Gents,

I would like to know why a standard SPEEDTRONIC MKV simplex\TMR not configurable for LM turbine operations?

 

Paul,

GE Aircraft Engine in Evendale, OH, USA, determines the control systems which are acceptable for the different types of LM units--they are NOT all the same, even within a model line (LM2500; LM5000; LM6000; etc.).

In general, it's the speed of the processors ("scan rate") which determines which control system can be used for which type of LM unit. And, <i>In general,</i> SIMPLEX Speedtronic panels are not used for LM units. Non-LM TMR Mark Vs are acceptable for use on SAC units, but not on DLE units.

If the engine is under warranty by or leased from GE they have the last word about which control system can be used.

All that being said, Mark V Speedtronic control systems--while good control systems--have limitations (the HMIs) and are no longer manufactured and are poorly supported by GE--though there are many after-market suppliers which are very good at supporting and troubleshooting the Mark V. Finding someone to reprogram a Mark V for a different turbine is a different matter though, and would consume many man-months of engineering and testing/simulation and commissioning.

Hope this helps!

 

Hope this helps!

CSA, Thanks for responding.

I wanted to know more specifically about the function of the winSys control module. What is it and its functional role?

From my perspective, reprogramming a MKV to operate a LM turbine does not seem that complex. That's why I am trying to understand the winSys module. Could it be that a program slice specific to a particular LM engine is executed in this module?

What possibly could be these special algorithms that needs this module for its execution.

I understand the scan rates of various processors, but I don't think that's the complete story.

 

Hi,

I would like to share a supplementary information to CSA's replied and to your question.

GE defined how different type LM machine should be controlled, but does not defined that control system should be simplex, dual or TMR. The document is called IDM (Industrial Design Manual). Each type of machine has a dedicated section that includes various specifications. For example, in 'LM2500 Plus SAC' section, there is a document number M50TF3805 defines how it is should be controlled. All the packager who provides control system including GE must strictly follow all requirements.

For 'LM2500 Base SAC' model, the requirement is at minimum (relatively to other machine), the document only describes protection/limits and briefly on control scheme. In fact only protection, fuel control, starting sequences and some sequencing for auxiliaries are handling by control system. VSV (Variable Stator Vane) on this type of machine are mechanically control. Personally, I worked on few machines (LM2500 Base SAC) that controlled by Mark V TMR with SRV and GCV. However, my statement in this paragraph is not valid for 'LM2500 Base DLE' and another model. There is a stringent requirement defined in a dedicated document and it is beyond the capability of an ordinary Mark V to handle.

To cope up with more stringent requirement, GE came up with Mark V LM with TMR configuration. The name is Mark V, but it is totally different animal as CSA mentioned many times in this forum. Even though Mark V LM's speed is much faster than ordinary Mark V, it has its own problem know as sequencer buffer overrun. GE provided Mark VI as a replacement for Mark V LM, Mark V and other old products. Again, scan time of Mark VI was not qualified as per IDM's requirement to operate in TMR configuration. Prior to the release of Mark VIe, majority of LM machine equipped with Mark VI in Simplex configuration or Woodward NetCon/MicroNet due to mentioned reason. During that period, Woodward is one of a few system that passed IDM qualifications to operate in redundant configurations.

For Mark VIe, it passed IDM qualification to operate in redundant configurations. Initially GE always proposed either Mark VIe in Simplex or TMR, it is up to the customer's desired. Later GE added Dual configuration to a proposal for customer. Personally, this seems to be the most sensible configuration. Most of sensors on LM machine are dual sensors unlike Heavy Duty machine that have triple sensors (for example speed sensing devices).

At the earlier state of LM machine with Mark VIe, the critical control/protection loops for the machine itself aka. 'Core Engine', sequencing/protection for auxiliaries, and interface to external devices including to HMI are in one processor. It seems that Mark VIe on LM machine type experienced the same problem with Mark V LM. Later on(around five years ago), GE came up with an improvement on Mark VIe for LM machine. It has two set of processors, one dedicated for 'Core Engine' and the other for auxiliaries and interface. Since that time Mark VIe is the most awesome control system for turbine control (except bug here and there which is now identified and eliminated throughout the upgrade of firmware).

Now, back to ordinary Mark V with 'LM2500 Base SAC' with an upgrade of EPROM and HMI. It can provides interface to OPC Client i.e. PI Historian or DCS by 'GeCssOPC DA Server' running on a WorkstationST resides in HMI. With the new HMI, Cimplicity is almost fully integrated with WorkstationST. With a new WorkstionST/HMI, the alarm management is much easier, better trending capability. So far, everything is looking good to me, except a grouping in Cimplicity graphic files causes a lot of headache for field engineer to make any change.

 

Green_Man, that document shed some light.

As I dig into this last night, I came across a document showing the IO interface on LM machines, and I noticed that the electrical characteristics of the sensors on some LM engines would not interface the ordinary MKV/IO boards. I noted boards of similar function where the signals passes through, and could only conclude its got to be a scaling issue.

Your document also solidified some concerns, one being the sequence buffer overruns,I could see that happening.

I have concluded its the core sensor, electrical characteristics and count that's driving the selection of the winsys module.

 

Green_Man's comments are excellent.

As for WinSys, I can only surmise that it's some kind of MS-Windows service that's required for communication between the HMI and a Mark LM panel.

The Mark V LM panels use a Real-Time OS (RTOS) which is not MS-Windows (it's some flavor of QNIX, I believe).

Sorry I can't be more help; I successfully avoided LMs for most of my career and would like to continue doing so.