Modelo Virtual de Vazão de Gás para Medidores do Tipo Placa de Orifício

Abstract

The accurate measurement of flow rate is crucial in primary processing on oil platforms and for compliance with legal standards. Although orifice plate flow measurement is still widely applied, it lacks robust monitoring methods. The application of virtual flow models presents an alternative for monitoring and diagnosing this technology. This study implements a virtual model based on the formulations of ISO-5167 and AGA-8 standards. Additionally, a data-driven virtual model is implemented using a multi-layer perceptron (MLP) neural network. The results demonstrated high accuracy in the virtual model based on ISO and AGA, with average deviations of 10−3% for mass flow rate. The MLP-based virtual model yielded consistent results with similar studies.