I am trying to perform fuzzy logic control of a DFIG based wind energy conversion system in which for the control of the Rotor side converter I am using Fuzzy logic controllers (whose inputs are active and reactive power errors and the rate of change of these errors.)to generate the reference signals of the rotor direct and
quadratic axis current components These reference currents are then compared to their actual values and again given to FLCs along with their rate of change of error to give reference Vdr* and Vqr*. But I do not know how to choose the scaling factors or gains to normalise the input error, change of error and output. How to decide these scalin factors?
Actually there is no method to decide what the value of gain you should choose. May be trail and error method can be chosen. From my little expert normally the scaling factor gain of the change of error should be chosen to be one tenth of the scaling factor gain of the error.
another way you have to use an optimization such as PSO, SA ...etc, method to tune scaling factor gains, in order to obtain the appropriate desired performance.
Hi Swati,
If you want to re-scale the error signal to a range between 0 and 1 by the normalization technique, the formula is given as follows:
enormal = (e − emin)/(emax − emin).
If you have the defined range of premise universe, having the normalized data on the bound [0, 1] allows you to determine the input gain easily.
In Simulink model if you are using the fuzzy logic controller with rule viewer than by simply observing the error and change in error in the ruleviewer you can set the values of the two input gains. For maximum possible error the value of gain should be such that the input to fuzzy controller remains in the selected range (generally -1 to +1 or 0 to 1) of membership functions.
Dear sir,
the values if the scaling factors can be selected as a design parameters and by using an optimization algorithm such as PSO, GA to find there values for optimal performance index .
- Badges
- Science topic