Forty years of numerical modelling
Cunge, J.A. (2003). Forty years of numerical modelling, in: Waterbouwkundig Laboratorium en Hydrologisch Onderzoek (Ed.) Colloquium: Numerieke oppervlaktewater modellering, mogelijkheden en beperkingen, Antwerpen 23-24 oktober 2003. pp. 1-11 In: Waterbouwkundig Laboratorium en Hydrologisch Onderzoek (Ed.) (2003). Colloquium: Numerieke oppervlaktewater modellering, mogelijkheden en beperkingen, Antwerpen 23-24 oktober 2003. Waterbouwkundig Laboratorium en Hydrologisch Onderzoek: Antwerpen. , more |  |
| Available in | Author | | Document type: Conference paper
|
| Keywords | Modelling
Numerical models
|
| Abstract | In the presentation three main aspects of numerical modelling technology, as it has been developed and introduced in hydraulics and hydrology will be mentioned and characterised. Namely: historical development and conditions that allowed such development; scientific and engineering aspects of applications, including market availability of the software and influence of the market on developments;reliability and limitations of the technology for engineering purposes. Last but not least some still unsolved theoretically problems and difficulties in that domain will be mentioned.The presentation is engineering-oriented, stressing the applications point of view and concerns essentially deterministic modelling. The history sketched in the presentation goes back to XIXth century in order to show interdependence between available theory, algorithms and informatics tools,including traditional and psychological obstacles. Then the history of four generations of models used from 1960 to day as well as current situation of "mass-market" for software are shortly summarised. Present situation of some confusion between application domains of correlative modelling (Artificial Neural Networks, Genetic Algorithms) and of deterministic models is mentioned with a frontier drawn between two categories.A chapter is devoted to engineering applications and related use and misuse of the tools as well as to the so called "good practice" generally admitted paradigm of modelling; it is based on four steps: setup-calibration-validation-running. Criticism of the paradigm from the point of view of current state of knowledge and proposal for new paradigm is given.Should one believe in models results? The limitations can stem from lack of understanding of physics,lack of formulations, blind application of unsuitable schematisations, etc. What can do an engineer,who must solve a problem even if there is no "good practice" validated tool available? Example of river and estuary hydro- and morpho-dynamics is taken to show limitations and doubts one can have about the results of modelling.N.B.: Present text is accompanied by presentation of figures (transparencies) as used by the author during the Colloquium. This presentation is recorded as separated Power Point File that can be printed out in colours or under black and white form. Numbers of figures referred to in the text correspond to order numbers of the transparencies in Power Point file. |
|
