When it comes to security of electrical power supply, regular inspections of power plants, substations and power lines are key. So far, it’s been humans performing this task. But this may soon be a thing of the past. Autonomous robots and drones are getting ready to support them.
By Hubertus Breuer
When ANYmal is off the leash, its curiosity knows no bounds. It hurries through halls and corridors full of machines and pipes exploring every corner with its cameras, sensors, and microphones. The four-legged robot from ANYbotics has been specially trained to autonomously go on inspection rounds at Vattenfall’s combined heat and power plant in Berlin-Marzahn.
Enhancing autonomy with a special brain
ANYmal's trainers are Jan Weustink and Jaydeep Naha from Siemens Energy, the former a research and development expert in power plant automation and digitalization, the latter general manager for robotics and industrial metaverse. “Autonomous robots like ANYmal are perfectly suited for ensuring the operation and thus the supply security of a power plant, especially in times when fewer personnel are available,” says Weustink, explaining the reason why Siemens Energy and Vattenfall decided to test the robot in Marzahn.
ANYmal’s inspection rounds are controlled by its special “brain”. It’s based on an indoor navigation system for smartphones developed by Weustink and Naha that uses augmented reality to guide service technicians to the site with pinpoint accuracy, even without GPS or internet access. “This Dynamic Virtual Plant Insider, as we call it, not only enables the robot to move autonomously around the power plant. It also tells it where to find all the valves, pumps, gauges, etc. that it needs to check,” explains Naha. “It’s a capability that we can implement on any robot that works with the standard ROS operating system or supports comprehensive APIs,” Weustink points out. “Meaning – we can also choose to equip other robots with this digital brain.”
Domain knowledge and digitalization
The quality of the Dynamic Virtual Plant Insider (DVPI) is based first and foremost on the comprehensive domain knowledge that Siemens Energy has about power plants in general and specifically about the power plant built in Marzahn. “That’s knowledge that ANYmal was equipped with,” Naha points out. “In addition, as with ANYmal’s case, robots with DVPI have the ability to constantly update their knowledge concerning their surroundings, just as natural brains do.” Recent advances in sensor technology, real-time algorithms, AI, edge and cloud computing make this possible. So ANYmal also works if its Wi-Fi or 5G connection to the control system were to be interrupted.
Ensuring security of supply by air
In its search for autonomous vehicles providing greater security of supply in the energy industry, Siemens Energy isn’t just limited to robots. “We also work with autonomous drones,” Weustink says. “For the combined cycle power plant Alon-Gat, Israel, a team of digitalization experts at Siemens Energy Ventures and drone manufacturer Percepto have partnered up to develop a solution that enables autonomous external inspection of the power plant.” The drone regularly flies over specific parts of the plant, examining them with its sensors and cameras as well on-board software with sophisticated AI and deep-learning algorithms to check for anything unusual. The drone sends the collected data to the control center, and if any anomalies are detected, it’s examined in more detail and any necessary maintenance or repair measures can be initiated.
“Drones are also available for required inspection flights on high-voltage overhead lines”, explains Naha. That was shown by a demonstration flight in September 2022, which Siemens Energy and Bayernwerk carried out on a section of an overhead line in southern Germany. For this task, a drone was equipped with Siemens Energy’s multi-sensor system SIEAERO. It allows to capture in one flight all the inspection data the later AI-based analysis needs, for example, to identify, e.g., damage to overhead land pylons since the last inspection.
To boldly go where no humans can go
For Weustink and Naha, it’s clear that inspection robots and drones will be commonplace in the near future and even have more capabilities than today. For example, Siemens Energy, Elia Group and Nemo Link together with Ross Robotics are currently co-developing an inspection robot that is able go inside high-voltage direct current converter halls during operation.
When operating such facilities are inaccessible to humans due to the risk of electrical flashovers. Until now, they have to be shut down for inspections. With a robot capable of withstanding strong electromagnetic fields, inspections can be carried out without costly and time-consuming outages. In recent tests conducted in a high-voltage test laboratory, where electrical flashovers can be ignited with over 1.000 kilovolts, the robot has already proven that it’s highly capable and resilient for its future job. With its thermal imaging camera, for example, it will soon deliver images from converter halls during operation in Belgium and Germany, showing leakages or hotspots in order to predict necessary maintenance works for the next planned outage.
Towards autonomous power plants
In the future, autonomous robots and drones will also play a decisive role in ensuring that gas-fired power plants can be operated autonomously. “This is indispensable for the success of the energy transition,” emphasizes Weustink. “If renewables aren’t available on cloudy, windless days, gas-fired power plants will continue to be needed as an ad-hoc reserve for the foreseeable future.” However, staffing such power plants 24/7 with service technicians for occasional ramp-ups will hardly be possible due to the increasing shortage of skilled workers. It’ll therefore be inevitable that they’ll be controlled and maintained remotely with the support of autonomous robots onsite. “We're sure that this future isn’t too far away,” Naha and Weustink affirm, “We have all the technologies to make it a reality.”
Hubertus Breuer is a science and technology journalist based in Munich.
Combined picture and video credits: ANYbotics, Vattenfall, Ross Robotics, Percepto, Siemens Energy