Kuldeep,

Since your post mentioned mains drop and load variation, I would look, initially, to these items as primary causes of disturbance in your system. Since you have identified them as potential causes, can the controller measure, or derive, them and adapt its output directly due to those measurements instead?

As a suggestion, take the partial derivative of I with respect to V and load (R?) to obtain an algorithm that will get you toward a decent current control as long as the controller can obtain measurements of V and I.

A PID may be neccessary to follow slowly varying disturbances.

If you are dead set on implementing a PID for this purpose, be sure to use as the integrator initial condition, a value correponding to the desired current when nominal values of mains voltage and load exist. The PID will have less 'distance' to travel to settle to a new operating point.

Cheers,

Bruce