Hi All,

I have a TIL from GE about Mechanical Overspeed.
With the Speed Tronic TMR MKV or MKVI no need to test the mechanical over-speed after a M.I, what is right. What is your opinion or experience on this

Thanks,

 

Dear BB,

I suspect you will find a lot of opinions, especially on this subject. My personal opinion is that if you have a TMR MKV or MKVI panel then testing of the overspeed bolt should be unnecessary. My reasoning is that you basically have overspeed protection from the main core, either MKV or MKVI, and you also have overspeed protection from each protection core X, Y and Z. The main reason I dislike testing the overspeed bolt is the stress on the machine of a trip from the overspeed bolt setpoint of 110-113% of speed. I would much rather prove the logic and electronic trip functions using a frequency generator, or setting the control constants of overspeed to a lower value for testing and seeing that logic and trip mechanisms function properly.

Correct testing and verification of these overspeed devices should be protection enough to prevent an overspeed event. I think the issue typically is with insurance wanting proof that all overspeed devices are functionally tested and proven.

I await answers from others on this subject!!

 

I agree with everything MIKEVI said, but I want to make a very important clarification.

Units originally built and installed with Mark IVs, or Mark Vs, or Mark VIs most likely have six independent speed pick-ups, one for each of the control- and protective processors. And, many of these units don't even have overspeed bolts. So, we're not referring to these machines in this discussion.

Some machines with mechanical overspeed bolts, especially those retrofitted with Mark IV or Mark V or Mark VI control systems, were not originally equipped with sufficient speed pick-ups to provide an independent speed pick-up for each processor (control- and protective-). So, in some cases, speed pick-ups were jumpered to various processors, and in some cases it wasn't done in the best manner. To my way of thinking, units without an independent speed pick-up for each processor and an overspeed bolt should probably conduct regular, periodic overspeed tests.

In the end, it's really up to the insurance company, as MIKEVI says, because they have the "liability".

As for the stresses present during overspeed testing, I've always felt that GE has a double standard. For GE steam turbines, they have overspeed testing (executed in the Speedtronic) that automatically ramps speed up in a controlled fashion until the unit trips. All the operator does is select and enable an overspeed test (primary or emergency), and the control system does the rest. Nice and simple and easy for everyone involved.

However, in true "If it ain't broke, don't fix it!" fashion, they have continued to have manual overspeed testing in gas turbine control systems that require operators to manually raise speed. This can be tortuously slow on many sites because the operators don't understand what they're doing and don't regularly perform overspeed testing. The result is that the unit is running above rated speed for much longer than is necessary, in some cases <b>MUCH</b> longer they should be. It's a real shame that GE hasn't made overspeed testing the same for both gas- and steam turbines, especially for sites where there are both GE gas- and steam turbines! It just makes operators even more uncomfortable and gives them even more fodder for complaining (as if operators needed any more excuses!).

Lastly, overspeed testing really involves testing the fuel shut-off system in the event of an overspeed condition. Electronic testing using a frequency generator can be done to eliminate the stresses on the machine, but that doesn't really test the fuel shut-off system (dump solenoids, etc.). GE could make overspeed testing really simple and easy to perform without even firing the gas turbine (even steam turbines) if they would devised a scheme that allowed the operators to energize the fuel (and steam) stop valves and have them de-energized when the frequency exceeds the overspeed setpoint. That would make overspeed testing simple (just need a frequency generator!) and completely reduce stresses on the machine.

It's not clear how the insurance companies would view this (after all, it was insurance companies that insisted machines be equipped with mechanical overspeed bolts for years after the electronic overspeed systems were proven reliable), but every journey begins with a single step.

 

The current technique used by safety professionals to determine the testing frequency of safety systems is to perform a study to determine the required safety integrity level (SIL). When the SIL is determined, the redundancy and testing frequency will be defined by the required SIL.

A basic tenet of safety systems is that a control system whose failure could place a demand on a safety system cannot be used for the safety system. So, since the governor is actively controlling the speed and could drive the speed to unsafe levels during a failure, a completely independent device is needed to protect against overspeed. That means the independent overspeed trip system must have separate sensors, logic devices and final elements (fuel stop valve). A Mark V or Mark VI implemented as GE supplied them on new systems meet these requirements.

If a Mark V/VI is installed as a retrofit on an older machine AND the mechanical bolt is retained, the requirements for testing the bolt become blurry. If the retrofit is done just like the new machines (complete separation of overspeed from control - separate probes, etc) the bolt is no longer required, so testing of the bolt is not strictly required to validate its integrity as a safety device. On the other hand, if the installation of the new controls do not provide the complete separation of devices between controls and safety, the bolt operational precision must be periodically demonstrated.

How often should you test? Before any scientific techniques were devised such as those prescribed in IEC-61508 and IEC-61511, the frequency was empirical, once or twice a year.
If it was my turbine, I would make sure the installation complies with the new standards (separation), then lock down the bolt and conduct the testing of the overspeed trip system at lower speed or with a frequency generator. Most overspeed test accidents happen during the testing of the bolt, so eliminating the bolt will make your turbine much safer.