In this paper the Han-Powell algorithm is proposed as a means for optimizing the reactive power generation in a large scale system. The algorithm is based upon the iterative resolution of quadratic programming subproblems. Other approaches based upon recursive resolution of linear programs seem not to be suitable for the strongly non linear characteristics of the problem.In the adopted model, a reduced and compact problem is solved with only inequality constraints.The equality constraints (load flow equations) are utilized to eliminate a large number of variables (dependent variables). In this way the optimization model is expressed only in terms of the control variables (reactive power generations or generator terminal voltages and LTC transformer tap positions). This approach allows ns to perform the reactive power scheduling of the Italian electric production and transmission system (with about 500 nodes) with reasonable computer requirements (CPU time and central core). The main advantages of the Han-Powell method are:-it is robust and reliable-it shows a sure and fast convergence-it is easy to implement provided a good quadratic programming subroutine is available.
Read full abstract