@GabrielKS @jd-lara The bug in the Q redistribution was that the units was hitting limits in iterative process of the remaining Q, but it was not correctly substracted to the residual in that case. This was fixed by replacing reallocated_q += q_frac by reallocated_q += q_set_point - current_q and properly checking the condition of the limits by using q_frac + current_q rather than the clamped value.

Regarding the tests: Tests are failing for the net_flows with the update of #41.

@GabrielKS @jd-lara Let me know if we are OK with adding the symmetrical and I will update the tests to change the zeros for negative values if needed, or going back to only having one direction and the other zero.

Regarding the tests: Tests are failing for the net_flows with the update of #41.

@GabrielKS @jd-lara Let me know if we are OK with adding the symmetrical and I will update the tests to change the zeros for negative values if needed, or going back to only having one direction and the other zero.

I like the new behavior you have implemented.

@GabrielKS It is ready for review.

There is now one kwarg when calling solve_ac_powerflow!: max_redistribution_iterations that controls the max number of redistribution iterations