FlickrCollapsing bridges, crumbling aqueducts, and corroding highways: Infrastructure is big (and increasingly urgent) news in the U.S. A new class of robots and inspection drones from a variety of vendors are beginning to scale structures and perform acrobatic flybys to discern the real state of affairs.
With federal dollars for infrastructure increasing, the technology could have additional public sector applications. But how big a difference can robots make? Is it too little too late for long-overlooked infrastructure problems? And is the regulation of autonomous systems keeping pace with the technology?
To answer these questions and more, I connected with Jake Loosararian, CEO and co-founder of Gecko Robotics, which leverages robotics as a service to perform industrial inspections in critical industries such as power, oil and gas, and defense. Gecko’s wall-crawling robots scale along industrial assets inspecting for damage while collecting 1000x more data points than traditional inspection methods at 10x the speed. They are furnished with cameras and sensors to detect corrosion, pitting, cracking, blistering, or laminations, and they encompass an impressive intersection of technologies, including robotics, AI, and machine vision.
GN: Let’s start with the headline: What’s at stake if we don’t get infrastructure right?
Jake Loosararian: In short, our standard of living. Damaged infrastructure can mean increased power outages, more expensive utilities and transportation, shortages of essential goods, and production and supply chain disruptions affecting many other aspects of the economy. There is also the risk of catastrophic disasters affecting workplaces and communities alike. Longer-term, it’s also important to recognize that we are still incredibly dependent on traditional energy sources, and if we can’t meet current needs at a cost that societies can bear, it will dramatically set back our ability to transition to cleaner forms of energy. And this will only get tougher in the face of inflation, geopolitical unrest, and global economic uncertainty.
GN: Okay, so pretty important stuff then. Let’s talk solutions. Why are robots becoming a critical tool for infrastructure inspection?
Jake Loosararian: Globally, nationally, and locally, we’re facing a looming energy crisis, and while there is a lot of focus on fuels and other inputs, it’s also an infrastructure crisis. In the near term, we need to maximize both the production capacity and useful life of the assets we have, which are being strained as never before by age, as well as extreme weather and surges in demand. And while energy is the most noticeable example, the same is true of the infrastructure underpinning many other aspects of daily life, including transportation, water, chemicals, paper and pulp products, etc. Maximizing the capacity and health of these assets requires ground truth in the form of accurate, rich, and representative data and robots can capture 1,000x as much data as traditional methods, 5-10x faster, with ~75% fewer personnel. In conjunction with the right software, robotic data capture and processing can drive the analyses needed to perform repairs, replacements, preventative maintenance, and operational adjustments, as well as guiding investment in new infrastructure where appropriate.
Safety is also a major benefit: robots can operate in places humans either can’t or shouldn’t and can reduce or even eliminate the need for inspectors to work high off the ground, near hot surfaces (up to 240F) or in confined spaces with potential exposure to hazardous chemicals. At the same time, robots can actually make human inspectors more valuable by informing their work with better, richer, and bigger data and allowing them to focus on their unique value add.
Robotic inspections are a vital safeguard against environmental damage, whether in the form of a catastrophic accident or everyday leaks and contamination.
GN: Sounds alluring. So how do your robots work, and what sorts of use cases are we talking about?
Jake Loosararian: Our robots are equipped with ultrasonic transducers, localization sensors, lasers, and HD cameras. They climb vertically and horizontally, adhering to surfaces magnetically, to inspect tanks, boilers, pipes, pressure vessels, digesters, hydropower equipment, dams, ships, and other assets for changes in thickness, cracks, corrosion, blistering, and other forms of degradation. Our robots are remote-controlled and use multiple types of ultrasonic non-destructive testing techniques, including Rapid Ultrasonic Gridding, Rapid Automated Ultrasonic Testing, Phased Array Ultrasonic Testing, and our latest development, Trilateral Phased Array. Our robots acquire data at a previously unknown scale and fidelity, and our data systems and software allow human experts to turn that information into vital operational and strategic decisions. In terms of use cases, Gecko is currently focused on critical industries, including power, oil and gas, chemicals, pulp and paper, heavy manufacturing, water storage, transportation, and defense.
GN: We’ve heard a lot about inspection drones recently as well. Would you consider that a competitive technology, and if so, what advantages do wall-crawling robots have over drones?
Jake Loosararian: It’s not an apples-to-apples competition because they’re good at different things and tend to be used accordingly. For our use cases, wall-crawling robots have several distinct advantages over drones. Robots have much more powerful ultrasonic testing (UT) data acquisition capabilities than drones, making them better suited to collect data at a scale needed to reliably inform sophisticated models, whereas drones are generally used for visual representations. Robots can have over 100 transducers on them, making them ideal for capturing metallurgical data around thickness, corrosion and cracking. Robots also collect massive amounts of ground truth data. Drones are best at image capture, which is useful but doesn’t tell you what’s going on in the world of the unseen. Robots and drones are two different tools at two different levels of complexity – think welding torch vs. screwdriver. (For a closer look, see Gecko’s December 2021 blog post comparing crawlers and drones).
GN: One of the cool things for me is that this technology isn’t hypothetical; it’s actually being used. Any customer success stories you can share?
Jake Loosararian: We were able to revolutionize Shell’s inspections, starting with an inspection of the tank roof of its Convent Refinery. During that inspection, the robot was able to collect 1,000 times the number of measurements as a human inspector would in a tenth of the time.
In our work with Siemens Energy, Gecko’s robots enable inspections to be conducted 5 to 10 times faster with 99.6% more coverage than manual inspections. We also eliminated the need for scaffolding and confined space entry. We work with BP as well, starting with an inspection at BP’s Blackhawk central gas-treating center in late 2020. That first inspection took 4 days, as compared to the previous inspection time of 4 weeks, collected 7.3 million measurements, and cost 75% less than previous vessel inspections.
In Oklahoma, we did an inspection of the above-ground storage tank of a 60,000 bpd refinery. They switched to our services not only because we could provide orders of magnitude more data than a human inspector but because of the unnecessary risk having an inspector on the tank roof posed. People walking around on the tank roof could not only damage the tank, but they could also easily and severely hurt themselves by accident. We’ve also had extensive success inspecting power plant boilers, extending maintenance windows and, in some cases preventing costly forced outages.
At one of SE Asia’s largest biomass plants, a Gecko inspection team of 4 personnel was able to collect 500x more data than the previous inspection team of 20 and make preventative maintenance recommendations that reduced the frequency of planned outages equating to tens of millions of dollars in potential savings.
