淮河沙颍河支流水质预报模型的参数率定

Choosing Parameter of Shaying River Water Pollution Prediction

水质模型的参数率定一直是水质预报的难点。由于要求模型计算结果准确，故必须对每个参数反复多次计算才能给出一个较为可用的参数值。用传统的串行计算方法来处理参数率定问题，其计算时间过长，无法快速给出准确的模型。文章利用已有的数据和经验，根据沙颖河具体的水质和应用特点，在国内首次提出了采用基于并行计算技术的高性能计算方法，对沙颖河支流水质模型进行多参数率定的方案，并在曙光系列并行机上得以实现。结果表明：其率定的准确性大大优于常规率定，并缩短了整个模型的建模时间，由此为模型的快速应用提供了可靠的参数依据。

The choice of good parameter of water pollution model is a nodus in hydraulics. Because of the requirements of exactness result,an available parameter is always achieved by very large scale and many times repeated computations.The traditional serial computational method might meet with problems of needing too much time or resulting with an undetermined parameter. Using existing data and experiences, considering the character of Shaying River, and based on the water quality forecast system of Shaying River,in this paper,we firstly put forward a high performance parallel computing method of multi-parameter choice in China,and implement it on Dawing High performance parallel computer for the choice of parameter for Shaying River. Experiment shows that the exactness is better,the time for building the model is shortened, thus supplies more credible parameter data for the application of the model.