Sie sind auf Seite 1von 2

Just about everybody now knows that the Number 4 turbine at Eskom s Duvha power st ation has gone

to heaven. The Internet is full of it: They were doing a test of t he turbine overspeed protection system and, in short, the protection did not kic k in . . . The turbine has a governor valve which controls the amount of steam coming into t he turbine in order to keep it running at the right speed (3000 rev/min for our grid frequency). It also has a main isolation valve to shut the steam off comple tely. The protection systems (of which there are independent systems, and a dude with his finger on the emergency button) are supposed to close this main isolat ion valve in a fraction of a second when the turbine overspeeds. So they get ready for the test, dump a helluva lot of steam onto the turbine and speed starts going crazy it went from 3 000 rev/min to 4 500 rev/min in ten seco nds (they are generally designed for only 10% to 15% overspeed). All three prote ction systems should kick in by the time you get to 110%. Anyway, I don t know why , but all three systems failed . . . What do we make of this, my dear Watson? First, what we can make of it is that t he above is not written by an engineer. Or somebody even close. It is very likel y that, in fact, there was no operator error at all just that the turbine just f ailed like that. Like the Concorde. Like the Titanic. Like Piper Alpha (yes, go look it up). A review of the test process: to conduct an overspeed test, you run the turbine and do not synchronise with the mains. You use the governor control to raise the turbine speed to a set point where the electrical overspeed trip i s supposed to operate. You can do this as slowly as you want. If all is well, it does operate and trips the turbine. You restart and then bypa ss the electrical trip and raise the speed to the point where the mechanical ove rspeed trip operates. Before you do all this, you will have already established that the emergency stop button trips the whole thing. All with me, so far? What you should see is that, unless something incredible happens, you just cannot dump a hellava lot of steam onto the turbine; and the turbine is big you would have t o have a vast amount of steam to get it up to 1,5 normal speed in ten seconds. So what really happened? Let us consider what happens when a turbine is connecte d to the grid and running at full load. Around 800 MW of steam is running throug h the tur- bine and 600 MW of electrical power is coming out of the alternator i nto the electrical grid. The rest of the steam is recycled or lost as heat. What happens when the circuit breaker to the electrical grid trips? Suddenly, you ha ve 800 MW of steam still going into the turbine but no balancing electrical powe r coming out. The turbine will overspeed and either trip on overspeed or the ste am safety valves will lift and all is safe. What I think happened was that Eskom was doing a full-load rejection test, havin g tested all the overloads . . . and the turbine failed. Sad but true such thing s do happen. If true, the news is not good. Incompetent engineers can be fired b ut faulty turbines cannot be fixed swiftly. If one is wrong, so are others. This would be a disaster of major proportion. But I just cannot see that Eskom would do such an important test and cause the t urbine to fail. The State-owned enterprise does have very good technicians and e ngineers and historically good test procedures. What next? Well, in the worst ca se, all the Duvha turbines are faulty and the African National Congress company, Chancellor House, buys 25% in a turbine company to replace them at five times t heir actual cost. In the best case, it is an isolated incident and Eskom asks me to write a report on the failure. Nice idea, but I m not holding my breath. Terry Mackenzie-hoy First published in EN.

http://www.dailytenders.co.za/Global/News/Article/Article.asp?ID=6534

Das könnte Ihnen auch gefallen