| The NetWork - A community newsletter for DHI Software users |
February 2005 | |
|
|
Flood forecasting models As part of a major flood forecasting project in China a comprehensive comparative study was carried out for five well established hydrologic models and three well established hydraulic models to assess their performance and suitability for use in real-time forecasting. The model evaluations were undertaken using common data sets and technical performance criteria specified in the Chinese National Standards for Flood Forecasting Models. For the hydraulic models the conclusion was after the consideration of all evaluation criteria that MIKE-11 was found to be most suited for use within the flood forecast project. The comparison study was described in a paper presented at a national conference in Australia in July 2004.
Links: Evaluation of hydrologic and hydraulic models for real-time flood forecasting (PDF-305KB) Link to the paper in PDF format Link to Web site of flood forecast project | |
|
|
|
Announcing AUTOCAL OfficeGRIDAUTOCAL is a generic tool for performing automatic calibration, sensitivity analysis, and scenario management with MIKE SHE, MIKE 11, MIKE 21 and MIKE 3. It is an extremely powerful tool for modellers when setting up and calibrating a model, assessing the sensitivity of model results with respect to different input variables, and for simulation of different modelling scenarios. Sensitivity analysis and parameter estimation in complex models is often hampered by the high computational demand associated with the model simulations. With running times measured in hours such applications may take several days, weeks or even months. However, there is away around this and that is AUTOCAL OfficeGRID. AUTOCAL OfficeGRID provides distributed computing facilities that enables sensitivity analysis and parameter estimation in computationally demanding models by effective utilisation of multi-processor PCs and PCs connected in a network. The workload is distributed to a number of PCs whereby the computation time can be reduced significantly. AUTOCAL OfficeGRID does not require execution on a well-defined cluster of PCs but can be used on the existing network of PCs in the organisation. The OfficeGRID network is dynamic and PCs can join and leave the network while an AUTOCAL application is running. Thus, AUTOCAL OficeGRID allows for effective utilisation of the existing computational resources, which are otherwise unused outside normal working hours. Further details and some results are available here. |
|
|
|
| Modeling tips & hints - Getting topographical information on lateral linkage couplings in MIKE FLOOD the graphical approach. |
Service pack 2 of MIKE Flood version 2004 includes a feature for analysing the invert levels of lateral links. The feature works by generating an ASCII file containing the invert levels of the lateral links. This file may be directly imported into the cross section editor of MIKE 11. Internally within MIKE Flood lateral links are treated as follows:
- The h-grid points within the linked MIKE 11 reach are identified.
- The MIKE 11 reach is normalised so that the reach runs from 0 to 100 %.
- The linked MIKE 21 cells are given a chainage based on the cell size. That is the chainage runs from cell centre to cell centre plus additional contributions from the first and the last cell.
- The MIKE 21 chainages are normalised so that they also run from 0 to 100 %.
- Depending on the method used the lateral structures will be set up according to the M21 cells or the M11 grid points. The chainage of the structures will be based on either the chainages in m11 or the cell defined chaniages in m21.
- The chainage of the structures are scaled to run from 0 to 100 %
- The code uses the location based on the scaled chainages (0 -100%) to locate which cells and h-points match up with which lateral structure.
- If the HGH method is used the invert is checked so that the greatest of the M11 and the M21 invert is used.
- The ASCII file is exported in the first time step of the run.
To generate this ASCII file the following procedure should be followed:
- Set-up the MIKE Flood model
- Place an empty file named "MFLateral.txt" in the same folder as where the *sim11 file is located.
- Run MIKE Flood.
- Once the first time step of the simulation has been completed the file in step 2 will have been overwritten with values describing the inverts of the lateral links.
- Open a new cross section file (*.xns11) and import the " MFLateral.txt" using File -> Import -> Import raw data.
- The ASCII file has 6 entries for each lateral link given by:
- M11 chainage, invert, percentage (in the n column) - M11 percentage, invert, percentage - M21 chainage, invert, percentage (in the n column) - M21 percentage, invert, percentage - Struc chainage, invert, percentage (in the n column) - Struc percentage, invert, percentage
- The inverts of each of the links may now investigated for possible inconsistencies.
|
|
|
|
|
|
|
|
| Contact - Europe |
|
|
|
|
|
|
|
|
|
|
Copyright © DHI Water & Environment. All rights reserved
 This newsletter was sent to you by DHI Water & Environment in response to your registration. All details are treated confidentially and will not be transferred to third party.
|