Algorithmic Aspects of a Commercial-Grade Distribution System Load Flow Engine

The increasing penetration of PV requires more and more load flow calculations to assess the corresponding impact on planning and operation of distribution systems, e.g., during high-resolution quasi-static time-series calculations. There is therefore a strong incentive to optimize the speed of distribution load flow engines as much as possible without compromising on solution accuracy. This paper details the structure of a commercial-grade topology-independent Newton-Raphson distribution load flow engine, the reasoning behind some of its design choices, and recent improvements made to three rarely discussed aspects: initialization, calculation and discretization of tap changer positions, and solution of linear equations. Computer studies demonstrate the reduction in computational time on various distribution systems, allowing for faster interconnection studies with high PV penetration.