturing data, identifying defective assets and scheduling maintenance will minimize interruption of power to their consumers. But the challenge now amplifies to training and integrating advanced computer vision models into their evolving transformation. Shortage of data scientists further delays the anticipated pace of adopting technologies.
Accelerate AI solutions with IBM PowerAI Vision
We introduce IBM PowerAI Vision, a tool to train models for accessing health of assets. A tool build for subject matter experts who are knowledge-able about assets can now bring their data to train and deploy models with no coding. The data could be images and videos.
Features like manual labeling, auto labeling from videos helps minimize time on curating the datasets for training. It is also difficult to find training samples on assets which might have naturally weathered over a period of time. IBM PowerAI Vision provides abilities to augment data from smaller set of samples and minimizes need to schedule piloting of the unmanned aerial vehicles (drones).
IBM PowerAI Vision trains models to classify images into pre-trained classes. In addition to classification, models can be trained to identify localized portion of defective components (aka Object detection). These insights are valuable for organizations to schedule maintenance on such assets. Following is a brief video of IBM PowerAI Vision at play on training and classifying images into various classes.
IBM PowerAI Vision is addressing requirements from Energy and Utilities with tools that accelerate transformation of their businesses with artificial intelligence. We welcome you to take IBM PowerAI Vision for a spin on our cloud. Please feel free to review our videos on user experience for training models for classification and object detection. You can find more indepth details of our product on the IBM marketplace. For developers, we got code samples to build the next revolutionizing solution. Early adopters can download the SW free for 30 days. Looking forward to hearing your valuable feedback.