I have a question regarding the isochronous load sharing operation of synchronous generators. I very much appreciate your comments and feedback.

Assume we have three generators with the rated powers of Pmax1, Pmax2 and Pmax3. We want to operate these units in an islanded system in isochronous load sharing mode. In order to preclude fighting between them, we take their corresponding speed settings through load sharing line, we develop an average speed setting and use it as the new speed setpoint for all three generators. In this case, the generators simultaneously go up and down in response to any changes in demand to achieve a fixed frequency in the islanded system. I think this is more or less the mechanism of isochronous load sharing operation. My question is about the relation between loading powers on these units? Somewhere I read that in this mode, each unit shares the demand with the ratio of its rated power to the rated powers of all generators; is this correct and why? For example, if P1, P2 and P3 are the loadings on these three units, what would be the relation between P1, P2, P3 and their rated values? Is the following relation correct?
P1/(P1+P2+P3) = Pmax1/(Pmax1+Pmax2+Pmax3)

Many thanks in advance. I also appreciate any other comments/clarification regarding the general concept of isochronous load sharing mode.

 

dear Sam,

We are working at the plant on the island. We have four units with speedtronic mark4 (f6). isochronous system we're using. Every time the three units are operation, only one unit is selected in isoch. If we choose two or three units. Not sustainable network frequency. Usually the selection of works on 3/4 power. Choice allows us to be able to have a sudden change of 9 MW. Sudden decrease in time (more than 9 MW) Operator response reactions. He keeps a constant load the unit to other units. (The reduction load in other unit). If the reduction is more than one unit (30 MW), it is a loss of load. These actions are depending on the behavior of the empirical the distributed network.

 

mohamad noruzi,

I don't know that this could be called "isochronous load sharing" per se. It's just typical island mode operation, where one unit (the Isoch unit) is used to maintain frequency as load changes and the operators are responsible for making sure the Isoch unit can deal with any expected load changes by increasing or decreasing the load on the Droop units.

The concept of isochronous load sharing seems to mean different things to different people. Or, it is implemented in different ways by different people on different sites. What the original poster is describing is really some kind of PMS (Power Management System) that is trying to control frequency by adjusting the load on several machines all operating in Droop Speed Control. This works reasonably well if the load doesn't increase or decrease by very much, or the load doesn't increase or decrease at a fast rate.

What the original poster seems to be doing is to try to develop a formula or formulae that can be used to monitor the change in frequency (which would correspond to a load change) and then calculate exactly how much to increase the turbine speed setpoints of several units operating in Droop Speed Control to make the exact change to make the frequency return to nominal.

In my experience, this is very futile. For many reasons. Droop Speed Control usually has pre-programmed ramp rates at which the fuel is actually increased or decreased and the PMS doing the frequency control doesn't take this into account. (Isoch Speed Control doesn't have this "lag"; because it's intended to maintain the frequency (speed) very closely it changes fuel as fast as required to keep the frequency as close to nominal as possible.) So, the PMS ends up increasing the setpoints too much because of the "delay" caused by the loading/unloading ramp rates of the Droop Speed Controlled-machines. And, if the load can change by a large amount and/or it can change very quickly then frequency isn't controlled very well and there can be a lot of problems--all of which get blamed on the turbine control systems.

The concept of having one unit adjust its load as necessary (as fast or as slow as required) to maintain frequency just makes sense. The problem comes in when untrained operators and plant managers think they don't have to monitor the load on the Isoch unit and use the Droop machine(s) to keep the load on the Isoch unit in a range the Isoch unit can respond to.

For example, if the Isoch unit is rated at 25 MW and it's operating at 22.5 MW and suddenly the load on the island increases by 4.0 MW, the Isoch unit can't increase its load above 25 MW, so the frequency will decrease until someone loads a Droop Machine by at least 1.5 MW to return the frequency to nominal.

Or, if the load on the Isoch unit is down to 1.5 MW and suddenly the load on the island decreases by 3.0 MW, the Isoch unit cannot decrease it's load to -1.5 MW (well, some prime movers can, but that's not normal), so the frequency will increase until someone reduces the load on a Droop machine by 1.5 MW.

That's what a good PMS should be capable of doing: Adjust the load on Droop machine(s) to keep the load on the Isoch unit in a range the Isoch unit can respond should the load change suddenly or very quickly. And that's where most PMSs fail.

A lot of people just don't want to use Isochronous Speed Control on any machine because they are uncomfortable with it. There are all manner of myths and falsehoods about Isochronous Speed Control (almost as many as there are about what Droop Speed Control is and does) and these same people think they can design a system to control frequency by keeping all the machines in Droop Speed Control. And, if the load on the Island is relatively stable and doesn't change very quickly, it will work reasonably well.

I may be wrong about what the original poster wants or intends to do, but I'm pretty sure the way the island is being operated at your plant, mohamad noruzi, is the way an island should be operated.

Unfortunately, it's NOT the way most islands are operated.

 

sam,

I think what your equation is getting at is that the scheme is implemented based on each unit sharing an equal percentage of its capacity. The equation you gave would determine a bias factor by which to bias the set point to maintain balance between the units.

If p1=p2=p3 and p1max=p2max=p3max, then:

p1bias = (p1max/(p1max+p2max+p3max))/(p1/(p1+p2+p3)) = (1/3)/(1/3)= 1

the same would be true for p2bias and p3bias

But, if (1/2)p1=p2=p3 (p1 is twice the power output as p2 or p3)and p1max=p2max=p3max then:

p1bias=(1/3)/(1/2)=2/3... because the bias is less than 1, this will cause gen1 to produce less power so that p1 decreases. Imagine if the setpoint was 60Hz, and now you applied a bias factor of 2/3 to the set point. Now the set point =40Hz. If gen1 now tries to drive the system to 40Hz, it will reduce its power output in a hurry. It's probably obvious that a control system would never intentionally do this, so there are other factors that determine how it would actually unload itself.

Similar procedures tell us what would happen to gen2 and gen3:
p2bias=p3bias=(1/3)/(1/4)=4/3... because the bias is >1, gen2 and 3 will increase power output.

The biases will continue to be <>1, until the machines are balanced again based on the percentage of their capacity, then the bias will return to 1.

Please note, this is all theoretical. I've created such a system in a lab environment, but the real world is a different animal. You are often subject to certain constraints and other limitations that are not obvious or intuitive when you get into the field. This is where the experience from those such as CSA is invaluable. Textbooks only get you so far. Getting your hands dirty and making mistakes along the way are what make you an expert.

Hope that helps,
nic

 

Mohamad, CSA and nic, thank you all for your comments and feedbacks; much appreciated.

Just as kind reminder to CSA, the machines in the described scenario are all running in isochronous mode not droop.

Sam

 

Just as a kind reminder to Sam:

"Droop Speed Control usually has pre-programmed ramp rates at which the fuel is actually increased or decreased and the PMS doing the frequency control doesn't take this into account. (Isoch Speed Control doesn't have this "lag"; because it's intended to maintain the frequency (speed) very closely it changes fuel as fast as required to keep the frequency as close to nominal as possible.) So, the PMS ends up increasing the setpoints too much because of the "delay" caused by the loading/unloading ramp rates of the Droop Speed Controlled-machines. And, if the load can change by a large amount and/or it can change very quickly then frequency isn't controlled very well and there can be a lot of problems--all of which get blamed on the turbine control systems."

(From my original reply , Dear.)

 

Came into this discussion late. I don't know if it is any help to you but we are running an islanded system in an LNG plant where our 4 X MS6000 machines are using a PMS system (ABB ENMC). Our four machines all run in Droop with PMS doing the frequency control. I would think that this an easier way than running multiple Isoch. machines which is very difficult to control.

 

Aero-derivative units can be loaded and unloaded MUCH faster than heavy-duty units. Most other gas turbines (heavy duty, industrial "Frame" units) usually have slower loading and unloading ramp rates than LM units in Droop Speed Control mode. It's not exactly an apples-to-apples comparison.

However, for more than one unit to be operated in Isochronous Speed Control mode at the same time in island mode the governors of both those machines would have to be de-tuned to prevent oscillations--even if there was some kind of PMS.

Either the Isochronous machine (single machine) or the PMS is going to control frequency. It's just not possible for two or more governors to operate in Isochronous mode at the same time with the same (PID) settings that would be used for operation as a single Isochronous machine.