Método ativo virtual para detecção de ilhamento em geradores síncronos distribuídos

• Ano de defesa: 2019
• Autor(a) principal: Muraro, Matias Rossato
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; Engenharia Elétrica; UFSM; Programa de Pós-Graduação em Engenharia Elétrica; Centro de Tecnologia
• Palavras-chave em Português: Geração distribuída; Detecção de ilhamento; Gerador virtual; Gerador síncrono; Distributed generation; Islanding detection; Synchronous generator; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA
• Link de acesso: http://repositorio.ufsm.br/handle/1/22493

This work presents a local and passive islanding detection method for synchronous distributed generation (DG) directly connected to the distribution power system. Due to the characteristics of these generation system, such the electromagnetic coupling between network and the rotating mass of the rotor and governor system and the synchronous generators self sustaining voltage, limitations on local and passive islanding detection techniques can occurs. Thus, an islanding detection technique based on the point of common coupling (PCC) electrical frequency and DG mechanical frequency comparison after a disturbance occurrence is proposed. In case of this frequencies have similar behaviors, is considered that an islanding has occurred. It is used the needed energy to match the DG mechanical frequency with electrical frequency to verify the frequencies similarity. To avoid a active protection in the DG, a virtual DG that has a similar dynamic behavior to the real DG is developed. The PCC electrical measures are the inputs of this virtual DG. The virtual DG is obtained through the synchronous generator and approximated governor dynamic models, which take account the real DG parameters. The modification in virtual DG frequency is performed by an active virtual protection, that is obtained in order to allow stability for the proposed methodology. The host power system and the DG electrical and mechanical parameters are the input data to proposed method configuration. In this way, simulations and user know how are not necessary to use it. Simulations results show that the islanding detection proposed method has a small non detection zone and appropriated detection times, despite more than one DG are connected to the host power system. The results also show that que proposed method does not cause false detections for other events occurrence in distribution power systems. Furthermore, the necessary time to islanding detection is obtained from DG and host system parameters, this allows a direct and objective fast reclosing times determination.