The relation between the Zg which equal to R±jX and the reactive power of the generator is much closed. The reactive power of the generator is calculated from the 

imaginary part of the resultant of voltage multiplied with the current conjugate. The current value is obtained from the voltage and impedance. In general, the resistance in the generator is very small so it can be ignored, and the total impedance is the reactance. So, the reactive power of the generator is depending on the generator operation mode. I mean that the power factor operation type unity, lagging and leading. For more information, please read the power system analysis book which was written by Hadi Saadat, and we found more details.

Complex power generator S_1=E_x I_*. Complex power OUTPUT load S_2=Z_*|I_|^2. Reactive power genepatora Q1=Im(S_1). Reactive power load Q2=Im(S_2)communication is a Q1=Q2.

Mr Valery Frisk

my opinion is not that you think

Very simple. If Q is the reactive power of the generator and X its internal reactance. If V is the line voltage of the grid and E the internal voltage of the generatr with delta the phase shift in between E and V, the reactive power is about:

Q=3.E.V.cos(delta)/X

thanks professor André Capolino

But is there any power factor such az lag and lead that are can leading to reactive power circulation between Generators?

delta is somehow related to power factor cos (phi) and can be larger than 90° or smaller than 90° depending of lagging or leading. Reactive power can be negative or positive depending of the generator power factor. On the other hand, the reactive power Q has also to energized the reactance X.

Although my background is more in the thermodynamics of power systems, such as CBC (Closed Brayton Cycle) systems with nuclear heat sources, I agree with Prof. Capolino's explanation. Of course, to compensate or counter-balance inductive reactance capacitors can be added in series.