By Mu Li, Solutions Architect – AWS Energy
By Subrahmanyam Madduru, Sr. Global Partner Solutions Architect – AWS
By Sacin Porwal, Practice Head, AI Automation Solutions – Wipro
By Swapnil Belhe, Chief AI Solutions Architect – Wipro

Wipro-AWS-Partners

Asset integrity management is a key activity for energy companies. It ensures safe operations, avoids unplanned downtime, and prevents incidents.

This extends across almost the entire energy value chain. Be it onshore or offshore production, pipelines, refineries, or wind turbines in renewables, asset integrity management is important to all of these businesses and operations. However, there are common challenges in today’s inspection processes.

Take an offshore rig for example; it may involve gathering data via multiple offshore visits throughout the year. Once the inspection companies make those trips, data gathered can be in a variety of formats, such as images, videos, drawings, or even notes.

Inspectors look at images and watch videos. They then use a process to identify, measure, and categorize damages like corrosion, broken objects, and leaks. Depending on severities, this process can be manual and time consuming.

Yet, despite best efforts and intentions, there still could be missed opportunities for preventative maintenance due to human eyes’ perception of image quality, lighting conditions, depth, or other factors.

With recent advances in the field of machine learning (ML), specifically computer vision, there are digital technologies that can enhance customers’ existing workflows and help plan preventative work.

Challenges in Building Digital Inspection Solutions

Energy customers have traditionally used surveillance and remote control tools to handle inspections. Many industry-leading energy companies have already attempted digital inspection of their assets.

The acceleration of this process towards remote and virtual operations marks a critical step forward. However, there are challenges such as constantly changing technology, fear of vendor lock-in, non-standardized methods for data gathering, and limited inspection coverage. As a result, gaps may exist for customers while scaling solutions across their assets.

Machine learning can lead to wider adoption of digital virtual inspection. Customers are looking for flexibility in development and integration of their own ML models into the inspection applications. Specifically for offshore oil and gas operations, there is an opportunity to transform inspection procedures to ensure asset integrity.

In this post, we discuss how InspectAI can help customers deploy a cloud-based solution and transform their inspection process. InspectAI is Wipro’s digital visual inspection and integrity management solution built on Amazon Web Services (AWS). It helps customers achieve business goals by reducing non-productive time (NPT) and operational costs.

Wipro is an AWS Premier Consulting Partner with six AWS Competencies, including Industrial Software Consulting and Data and Analytics Consulting. Wipro is also a member of the AWS Managed Service Provider (MSP) and Well-Architected Partner Programs.

Wipro’s InspectAI Solution Benefits Energy Companies

Wipro envisioned InspectAI to address customers’ specific requirements of asset exterior virtual inspection. By leveraging AWS Cloud services for artificial intelligence (AI), machine learning, and serverless technologies, InspectAI brings advanced digital technology to the inspection process.

InspectAI has modules like assent monitoring, predictive analytics, and vegetation monitoring for energy industries like oil and gas, wind turbines, and utilities.

Wipro-Inspection-Management-1.1

Figure 1 – Wipro InspectAI solutions for energy industries.

InspectAI helps energy customers with the following:

  • Increase efficiencies: Standardization of the inspection process across all asset types will reduce costs and increase compliance as best practices are shared between assets. Increased reliability and availability improves efficiencies across the production and processing value chains.
  • Lower operating costs: Inspection cost is reduced and inspections are only carried out as needed.
  • Reduce risk and improve Health Safety & Environment (HSE): Customers have an improved view of platform and asset conditions. This leads to effective response and reduces HSE incident risk associated with corrosions.
  • Reduce emissions: Reduced carbon footprint from remote and autonomous inspections.
  • Achieve Net Zero ambition: InspectAI applies both clean energy and hydrocarbon facilities.

InspectAI Solution Overview

InspectAI uses a modular approach for customers to manage their end-to-end inspection process. The solution consists of three major components: Data Ingestion, Inspection Management, and Data Analysis.

Using an example of offshore asset inspection, the Data Ingestion module streamlines drone-based data capture. The Data Analysis module automatically detects corrosion anomalies and accelerates inspection. The Inspection Management module schedules flight paths and enriches images for analysis.

Wipro-Inspection-Management-2.1

Figure 2 – InspectAI solution overview.

Wipro built InspectAI using AWS cloud-native AI/ML services like Amazon SageMaker Ground Truth, and serverless technologies like AWS Lambda, Amazon Simple Queue Service (SQS), and AWS Step Functions. It also utilizes Amazon Elastic Compute Cloud (Amazon EC2) instances, including both CPU and GPU computes.

We’ll describe each of those components below and how various AWS services are leveraged.

Data Ingestion

Depending on the type of asset and inspection requirements, various forms of sensors are deployed like cameras, LiDAR, and thermal infrared (TIR). It’s also possible some of these sensors will be used on a transport vehicle like a drone or crawler.

Data format from these sensors is disparate and challenging to process due to a lack of data standardization. InspectAI’s data ingestion module applies best practices and handles diverse data sources. This module has a flight planning interface to help drone operators plan the image-capture path and replay previously recorded inspection paths.

Finally, there’s a component called Data Scan, which is an application deployed to the edge and removes the distortion and duplications during image capture process in the field.

Inspection Management

The Inspection Management module has an image enrichment layer, as ingested data can still be noisy. There’s also no linkage between the captured data and metadata such as actual asset and naming conventions.

The Inspection Management module removes noisy images and matches the incoming images with the corresponding engineering drawings. This module also helps schedule inspections.

Data Analysis and Reporting

The Data Analysis module processes images and automatically identifies anomalies like corrosion, substrate conditions, and broken objects. It leverages data from the Inspection Management module, clean images, and engineering drawings stored in an Amazon Simple Storage Service (Amazon S3) bucket.

This module must also identify the component or location of the anomaly and grade the severity. The solution creates a base ML model and uses Amazon SageMaker Ground Truth for labeling anomalies such as corrosion using semantic segmentation. The base model can be augmented further with specific customer images to improve the accuracy.

Amazon SQS is used to decouple and scale the individual services for the incoming images, while AWS Step Functions manage routing of requests to multiple services.

Various image analysis services are invoked in a predefined sequence for each incoming request. The image in each request is processed for image enrichment, component identification, and corrosion detection and grading. The component identification model uses AWS compute services like Amazon EC2 instances with P3 GPU instances for inferencing.

Wipro-Inspection-Management-3.1

Figure 3 – Wipro Data Analysis module using AWS services.

The corrosion detection and grading is based on Amazon SageMaker GroudTruth models and services, and customers can use this service to create a workforce to label data.

The teams in the workforce, whether internal to the customer, third-party vendor, or general public, will use the Amazon SageMaker Ground Truth labeling jobs to semantically label the corrosion levels.

The ML model is trained on labelled data in Amazon SageMaker, and the model inference endpoint is used by the orchestrator to process the incoming request for corrosion detection and grading.

Amazon SageMaker Ground Truth Semantic Segmentation

The semantic segmentation algorithm classifies every pixel in an image, which helps in identifying degradation. Wipro has tested multiple Amazon SageMaker Ground Truth semantic segmentation algorithms like the Fully Convolutional Network (FCN) algorithm, Pyramid Scene Parsing (PSP) algorithm, and DeepLabV3. The solution uses the DeepLabV3 as the backbone for encoder.

InspectAI uses multi-level segmentation using DeepLabV3 Wipro ML models. These models identify components, enrich images, detect anomalies, and produce the desired outcome with semantic segmentation.

Wipro has worked with a major oil and gas customer and used the InspectAI solution for asset inspections and report generation. It improved the customer’s speed of inspection and streamlined their data gathering and report generation.

Summary

InspectAI is a solution that can help customers with digital exterior visual inspections. It can identify surface condition anomalies and measure severities. The overall solution provides visibility into asset condition, facilitates effective inspection management, and generates comprehensive inspection reports.

InspectAI was built on AWS using Amazon EC2 instances, AI/ML services, and serverless technologies. It uses compute-optimized GPU-based instances for deep learning and image analysis. Amazon SageMaker Ground Truth powered by deep semantic segmentation helps create pixel-based accurate segmentation for each object.

Customers can integrate the InspectAI platform with their inspection management tools. For customers who prefer to work with partners, Wipro offers digital asset management services using AI-based inspections and asset lifecycle predictions. Benefits to customers include remote asset condition monitoring, better communication between equipment owners and inspection teams, and quicker decision making on preventative maintenance.

Learn how you can transform the core and build the future of your energy business on the AWS Energy website.

.
Wipro-APN-Blog-CTA-1
.


Wipro – AWS Partner Spotlight

Wipro is an AWS Premier Consulting Partner and MSP that harnesses the power of cognitive computing, hyper-automation, robotics, cloud, analytics, and emerging technologies to help clients adapt to the digital world.

Contact Wipro | Partner Overview

*Already worked with Wipro? Rate the Partner

*To review an AWS Partner, you must be a customer that has worked with them directly on a project.