Application of Neural Network Models for Estimating the Flow Rate of Components in a Multiphase Flow from a Well

Due to high productivity and demand for products, oil industry holds a prominent place in contemporary industry. Increased attention is paid to the monitoring of its state and the production capacity of fields under operation. The exploration of new oil fields and the drilling of producing wells require high time and financial expenditures. To reduce them, some advanced technologies including different artificial intelligence methods are currently used, e.g., to estimate the percentage of fractions in the oil flow from a well. Some results of studying a number of models for predicting the flow rate of components in a multiphase flow from a well on the basis of neural networks of different architecture are presented. The study was carried out by using the readings from the sensors of multiphase flow meters based on a Venturi tube and an X-ray flow meter.