A new one-dimensional control valve model for the local flow coefficient based on local analysis of CFD data

  • Qingyun Bu School of Computing & Engineering, University of Huddersfield, Queensgate, Huddersfield
  • Aliyu Aliyu School of Computing & Engineering, University of Huddersfield, Queensgate, Huddersfield
  • Rakesh Mishra School of Computing & Engineering, University of Huddersfield, Queensgate, Huddersfield
Keywords: Control valves, flow coefficient, numerical simulation, valve design, pressure distribution, computational fluid dynamics, local flow analysis

Abstract

Control valves are essential components of the energy, nuclear, process, and the oil and gas industries. They are needed for the accurate control of flow rate, pressure, and temperature crucial to the success of reactions in chemical processes and delivering the right amount of fluid to downstream facilities for storage, transport, or sale. As a result, the optimal design of control valves for being fit for purpose and for energy efficiency is of crucial industrial and commercial importance. To achieve optimal design, computational fluid dynamics is used as a powerful tool for the analysis and design of many process units including control valves. The full three-dimensional CFD simulations are useful but can be computationally expensive and time consuming if many designed needs testing. Furthermore, rapid solutions may be needed for testing many initial design configurations or for digital twin purposes where real time or near real time solutions are needed. One way of achieving quick solutions is the use of one-dimensional approximations of the high-fidelity simulations 3D flow within the valve. In this study, CFD simulations were carried out and used to establish a 1-D numerical model to predict the flow within a control valve having a trim with a complex geometry. Five valve opening positions (20, 40, 60, 80, and 100%) and two inlet flow velocities of 1 and 2 m/s were simulated. The valve flow coefficient Cv for each of the 10 cases was collected along the valve–pipe system on fifteen local planes and used to build the 1-D model in which the Cv is correlated with the VOP, Reynolds number, and length-diameter ratio. It was shown that this data driven model resulted in a coefficient of determination R2 value of 0.82 when compared with the underlying data and can be used with confidence for initial design and digital twin applications.

Published
2022-02-21
Section
Articles