Published in IET Generation, Transmission & Distribution Received on 13th January 2011 Revised on 2nd September 2011 doi: 10.1049/iet-gtd.2011.0429 ISSN 1751-8687 New network sensitivity-based approach for real-time complex power flow calculation W.-T. Huang K.-C. Yao Department of Industrial Education and Technology, National Changhua University of Education, No. 2, Shida Road, Changhua 500, Taiwan E-mail: vichuang@cc.ncue.edu.tw Abstract: This study proposes a novel network sensitivity-based approach to solving complex power flow calculation problems in real time. A new sensitivity factor, named Jacobian-based distribution factor (JBDF), is used for the calculation of active and reactive power flow in transmission systems. It is derived from the Jacobian matrix of the base case Newton– Raphson power flow solution, and kept constant during real-time line flow calculation. Unlike well-known distribution factors, such as generation shift distribution factor (GSDF), generalised generation shift distribution factor (GGDF) and Z-bus distribution factor (ZBD), this approach reflects changes in complex injection power. Changes in load conditions from base case loads, with either conforming or non-conforming changes in complex power in each bus, can be used to rapidly compute active and reactive power flow without iterations. The proposed approach was tested on IEEE 14-Bus and 30-Bus systems. Numerical results demonstrate that the proposed approach is not only superior to previous distribution factors, but also compares favourably with the Newton– Raphson power flow method. It is well suited to real-time applications in steady-state security control and optimal dispatch. 1 Introduction In AC power systems, power flow analysis is vitally important in the planning and operation stages. Particularly in the operation stage, power flow solutions including active and reactive power flows in transmission lines, together with bus voltage profiles, indicate the present system state. Reactive power plays an important role in power systems, in maintaining bus voltages within specified limits. As for prior contingencies, drops in voltage related to reactive power contributed to blackouts in the western USA in 1996 and France in 1978, and significant voltage swings because of reactive power in the mid-west and northeast USA in 2003. In addition, the major purpose of reactive power dispatch is to improve voltage profiles and minimise real power transmission loss, while satisfying unit and system constraints. In addition, active power balance is the dominant factor in maintaining the system in a stable state. Active power generation changes according to load demands and system losses. This involves the most important information with which operators grasp present active and reactive power flow in power system, representing the basis of real-time economic dispatch, security assessment and contingency analysis. Computation speed and the accuracy of solutions are key problems in ensuring that power systems are operating under secure conditions. Until now, the rapidity and accuracy of computation in power flow studies have been the goals of researchers of such system. However, full Newton power flow for real-time line flow calculation is computationally expensive. This study proposes a fast, reliable method, comprising the Newton – Raphson algorithm and Jacobian- based distribution factor (JBDF) for line flow calculation in real-time power system applications to speed up computation. Conventionally, line flow has been calculated by executing an AC power flow program. The power flow program can be modelled by a set of non-linear simultaneous equations, approached by numerical iterative methods. Well-known approaches include the Gauss–Seidel [1], the Newton– Raphson [1] and the Fast Decoupled method [2]. Owing to the many iterations required to converge, the Gauss–Seidel method is not suitable for real-time applications. The Newton–Raphson approach can solve the problem with few iterations, but convergence is always a problem, because of its dependence on initial values. Executing the Gauss – Seidel program for several iterations to identify suitable initial values, and then performing the Newton – Raphson program is a feasible approach in some applications involving complex power systems. Stott and Alsac [2] proposed the fast decoupled method to overcome the computational burden of the power flow problem. Based on the concepts of DC power flow, this approach solves simultaneous power flow equations by B ′ and B ′′ constant matrices, instead of the Jacobian matrix of the Newton method. Because there is no need for complex computation in the formation of the Jacobian matrix, it rapidly solves power flow equations; however, it still requires more iterations than the Newton method. In recent studies, specialised algorithms have been proposed to enhance robustness and efficiency to improve the problem of IET Gener. Transm. Distrib., 2012, Vol. 6, Iss. 2, pp. 109–120 109 doi: 10.1049/iet-gtd.2011.0429 & The Institution of Engineering and Technology 2012 www.ietdl.org