An Extended Conic Quadratic Program for Load Flow Control

Power flow solutions form the basis for computer-based steady-state network operation and planning. In modern power networks, accurate power flow computations need to account for models of controllable devices such as tap-changing transformers, phase shifters, and unified power flow controllers. This paper presents a new load flow control optimisation framework based on an extension of the conic quadratic programming format. Its objective is to minimise the L1-norm of deviations from prescribed control target values and thus produces feasible solutions even for non-attainable targets. In the proposed framework, the classical non-linear power flow equations are transformed into linear constraints. The nonlinearity which is inherent in the power flow model is transformed into a set of conic quadratic and arctangent functional equalities. This results in a load flow control model that is amenable to straightforward integration in efficient primal-dual interior-point methodologies. Moreover, the linearity of the power flow equations allows the use of scaling techniques for improving the numerical conditioning of the problem. The performance of the proposed method is validated by comparing with previously published Newton-Raphson based power flow results. Computations on a 2383-bus test system are also reported.