Power Station with real-time rejuvenation

The Jebel Ali power plant of Dubai Electricity and Water Authority (DEWA) has impressively demonstrated how combining a digital twin with artificial intelligence (AI) can achieve a lasting improvement in the operating result. Five gas turbines in Jebel Ali’s M Station, which is the biggest power station unit in the United Arab Emirates, were recently equipped with this type of upgraded controller.

Thanks to the upgrade, age-related performance losses in the turbines can be automatically offset. It also ensures very stable combustion, so the turbines operate even more efficiently and with lower emissions overall. The benefit for DEWA is twofold: The performance of the five turbines has been increased by up to 3.5 MW each, and NOx emissions have fallen by as much as ten percent, depending on the operating mode.

The five turbines probably won’t be the only ones at DEWA with an upgraded controller for long. “We are already preparing to upgrade other DEWA turbines in the same way,” says Kai Süselbeck from Siemens Energy. He has been in charge of developing the upgrade as well as the pilot testing and implementation in Dubai since the project began in November 2017.

At that point, Siemens Energy had just completed a major thermodynamic examination of the turbines in M Station in order to offset age-related performance losses. “It was all done manually at the time,” Süselbeck says. In fact, a specialist in thermodynamics from Siemens Energy still analyzed all the data directly to calculate the compensation factors. “And with an excellent result,” Süselbeck affirms. “It was no surprise that we immediately got a request to automate the process.”

It didn’t take long for Süselbeck to find an automation solution. Co-workers at Siemens Energy in Orlando had recently developed a solution that was perfect for what he wanted: It could calculate a turbine’s operating status in great detail. It’s based on a digital twin, which produces a precise physical simulation of the turbine’s thermodynamic behavior. It just needs to be fed with the latest readings from the turbine and its environment.

In short order, everything’s in place to enable a precise snapshot of the turbine’s operating condition. “Of course, we still made a few adjustments to suit the specific circumstances of Turbine 11,” Süselbeck recalls. “But then we were able to put it into operation.”

The calculations using the digital twin provide values at a level of quality that wasn’t previously available, especially for the turbine inlet temperature of the gas, which is a key parameter for turbine control. Not only does the twin make it possible to determine the aging of the turbine with great accuracy, but because the temperature readings are available in real time, age-related performance losses can now also be offset in real time.

“Of course, it would be even better if we could measure the temperature directly,” Süselbeck concedes. “But that isn’t possible. We still don’t have a measuring system that can cope with such high temperatures. With these gas turbines, the turbine inlet temperature can be as high as 1,500° Celsius.”

Süselbeck also benefited from earlier efforts by co-workers when it came to the AI that now makes up part of the DEWA controller upgrade. A project by Siemens Energy in Berlin and Siemens Corporate Technology in Munich tested what were then new learning-capable algorithms that assist with turbine control in complex control tasks. “Strictly speaking, this is a special type of AI known as ‘reinforcement learning," Süselbeck explains.

In this case, the algorithms imitate a learning behavior that’s found in activities like children who are learning a new game by trial and error and continuously improving their playing strategy. “By feeding an AI system with genuine measured operating data from the turbine and the temperature readings supplied by the digital twin, the system can develop control strategies that were previously unknown and are many times better than the other traditional, model-based control solutions available on the market,” Süselbeck observes.

A number of patented innovations by Süselbeck’s co-workers in Munich made reinforcement learning suitable for deployment in a power station controller. One important element is a method that allows the algorithms to be structured so they can develop their strategies using far less prior empirical knowledge than was previously needed.

“Without this step, which we’ve also protected with numerous patents, we wouldn’t have been able to include AI in the DEWA controller upgrade,” Süselbeck notes. “But then we’d have been lacking an important element needed to automatically compensate for the aging process. AI doesn’t just ensure more efficient combustion with lower emissions. It also watches to make certain that compensating for the aging process never jeopardizes combustion stability.”

The fact that the controller upgrade was developed to the market-ready stage in just one year isn’t due solely to the innovations from Orlando and Munich. As one of the world's leading drivers of innovation for the digital transformation of utilities equipped for Industrie 4.0, DEWA has also played a very important role. It has been an extremely receptive partner for Süselbeck in the pilot testing process, especially the Emirati experts.

Following successful collaboration in the development of the control upgrade, Siemens Energy and DEWA are therefore already working together to advance further digital innovations. The objective is to extend the control upgrade to further components of Jebel Ali M Station in the next phase. As a result, controller upgrades won’t benefit only the gas turbines in the future, but will also improve the subsequent steam production process, which is used at Jebel Ali both to generate electricity and to recover drinking water from seawater. The goal is to optimize the overall operation automatically.

Süselbeck is certain that, “In power plants like Jebel Ali, which have to meet rapidly changing demands, an optimizer like this is crucial for becoming more efficient and reducing emissions. I have no doubt that with the amazing collaboration between Siemens Energy and DEWA, we will achieve this goal.”