B6.1 Improved Gas Liquid Flow Meter Using a Neural Network
- Event
- SMSI 2021
2021-05-03 - 2021-05-06
digital - Band
- SMSI 2021 - Sensors and Instrumentation
- Chapter
- B6 Flow Measurement
- Author(s)
- M. da Silva, A. do Nascimento Wrasse, E. Nunes dos Santos - Federal University of Technology - Paraná, Curitiba (Brazil), R. Morales - Multiphase Flow Research Center (NUEM - UTFPR), Curitiba (Brazil)
- Pages
- 137 - 138
- DOI
- 10.5162/SMSI2021/B6.1
- ISBN
- 978-3-9819376-4-0
- Price
- free
Abstract
This paper presents an improved gas-liquid flow rate meter based on gas fraction, pressure drop measurements, and an Artificial Neural Network (ANN). The available database comprises 416 sets of data, in which 50% is applied to train, 10% to validate and 40% to test the ANN. We evaluated the performance of the network considering 10 neurons at the hidden layer. Quantitative results are presented, regarding gas and liquid flow rate estimation, showing significant improvements compared to a previous methodology based on algebraic approximations.