Now imagine you are controlling a microgrid with a solar plant, five battery containers, an electrolyzer, a fuel cell and three combined heat and power plants. Your task is to provide energy as sustainably as possible, yet of course reliably and as economically as possible. To do this, you need the current and future weather data, because you need to know when you can expect electricity from the PV system. But you also need to take into account the prices on the electricity exchange - and of course your own fluctuating energy requirements. And of course, you need all of this ahead of time. Difficult? Yes. Feasible? Yes, with artificial intelligence.
mtu EnergetIQ, the Energy Management System (EMS) from Rolls-Royce Power Systems, uses artificial intelligence to control what is described as a micorgrid. This predicts the user's load profile. Based on past data, it can accurately predict when how much power will be needed. And from that, the system then calculates exactly which energy source to use at what time to optimally meet the energy demand. "Based on neural networks, our system is self-learning. It learns about our customers' energy consumption and uses it to derive future demand," explains Jan Henker, Senior Expert Automation and Controls at Rolls-Royce and responsible for EnergetIQ.
Artificial intelligence supports maintenance
mtu EnergetIQ is a predestined field of application for artificial intelligence, but it is not the only one. Another example is the maintenance of motors. Even today, there are often rigid maintenance schedules: after a certain number of operating hours, a motor part must be replaced or serviced-no matter what its condition. This is simple, but not sustainable. The goal is for sensors to detect the condition of individual parts on the engine and provide indications of when they need to be replaced. But artificial intelligence opens up whole new possibilities: It can detect anomalies in an engine's operating data before an error message or even damage occurs. The prerequisite for this is data loggers that bundle the sensor data and send it to Rolls-Royce Power Systems.
"Artificial intelligence is no longer a future topic for us, but has long since arrived in many of our products."
Artificial intelligence detects anomalies before damage occurs
A major goal of the Rolls-Royce AI team: every mtu product, but also every machine used to manufacture mtu products, generates a model of itself. This model learns more and more about its own operation and how it should actually run. If it runs differently, it immediately reports this anomaly so that it can be reacted to. His team has already successfully demonstrated that this works. "Ideally, the AI will also provide the diagnosis and suggested remedial measures, so that the repair can be carried out in a targeted manner before a failure occurs," says Tobias Weiß, looking to the future.
"AI Companion" thanks to artificial intelligence
A future that is no longer far away for him. "The ChatGPT program shows us quite clearly how rapidly artificial intelligence is developing and finding its way into our everyday lives. We follow the latest developments very closely with the aim of applying them in concrete solutions for our colleagues and customers," says Weiß. One initiative he is currently driving forward is the "AI Companion" - a kind of digital companion in the company that every employee is to receive. Just like the model of the engine or machine, this one is intelligent and learns with each working day, just as the employee wants. They can ask their Companion questions and get support at work. Sounds like an AI geek from Silicon Valley is talking? "This is already in the works," Tobias Weiß reports. "It won't come in a few years," he is convinced. The freaks from Silicon Valley will certainly agree with him.