Decomposition algorithms for multi-area power system analysis

Abstract

A power system with multiple interconnected areas needs to be operated coordinately for the purposes of the system reliability and economic operation, although each area has its own ISO under the market environment. In consolidation of different areas under a common grid coordinator, analysis of a power system becomes more computationally demanding. Furthermore, the analysis becomes more challenging because each area cannot obtain the network operating or economic data of other areas. This dissertation investigates decomposition algorithms for multi-area power system transfer capability analysis and economic dispatch analysis. All of the proposed algorithms assume that areas do not share their network operating and economic information among themselves, while they are willing to cooperate via a central coordinator for system wide analyses. The first proposed algorithm is based on power transfer distribution factors (PTDFs). A quadratic approximation, developed for the nonlinear PTDFs, is used to update tie-line power flows calculated by Repeated Power Flow (RPF). These tie-line power flows are then treated as injections in the TTC calculation of each area, as the central entity coordinates these results to determine the final system-wide TTC value. The second proposed algorithm is based on REI-type network equivalents. It uses the Continuation Power Flow (CPF) as the computational tool and, thus, the problem of voltage stability is considered in TTC studies. Each area uses REI equivalents of external areas to compute its TTC via the CPF. The choice and updating procedure for the continuation parameter employed by the CPF is implemented in a distributed but coordinated manner. The third proposed algorithm is based on inexact penalty functions. The traditional OPF is treated as the optimization problems with global variables. Quadratic penalty functions are used to relax the compatible constraints between the global variables and the local variables. The solution is proposed to be implemented by using a two-level computational architecture. All of the proposed algorithms are verified by numerical comparisons between the integrated and proposed decomposition algorithms. The proposed algorithms lead to potential gains in the computational efficiency with limited data exchanges among areas.