Avaliação experimental da modelagem de cargas residenciais utilizando a representação ZIP em estudos de fluxo de potência em sistemas de distribuição

Abstract

This undergraduate thesis aims to propose a residential load modeling approach more closely aligned with real-world conditions, based on the ZIP representation implemented in Model 8 of the OpenDSS software. The ZIP model is composed of three components associated with constant impedance, constant power, and constant current behaviors, which define the relationship between active and reactive electrical power and the variation of the voltage applied to the load. In order to obtain a representative composition, the P×V and Q×V behaviors of equipment commonly found in Brazilian households are experimentally analyzed, allowing not only the verification of the adherence of such equipment to classical ZIP load models, but also the quantification of the degree of participation associated with each model component (constant impedance, constant current, and constant power) whenever such adherence is confirmed. Based on the experimental results obtained, an aggregated load model more representative of the actual characteristics of Brazilian residential consumption was proposed, explicitly incorporating the predominant contribution of each ZIP component observed in the different analyzed devices. This approach not only improves the electrical representation of loads in computational studies, but also enables a critical assessment of the impacts of the adopted modeling on the calculation results of technical losses in distribution systems. Finally, the proposed modeling approach is compared with the methodology currently adopted by ANEEL in the process of calculating technical losses in distribution systems [Module 7 of PRODIST], highlighting some limitations of the regulatory approach, such as the use of generic load compositions and the greater complexity associated with its implementation. The obtained results demonstrate that the proposed model provides a more representative characterization of actual load behavior while requiring lower computational effort, thus constituting a more suitable and efficient alternative for technical loss studies in distribution systems.