GE gas turbine Start/Shutdown sequencing


Have you received the information in the format you requested?

Decades ago GE used to provide such a flow-chart of start-up operation as part of the Control Specification group of drawings. But, that practice ended in the early 1980's, though there is a rumour that the Belfort and Fireze divisions of GE may have started something similar again with turbines produced in their factories or under their guidance.

Flow-charts would be useful for understanding the basics of the sequence of a start-up, or a shutdown--occurring under normal operating conditions. However, many heavy duty gas turbines these days have so many more auxiliaries than they used to 30 or 40 years ago, as well as the capability to burn multiple fuels and may also be equipped with low emissions combustion systems and their auxiliaries. A detailed flow chart would have lots of branches and be pretty difficult to follow in a lot of cases, particularly if the unit were experiencing problems during a start-up or shutdown.

If you are experiencing problems during a start-up or shutdown, please describe the issues you are seeing here--including listing all the Process Alarms (and Diagnostic Alarms) which are being annunciated during the problem situation(s). Try to give as much information as possible about the speed and/or load of the unit, the IGV angle, the fuel being burned, the fuel flow-rate (or FSR, Fuel Stroke-Reference), and fuel supply pressure and stability. We can try to help you understand what may be happening if you provide good information and are willing to get more data when requested.

Also, if your unit has a diesel starting motor please tell us if it uses a hydraulically-operated jaw clutch or a SSS (Synchronizing, Self-Shifting) clutch between the torque converter output shaft and the Accessory Gear Box starting means input shaft (which is directly connected to the turbine shaft via the Accessory Coupling).

Looking forward to hearing back from you!
Hi CSA, many thanks for your prompt and detailed response. In fact I have requested for the Flow chart to understand and help in Engineering and developing the sequencing to replace GE mark-v Control system with DCS by one of the DCS manufacturers (may I have your email ID so that I can share the information on DCS manufacturer as I am not supposed to reveal this information on a public forum). The unit is a Fr-6 gas turbine with Diesel Engine Starting device and Hydraulically operated Jaw Clutch with Self-sequencing Ratchet mechanism, Duel fuel (HSD & Natural Gas) firing. Unit is installed and commissioned in 1990s

Any control system integrator which has agreed to perform this control system replacement MUST have considered they were going to have to learn how to read the Mark V control sequence program. There is a section in the back of the Mark V Application Manual, GEH-6195, that describes how to read control sequence program blocks--Primitive Blocks and Algorithm Blocks. One of the really excellent things about GE's blocks is that they are based on relay ladder logic (which is pretty basic) and they (at least the Mark V blocks) use graphical representations of what happens inside the blocks so that one can follow an input through the block to an output or outputs and see how the values are operated on inside the block. GE calls it "Big Block Language", or BBL, but it's really just a relay ladder diagram-based graphical method of representing operations. There are automatic parameters, and there are passed parameters (see the descriptions in the manual--or ask here if you need help on a particular block).

You're asking for a start-up and a shutdown sequence. Next you'll be asking for a loading sequence, and a description of Droop Speed Control, and how the gas control and liquid fuel bypass control valves operate. Everything the control system integrator who will be programming the DCS-based control system needs to know is in the Mark V control sequence program (also called the CSP). At some point, someone is going to have to give up and learn how to interpret that program to "duplicate" it in the DCS. And, how to incorporate LVDTs and flame sensors and speed pick-ups and vibration sensor inputs to the DCS, and then how to configure bipolar mA outputs to drive the electro-hydraulic servo-valves used on the fuel control valves and the IGVs. And, none of these things are simple matters for most DCS's, because they weren't designed or built for high-speed, rotating equipment control and the types of inputs and outputs used for a very large portion of the gas turbines used in the power generation (or even mechanical drive application) fields. Most control system integrators end up using a long of converters to convert LVDT signals to mA inputs, because the DCS doesn't have input cards capable of working with differential voltage inputs from LVDT outputs--not to mention most DCS's don't have the capability to output 7.0 VAC RMS at approximately 3.0 kHz to power the LVDTs. And, then there's flame sensors (some of which operate at 335 VDC--try finding that on a DCS system), and there are vibration sensors that operate at 100 mV/inch.

Nope; this ain't for the faint of heart. And, while you might find a flow-chart for a particular turbine, you're still going to have to refer to the Mark V CSP for the details of how the turbine being fitted with a new control system has to operate. There's just no way around it. If you have specific questions, ask them here--we are happy to help.

I am personally not a fan of using DCS-based systems for turbine control. GE Mark* turbine control systems are purpose-built control systems--designed to control high-speed, rotating equipment using the types of sensors and devices required to do so safely and without the need for converters and additional equipment. There are a LOT of control system integrators who are caught unaware of the intricacies of turbine control, thinking turbine control is just a "process" that can be automated. To a certain extent that is true--if one knows how a turbine is supposed to operate and be protected then one can use just about any control system to do so, PROVIDED that control system has the ability to either interface with the existing controls devices and instrumentation, the existing controls devices and instrumentation are changed to match the new control system's capabilities, or converters are used to interface the existing controls devices and instrumentation to the new control system's capabilities. The latter is usually what happens, and it can get very, Very, VERY messy. The control system integrator usually fails to write the necessary documentation to allow turbine technicians to be able to perform the extra steps involved in calibration or adjustment because of the use of converters. And, the wiring drawings can also get quite messy, especially when converters have to be changed during commissioning when it's discovered they don't work as expected.

As with any such endeavour, the key is to choose the right control system integrator. One with experience in turbine controls and upgrades--demonstrated experience, not just once or twice, but repeated experience. Too many people choose a turbine control system by the control system equipment manufacturer (preferring a particular brand over others for whatever reason). Many control system integrators have a lot of experience with one or more control systems--but DO NOT have experience with high-speed rotating equipment such as turbines and generators. I, and others, have been to sites where PLCs or DCSs were used to replace Mark* turbine controls with very bad results. It's clear from looking at the logic written and running in the control systems that some very intelligent programmers wrote various bits and pieces of the logic, sometimes over-writing other's work in the process. But, if they don't understand basic power generation fundamentals such as Droop- or Isochronous Speed Control, or modulated IGV control, they can make a mess of the logic and greatly reduce the reliability of the machine.

In some cases (many, actually) control system integrators have developed (or had to develop) proprietary control system input and/or output cards--defeating the major reason for using a PLC or DCS (the ready availability of off-the-shelf parts and cards). And, in some cases, special software is also developed and implemented--which defeats the other reason for using a PLC or DCS for turbine control--the wide use of a particular brand of control software meaning a lot of people know how to troubleshoot it and program it.

When deciding on a supplier (control system integrator), one needs to get references from the potential suppliers--and then do the hard thing: Contact those references and get their impressions and feedback on the knowledge and experience of the supplier in commissioning and supporting the equipment on a high-speed, rotating application (such as a turbine and generator). Failure to do so can be very costly in schedule over-runs and reliability issues. It's not about the equipment being used--it's about the experience and knowledge of the people configuring and programming the equipment being used to control and protect high-speed, rotating equipment used in power generation (or mechanical drive applications (such as compressors or pumps)).

Best of luck with your project. Choose wisely (if you're the purchaser). A LOT depends on doing the "due diligence" when choosing a supplier, and--again--it's not the equipment you should be focusing on, it's the knowledge and experience of the supplier in configuring and commissioning the equipment for your application.

Hope this helps! Again, if you have specific questions, we are probably able to help. Best of luck in your endeavor!

Oh, and make sure the DCS programmer also has copies of the P&IDs for the GE-design Frame 6 heavy duty gas turbine.
To add something which may be very important to understand, each CDB (Control Signal Database) signal in Mark V has a "longname" description associated with (well, almost every one--at least every signal SHOULD have one associated with it). There is a file, LONGNAME.DAT, which should be specific to the unit software (called application software). It can be found on the operator interface, in a folder called something like UNIT1 (for a single unit site). On older operator interfaces, the folder would be on the F: drive; on newer operator interfaces (ones running a more recent version of MS-Windows (such as XP, for example), it will probably be in a folder called SITE, or, sometimes, MASTER/SITE, or something similar.

A longname is an 80-character (or less) ASCII text description that someone thought was appropriate for a signal (it could be longer than 80 characters, but if it is, it gets truncated by most GE software). The thing to remember about longnames is: They are not guaranteed to be 100% correct. I know--that's hard to understand, but, there was never any kind of error-checking software or procedure or review to ensure, 1) that every CDB signal had a longname, and, 2) that every longname was an accurate description. I would estimate that about 98.519% of longnames are appropriate.

When looking at a longname description for the first time, it should be read for "intent." That is, it should be read as, "This longname is intended to be an accurate description of the signal's function." As written above, about 98.5% or so of longnames are true and correct. But, there are times when they are incorrect--and an unfaltering belief that EVERY longname is true and correct and accurate will cause a person to go mad. Literally. Sometimes you just have to look at the context of the signal, and decide for yourself if the longname is good or not. (That takes some experience and practice--admittedly, but any coder already knows that when they are looking at someone else's code these things are going to happen.)

There are also signals which do not have a longname--shouldn't have happened, but it did. Some jobs (units) with a lot of custom, site-specific software seemed to have more than other jobs. Sucks; but it's true.

Last thing to know is: Properly written--and about 97.1634% of CDB signal names are properly written (meaning the name was chosen properly and according to unwritten standards and using a GE-modified version of ANSI device nomenclature)--signal names are very good at describing the signal's function, and if it is a logic signal it usually describes when the signal is a logic "1".

Signal naming has been covered MANY times before on threads. A few years back, there was a very long few threads which went through some start-check and trip rungs and every signal was described in quite some detail, as well as when it would be a logic "1", and when it would be a logic "0". You can use the 'Search' function of to look for those threads; it would be very helpful for someone trying to discern GE's signal naming and usage. VERY useful.

Okay. Hope this helps! You can use your preferred World Wide Web search engine and search for 'ANSI device numbers' or something similar to find .pdf files of ANSI device numbers which are useful in understanding GE logic and programming.
Dear CSA,
I am sorry for the delayed response, which was because I couldn't work on the systems for some medical emergency at home.

Many thanks for your detailed explanation. The systems integrator who is building this control system seem to have done similar retrofits on Steam turbines successfully but on Gas turbine, this is the first time they are venturing. I was a Startup Engineer myself and I have been approached to help them during commissioning. However they also sought my help to understand how to read the CSP, Device Summary and Piping schematics and also the GE Speedtronic control system principles, viz. Startup, Acceleration, Speed/Load, Temperature, Shutdown & Manual controls and the Protection function like Over Speed, Flame Protection, Vibration, Over Temperature etc. which I could help to the extent what I know. When it came to Start up sequencing, I thought a Flow chart would be ideal to give at a glance an outlook of start / stop of various auxiliaries that come in during the startup sequence while coasting to FSNL. Thank you once again.