Hi All,
I am a bit of a newbie on control theory and could do with some advice on an application I am looking at.
The application is for a generator governor controller that runs in speed control in an island network (I.e not droop) and will experience frequent load steps. The problem is that it’s ability to pick up load is nonlinear and depends heavily on its existing load. For example at light load (<25%) it can pickup around 25% of its rating in one step whilst maintaining its frequency limits, but at medium load (25% to 40%) its ability to manage the same step load drops off linearly, and above 40% it drops off very quickly.
Most standard governor models just use a basic PID control with the gain (and other constants) that are tuned to the specific capability of the generator. In this it would work and would end up trying to overstep the capability, soI am wondering what would be the best strategy.
My first thought would be to replace the normal gain factor with a variable gain factor set by a feedback loop monitoring the generator loading. Then I could reduce the gain factor down as the loading increases. I could use either a number of different gain values or even a group of piecewise functions.
Is this approach common? I see it referenced as a gain scheduling controller, in a few textbooks, but I have never used it before. Is it practical? I can see lots of potential problems with tuning and instability. Would this be a form of cascade control? It seems to be , but I am not sure on doesn’t seem like it as, like it meets the criteria.
I am a bit of a newbie on control theory and could do with some advice on an application I am looking at.
The application is for a generator governor controller that runs in speed control in an island network (I.e not droop) and will experience frequent load steps. The problem is that it’s ability to pick up load is nonlinear and depends heavily on its existing load. For example at light load (<25%) it can pickup around 25% of its rating in one step whilst maintaining its frequency limits, but at medium load (25% to 40%) its ability to manage the same step load drops off linearly, and above 40% it drops off very quickly.
Most standard governor models just use a basic PID control with the gain (and other constants) that are tuned to the specific capability of the generator. In this it would work and would end up trying to overstep the capability, soI am wondering what would be the best strategy.
My first thought would be to replace the normal gain factor with a variable gain factor set by a feedback loop monitoring the generator loading. Then I could reduce the gain factor down as the loading increases. I could use either a number of different gain values or even a group of piecewise functions.
Is this approach common? I see it referenced as a gain scheduling controller, in a few textbooks, but I have never used it before. Is it practical? I can see lots of potential problems with tuning and instability. Would this be a form of cascade control? It seems to be , but I am not sure on doesn’t seem like it as, like it meets the criteria.