Optimization of X-ray tube voltage to improve the precision of two phase flow meters used in petroleum industry
To the best knowledge of the authors, in all the former studies, a fixed value of X-ray tube voltage has been used for investigating gas–liquid two-phase flow characteristics, while the energy of emitted X-ray radiations that depends on the tube voltage can significantly affect the measurement preci...
- Autores:
-
Alanazi, Abdullah
Alizadeh, Mehdi
Nurgalieva, Karina
Grimaldo Guerrero, John William
Abo-Dief, Hala M.
Eftekhari-Zadeh, Ehsan
nazemi, ehsan
Igor, Narozhnyy
- Tipo de recurso:
- Article of journal
- Fecha de publicación:
- 2021
- Institución:
- Corporación Universidad de la Costa
- Repositorio:
- REDICUC - Repositorio CUC
- Idioma:
- eng
- OAI Identifier:
- oai:repositorio.cuc.edu.co:11323/8974
- Acceso en línea:
- https://hdl.handle.net/11323/8974
https://doi.org/10.3390/su132413622
https://repositorio.cuc.edu.co/
- Palabra clave:
- Tube voltage optimization
Artificial intelligence
X-ray
Two-phase flow
GVP
Sustainable technology
- Rights
- openAccess
- License
- CC0 1.0 Universal
| Summary: | To the best knowledge of the authors, in all the former studies, a fixed value of X-ray tube voltage has been used for investigating gas–liquid two-phase flow characteristics, while the energy of emitted X-ray radiations that depends on the tube voltage can significantly affect the measurement precision of the system. The purpose of present study is to find the optimum tube voltage to increase the accuracy and efficiency of an intelligent X-ray radiation-based two-phase flow meter. The detection system consists of an industrial X-ray tube and one detector located on either side of a steel pipe. Tube voltages in the range of 125–300 kV with a step of 25 kV were investigated. For each tube voltage, different gas volume percentages (GVPs) in the range of 10–90% with a step of 5% were modeled. A feature extraction method was performed on the output signals of the detector in every case, and the obtained matrixes were applied to the designed radial basis function neural networks (RBFNNs). The desired output of the networks was GVP. The precision of the networks in every voltage and every number of neurons in the hidden layer were obtained. The results showed that 225 kV tube voltage is the optimum voltage for this purpose. The obtained mean absolute error (MAE) for this case is less than 0.05, which demonstrates the very high precision of the metering system with an optimum X-ray tube voltage. |
|---|