The Robots Are Coming — For Your Wind Turbine
An interview with Josh Goryl, VP, Business Development at SkySpecs. SkySpecs is an Ann Arbor, Michigan based robotics company focused on the wind industry. This conversation has been edited for length and content.
So, Josh, can you give me a bit of context on SkySpecs, the problem it’s trying to solve, and how it’s trying to solve that problem?
Definitely. We’re a robotics and data analytics company focused on automating wind farm operations and maintenance. Our fully autonomous drone inspection solution inspects a turbine in fifteen minutes or less and was the first of its kind when it came on the market four years ago. We’ve since developed a data analytics and software platform, that serves as the electronic health record platform for wind turbines. We work with energy companies, IPP’s (independent power producers), and public utilities, as well as manufacturers and service providers. The platform, Horizon, is the hub for their renewable energy operations.
Got it, so how has SkySpecs’ solution evolved since its founding? Was SkySpecs focused on a different problem initially?
Prior to developing our inspection solution, we were selling a drone-agnostic collision avoidance product. We realized early on that this solution seemed to work really well for infrastructure inspection applications. We had some early partnerships in the wind industry and learned what it meant to conduct a wind turbine inspection. We quickly realized that our solution solved some important challenges for the industry, which is the short version of how SkySpecs ended up in the wind turbine inspection business.
So I’ve heard you talking a bit broadly about inspections, can you walk me through how that process works from the pilot going out to the wind farm to the data flowing back to SkySpecs?
Good question. So, what happens is our pilots take our drones out into the field, they position the drone at the base of the turbine, and send it up to autonomously inspect the turbine. As the drone flies up, it’s creating a three-dimensional model of the turbine in real-time with cameras and sensors that are mounted on the drone. On average, the process takes about 15 minutes for a two-megawatt wind turbine. Once the drone lands, the pilot picks up the drone, puts it in the back of their pickup truck, drives to the next turbine, and repeats that process.
On a good day, we’re able to scan between 15 and 20 turbines using one drone. At the end of the day the pilot connects the drone and the tablet to the internet, and all the data is automatically uploaded to Horizon, where our customers can then interact with their data.
Wow, very cool. I know you mentioned the drone is taking photographs of the turbine, what other technologies are deployed on the drone to inspect the turbines?
We’re also collecting all sorts of laser and flight data on the turbine to help build a better 3D model of the turbine. This data is extremely valuable in our analysis process as it allows us to obtain precise location and damage size information. This becomes incredibly valuable over time as we can measure things like damage propagation. Ultimately, this rich data set enables our customers to make better repair decisions.
It sounds like you’re doing some really interesting things with lidar sensors and cameras mounted on the drone. How is that data analyzed within Horizon, is SkySpecs using machine vision to identify potential damage sites?
We are constantly looking for ways to further automate our processes and help our customers make better decisions. For example, in terms of automating damage assessment, we found that many of our customers don’t necessarily care about having that process automated or how that gets done as long as the results are accurate. One of the things that we are focused on is what we call Oscar, which is our rules engine. This engine allows us to map damage annotations to a customer-specific annotation or remediation.
We know that every company is different in terms of their risk tolerance and budget constraints. For example, a customer may decide that they want to automatically trigger an action — like reinspecting within a specific period of time — if a certain type of damage is identified. Oscar allows them to build these parameters and will continue to learn and improve over time.
So how do you work with customers to define those parameters? Does SkySpecs make recommendations based on the data it has collected?
We have our own recommended best practices, but what we’ve realized is that every customer is unique. They all have different budgets, risk tolerances, and operating models. We’ve realized that we shouldn’t be dictating what a customer should do, but work with our customers to tailor recommendations based on their requirements. Some customers have higher budgets than others and we can translate damage into repairs accordingly. Others want to send repair crews for a specific type of damage. That’s something that we work with asset management and engineering at each company to determine what those steps should be. The other thing that I’ll add is we do realize that a lot of these operating strategies are proprietary — much of how our customers run their wind farm is their secret sauce. And so, we’re not only giving them a mechanism that’s safe and effective, but also one that makes them more efficient while protecting their strategy.
What are some of the challenges that come with inspecting turbines and operating in these remote environments?
That’s right — many of these wind farms are in very remote areas. They don’t always have the best internet connection. Many are also in harsh environments — we do a ton of work offshore. So, for example, when it comes to these large turbines that are in the North Sea, there are safety limitations on how the vessels we fly off of operate. There can also be more downtime due to these factors. We also don’t fly in any precipitation as it affects the image quality, but we can fly in winds up to 38 mph.
What’s the turnaround time for a client after the inspection data has been uploaded to Horizon and what do they see when they get that report?
We get the data back to our clients via a Horizon dashboard in 7–10 days. In Horizon there’s a customizable table that customers can use to sort and filter through all the inspection data. From there they are able to create a repair plan. These plans can be shared with third-party vendors who receive granular access to the damage data through our collaborator feature in Horizon. These vendors are then able to go in and comment on the damage and upload relevant repair information on the platform. Our embedded business intelligence tool allows customers to create customizable dashboards depending on what a customer is interested in understanding.
Got it, how do those third-party vendors interact with clients in Horizon?
Another benefit of Horizon is that it supports the repair RFP process as well. Owner customers can create a repair plan in Horizon based on the damage reports and then send it out to all their approved vendors who can send a repair proposal back through the system. Horizon has been important in streamlining these processes, which can often be disorganized and difficult to track.
Given that SkySpecs’ inspection service is an added cost and your clients are very sensitive towards their operational costs, how do you track the value that you’re delivering to your clients?
The easiest business case to track is the reduction of catastrophic damage and blade failures. So anytime we find Category 5 damage (damage is often graded on a 1–5 scale in the wind industry), which can be catastrophic, and the customer takes action, we’re helping them prevent a blade failure, which can cost a customer anywhere from $150,000 to $500,000 to fix.
Here’s another anecdote — let’s say, for example, that we go out and find minor damage on a turbine, maybe a small Category 3 crack. That crack could cost approximately $4,000–6,000 to repair. We’ve had situations where we’ve inspected that same turbine a year later and found that the crack is now a meter-long Category 5. At that point, the customer is going to need to spend several hundred thousand dollars to fix the damage. If you look at scenarios like that on the macro level and damage propagation across an operator’s entire fleet, operators can be more proactive and save money on larger repairs in the long run. The cumulative savings are substantial.
So, rewinding for a moment, before you really even engage with a customer you have to sell them on the service. How do you sell your service to operators who haven’t employed something like this in the past?
First of all, it is very much a Solutions sale and there is a lot of education that we provide at the start. Our customers are large enterprises, many of which are going through a digital transformation. Generally speaking, the industry has adopted the use of drone technology pretty quickly — in large part due to the cost reduction and quality improvements that were immediately realized. In some cases, it does take patience and can be a long journey for a big customer to get comfortable with the transition. After all, we are bringing new technologies to a relatively young industry.
Given that SkySpecs does a lot of work with data and provides a service to organizations that haven’t used or been able to use the level of granular data that your drones provide, what advice do you have for startups/students/researchers who are trying to convince organizations to spend the time and effort to analyze data?
Always listen to your customers and provide real solutions to their problems. Don’t get too overwhelmed about trying to make sense of all of the data because there is a lot available. Focus on one problem that you can solve. At SkySpecs, we stayed focused on the wind turbine blade problem instead of expanding to other verticals, and as a result, we’ve become an integral part of our customers’ teams, working to automate processes and understand their challenges.
Lastly, what excites you the most about the future of SkySpecs?
It’s really exciting to make an impact in the wind industry and help customers lower their operations and maintenance costs. As a team both in the US and Europe, we are committed to making renewable energy ubiquitous and we are passionate about the role SkySpecs is playing in that mission. We truly believe that we can help build the robotically run renewable energy plant of the future.
