Uncertainties in Power Flow Analyzed Via a Fuzzy Number Based Method

The power flow is one of the most important tools available in a supervisory center in order to help the operators in terms of the electric system operation. The power flow analysis provides the system voltage profile, which is the starting point for the network operation planning. This analysis is based on the knowledge of the real and reactive power demands at all buses in the power system. These values are measured by mechanical instruments or in some cases they just are estimates of the predictive load. This way, the goal of this paper is to present and discuss a computational tool for the load flow analysis that takes into account the influence of both the measurement errors and the errors generated by the numerical computation. To represent uncertain data in scientific computation, there are basically two solutions: (i) interval mathematics, and (ii) fuzzy logic. Interval mathematics is highly efficient in controlling several types of computational errors. It controls automatically all soundings generated by the floating point system, and its final result is the maximum accuracy from the computational point of view. Fuzzy logic is a decision support system with uncertain knowledge which allowing the handling of vague information, in general described in natural language, converting it to a numerical format. The current computers can manipulate this latter format in a very easy way. Using both former mathematical approaches, a robust software was developed to deal with uncertainties and vague information. The program was built in C++ language aggregating the interval library C-XSC for Linux operational system. To test and validate the proposed method, the methodology was applied to a real equivalent electrical system from South-Southeast region of Brazil.