江净超, 余洁, 秦承志, 刘军志, 李润奎, 朱良君, 朱阿兴. 知识驱动下的水文模型参数智能化设置方法[J]. 武汉大学学报 ( 信息科学版), 2017, 42(4): 525-530. DOI: 10.13203/j.whugis20150044
引用本文: 江净超, 余洁, 秦承志, 刘军志, 李润奎, 朱良君, 朱阿兴. 知识驱动下的水文模型参数智能化设置方法[J]. 武汉大学学报 ( 信息科学版), 2017, 42(4): 525-530. DOI: 10.13203/j.whugis20150044
JIANG Jingchao, YU Jie, QIN Chengzhi, LIU Junzhi, LI Runkui, ZHU Liangjun, ZHU Axing. A Knowledge-driven Method for Intelligent Setting of Parameters in Hydrological Modeling[J]. Geomatics and Information Science of Wuhan University, 2017, 42(4): 525-530. DOI: 10.13203/j.whugis20150044
Citation: JIANG Jingchao, YU Jie, QIN Chengzhi, LIU Junzhi, LI Runkui, ZHU Liangjun, ZHU Axing. A Knowledge-driven Method for Intelligent Setting of Parameters in Hydrological Modeling[J]. Geomatics and Information Science of Wuhan University, 2017, 42(4): 525-530. DOI: 10.13203/j.whugis20150044

知识驱动下的水文模型参数智能化设置方法

A Knowledge-driven Method for Intelligent Setting of Parameters in Hydrological Modeling

  • 摘要: 水文模型的参数设置涉及专业领域的知识和繁琐的操作步骤,是水文建模面临的一大难题,在一定程度上限制了模型的应用和推广。基于知识驱动的方法,将水文模型参数设置知识分为参数提取知识和取值范围知识两类,分别对其进行形式化表达和自动推理,初步实现了水文模型参数智能化设置;结合Web Service、工作流和参数自动率定等技术,研发了水文智能建模原型系统;最后,以TOPMODEL模型的参数设置为例,对知识驱动方法进行了验证。结果表明,知识驱动的方法能在保证模拟精度的前提下有效简化水文模型的参数设置流程,降低水文建模难度。

     

    Abstract: Setting parameters for hydrological models requires not only specialized knowledge but also tedious operation steps. This is a major difficulty in hydrological modeling that largely constrains the ease of use of hydrological models. Using a knowledge-driven method, the knowledge on parameter setting was divided into the knowledge on parameter extraction and the knowledge on parameter value range. The above knowledge was formalized and inference engines were designed for setting parameters in hydrological modeling intelligently. A prototype system for intelligent hydrological modeling was implemented using web service, workflow, and automatic calibration of parameters. A case study of automatic intelligent parameter setting was conducted for TOPMODEL in a real watershed. The results showed that the knowledge-driven method was able to conduct parameter settings automatically and achieve satisfying modeling results. Therefore, the proposed knowledge-driven method and the intelligent system have great potential to simplify hydrological modeling processes.

     

/

返回文章
返回