Vectorization implementation of optimal power flow in rectangular form based on interior point method

This paper presents a vectorization implementation of optimal power flow (OPF) with tie line transmission constrains in rectangular form based on primal-dual interior point method (PDIPM). The vectorization of OPF is established by arranging the control variables and state variables according to the variable types. Vectorized formulation is demonstrated in detail. The indefinite coefficient matrix of correction equation is dealt approximately; hence correct equation can be solved by LDL^T decomposition. AMD reordering algorithm is used to reduce the fill-in elements in LDL^T decomposition, which improves the calculation speed considerably. The C++ power system vectorization calculation library and OPF program are developed. This objected-oriented library implements basic linear algebra equivalent to BLAS (Basic Linear Algebra Subprograms) level 1-3. The dynamic sparse formats transfer strategy in this library makes OPF program run much faster. Numerical simulations on test systems ranging in size from 118 to 2168 buses validate the correctness of the proposed method.