Factors causing slow convergence in the Newton optimal power flow (OPF
) approach are presented in this paper. The solution speed of the appr
oach is evaluated by the numbers of the main and trial iterations. The
OPF discussed in this paper deals with minimization of the system MW
losses while satisfying system security constraints by adjusting react
ive power controllers. The factors causing slow convergence include (a
) insufficient sources of reactive power, (b) a solution space with a
flat bottom, and (c) the load model. This paper helps to clarify the a
mbiguous phenomena of slow convergence. The main phenomena, explanatio
ns and termination criterion/enhanced treatment are presented for each
factor. The IEEE 30- and 118-bus systems and a practical 217-bus syst
em, namely, the Taipower system, are used to serve as samples in this
paper.