Robotics technology and artificial intelligence (AI), have been a major ingredient in the oil and gas industry’s success in bringing down costs and increasing efficiency in their operations. It has also significantly improved the safety culture in its key processes, especially when maintaining assets in or involving hard-to-reach areas such as storage tanks, offshore platforms, etc.
Now that the world is looking to transition to renewable energy, will robotics technology have the same key role and great contribution to the efficiency improvements, and consequently cost reduction in this industry?
To answer questions surrounding this topic, we invited Adam Serblowski, VP of US-based AI company Airtonomy. Adam is a veteran of the industrial robotics industry, having spent a decade building, deploying and productising robotic solutions for the Oil & Gas industry. By focusing on solutions, instead of technology, he has been able to deliver transformational change as demonstrated by his delivery of the world’s first IECEx Zone 1 Mobile Robot and the first commercial BVLOS permit in the USA.
At Airtonomy, Adam is now using this experience to solve the challenge of scaling robotic deployments for large enterprise users which will lead to improved operational efficiency and lower human exposure to potentially hazardous situations.
Read his insights below.
Looking for Robotics technologies? Find them on our platform here.
1. What are the popular applications or use cases for robotics in the energy sector? What are some examples of cross-application robotics technologies that were proven and tested in the conventional energy and other major industries and can be replicated in the renewable energy sector?
Generally, industrial operators have largely similar challenges and people are looking at the same basic things: what is the condition of my asset that we can make plans for and do repairs accordingly. The geometries are just a little bit different, with different environmental factors in the areas they operate which may cause some variations in how we deploy the robots.
So, I think a lot of the robots are actually portable from where we use them, such as some of the non-destructive testing (NDT) robotic crawlers that can be used on vessels or on towers for wind turbines or drones with cameras and sensors that can be used for any number of applications.
To answer the question, are these robots completely different? Yes and no, it is often the same basic platform with just different sensor/hardware packages on them. For general inspection activities, we will see the same robots used across many different use cases. But, we will of course see some specialised robots, such as blade repair robots, which are highly specialised and are going to do just one task or solve just one challenge.
In terms of process, it is going to be all the same. When we look at a robotic inspection, people are searching for evidence of specific failure mechanisms that are specific to their assets. We need a robotic tool equipped with some combination of sensors, and mobility to navigate to the area of interest and take a reading. We need a process in place that is going to help us execute that mission, capture the data and push reports out. Ultimately, that part of the process is exactly the same, then the technology at the front end, how that actually changes a little bit, will be very portable.
So, when we people start saying “oh we need to develop robots for the renewable energy industry”, we need to see that there are already a lot of robots in the industry that can be reused. Then, overtime, we are going to find out what are the specialties—what do we need specifically for solar, hydrogen, etc.
2. What are the primary considerations for energy companies when deploying robotics technology in their operations? What are the vulnerabilities and opportunities surrounding its deployment?
There are two big things, which are largely the same for any business. First, there are assets that were distributed or deployed in remote areas and managing these assets is inherently risky. So, if I can keep someone from climbing that wind turbine or solar panels on roofs by using drones, then that would be one primary consideration.
And the other part is cost and efficiency. Right now, finding skilled people to carry out difficult tasks needed in renewable energy is really tough. So, robotics fills that gap in a cost-effective way.
The Oil & Gas industry does have some unique challenges owing to the potentially explosive atmospheres that exist in their facilities., The industry has really adopted a lot of robotics over the last decade by focusing on activities in which this risk can be eliminated during the work, or by staying outside of these zones and looking in. But for robots that can get into the core of their facilities, there is a very limited market for that. And this slows down the rate by which these types of robots develop.
For new energies, on the other hand, is location—a large number of assets are dispersed across multiple, mostly remote locations. How do we have one robot in a remote location controlled by a crew of people from another city? And this is where regulations would also play an important role.
But overall, I do not see limits of what robots we can use for new energies like how we use it for the oil and gas industry.
3. Part of the hesitation to deploy robots in complex and unique operating environments is the lack of a framework that could quantitatively assess the reliability and resilience of robots throughout their life cycle. What do you think should be done to 1) address this hesitation from industries and 2) put in place an accurate process assessing reliability and resilience of robots, especially in harsh environments where renewable energy companies operate?
Actually, I do not feel that hesitation anymore as this technology has been in use for over a decade now. When you go to trade shows these days, you will see more and more energy companies sharing their own robotics programs. Now, it is just providing that glue that will embed these robots into their own enterprise.
So, there is no roadblock to convince people anymore. There is enough experience that people believe and understand the value of robots. However, the challenge now is to show the value of robots as a tool that is embedded into daily operations supporting regular workflows as opposed to a specialised one-off tool.
In terms of regulatory or legal frameworks, there is no formal certification or bodies regulating robots where you can have them approved or certified as ready to use. However, there are different guidelines and processes depending on what type of robots we are looking at which we can use for guidance. There are general guidelines on how to operate drones and industrial robotics safely, and industry group such as SPRINT or ASME that are working to build more industry specific guidelines through contributions of their members.
Just as valuable as formal certification is the number of deployments and successful examples these individual robots have and how portable they are across industries. Each successful deployment is a data point which shows the value and the safety of these tools, and that is invaluable to both end-users and industry groups to build trust. And with this, people learn to inherently trust them to complete the task just like we do with an arm in a factory today.
4. Scalability of robotics technologies is still a challenge, resulting in parallel issues in the interoperability of robots sourced from different suppliers. How do you think we can address varying systems in the supply chain?
This has been a passion of mine for a while in my career in robotics. Today the industry is largely focused on tools for task, in other words, lets match a robot to a task and when the job is done the robot leaves. But there’s a whole lot of smaller activities that you cannot actually justify bringing in robots for and so then you start to look at embedding robots into operations. That is where interoperability and scalability become so important.
For instance, if I want to embed a drone into an operating facility, it has to be deployed by existing staff who have all sorts of tasks already and don’t have time to take on new work. Hiring a dedicated staff to fly a drone all the time isn’t an option easier as drones just can’t fly all the time and we don’t want staff sitting around idle waiting for a weather window. So, that’s when you are going to need that suite of tools that are going to make operating robots so easy that they can be a part of the workflows and executed by the regular staff. In doing this we are allowing existing staff to do the same job they have now, but in a safer and more efficient manner without the need for extensive reskilling programs. Just as important to an enterprise user is scalability which is achieved not only by ease of use, but will also include integrating to the data flows with existing asset management tools and making it easy to package up the entire solution in an easy to replicate package, eliminating the need to treat each new deployment as a brand new investment.
So, let’s say I create an inspection for a refinery here in the United States, I should be able to access that in any other refinery anywhere else in the world for that matter. By doing that it enables us to reduce the training needed for individual pilots, and robots then actually stand as an everyday tool.
To do this, we are going to need a software that ties all these robots together and creates a framework to enable an enterprise to access all of this data in a consistent fashion, regardless of source, operator or location. And that is, in essence, Airtonomy. I was attracted to Airtonomy is because I have been looking at this challenge for years now, and have never found a good solution. Time and again talking with industry peers this was identified as a roadblock which prevented scaling, but all automation solutions I found were always focused on solving a very specific vertical challenge or were so broad that you were effectively developing your own solution. With Airtonomy UNITI, regardless of the drone you are using, the pilot experience is identical and the data is stored, organised, contextualised and presented in the same way. This allows enterprises to build up workflows around this data as they can now access it in a predictable way without relying on pilot experience affecting data quality. For example, for tank inspection, we have built a rules-based flight plan in UNITI which describes what a tank is and what it looks like. For a pilot to inspect a tank for the first time then, they just need to provide some common data points such as the GPS coordinates, the size and height of the tank, and then the software using these rules and data adjusts the flight plan to be specific to that asset. This is transformational as it essentially allows us to copy and paste flight plans across any tank rather than manually creating a custom flight plan for every single new tank, we inspect. Just as important, because of the rules-based flight, we know at any point in time where the drone is and what it is looking at so we can add context to data at the moment of capture eliminating the post-processing work that often goes into these programs. I have literally seen examples where we have been able to go from thousands of images per inspection down to less than a hundred while still getting the same quality localisation of alternate methods. These are the problems that have prevented scaling drone and robotics programs, and I am excited to be at a company that is actively solving them.