# Increment/decrement rate of TNR in Speedtronic MARK IV control system

P

#### Pirate

I'm working in a gas turbine power plant, GT6541, control System Speedtronic MARK IV.

We supplied load through national grid. Our grid frequency varies from 51.25 Hz to 48.75 Hz. During peak hour frequency varies drastically. We run the turbine in preselected mode. In preselected mode, MARK IV controls the TNR maintaining a constant difference between TNR and TNH, increases TNR when frequency increases and decreases when frequency decreases.

The problem is that, when frequency decreases drastically Mark IV cannot decrease TNR drastically, as a result load increases above the preselected load. Our turbine is little bit older and we cannot produce more than 33 MW in rated rpm, 50 Hz frequency. And we cannot produce more than 29 MW when frequency is less than 48.9 Hz. Suppose, we preselected the load in 32 MW. When frequency decreases from 51.2 Hz to 50.5 Hz load increases to above 34 MW and then machine goes to temperature control mode. The further drastic frequency drop may cause the turbine trip due to exhaust over temperature. When our frequency decreases drastically and turbine reaches to temperature control mode then we decrease TNR manually by pressing the manual frequency decrease button.

Now I want to know, Can we change the decrement rate of TNR for only Turbine operating at Temperature Control mode? Is the decrement rate available in control constant?

C

#### CSA

Gas turbines don't behave very well during frequency excursions if they are at or near Base Load when the frequency changes, or if they are being operated on Pre-Selected Load Control during frequency changes.

The power output of a single-shaft heavy duty gas turbine is a function of the air flow through the machine. (Gas turbines are mass-flow machines.)

The speed of a single-shaft heavy duty gas turbine driving a generator is a function of the grid frequency. When the grid frequency is less than 100% of rated, the air (mass) flow through the machine is less than rated. This means the power output will be less than rated.

If the turbine is being operated at or near Base Load and the frequency drops, then the air flow through the machine will drop because the compressor speed will drop and that means if the fuel were held constant that the exhaust temperature would increase excessively. So, the turbine control system must reduce fuel as compressor speed is reduced so as not to exceed the rated firing temperature. Which means that load will decrease.

But, there's very little that can be done (short of over-firing the unit!) to make a gas turbines' power output increase during a frequency decrease when at Base Load.

And the opposite is true when the frequency increases when at Base Load. The air flow through the compressor increases and so more fuel can be burned for the same firing temperature so load increases, which is exactly the opposite of what you want to have happen.

Consider a single-shaft heavy duty gas turbine operating at Part Load <b>without</b> Pre-Selected Load Control enabled when the frequency dropped. Droop Speed Control would increase the fuel to increase the load to try to support the grid. It doesn't change TNR to do this; because TNH changes the error between TNR and TNH changes and load changes to try to support the grid.

If Pre-Selected Load Control <b>is</b> enabled when frequency drops, as load increases when frequency drops, the load control will cause TNR to be reduced to decrease load, which is exactly the <b>opposite</b> of what should happen during a frequency decrease!

When a unit is at Part Load on Pre-Selected Load Control and the frequency increases, the load tends to decrease but Pre-Selected Load Control increases TNR to try to keep load equal to the reference which is not what should be happening during a frequency increase.

If a unit were at Part Load without Pre-Selected Load Control enabled and the frequency increased, Droop Speed Control would respond by decreasing load to try to help the grid to reduce frequency.

GE now sell an option called "Primary Frequency Response" (an absolutely <b>horrible</b> name!) to overcome this little problem with the way Pre-Selected Load Control behaves during frequency excursions. (Imagine, paying to fix a problem of the manufacturer's making.)

So, if you're operating your unit at Pre-Selected Load Control at or near Base Load and the grid frequency is changing as much as you say it is, you probably should not be on Pre-Selected Load control as you are contributing to the grid frequency disturbance because the normal Droop Speed Control response is being over-ridden by the Pre-Selected Load Control feature.

I don't think you want to be trying to control a steady load when the grid frequency is unstable. You should be operating at Part Load without Pre-Selected Load Control enabled to try to support the grid. If you are trying to control a steady load when the grid frequency is changing, you are NOT supporting the grid and are likely contributing to the grid instability.

P

#### Pirate

Thanks for you post...

> You should be operating at Part Load without Pre-Selected Load Control enabled to try to support the grid. <

We Operated our turbine at Part load without Pre-selected load and got some horrible experiences. In Part load, synchronized with grid, TNR doesn't changes only TNH changes. When grid frequency increases the error between TNR and TNH decreases and hence FRSN decreases which decreases power output.

FSRN= (TNR-TNH)*FSKRN2+FSKRN1

We took some experimental observation and found that for unit frequency change our load changes for 16-17 MW and our grid frequency is very unstable varies within 2.50 Hz(approx). So with the grid frequency response at Part load our machine load can be vary from 0 MW to (2.5*16=)40MW. But our turbine rated load is 33-34 MW.

And we have experienced a trip for back power when turbine was operating at Part load and frequency increasing. And very often turbine gone to Temperature Control Mode when frequency was decreasing. In order to maintain a steady load in Part load mode we changed the TNR manually by pressing manual frequency increase or decrease button.

So its not possible for us to support the gird with such a small gas turbine.

Now will you please let me know how the "Primary Frequency Response" works and will it be enough to solve our problem as we want to maintain a constant load and want to protect turbine from Temperature Control Mode?

If "Primary Frequency Response" fails to solve my problem then I want to know the TNR decrement rate of speedtronic for only Temperature control mode. Here I wanna add that when turbine goes to Temperature Control Mode, the normal decrement rate is very slow and exhaust temperature goes further as our frequency reduces drastically. In that time we press the manual frequency decrease button and then TNR decrement rate increases, we observed, and we can retrieve turbine to Preselected load or Part load. If we can increase the TNR decrement rate for only Temperature Control Mode, I think our problem can be solved. Is my idea logical or possible? Or what might be the best solution?

C

#### CSA

I would be extremely surprised to learn that the operators of the national grid to which you are connected would want your unit to be operating at a constant load during grid frequency excursions. Grid operators rely on units operating in Droop Speed Control to respond to changes in frequency to support the grid.

Droop Speed Control has two major functions: To allow generator prime movers connected in parallel to operate stably with one another, and, second, to allow units to respond to changes in grid frequency by automatically adjusting their power output to try to maintain grid frequency.

You can think of grid frequency and Droop Speed Control in the following manner. Consider a tandem bicycle with two riders each pedaling the cycle and they are instructed to maintain a constant speed while riding on a level plain. Now if one rider increases the torque he is providing and the other rider does nothing to change the torque he is providing then the speed of the bicycle will increase above the desired speed. Conversely, if one rider reduces his torque output and the other rider does nothing to change his torque output then the speed of the bicycle will decrease below the desired speed.

Now consider what happens when the bicycle has to go up a hill, requiring more torque to maintain the same speed. If one rider increases his torque output to his maximum ability and the other rider does nothing to change his torque output the bicycle might slow below the desired speed. If both riders increase their torque output to their maximums then the bicycle might travel faster than the desired speed. If the two riders don't coordinate their outputs under such a condition, then the bicycle speed will not be constant and will likely vary greatly.

When a grid's frequency decreases, it's because the prime movers can't supply enough torque to keep the generator rotors spinning at the desired speed (frequency). Droop Speed Control will sense the change in unit speed and adjust fuel to try to maintain grid frequency.

If any unit does not respond to grid frequency changes to try to help maintain grid frequency by picking up load when grid frequency decreases or by dropping load when grid frequency increases then it's not helping to support the grid. If it just tries to maintain it's load, it's still "along for the ride" because it can't go any faster or slower than the grid, but it's not helping the grid.

The horribly named 'Primary Frequency Response' is not going to help you.

When grids experience these kinds of disturbances it's usually because there aren't enough units operating at Part Load without Pre-Selected Load Control enabled to respond to drastic increases or decreases in load. In other words, most of the prime movers are running at or very near rated power output.

What's worse, if there are a lot of single-shaft gas turbines driving generators connected to a grid that's experiencing frequency fluctuations, they can exacerbate the problem unless they have special controls such as are mandated in some parts of the world (like the UK). As you have experienced, when the grid frequency is low, your "Base Load" power output is less than rated--which is exactly the opposite of what should be happening. Understandable, but not desirable. Any prime mover, actually, will struggle to increase it's power output above rated when the grid frequency is dropping, but if there are enough prime movers operating at part load to take the additional load without reaching "Base Load" then the grid frequency won't drop by much.

There is no TNR decrement rate for exhaust temperature control. It's conceivable that you might be able to program the Mark IV to do what you want. But, there are two problems: First, there aren't many people left in the world who can program Mark IVs, and second, the Mark IV had limited memory and a very crude set of programming functions. So finding someone to do what you want, being able to pay them what they'll want for the work, and having sufficient memory and programming "blocks" to do what would be required are all unknowns and stumbling blocks.

You're really between a rock and a hard place, as they say. When you're providing power to a grid that's experiencing the kinds of fluctuations you're talking about it is very frustrating.

I would also be very surprised to learn that you are able to control the speed of your turbine when grid frequency is changing by changing TNR. I think all you're doing is changing the fuel flow to stabilize the load, and frequency is still pretty unstable. As said earlier, there are "soft" spots in grids (very difficult to describe all the factors and possibilities) in which individual units can seem to have more stability than others but it's kind of an illusion and the power system dynamics are very unstable unders these conditions.

I don't know how to tell you to proceed. I don't have any experience with these kinds of frequency excursions, and I also wonder if there are other factors at work here we don't know about. Like, what kind of fuel is the unit burning? If it's liquid fuel, what is happening to the liquid fuel forwarding supply pressure during these load swings? Is the L.O. clean and are the servo-valves operating and responding properly? You seemed to indicate that these load swings have gotten worse "recently" and there are a lot of things which could contribute to problems which might be entirely related to grid frequency disturbances, though they are probaly the "exciting" factor to the problem(s).

So, as with all troubleshooting, the first question to be asked is, "What has changed?" And, one also has to remember that sometimes "a" problem is really multiple problems all acting together to seem as one problem, excited by one condition.

P

#### Pirate

Many many thanks for your explanation.....

But I'm sorry that I failed to clarify my thoughts....

> I would also be very surprised to learn that you are able to control the speed of your turbine when grid frequency is changing by changing TNR. <

I also know that its not possible to change turbine speed when synchronized with grid.

Our turbine runs on gas... Still now we didn't face any problem with servo valve.

Thank you...

M

#### MOB

Well that was a great piece of discussion. Some really hard brain storming to clear the facts. It is not very uncommon to have heard of load swings during frequency excursions but what was interesting was the understanding the article provided on the part load operational behaviour vis a vis the load select function.

Frequency excursions to the tune mentioned in the beginning is fairly unheard of but considering the fact that the happenings are facts since have been witnessed, the augmentation of load during a decremented frequency condition is relatively common for the simple reason that actual turbine speed being first affected lowers before reference can be reduced.

Owing to this there is an increase in the difference of the above (I am talking of units closer to rated or on rated load). The difference simply put, a higher than the allowable droop (for which the machine is designed) would obviously mean an overshoot (of real power exceeding rated). So the first thing to happen is the unit going for high exhaust temp. condition. If the reduction in frequency is reasonably slow (no figures to support) then the high exhaust condition can be avoided since then the reference speed would move in tandem with the actual and the speed control would swiftly reduce the load without any temperature overshoot.

One more thing, is the manual reduction of reference speed, what is being done is the overriding of the auto load decrement/increment function (every unit would be having a defined value of MW/Sec). By manual intervention, the manual loading is being called for which is several times faster (around eight times) for the 6541 type of unit, that you have mentioned (check TNKR* values). Check L83JDn values, n=2, 3 & 4. For your query on decremented rate change, check TNKR1_2. Also check the logic L3TNRERRX ('DWATT TOO LOW TO SUPPORT TNR, TNR LOWER').

Hope you would be able to gather some information from the above.