Voltage regulator in power station is normally a composite feedforward and feedback regulator. This regulator has the best performace and stability at this application. Is there any theory for composite feedforward and feedback regulators? What's the difference between pure feedforward and pure feedback regulator?
A purely feedforward regulator is allocated in the feedforward loop prior to the controlled plant.
A purely feedback one is allocated in the feedback loop.
But if the control is a "regulator" ( typically, the external reference is zero) and the plant G and controller K are linear then the closed - loop characteristic equations in both cases are identical since the feedback error is(minus ) the output. The closed-loop transfer functions have the same characteristic equation but they are not identical so that there are some different properties, signals etc. (although not different stability properties in the purely theoretical case ) for tracking of arbitrary nonzero reference signals:
KG/(1+KG)- feed-forward versus G/(1+KG) - feedback
To my mind, the best explanations always come from a examples. Later, they could be generalized. So, my example would be a simple power converter, either DC/DC or DC/AC. A classic feedback control assumes a stiff source, such as DC bus voltage in voltage-source DC/AC inverters or DC/DC converters, for example. Consequently, it counts on the source to be a constant part (gain) of the overall loop-gain transfer function of the plant-controller system. However, when this input/source is not so stiff, a compensation of the gain is necessary to avoid or suppress a dynamic change of the system loop-gain transfer function due to dynamic change of the source. This compensation is provided by a feedforward loop, commonly placed as the inner-most control loop, although it can be placed as a cascading component to th eentire closed loop system if its dynamics is expected to be very slow (approx. 5x slower than any expected dynamic change in the system, defined by its time constant). One example of such source dynamics would be a response to sudden, steep load change, which would create current inrush of such a magnitude that will cause DC bus voltage to decrease. Another example would be a sudden motor brake, which would cause the opposite effect. The third example would be a predictive control based on look-up tables created from historic behavior data. The output of this look-up table could go to the correction of any signal or parameter of the system, including command/reference point and feedback signals.
Good reading references would be:
"Fundamentals of Power Electronics," by R. W. Ericson and D. Maksimovic,
Wikipedia references from http://en.wikipedia.org/wiki/Feed_forward_(control),
free pdf downloads from the internet, such as university course notes from Stanford or MIT (http://web.stanford.edu/class/archive/ee/ee392m/ee392m.1034/Lecture5_Feedfrwrd.pdf and http://ocw.mit.edu/courses/chemical-engineering/10-450-process-dynamics-operations-and-control-spring-2006/lecture-notes/8_cascade_and_ff.pdf, respectively).
Dear @ Zoran, thank you for your choice of examples and the link of university course! With kind regards Christian
Generally, feedforwad control have faster response than feedback control because in ffedforward control we are measuring process variables before control elements. Their response is different because the zeros of the transfer function are different in both case but closed loop poles are identical.
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Nucleic Acids
- RNA