Report Exercise 3.3 - -Mike21


Team 1 - HydroEurope

Deadline: January 13th 2016


Benoît BESSEAS(1), Océane CALMELS(1), Laura DAUL(1), Guillaume HAZEMANN(1), Quentin MOLIERES(1), Kamila FUZIEK(2), Arianna VARRANI(3), Taline ZGHEIB(4), Zhengmin LEI(1)

Supervisors : Philippe GOUBERSVILLE(1), Olivier DELESTRE(1)

(1) University of Nice Sophia-Antipolis, France

(2) Warsaw University of Technology, Poland

(3) Brandenburg University of Technology Cottbus-Senftenberg, Germany

(4) Vrije Universiteit Brussel, Belgium


I.                   Introduction

    Within the HydroEurope Project for modelling the Var flood events of 1994 and 2015, 2D hydraulic modelling tools have been presented. Main aim for the following exercise is to introduce and get familiar with MIKE 21 software for the solution of 2D SWEs. Two different tasks have to be accomplished: the first is to set up a simple model to understand the main steps of model setup in MIKE 21 and how to visualize results, and a “real data” model for the Var River mouth.

II.               Methodology

    Being Mike software commercial, there is the need, as for many other licensed softwares, to prepare and pre-process the input data in to create specific files with format readable by Mike. To set up a model then means to put all files together, via the use of a GUI, and after defining the solver’s parameters to run the model (Figure 1).

Figure 1: Main Steps to setup a model in Mike21

    In the following two cases are being investigated, via Mike 21 software: (i) a first simple model (Exercise 1) setup from scratch and (ii) a more complex model for the mouth of Var River (Exercise 2).

III.            Results and discussion

1.      Exercise 1

    In this first exercise, we became familiar with the MIKE 21 software. A simple case was set up from scratch: an open rectangular channel with simple bathymetry. The input data were created from blank files in Mike Zero environment

                 a.       Bathymetry

    In order to run a MIKE 21 simulation, we first need to create the bathymetry file that can define our rectangular channel.

    We start by creating a blank Grid that we set as a 2-dimensional grid with a map projection in the UTM-30 system. Then we set the geometry properties as follow, 100 cells on the J-direction, 10 cells in the K-direction with a cell size of 5m. To finish we set the definition of no-data value at -999m and dry land value at 25m.

Figure 2: Domain and boundary conditions in red (land elevation)

    The last step was to create the borders of the channel. For that we input the values of 25 (defined land value) in Row 0, Row 9 and Column 99 to define them as the wall of the rectangular channel. The thickness of channel is assumed as the cell size 5m. We obtained the following bathymetry (Figure 2).

If we wanted to create a slope gradient of 1:500 from 0 m at the opening to 1m elevation at the end of the channel, we should add this change for each grid, this means the elevation value for each grid will be elevated little by little as the slope exists.

To realize the bed slope in bathymetry file, we need to use interpolation tool. We change only the value in column 98 (except for the value in row 9 and row 0 which are the walls of channel) by inputting value 1. Then select the grid we need to interpolate value enter 0 and 1. The interpolation can be found in Tools. 

b.      Flow model

     Now that our bathymetry is created, we need to set the flow model that define the simulation. For that we create a Mike 21 Flow model

Firstly we need to enter the “Basic parameters”. We chose the simulation as “Hydrodynamic only” (other modules can be select for more advanced simulation) then we set the bathymetry created previously as the bathymetry of the simulation, we chose a “Cold start” and enable the Coriolis forcing. Then we set the simulation period, from the 01/01/2016 with a time step range of 1000 every 1s and we set the boundary by using “Program detected” that can directly detect where is our boundary (left border of the rectangular channel).

Secondly the flow model requires the “Hydrodynamic Properties”. We set the initial surface elevation of 5m. The boundaries conditions are very important, they define how the water move at the boundaries. For this exercise, we set the water level constant at 10m but we could create a water level time series file and load it to make the level variable during time. We set the Manning number that define the roughness of the channel to 32, this number characterize a smooth outdoor river channel.

Before running the simulation, we needed to specify which kind of results files we wanted to create. Three options were available, point series that give the information of one cell, lines series that give the information of multiple points (can be used for see what happen on a section) and area series to get information on the cells in an area. We chose to create an area series that could allow us to see the water level and the fluxes on the whole bathymetry. Then the simulation was set, we could run it and check the results.

c.       Results visualization

We obtained the following result file (.dsf2) as follow.

Figure 3: Results: wated depth

    We can observe for each time steps the level/fluxes of each cells.

There are three type of simulation result files in MIKE 21. They are type 0 which is point series, type 1 which is line series and type 2 which is area series.

Figure 4: Animation window

    If we want to make an animation, we need to select the Data Viewer for dfs2 filetype, open the results, and click on the camera forward button. We choose a codex to make the video and we run it.

2.      Exercise 2

To study the behavior of the mouth or the Var river in case of high discharges from the catchment, and analysis of different situations (with different initial conditions) was performed via Mike21.

Figure 5 Comparison of results for the three simulations at 6 :00 on 4/10/2015 a) constant IC, b) hot start, c) varying IC and reservoir downstream

    Three cases were analysed: i) constant initial conditions equal to 0 at elevation 0, ii) “hot start” with initial condition coming from the final results of the previous simulation and iii) “open boundary” at the river mouth (condition imposed by simulating a reservoir at the mouth, for open boundary conditions.

    As can be seen from Figure 5 for the last time step there is no significant difference between the cases i) and ii) while in case of a fictitious open boundary, case iii), a little more widespread flooding occurs. If one observes the results at other times, the situation is the same.

    Comparing the velocities at the downstream end (Var River mouth), the results show a similar tendency, cases i) and ii) almost overlap, while case iii) shows higher values.


IV.            Conclusions

    Exercise 1 was proven very useful in learning to handle Mike 21 data types and understanding how it generates results.

    As for Exercise 2 the trick of setting a reservoir as downstream condition allows more accurate results at the mouth, as it could have been guessed. On upstream zones it is not clear whether this condition offers better results, for no measured data are available.