Commissioning Speedtronic Mark-VIe

M

Thread Starter

Mustapha

What are the disadvantages of Speedtronic Mark-VIe? and what are the main issues to be followed during commissioning of Gas Turbine with Mark-VIe?
 
> What are the disadvantages of Speedtronic Mark-VIe?

One "disadvantage" of the Mark VIe is that one needs to understand the concept of critical vs. non-critical I/O, particularly with TMR configurations.

Also one needs to understand the concept of "primary" and "emergency" trip circuits.

Lastly, one needs to understand the concept of normal and inverted inputs. It is <b>NOT</b> enough simply to observe a change of state when testing contact input circuits; one needs to be certain that the input is properly indicating the state of the input (in other words, the logic is a "1" when is should be a "1" and a "0" when it should be a "0").

> and what are the main issues to be followed during commissioning of Gas Turbine with Mark-VIe?

Commissioning a Mark VIe control system is <b>NO</b> different than commissioning any other control system.

As with the commissioning of <b>ANY</b> control system, one needs to ensure all of the inputs to and outputs from the control system are working properly, providing the proper indication and performing the proper functions.

As with the commissioning of <b>ANY</b> control system, one needs to ensure that events that occur at a particular point during operation do occur at the proper instance, and that functions (such as droop speed control or exhaust temperature control or overspeed protection or vibration protection, etc.) work properly.

Whether you're commissioning an Allen-Bradley-based PLC or a Speedtronic or a Siemens control system, you need to know what's supposed to happen when, and to ensure that all of the inputs and outputs provide the proper indication and produce the desired results.

So, in addition to understanding how a particular control system handles inputs and outputs, one needs to understand how the process (turbine and driven device and auxiliaries, etc.) and what is supposed to happen when. By understanding how the control system handles inputs and outputs (I/0) and how the process works, one can properly commission any control system.

Speedtronic or otherwise.

The problem with Speedtronic control systems is <b>NOT</b> the control system--it's the people who configure the control system for a particular application. They don't properly document how the process is supposed to work (other than the application code, or sequencing or logic, used to program the control system). They just assume that if someone can interpret the application code (be it in relay-ladder logic format or function block code format) that the reader can intuit how the process is supposed to work.

Reading GE's relay ladder logic and algorithms, or blocks, or big blocks, or function blocks, does take some getting used to. However, at least the blocks have a graphical description (if not a written one), whereas many "blocks" employed by other control system vendors have only a written description of internal operation which may or may not match the actual operation of the block's internals (I've found this to be the case particularly with many Woodward GAP blocks over the years).

Again, the real "problem" with GE Speedtronic control systems isn't the hardware or even the programming and configuration software. It's the lack of description of how functions like droop speed control or exhaust temperature control or IGV exhaust temperature control or DLN control work. If GE had to document the operation of their heavy duty gas turbines the way they had to document the operation of their aircraft engines (for military and civilian aircraft operations) they would be very easy to work on and understand and troubleshoot--with any manufacturer's control system.

If GE documented how the heavy duty gas turbines were supposed to operate from START back down to zero speed, with all the functions in between, we wouldn't be having a lot of the questions and discussions we have on control.com.

Or in any other forum for that matter.
 
P

Process Value

Commissioning a Turbine control system

Commissioning a turbine control system is more or less like commissioning any other DCS system. The most important thing is to know how the control architecture is implemented (in mark VI and mark VIe, the TMR is an example). Typically the commissioning activities start after the plant first trial run. This is when the erection team hands over the plant to the commissioning guys. The commissioning time will depend on how well the erection was done. Have a look at the following post it gives you a brief idea of how commissioning is done in a power plant on the whole.

http://control.com/thread/1298089201

Now back to the speedtronic commissioning. Over the years from experience and training; myself and my team do commissioning/maintenance with a framework , as they say its 90% planning and 10% execution. This is a general commissioning procedure for any turbine control system and can be adapted well for speedtroninc

also. * note - I believe that musthapa is a experienced guy and is looking for specifics into Markvie systems but it is hard to give specific answers for general questions so this is written for anyone who wishes to know how commissioning/maintenance is done. Hope Mustapha will find it useful too. The total commissioning period can be broadly classified into two.

A. Plant Instrumentation and control system commissioning
B. Plant control logic testing and tweaking
 
P

Process Value

Phase 1 - Plant Instrumentation and control system commissioning
This mostly involves with the field instrumentation and the actual controller. Most of the critical systems and some of the non critical systems will be operational, commissioning and testing guys make sure that they work perfectly ( or fake it :p in some cases ).

Step 1 - Divide the turbine control into different subsystems - this is when the plant PID comes into play. The PID's make the job easier and does the subdivision for you. In a Gas Turbine system the different subsystems will look like this.


Scheme for lube oil
Scheme for cooling and sealing air
Scheme for trip oil
Scheme for cooling water
Scheme for starting means
Scheme for liquid fuel
Scheme for GCV/SRV
Scheme for atomising air
Scheme for Hydraulic oil supply
Scheme for IGV
Scheme for flow inlet and exhaust
Scheme for compressor washing
Scheme for control devices
Scheme for Fire Protection


Step 2 - Make a list of all instrumentation points (DI/DO, AI/AO, control actuators, etc) connected to each subsystem - Again its PID to the rescue. PID will have all field instrumentation (and also control room instrumentation) laid out. All you have to do is to segregate it into DI/DO, AI/AO, etc. A sample list for lube oil is shown below

Scheme for Lube oil
Lube oil pressure switch
Lube oil temperate switch
lube oil pressure indicator
Lube oil temperature indicator
Lube oil tank vacuum indicator
Lube oil level indicator
Lube oil low switch
Lube oil high switch etc ....

Step 3 - Make a "protocol sheet"/"checklist" - Once you have all the instrumentation points segregated the next important thing is to make a protocol sheet. A sample protocol sheet "row" for a pressure switch is given below.

SLNo Field ID Control system Tag Field Inspection connection health Loop health Point health Remarks

Here Field ID - Field ID as given in PID
control system tag - point tag corresponding to the field instrument as in the control system
Field inspection - visually inspect the filed instrument for any damage and tick if it is ok
Connection health - this is optional for new commissioning but important for maintenance. Most of the problems are because of bad wiring. Most if not all instrumentation have a standard connection specification. Checking the health of the connection wires though important is often time consuming and people skip it. Once again this is recommended.
Loop health - This means that you test the control loop till the control system without the field instrumentation. ie if a digital point you short the potential free contacts or if analog inject a 4-20 ma signal , or if you are using a field bus device , you can do this without disconnecting the instrument with the help of fieldbus testing devices.

Point health - This the over all point health , ie you test the loop with instrument. sometimes it is possible to test the instrument on the field itself. like a pressure switch where compressed air is injected till it turns on or a limit switch where all you have to do is to is to push the small switch actuator, but in some cases it is not possible , like a temperature switch which needs a external heat bath calibration. In either case you declare this ok if either of the above is done depending on which point you are testing.

remarks - Boast you success of having tested a difficult loop here :)

I gave a sample for Digital input, the same can be applied to analog input. In output points you do the same expect that instead of looking at the control panel for the input to come , you force the signal in the control room and look in the field if it is working or not. the rest is the same. for actuators , the process is essentially the same , you can consider a an actuator as a output with a feed back input so you check both at the same time. for example in a servo you force a current ( output point) and see the servo movement (an input point) etc.

by the end of this step , you will be having a lot of lists. if there are 10 subsystems and each subsystem has at least one DI , DO , AI , AO and actuator you will be having a total of 50 separate sheets. This seems like a lot of work but believe me this is what is going to save you in the end.

step 4 - execution - Yes once you have completed the first three steps and have gathered a good group who are willing to do whatever it takes to get the job done ;) you are ready to proceed to execution of the plans. as i said you have 50 sheets left to fill up , proceed sheet by sheet. If you have the liberty of a lot of manpower most of it can be done concurrently. at the end of the execution of the 50 blank sheet you started with , thumb rule is at least 90% should have "success / checked ok" marked in the remarks column ;)
 
P

Process Value

Phase 2 - Plant control logic testing and tweaking

by the end of phase 1 , you have almost if not all the plant instrumentation working , feeding picture perfect data into your control system to munch on. This phase is to make sure that all the control philosophy is implemented correctly.

Step 1 - Divide the control logics into different subsystems - Again in a well designed system the Field PID subdivision and control logic subdivision will be the same. And ideally each subsystem should be independent of one another with any dependencies explicitly defined. This is where GE systems comes crashing down on you like a boulder off a cliff. They do a reasonably good job of dividing the control system into different subsystems , but the dependencies are not well defined and a single point will be used in a 100 different places that it becomes hard to keep track on. But commissioning engineers do not have
the luxury to revamp the system , you just have to live with it and make it work. Which leads us to the second point

Step 2 - " STUDY " the control philosophy for each subsystem - Though this is a generalization, i believe that this is where problem starts no one bothers about studying the control philosophy of the equipment they are commissioning. The end result is you do not know why the code is behaving in a way and worse you do not know how to fix it. GE provides a rudimentary documentation of the control philosophy, but plant auxiliaries will differ from site to site and as far as i know not documented anywhere. Yes this is a inherent disadvantage, but taking a look at the code will help. you can also consult site engineers/operators if they have a prior plant running , over the years i have come to know that they have a intuitive wisdom of what is good as they are seeing these machines every day. but the main thing is to know how the machine is supposed to to controlled. GE control can be subdivided into the following

Startup and shutdown control

Core control
flame control
compressor discharge control
Speed and load control
generator load control
Exhaust control
IGV control
FSR control
FSRN control
FSRA control
FSRSU control
FSRSD control
FSRT control
VIbration protection control
steam/water injection control
DLN burner control
DLN fuel mixing control
AUxillary control
Lube oil control
trip oil control
Hydraulic oil control
Starting means control (diesel engine)
Hydraulic ratcheting control
fuel pump control (site specific)
vent fans control

generator and excitation control

Trip and protection logics

Step 3 - Once again make a Logic check "protocol sheet" - Once you know the logic subsystems and the control philosophy behind them , it is time to make a check list. This can be tricky as this is not a straight forward one. for example lube oil control subsystem will have the following important logics

1. AOP start
2. EOP start
3. Lube oil temp high trip
4. Lube oil pressure low trip
5. Lube oil tank low alarm, etc .....

if you have studied the control philosophy well you should be able to differentiate the important logics well. the check list can look like this

Logic Description Signals involved Logic health Modifications needed Remarks

here
Logic description - a brief description of the logic under test
Signals involved - tag Nos and descriptions of the important signals involved in the logic
Logic health - How well you think the logic is performing up to the control philosophy , Individual testing may not be possible always , it is
numerous to list here all the things that can be done around it , but once you know the control philosophy you can intuitively
figure out how to test the logics. Again testing can be done in many ways. One is pure virtual testing, ie you force signals
and test the output , this is is mostly the first stage. Given that you have already throughted the field signals in the first
phase " Plant Instrumentation and control system commissioning " , it should work. Some times client will ask you to test the logic
with the field instruments , they pay you for it so you better keep them happy :).
Modifications - If you or the client think that the logic is not performing well, jot down a few modifications necessary to make it work
Remarks - Once again boast a success story of a difficult logic or rue over impending modifications : P

By the end of this phase you will again have a lot of papers , assuming 20 main subsystems and 5 main logic per subsystems , you have around 100 logics to test and tweak. But many are tried and tested logics which you will not need to do anything. Once again i stress the importance of keeping this record , when you do this for a few sites you will see that you would have covered most of everything you need to see , as always getting by the first few sites is always the difficult task.

Step 4 - Testing and tweaking - This is an iterative process, but i have seen that most of the turbine systems come with a tried and tested code and you seldom tweak anything, its mostly testing to make sure that it works. The conditions is different for boiler control so it it with the turbine auxiliaries. Take into account what the operating and maintenance personnel have to say , at the end of the day you need a stellar reputation and it only happens when you keep the client / operating personnel happy. Over the years i have seen that most of the request for tweaking does not come to modifying the control code but the graphics in the HMI , so make sure that you not only know the control system but also the HMI part of the controller. The final test in tweaking is always a battery of performance tests. This is again platform specific, usually done to test the redundancy and robustness of the controller. you do not have to make a list of battery tests , it will be given to you (much to your displeasure) by your superiors and it is mostly based on the initial contract.


Phase 3 - Optional , always recommended and usually mandatory "End of Commissioning party "!!! where booze flows in rivers but always ends with you getting home for a good rest but still contemplating weather "l63p21%$#" works ;). sigh..

SO that brings to an end of a crash course in commissioning a turbine control system. hope you guys enjoyed it :)
 
Top