Localização de fontes harmônicas utilizando técnicas de processamento estatístico de sinais

Abstract

Research in Electric Power Quality (QEE) is a field of electrical engineering responsible for providing improvements in energy supply and consumption. The term QEE works with the prevention and mitigation of electrical disturbances. Non-linear loads connected to the Electric Power Systems (SEPs) promote undesirable changes in electrical signals and contribute to the disturbance generation. This work presents a methodology for detection of harmonic current sources as a possible contribution, identifying the locations of the harmonic sources. In this document, the source estimation problem is treated as a Blind Source Separation (BSS) paradigm. This philosophy assumes that the sources are not known. Consequently, the voltage measurement is the only available parameter for network monitoring. Additionally, the initial premise assumes the power grid topology is unknown. In this case, situations with large volumes of data must be considered. Thus, Principal Component Analysis (PCA) is implemented as a tool to reduce dimensionality and aid in computational complexity for the estimation step. Independent Component Analysis (ICA) is a BSS tool and it is responsible for signals estimation, in this case, the profiles that characterize the sources. With the estimated profiles, the joint entropy is used to determine the probable locations where the harmonic current sources are connected. In this step, the minimum value of joint entropy is used as a decision parameter for detection. The methodology effectiveness is verified through the use of computer simulations and test systems, such as the IEEE 14 bus and the IEEE 33 bus.