This paper presents an approach to calculating the solution to the Power Flow Problem (PFP) in ill-conditioned large-scale system models. The proposed procedure is based on the introduction of a conditioning step applied to the initial estimate attributed to the PFC.
This procedure is performed by adding a low-norm diagonal matrix (small disturbance) to the Jacobian matrix of the PFC. The result obtained from this step is used in subsequent iterations of an iterative method. The methodology was investigated considering the methods of Newton-Raphson (NR), Heun-King-Werner (HKW), and some of their variants. The particularity is
that when an estimate of type flat start is used, the NR method does not converge without the proposed conditioning step. Also, the HKW approach fails or the user has difficulty in adjusting its parameters. Simulations performed in four ill-conditioned large-scale power systems models demonstrate the methodology’s effectiveness.