Transform industrial inspections with AI-powered Autonomous Inspection

Industrial facilities rely on routine inspections to ensure optimal asset performance, employee safety, and compliance. Industrial inspections are often manually intensive and data heavy. Completing all of the inspection steps involved – from initial data collection, documentation, analysis, to reporting findings –consumes significant amounts of valuable time and resources. On top of this, scheduling and initiating necessary actions is cumbersome. Large inspections usually take days or even weeks to complete depending on the scope, available resources, and other factors. Besides their resource-intensive nature, such inspections can involve considerable safety risks associated with extreme heights, cramped spaces, and high voltage equipment. Moreover, manual inspections are prone to human error due to manual data entry and variations in interpretations of inspectors’ with different experience levels.

As the energy landscape rapidly transforms, relying solely on traditional inspection methods is fast becoming unsustainable. Here's why:

• Shrinking workforce: Attrition, ageing populations, cost pressures – whatever the reason, the pool of available inspectors and experts is steadily drying up.
• Rise in renewables: Sprawling solar and wind farms pose a logistical nightmare for manual inspections. Inspectors have to travel greater distances between assets and sites and understand newer equipment issues – multiplying the effort, time, and complexity.
• Dynamic energy requirements: Power generators are under increasing pressure to adapt and respond faster to fluctuations in energy supply and demand. Addition of renewables to the generation fleet has aggravated the challenge due to their intermittent nature. Slow and resource-intensive manual inspections simply can't keep up with the agility and performance needs of this scenario.

Sensor technologies can prove to be inadequate.

A common alternative to manual inspections involves using sensor-enabled data, often considered the “gold standard” of asset performance monitoring. However, sensor-based inspections fall short in addressing several inspection needs, especially those associated with visual inspections. For instance, sensors struggle to catch surface defects such as corrosion or cracks. They fail to offer the visual context needed for contextual awareness and seeing the bigger picture. Also, installing sensors could be expensive and might require shutting down the equipment.

The traditional manual inspection process is resource intensive, prone to safety risks and human error, and overall difficult to sustain. It presents a bottleneck to organizations in realizing the benefits from their other digital transformation initiatives, such as optimizing maintenance with Asset Performance Management software. There’s an urgent need to transform the traditional inspection processes.

To address these stated challenges, GE Vernova has introduced Autonomous Inspection – an industrial computer vision software that uses cameras and Artificial Intelligence/Machine Learning (AI/ML) algorithms to enable smarter inspections and monitoring. Depending on the use case, the application can partially or fully automate the inspection processes helping deliver faster, safer, and cost-effective inspections versus manual inspections. Its integration with select GE Vernova’s Asset Performance Management (APM)  applications ensures the visual data and insights are available to these applications to further improve asset and operational performance.

How does Autonomous Inspection work? 

Autonomous Inspection uses AI-powered image analysis and cameras to inspect and monitor assets. Images captured by fixed or mobile cameras (RGB/thermal) are sent to the application running in the cloud. Purpose-built AI/ML models analyze the images to extract or estimate desired inspection data, such as pressure gauge readings, temperature, or corrosion levels. The application also converts visual data into time series data and makes it available to integrated APM applications. Whenever user-defined thresholds are crossed, automated alerts or recommendations are generated and sent to users. Additionally, if a person’s approval or feedback is required, AI findings are shared with them. This human-in-the-loop approach helps continuously improve the AI/ML models’ accuracy further. 

Autonomous Inspection features several powerful capabilities to automate inspection workflows including image data collection, processing, integration, analysis, alerting, and management. With its human-in-the-loop approach, it combines the power of human expertise alongside AI to reduce inspection bottlenecks. Engineered to be industry-agnostic, Autonomous Inspection assists users across various sectors including power generation, oil and gas, manufacturing, chemicals, and mining.

Now let’s discuss a few examples of how this software is transforming industrial inspections:

1. Faster and cost-effective detection and response – go from days/weeks to hours/mins

Autonomous Inspection helps significantly reducese the time and effort involved with manual-intensive inspection workflows - from data collection and analysis to initiating maintenance actions. Parts of the manual intensive inspection workflows that usually take days or weeks can now be automated, completing the process within hours or even minutes. Capabilities such as automated defect/anomaly identification and alerting enable faster and timely interventions to help prevent any asset downtime and related losses. This leads to notable productivity gains and costs savings.

Autonomous Inspection employs cameras providing visual context around defects which are typically more economical and non-disruptive (requiring no asset shutdowns) than sensors.

A major Oil & Gas firm that used Autonomous Inspection during a Proof of Concept (POC) engagement reduced inspection review time for corrosion images from two weeks to 30 minutes.

2. Fewer hazardous inspections - Improved worker productivity and safety

Experience several tangible benefits to workforce productivity and safety by automating inspections. An impactful benefit of Autonomous Inspection is the ability to help reduce or reallocate the inspection workforce, allowing them to focus on more value-adding activities. By leveraging advanced cameras and AI/ML models, Autonomous Inspection can mitigate the safety risks associated with performing inspections at elevated heights, in difficult-to-access areas, and around electrically charged assets. Further, it helps reduce the need for inspection-related travel, as gathered data and insights can be accessed by users remotely via the cloud.

A major European utility firm is using Autonomous Inspection to inspect critical electrical assets including transformers and switchyards. The application is helping them transition to a remote-enabled, safer, and more efficient asset inspection model.

3. Envision your APM - Achieve superior business outcomes with visual insights

Autonomous Inspection offers a distinct advantage over sensor-only inspections by enhancing situational awareness with visual data and insights. Furthermore, organizations can integrate visual and processed (timeseries) data from Autonomous Inspection with their APM applications, enriching them with valuable visual insights. This integration unlocks a range of potential benefits, as highlighted in the following use cases:

• Inspectors can use mobile devices to take pictures of corroded equipment and send them to Autonomous Inspection for analysis. They can then access automated alerts/recommendation within the APM Integrity application and initiate the needed action.
• Timeseries temperature data of monitored electrical equipment, such as transformer, can be fed into APM Reliability Plus/SmartSignal application to help better predict transformer overheating.
• Inspection images can be consolidated at one location within APM to filter and be used for future reference and analysis. Thus, it eliminates the need to access them on separate systems and/or multiple sign-ins.

4. Reduce data entry errors and perception variations

Traditional and manual-intensive inspections are pierced with data entry errors and differences in perceptions. Automated data extraction from images using Autonomous Inspection helps reduce data entry errors. Through AI-driven analyses, inspection data interpretations are standardized across workers of varying experience levels, improving consistency in inspection findings.

Furthermore, Autonomous Inspection helps scale inspection process across one or multiple sites. By simply integrating additional image capturing devices/cameras, you can easily expand your inspection capabilities – say goodbye to the scalability hurdles of manual inspections.

Industrial firms realized that transformational technologies, such as AI’s computer vision, can help them gain a competitive advantage and sustain growth. Per a survey by research firm Verdantix, 38% of respondents across twenty industries mentioned that they already have or are currently implementing computer vision technology. Another 16% of them had implementation plans beginning in the next 12 months. Autonomous Inspection is among the transformational digital technologies helping organizations automate and accelerate their asset performance and operational goals. It delivers an impressive set of smart inspection capabilities. Driven by innovation and customer-centricity, we at GE Vernova are continuously striving to enhance and expand Autonomous Inspection’s capabilities and applicable use cases. In the future, Autonomous Inspection is likely to support unmanned systems like drones for image capture and expand deeper integration within APM applications.

Want to know more about how you could leverage GE Vernova’s Autonomous Inspection for your organization? Get in touch.