Hitachi Image-Based Inspections

โดย Hitachi Vantara

Hitachi’s image-based inspection is an AI-powered image-analytics solution.

Hitachi’s image-based inspection is an AI-powered image-analytics solution. It identifies an asset from its image and can detect defects in the asset using prebuilt Machine Learning models. It is currently being used by Energy Utilities (specifically, their transmission and distribution lines inspections and maintenance teams). It has wide capabilities in other asset-heavy industries as well, such as manufacturing, buildings and constructions, and transportation infrastructure.

Hitachi’s image-based inspection solutions replace dangerous, expensive, and time-consuming maintenance processes for utility companies by combining the industry’s most sophisticated artificial intelligence (AI) and machine learning (ML) capabilities in one fully automated and integrated platform.

  • Climate change and aging power distribution infrastructure are leading to increased asset failures that are causing increasingly deadly fires. In California, for instance, out of the twelve deadliest forest fires, eight have happened in the last five years.
  • Asset managers in the energy and utility industry bear the responsibility for keeping thousands of miles of these transmission lines, along with other equipment, such as insulators and dampers, safe and operational.
  • Conventional inspection and Operations and Maintenance (O&M) methods are fraught with risks. Technicians today use techniques such as hanging from helicopters or climbing the tall electricity transmission towers. While some companies use drones to collect photographs or videos to lower this risk, the images are still analyzed manually. This is a time-consuming process fraught with time-consuming inefficiencies and human subjectivity.

Hitachi Image-Based Inspections replace these dangerous, expensive, and time-consuming approaches by combining sophisticated artificial intelligence (AI) and machine learning (ML) capabilities in one integrated platform. The fully automated platform can analyze up to 5000 images per second, creating the scale needed to keep up with timely O&M goals. It also links automatically to commonly existing Field Service Management (FSM) and Enterprise Asset Management systems, creating a unified approach without needing multiple toolkits or data silos.

Hitachi Image-Based Inspections provide scalable image inferencing and processing to help asset managers classify assets, detect defects, and categorize defect severity. The solution’s human-in-the-loop capabilities enable continuous retraining of the model-based subject matter expert (SME) input for a wide range of asset classes and image quality. Image-Based Inspections provides real-time indication of moderate and severe problems, allowing asset managers to implement a priority-based maintenance strategy to improve reliability. This approach makes planned repairs easier across the entire process, providing inherent operations and maintenance savings (proactive maintenance versus reactive maintenance).