Adapt Your Pipeline Management to the Challenges of COVID with AI Posted on September 4, 2020 COVID-19 has put the oil and gas sector in a uniquely-difficult situation. Amidst a one-two punch of extreme market forces – market demand decrease and supply-chain disruption – the industry is staggering, with 23 oil producers and 18 oilfield service firms going bankrupt in 2020 alone.1 With such chaos in the upstream, dire tidings now loom for midstream firms, as upstream customers reduce scope and renege on pipeline contracts.2 If that weren’t enough, the health risks of COVID have forced many companies to quickly shift to an unfamiliar work arrangement – shifting numerous staff to remote worksites and having to re-think the nature of field operations entirely.3 Here’s the conundrum: How can you improve performance in a difficult market, while having fewer resources at your disposal than ever before? AI Can Help You Come Out of COVID-19 With Newfound Momentum AI is the central means through which you can have your slice of the $100 billion pie that the digital transformation is poised to deliver to midstream oil & gas firms. In a recent World Economic Forum Report on the Digital Transformation Initiative in the Oil & Gas Industry, it’s estimated that the implementation of automation along with advanced analytics and modeling is poised to reduce incidents by nearly 10%, and the number of pipeline spills by nearly 100,000 barrels by the year 2025.4 So how does this value manifest for you as a pipeline operator? Develop an intelligent preventative maintenance program: AI can pinpoint critical points of data and use them to build predictive analytics which reduces downtime and maintenance costs. Quantify visual data to play well with other sensor systems: Transform visual inspections of your pipeline from subjective to quantifiable data that can be easily referenced against other sensors for holistic analysis. Empower decision-making of remote workforces: Distill readings and raw imagery from your remote sensing technologies into actionable points of data. AI allows your remote pipeline management teams to spend less time trying to make sense of disparate data sources and more time making decisions that prevent damage. Reduce demand on field resources and personnel: When paired with Unmanned Aerial Vehicle (UAV) technology, AI forms the backbone of a solution that provides true automation for pipeline integrity monitoring. Let’s walk through the process to see how robotics and automation can deliver on these results. Quantifying What You See on Your Pipeline Right-of-Way After a baseline inspection of your pipeline right-of-way, deep neural network models and algorithms can be applied for subsequent analysis. This primarily happens through two analytic techniques: Anomaly Detection: Through training on numerous data sets, a neural network can learn to detect and categorize people, vehicles, structures, and out of the ordinary occurrences along your right-of-way. SkyX’s models are trained to detect numerous anomalies associated with pipeline failures, ground movement, third-party activity, and damage to environmental health. Change Detection: By comparing subsequent flights against an initial baseline, change detection analysis can reveal actionable issues that are subtly developing your right-of-way. During this process, an algorithm normalizes ROW images from different inspections down to the same geospatial location for comparison, while a fully convolutional differential detector model identifies the changes between the two given images. This is great for catching issues such as gradual erosion over your pipeline or development/excavation that’s approaching your right-of-way. While the neural networks behind this analysis may mimic the human visual cortex, they lack the complex decision-making of a human being to determine whether these are really and truly actionable issues. For this reason, a final step of human verification takes place before data is ported into your GIS. At SkyX, we have our own in-house analytics team that verifies and discards all flagged anomalies before client delivery. Create Profiles of Your Pipeline Integrity Issues Intelligent decisions are made from more than a single source of data. Thankfully, AI can make meaningful connections between your various sensor systems – for holistic pipeline management. Take the case of pipeline failure due to corrosion along the line. AI can combine visual anomalies in the area (such as soil discoloration or pooling liquid), readings from the nearest SCADA point sensor, models from the most recent PIG inspection, and even useful record data such as the age of the pipeline and coating history – and create a profile for what corrosion incidents look like on your line. Not only are these critical points of data useful to your team in a real-time scenario, but they also provide the fuel for intelligent predictive analytics. Realize Preventative Maintenance Through Predictive Analytics As AI starts to recognize the trends and patterns along your pipeline, it will provide insights that can be used to predict failures and plan maintenance interventions before damage manifests: Hotspot areas for certain issues and the common threads between them Where and when future issues are likely to occur An incident forecast based on detected changes from inspection-to-inspection By combining a real-time visualization of your pipeline with advanced analytics that pinpoint and predict potential threats to your pipeline, your remote teams and on-site staff can collaborate with greater efficiency than ever before. More importantly, with predictive insights, you can implement an intelligent preventative maintenance program that eliminates disastrous accidents, reduces downtime, and ultimately puts less stress on field resources and personnel. Adopting AI into your pipeline integrity management program will allow you to capitalize on the promise of digital transformation for midstream oil & gas. If it seems like a daunting undertaking, data providers like SkyX are ready to help you through the journey. Have questions about how high-quality aerial data can elevate your organization? Contact our team to discuss your unique challenges and data requirements. References Wave of North American oil and gas bankruptcies to continue at $40/bbl crude: report, U.S. Legal News, Reuters, Liz Hampton, July 2020 As upstream bankruptcies loom, oil and gas pipelines brace for contract disputes, S&P Global Market Intelligence, Allison Good, June 2020 A safer, smarter future: Working remotely in energy and materials, McKinsey & Company, May 2020 Digital Transformation Initiative: Oil and Gas Industry, World Economic Forum, 2017