There is two gas turbine running in parallel (island mode) to supply process plant which is not connected to external grid.
In case, one running machine trips, PLC based block load shedding scheme acts, which sense the trip and accordingly cuts the downstream load (calculated amount) instantly.
The calculation is simple,
say GT-1 and GT-2 is running with 10MW and 5MW load, if GT-2 trips, plc calculates (5MW-3.5MW), i.e1.5 MW load to be shedded.
3.5MW is the step load capacity of GT-1.
My question is, how this 3.5MW is calculated. There is no documentary calculation available with us how 3.5MW is calculated.
The machines are frame-V and Frame-3 GE make gas turbine with mark-VIe speedotronic control.
I am trying to calculate the ramp rate 3.5MW/Sec.
Please help me to calculate the ramp load capacity from gas turbine manual/mark-VIe manual.
Manuals as well as system is available to check.

 

barindra75,

Where did you read the Frame 5 step load capacity is 3.5 MW? 3.5 MW per minute? 3.5 MW per second?

In general, most GE-design B/E heavy duty gas turbines (Frame 5s are considered to be in this group of gas turbines) have a manual loading rate of 30 seconds from no load to full load and from full load to zero load (when being unloaded manually). Tht loading rate is also typically used for what's called "FAST LOAD" (and fast unload) in automatic operation. I'm going to make even another distinction that the units I.'m referring to have conventional combustors (and not DLN combustors). Where do I get that rate from? It's the loading rate Control Constant for Manual Loading/Unloading and Fast Load--one of an array of rates that are used to control how fast the unit is loaded or unloaded. (Sorry; I don't recall the name of the array off the top of my head at this writing.)

But, someone must have been given this 3.5 MW step load maximum value by someone from GE--or some consultant.

THIS is but ONE of many problems when a third-party power/load control scheme (often called a PMS, or Power Management System) is used to control multiple prime movers operating in Droop Speed Control mode when supplying a load or loads independent of a large grid (this is usually called "island modle"--operating independent of a large grid with one or two or three or four prime movers driving generators).

The PROPER way to handle this--and by proper I mean for conventional combustor-equipped units--is to operate one unit in Isochronous Speed Control mode and the other unit(s) in Droop Speed Control mode. Automatic load-sharing/control schemes can be used to "balance" loads as they change, but typically the people programming them are HORRIBLE at doing this, and so they say that operating all prime movers and generators in Droop Speed Control mode is "better"--which means it's easier for them get away from the site after commissioning without having to properly tune the power/load control system.

A single unit operating in Isochronous Speed Control will respond to load changes, even fast load changes, and do so very well--AS LONG AS THE LOAD ON THE ISOCH UNIT ISN'T ALLOWED TO BE TOO CLOSE TO ZERO OR TOO CLOSE TO BASE LOAD. And, to do that--one has to adjust the loads of the DROOP unit(s) to achieve that. BUT, most operators ARE NOT properly trained to understand that and neither are the power/load control system programmers (they're usually pretty smart programmers for things like warehousing and cardboard box-making and chicken processing--but NOT when it comes to power generation control).

Get the programmers/suppliers/commissioning personnel who provided/commissioned the power/load control system BACK to site and work with them to properly balance load. The preferred way would be to put one turbine-generator in Isoch and let it respond to load swings while using the Droop unit(s) to keep the Isoch load away from zero and Base Load.

But, good luck with that. There's a LOT of people who just don't understand island load (frequency) control and are REALLY BAD at Isoch operation, and aren't even very good at Droop Speed Control operation of multiple units. With Droop Speed Control enabled and active, units will have a pre-programmed rate for loading and unloading--and if load swings are large it's going to be common for frequency to be abnormal when this happens. With one unit operating in Isoch it WILL RESPOND VERY QUICKLY to load (frequency) changes and do a much better job of keeping the island frequency closer to nominal. All the power/load control system has to do is try to anticipate possible load swings and keep the Isoch unit load from ever going less than zero or above Base Load. And, that should be possible after analyzing data from islanded operation.

Any other mode of operation is going to be fraught with problems--and guess what's going to be blamed? Why, the Mark* , of course!!! When it's really the lack of experience and understanding of the operators, owners, technicians and suppliers/programmers and commissioning personnel of the power/load control system.

Full stop.

Period.

Hope this helps! (But pretty sure it won't...)

[Get the idea that I don't think the problem is the "maximum GT step load" value? Good. Because I don't--think the problem is the "maximum GT step load value." The way the units are probably being operated, changing the maximum GT step load value isn't going to make the island operation more stable/reliable.]

 

Hi brinda75,

In my experience the target of 3.5MW will have been set based on a series of factors - not just the generation;
* Allowable frequency drop on a load shed, i.e.; final grid frequency can you tolerate 49.5Hz? 49.0Hz? (We are 50Hz)
* Actual speed of frequency decay - this is a function of grid inertia - i.e.; rotating load Vs resistive load, gas turbine generator inertia (aero derivative = not so good = avoid on island grids), steam turbines = always good, large DOL motors will regenerate etc
* PMS response time and load block resolution - i.e.; what is instantly? 0ms, 50ms, 150ms? Do you trigger the PMS from the trip Generator CB trip coil energising? From the CB status? Triggering from frequency drop is too late.
* What size load blocks can you shed - 100kW, 1MW? Do you need to separate load blocks to give you more loads that can tolerate a short interruption supply?
* Also considered is load block power recovery time (i.e.; pump set might take 10s, a reformer may need to be relit) and production impact $$$

Most importantly I encourage you to analyse ALL load sheds, use a high speed recording of grid frequency, circuit breaker operation, recovery times, production outcomes, check the math - did you over-shed (frequency barely moved - you could have reduced your load shed impact), under-shed (did under frequency become a factor?), what was the recovery time, consequential damage, production losses, do you have enough automation?

A good load shed is all about minimising production loss;
* Detect the generator trip, grid condition (inertia), remaining generation capacity and type - i.e.; steam turbine is quick to load, gas turbine maybe not, and determine resultant size of load shed
* Assess the production state - where is your power going to have to come from after the load shed, and what do you need to maintain (heaters? boilers? compressors, reformers? etc)
* Identify the path to best outcome - which load blocks must be maintained?
* Execute the load shed
* Re-assess the generator status and grid status - allow grid time to settle (2-3s) - re-establish grid frequency, does the remaining generation respond adequately (instantly change to isoc etc)?
* Grid is now stable in the best possible state
* Initiate backup power options and emergency power protocols (dominos can be a problem)
* Auto reclose load block circuit breakers (one at a time, nominally 1s apart) and auto restart critical utilities and plant equipment
* Plant is now stable in the best possible state
* Wait for CRO to catch up - sometimes they don't even notice ...

Write yourself a report of each load shed event and learn from each load shed event.
Feed the results back into the "determine resultant size of load shed" calculation.
Happy load shedding.

Regards, PB