Fault Detection Systems via a Novel hybrid Methodology for Fuzzy Logic Systems based on Individual base inference and Statistical Process Control

This paper presents a novel methodology to model fuzzy logic systems used as fault detection systems. This method is based in a hybrid model assembled via Individual Base Inference (IBI), Statistical Process Control (SPC) tool in the form of Control Charts (CC) and MacVicar-Whelan Methodologies. The novel method provides a simplification of the necessary calculus of inference engine module with the approaches SPC tool (CC: X-S and X-R). The inference engine is assembled via a novel methodology based in the concepts used to create CC and can be made with the specification for R charts or X-S charts. This methodology consists in the hybridization of Macvicar-Whelan method for the assignation of fuzzy labels to provide rules. The IBI is created with the wide spread of real historical data. This data forms the Universe of Discourse (UOD) and the SPC tool is used to provide the fuzzy rules and rule reduction. The aim of this paper is to test the novel method and to compare it with a classic method to prove that this method is an efficient form to rule deduction without lost reliability in the output approximation. The proposed method was tested in an electrical power system.