Pictures of the Future      Spring 2005

"Remote Services Ensure Maximum Availability"

What are the main benefits of remote maintenance?

Weiß: The most important thing is the time factor. In the case of plant components, we don’t always have a service partner available locally. And as the operator, we often lack the required nuts-and-bolts expertise. When something goes wrong, it’s usually vital to correct the fault promptly.

Do you improve plant availability if the service partner can respond faster?

Weiß: Absolutely! Availability is very important. We’ve got base load plants that are operating with an availability of over 90 %. That’s as good as it gets, and not just in Germany. Another advantage is that a service partner like Siemens can apply experience from all the plants that it services to help us detect problems early.

How do your diagnostic systems work?

Weiß: We have three of them. In offline diagnosis we sample data at regular intervals. Our experts analyze these data to determine when repairs are required, say for large pump motors. We use online diagnosis when availability is paramount. So major subsystems—like in the newly installed power units at the Schwarze Pumpe and Boxberg power plants—are connected to a central control room. We use the third version, the remote services, particularly in our control systems, and that’s been working very well with the Siemens Remote Expert Center in Karlsruhe.

Can you increase efficiency or perform preventive maintenance with this method?

Weiß: Yes, we can. Any increase in efficiency lowers specific carbon dioxide emissions, among other benefits. We can use highly specific diagnostics to detect areas that cause a loss in efficiency, or where a worn part needs to be replaced. As recently as ten years ago, we had to completely dismantle every turbine into its component parts at regular intervals and then reassemble it. With today’s diagnostics, we can see that such component-level inspections of turbines aren’t needed that often. As a result, we’ve been able to substantially lengthen the intervals between complete inspections in the power units of our base load plants. We used to conduct such inspections every three to six years. Now we do them every four to eight years.

What’s the sequence of events in your remote maintenance?

Weiß: With our permission, a technician at Siemens in Karlsruhe can visualize the entire control system of a power plant via an ISDN line on a display screen and discuss functions with our technicians, or even download new software.

Suppose a malfunction takes place on a weekend. What happens?

Weiß: We run our base load plants continuously with three shifts, even on holidays. If anything happens that’s out of the ordinary, the first person to be contact is a specialized engineer at Technical Service, via mobile phone. For special situations we have agreements with service partners. A call to their hotline might be sufficient, but we could, if necessary, even have someone link up remotely. But not every problem requires an instant solution. Many situations can wait until Monday, because we’ve got redundancies in many of our components.

Let’s talk about data protection and security. Doesn’t it bother you when a service partner like Siemens is looking over your shoulder?

Weiß: Not really. We’ve got nothing to hide, and the service partner is contractually required to handle the data on a confidential basis. I’m more inclined to expect problems in the area of responsibility or liability. We transmit our data without analysis to the service partner. Let’s say something is overlooked at the service center, and the plant goes down. Now who’s liable? It goes without saying that communication takes place through absolutely secure channels.

What’s your vision for the year 2015?

Weiß: By then we will have made good headway in machine, turbo set, and generator diagnostics, and we should have a service package for the entire plant from a single source. Maintenance steps should then be entirely status-related, and the data should be used to optimize processes. Neural networks, for instance, can be useful in detecting irregularities. Recently, we had a drop in the feed water flow rate in an operational power unit. The control system responded by engaging a second and even a third pump before somebody noticed that there was a leak in the water line. A neural network that analyzes all of the variables and checks them for plausibility would have detected this fault much sooner. That’s where the future lies. We’re researching these matters with several partners, including Siemens. But there’s one thing you can bet on: Even in the future, no neural network will replace our professional staff.

                                   This interview was conducted by Norbert Aschenbrenner