Numerical simulation of trans-critical flows in open channels using a FVM scheme
- 219 Downloads
The present study makes efforts to simulate the behavior of fully developed stationary shocks, caused by the incidence of supercritical flow with a cross-barrier in an open channel. The numerical solution of nonlinear governing shallow flow equations has been implemented by the application of a second-order Roe TVD scheme. The obtained results from numerical experiment are compared with some measured in a laboratory setup. It can be deduced by comparison of the flow depths in numerical and measured experiments in three different cases of cross-barrier width of 6, 12 and 16 cm that the numerical scheme of Roe is a robust and capable method for simulation of complicated stationary shocks in shallow water flow.
KeywordsShallow flows Riemann solvers Roe scheme Discontinuities and shocks
When an abrupt change occurs in flow depth and/or velocity, the shocks or bores will appear. The shocks normally appear in supercritical flows, while transition from super- to the subcritical regime, named trans-critical, can produce a shock (i.e., hydraulic jump). They are divided into two general categories: dynamic shocks whose characteristics are various over time (i.e., produced shock due to dam break and gate operations) and static shocks whose locations and other characters are permanent after the flow steady state (i.e., hydraulic jump or oblique stock in transition).
The governing equations on shock simulation problem are the shallow flow equations. According to their nonlinearity, the analytical solution is limited to simplified cases such as reflectance of some important terms; hence, the numerical methods perform as valuable tools to solve the nonlinear shallow flow equations although, on the other hand, due to sudden changes in flow patterns by shock adjacent, the simulation of their locations and their shapes are more complicated. Therefore, the conventional numerical methods such as Preissmann (or four-point scheme) (Fennema and Chaudhry 1990) are not capable of capturing shocks. According to the solution known as the Euler equation, Jameson suggests ‘artificial viscosity’ and ‘modified equations’ to eliminate spurious non-physical oscillations adjacent sharp discontinuities (Jameson et al. 1981). Consequently, such methods were applied to shallow flow equations. In 1985, Fenema and Chaudhry applied beam–warming, Gabutti and Mc Cormack schemes including artificial viscosity for the simulation of the dam-break problem (Fennema and Chaudhry 1990).
Since the early days of the 1980s, the idea of Godunov conservative method has been in application for the solution of Euler equation. These methods are not capable of capturing sharp discontinuities such as shock and bores without employing numerical viscosity or simplifications of governing equations. In order to eliminate non-physical spurious oscillations, the ‘total variation diminishing (TVD) methods’ developed by Sweby (1984) and subsequently many researchers applied these methods to the solution of shallow flows in the 1990s. Alcrudo and Garcia-Navarro (1993) applied the Godunov-type finite volume method for the solution of shallow flow equations. Zhao et al. (1996) used the approximate Riemann solvers of Osher and Solomon for the simulation of unsteady flow in rivers. Chippada et al. (1998) continued with the Roe scheme to the modeling of flow through transitions and tidal waves. Also, Yang and Greimann (2000) went on to apply the TVD version of Roe scheme for the simulation of the dam-break problem with sediment transport. Zoppou and Roberts (2003) employed weighed average flux (WAF) in dam-break problem. Tseng et al. (2001) used high-resolution conservative methods for the simulation of wave in channels. Calleffi et al. (2003) applied approximate Riemann solver of HLL for the simulation of extreme floods in rivers. Sheng Bi et al. (2015) applied improved unstructured finite volume algorithm in 2D shallow flow modeling. Martins et al. (2017) offered a detailed validation of finite volume (FV) flood models in the case where horizontal floodplain flow is affected by sewer surcharge flow via a manhole.
In this research, the approximate Riemann solver of Roe, a TVD type scheme, has been applied for simulations of the steady oblique shocks caused by the confrontation of a supercritical flow to cross-barrier in the open-channel flow. In this line of attack, several laboratory experiments were used.
The governing equation
Riemann problem is defined as an initial value problem. The Riemann solvers in conservative methods are divided into four main groups: (1) flux difference splitting methods, (2) flux vector splitting methods, (3) exact Riemann solvers and (4) approximate Riemann solvers (Le Veque 1997). The approximate Riemann solvers on finite volume frameworks are also known as ‘high-resolution finite volume methods.’ In FDS methods, since the shallow flow equations system is from the parabolic type and includes the inherent character of signal propagation, numerical fluxes are derived based on local information of wave propagation. To this extent, they have special capability for capturing the sharp discontinuities. The FDS methods are derived based on Godunov method. Therefore, they are known as ‘Godunov-type’ methods.
The Roe scheme
In Eq. (8), Φ(r) is a limiter function; for Φ(r) = 0, there will be an upwind first-order TVD scheme, and for Φ(r) = 1, the Lax–Wendroff second-order non-TVD scheme will come out.
In order to eliminate the spurious oscillations that occur in the adjacent of sharp gradients, any of the limiter functions may be used. Thus, a second-order TVD scheme, or the so-called Roe scheme, will be achieved.
Extension to 2D problems
The numerical experiments show that the first-order accuracy upwind scheme has dissipative behavior in a shock front. The second-order Lax–Wendroff scheme makes spurious oscillations in adjacent of sharp discontinuities. The TVD version of Roe scheme is capable of capturing sharp discontinuities as well as the smooth region (Alcrudo and Garcia-Navarro 1993).
Initial and boundary conditions
The initial conditions are the flow parameters at time t = 0. The boundary conditions are divided into two general categories of open and solid boundaries. For current simulation, which can be seen in Fig. 2, in the longitudinal channel walls there is a solid boundary as well as upward and downward of the barrier. On the other hand, due to the supercritical condition in the downstream end the boundary is fully open. In the upstream of modeling reach, there is an open boundary with one extra condition of constant discharge for subcritical flow regime.
Application of the model
In this research, the developed shocks in the mentioned case have been simulated in two dimensions with Roe TVD numerical scheme. The numerical results will be compared with some experimental data from a laboratory setup.
Laboratory setup and measurements
Comparison between experimental and numerical results
The relative error in measuring the depth and the flow discharge which is due to the extreme oscillation and turbulence of the flow, especially in the first downstream shock that includes a barrier.
Despite the fact that the actual thickness of the barrier is about 3 mm, in the numerical modeling it is considered equal to 0.02 m.
Although the flow condition is 3D around the barrier, 2D shallow water equations (averaged depth) are applied in the numerical model.
The applied numerical model includes a second-order accuracy. Although the previous experiments and also the latest results of this modeling show the capability of this numerical method to capture the shocks, applying more accurate equations will be more helpful.
The first step, which is created in the first downstream reflection (Sec 6 in Fig. 8), is produced from the extreme jet flow, which is faced to a wall. It should be mentioned that the present theoretical model is not able to simulate such flows.
The conservative methods based on Riemann solvers are capable of capturing shocks (Fig. 9). In this study, the approximate Riemann solver of Roe (TVD version) has been employed to study the shocks behavior which is produced by the attendance of a cross-barrier in supercritical open-channel flow. It is concluded that the applied numerical scheme is generally able to capture developed stationary shocks downstream of cross-barrier in the flume. Both measured data and numerical experiments show that the produced shocks are damping by faring from cross-barrier to downstream end. Due to the existence of flow conditions around the barrier, especially at its downstream, the flow pattern is three-dimensional; therefore, the two-dimensional numerical method is not able to capture shock height properly.
- Jameson A, Schmidt W, Turkel G (1981) Numerical solution of the Euler equations by finite volume methods using Runge–Kutta time-stepping schemes. In: AIAA 14th fluid and plasma dynamics conference, Palo-Alto, CA, AIAA-81-1259Google Scholar
- Le Veque R (1997) Numerical methods for conservation laws. Birkhauser-Verlag, BaselGoogle Scholar
- Toro EF (2001) Shock-capturing methods for free-surface shallow flows. Wiley, HobokenGoogle Scholar
- Yang CT, Greimann BP (2000) Dam-break unsteady flow and sediment transport. USBR, Technical Service CenterGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.