Saturday Dec 15, 2018

Iranian Model Development- 1st Phase

Regarding that the compulational core of some of the compulational softwares is available, one of the ways which seems to be useful in this development is to use the softwares like POM or SWAN which computational core is totally available; but the main core of such a softwares mostly have been designed for special aims and so the computational method used by which is not appropriate for designing a multiapplicational software. For instance, the numerical method used in POM model which is a flow model on a structured mesh network with non-uniform  dimensions, suffers from numerical fluctuations  and however, is not appropriate for simulating the flows with severe variations (e.g. Tsunamies and dam breake simulation). Regarding that the text of this software computational programs is available, it is possible to do some slight changes through the programs, but however for completely solving the numerical problems of the model, a major part of the computational core of the model must be rewrited and the time required for these fundamental changes is not much lower than developing a new model. Another example we can mention is the famous softwave TELEMAC which had been one of the few softwares extremely used for simulating the flow on unstructured mesh, before MIKE21FM was developed. The computational core of this software was designed just for simulating the flows due to dam break; this model also sufters from low computation speed and yet is not capable of conservative solving of the flow equaitions. Despite being several years that this software main core has bean developed, still it is impossible for SOGREAH compony which is responsible for sale and support services to solve the numerical problems.
Because of the reasons above, and considering the experiences of working on both of the above softwares and also the present experience of the model development team in developing various models, it was intended to develop primarily a new model in this project.
It should be noted that during all of the development processes, present and available experiences of numerical models which are useful, have been used. So the final model is a model with high computational capabilities and extensive performonce. Another important advantage of developing a new model is that all parts of the model are available and simultaneously all parts are well understood. Because of that, the future developments of the model is possible. However, during the time and by creating the new modeling needs, future developments of the present model will be provided.
This section has been presented with the aim of capabilities and abilities introduction of developed or developing models for simulating the fiow in horizontal plan on structured and unstructured grid and also spectral wave model. Another aims of this section is to introduce the ability
of these models and comparison of them with present commercial or research models. 

شبکه بدون‌ساختار در محدوده خلیج فارس، دریای عمان و دریای عرب

Figure unstructured grid in Persian Gulf, Omman Sea and Arabian Gulf

1- Important capabalities of two dimensional model in horizontal plane on structured grid.
Regarding the previeus sections deseription, the developed model is able to do the hydrodynamic simulations, with acceptable accuracy and in a reasonable time.
Some of the capabilities of this model that many of the commercial and research models dosen’t have, are as follows:
   1- In this model, the computations related to the advection parts of the momentum equations are solved accurately and in a directed way. In addition, the equations are selved in a conservative way in this step, that is one at the problems of commercial models.
   2- Regarding that the directed methods with high order of accuracy have been used for solving the transport equation, simulating the flows having shock waves like tsunamis with high accuracy is possible.
   3- It is possible to implement boundary conditions in irregular open boundaries and implementation of the boundary conditions in these irregular boundaries is easily done. This capability does not exist in any commercial model. In addition, with this capability the model will be more powerful to simulate the flow on the nested grids.
   4- There is no limit for simulating the flow on nested grids. For instance, unlike the MIKE software, the boundary of present networks in small layers can be totally irregular,  and the bed elevation of each layer does not have to be the same.
In addition, there is no obligation for observing a particular ratio through the mesh dimensions in different layers. Also regarding that the simulation in every layer is done through it’s own time step, the computations run time on nested grids is much shorter than the similar softwares. In addition by using this simulation technics simulation is done faster than softwares like POM which are capable of using a mesh network with non-uniform dimensions.
   5- in the present model, the computation are performed only in the parts which are located in the water body in structured grids, and because of that, in addition to the possibility of flow simulation in complex geometiries, the simulation will be performed with very high efficiency.

2- Important capabilities of two dimensional model in horizontal plane on unstructured grids.
According to the previous sections descriptions, the development model will be able to do the hydrodynamics simulations with high accuracy and in a reasonable time. Some of the important numerical capabilities of this model that many of the commercial and research models doesn’t lacking to have, are as follows.

   1- In this model, the computations relevant to the advection parts of momentum equations are solved accurately and in a directed way. In addition, the equations are solved in a conservatively in this step; unconservative solution is one of the problems of many commercial models. It is interesting that one of the output parameters of TELEMAC software, which is one of the most famous commercial softwars for flow simulating over unstructural grids, is the lossing mass through the computations.
   2- Regarding that the directed methods with high order of accuracy have been used for solving the transport equation, simulatiung the flows with shock waves like Tsunamics with high order of accuracy is possible usually for these kinds of simulations.
We must use the methods, like Reiman approximate method using Ro method, or accurate solution of Reiman. Because of the algorithm and solution methods used in the program, its run speed is much higher than the above methods.

   3- The algorithm used to simulate becoming dry and wet is executable like structured networks. Although this algorithm keeps the mass conservative in the media, it doesn’t affect the quantity of the courant number.
   4- Fourth order Runge-kutta technique has only been used to solve the gravitational part. Because of that, it is possible to do the computations, accurately and with a courant number of near one or slightly bigger than one. Unlike the present softwares, these computations are done only in a small part of the program (graritational pharse), which the most critical courant number is relevant to. Therefore, run speed of the program will increase considerably.
   5- The combination of this model with the flow model on structured gird can create a very powerful model, for mathematical simulations in local studies. The function method of this synthetic model is totally similar to the structured  flow model on nested grids. It means that the required boundary conditions for structured girds system, comes from unstructured model and however, each of the two mesh networks, are solved with their own time step.

3- The spectral wave model capabilities in comparison with similar models.
As mensioned before, we can compare the designed models with similar models in several aspects.
In this developed model, it is possible to use several types of regular, irregular and nested grid networks, while the present models are only copable of using a unique gird system; specially in the case of irregular mesh network, the only present model is MIKESW. This model is also weak in solving the energy equation e.g. the order of the solution method is one. This method will definitely cause a numerical diffusion error and will predict an unreal value for wave energy dispersion. However, in the developed model the higher order methods like Quickest which are more accurate, will be used in the case of irregular mesh networks (It should be noted that for eliminating the numerical fluctuation, appropriate limiters are used in this method). Generally, using synthetic mesh networks, is one of the unique advantages of the developing model, which makes it different and distinguished from the other models. In addition, using the methods like PPM and new PPM in structured mesh networks also shows the superiority of this model rather than the similar models such as SWAN and WWATCH.

شبکه تودرتو خلیج فارس در سیستم مختصات کارتزین 

Persian gulf nested network in Cartesian coordinate system


Last Update : Mar 14, 2012 08:02