• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Optimisation sous contraintes d'un modele hydrologique pour une representation de la physique des processus.
【24h】

Optimisation sous contraintes d'un modele hydrologique pour une representation de la physique des processus.

机译：在水文模型的约束下进行优化，以表示过程的物理性质。

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The objective of this paper is to develop a calibration method oriented on physical processes and provide a robust parameterization in the context of climate changes; our research hypothesis is that a set of parameters modeling hydrological processes with respect to the physical representation leads to a better solution than one which relies only on a better value of the objective function, modeling hydrological processes with complete freedom.;While two calibration strategies are put forward and are compared to the method currently used at Hydro-Quebec. The first strategy which is proposed focuses on the assessment of the actual evapotranspiration of the value offered by the Atlas of Canada through a constraint. The second uses the database offered a climate model by constraining the shape of the actual evapotranspiration cycle. This approach with constraints is refined using a sensitivity analysis tool available in the NOMAD package. Each constraint is thus treated to be relaxed or tightened depending on the sensitivity of the objective function and according to the confidence in the quality of the observed data, often marred by error. Thousands of optimization are performed by the MADS algorithm to compare different methods in terms of a good representation of hydrographs and evapographs (or average annual cycle of evapotranspiration), as well as the variability of "optimal" sets of parameters and the robustness of these solutions.;From these analyses, the parallel calibration was found to offer the best performance on many levels. A strong correlation is found between the solutions produced by the optimization and those produced by simulating the flow of a set of data not used for calibration, called validation. Thus, it is possible to conclude to a stronger robustness in the context of climate changes. The best solution obtained by the parallel calibration strategy is also proved capable of producing a better representation of the validation flows than the other strategies, while modeling the evapotranspiration process in order to respect physical process.;The most remarkable application of the method developed in this thesis is the adoption of parallel calibration strategy for the project ( CQ)2. This project, which brings together with Hydro-Quebec, the Ouranos consortium on regional climatology and climate changes adaptation, the Centre d'expertise hydrique du Quebec (CEHQ) and Rio Tinto Alcan, aims to qualify and quantify the impacts of climate changes on the hydrological regime in Quebec. To attain this goal, a single parameterization has been optimized using parallel calibration for each of the 324 watersheds covering the entire province of Quebec. At the end of this project, almost 160 000 hydroclimatic scenarios have been generated, from which 98 climate simulations using five different methods of post-processing deployed on 324 watersheds. This considerable database only has value if the model calibration is valid and it is thus a great vote of confidence from the project decision makers towards the results obtained in this research to have chosen our method for the calibration of the HSAMI Model.;The contributions of this thesis to the advancement of knowledge in hydrology are numerous. The most obvious is the development of a robust calibration method which can be implemented in the context of climate changes and can also be extended to several other hydrological processes for better physical representation. In fact, it is the first time that data from a climate model are integrated in the calibration process. This new perspective on the optimization procedure results in a significant improvement of the performance of the optimization algorithm through the appropriate use of the offered options. The use of a complementary sensitivity analysis module also opens new doors in offering the possibility to better take advantage of the user knowledge of a model in formulating appropriate constraints. The calibration constraint approach proposed in this research is thus enhanced and new avenues of research result from it. Finally, for the operational research field, this project provides a great opportunity for the application of optimization tools. (Abstract shortened by UMI.).
机译：本文的目的是开发一种针对物理过程的校准方法，并在气候变化的情况下提供可靠的参数设置。我们的研究假设是，与仅依赖于目标函数的更好值，完全自由地对水文过程进行建模的一组参数相比，一组针对物理表示法对水文过程进行建模的参数可提供更好的解决方案。提出并与魁北克水电局目前使用的方法进行比较。提出的第一个策略侧重于评估加拿大地图集通过约束条件提供的实际蒸散量。第二种方法是通过限制实际蒸散周期的形状，利用数据库提供气候模型。使用NOMAD软件包中提供的灵敏度分析工具可以完善这种带有约束的方法。因此，根据目标函数的敏感性以及对观测数据质量的置信度（通常会因误差而损坏），将每个约束条件视为放松或收紧。 MADS算法执行了成千上万的优化，以比较不同的方法，以更好地表示水文图和蒸发仪（或蒸发蒸腾的年平均周期），以及“最优”参数集的可变性和这些解决方案的稳健性从这些分析中，发现并行校准可在许多级别上提供最佳性能。在优化产生的解决方案与通过模拟未用于校准的一组数据流（称为验证）产生的解决方案之间发现了很强的相关性。因此，有可能在气候变化的背景下得出更强的鲁棒性。还证明了通过并行校准策略获得的最佳解决方案能够比其他策略更好地表示验证流程，同时为蒸散过程建模以尊重物理过程。本文是针对项目（CQ）2采用并行校准策略。该项目与魁北克水电局，Ouranos区域气候学和气候变化适应财团，魁北克水文专家中心（CEHQ）和力拓阿尔坎铝业联合起来，旨在鉴定和量化气候变化对加拿大的影响。魁北克水文政权。为了实现这一目标，已经对覆盖整个魁北克省的324个集水区中的每一个进行了并行校准，优化了单个参数。在该项目结束时，已产生了近16万个水文气候情景，其中在324个流域上使用五种不同的后处理方法进行了98次气候模拟。仅当模型校准有效时，这个相当大的数据库才有价值，因此，项目决策者对本研究中获得的结果充满信心，选择了我们的HSAMI模型校准方法。本论文对水文学知识的发展是众多的。最明显的是开发了一种可靠的校准方法，该方法可以在气候变化的背景下实施，并且还可以扩展到其他几个水文过程中，以更好地体现物理特征。实际上，这是第一次将来自气候模型的数据整合到校准过程中。通过适当使用所提供的选项，可以从最优化过程的新视角显着改善优化算法的性能。补充灵敏度分析模块的使用还为提供可能性，从而在制定适当的约束条件时更好地利用模型的用户知识开辟了新的大门。因此，本研究中提出的校准约束方法得到了增强，并由此获得了新的研究途径。最后，对于运筹学领域，该项目为优化工具的应用提供了很好的机会。 （摘要由UMI缩短。）。

著录项

  • 作者

    Cartier, Dominique.;

  • 作者单位

    Ecole Polytechnique, Montreal (Canada).;

  • 授予单位 Ecole Polytechnique, Montreal (Canada).;
  • 学科 Operations Research.;Engineering Environmental.;Hydrology.
  • 学位 M.Sc.A.
  • 年度 2013
  • 页码 157 p.
  • 总页数 157
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号