This paper presents an approach to solving the load flow problem (LFP) in large-scale electrical power systems, suitable for situations where the system is numerically ill-conditioned and solved in an iterative way. The problem consists of applying a conditioning step to the initial iterate of the LFP and solving it. The investigation uses a technique based on Tikhonov’s regularization for the conditioning stage. The regularization process is only applied to the LFP Jacobian matrix for the first iteration. Hence, the ’partial’ name is used here, considering all the LFP iterations. The technique is implemented for two situations, one of which avoids the product of the Jacobian matrix by its transpose. Through comparative analysis involving the Newton-Raphson (NR) method, alternative perturbation techniques, and the Heun-King Werner (HKW) method, the study highlights the effectiveness of the conditioning step for convergence, particularly in cases where the NR methods and HKW are unable to converge without this step. The proposed technique’s efficacy is demonstrated through simulations performed on five large-scale and ill-conditioned power system models, including a system with more than 109 thousand buses.
