具有历史洪水资料的期望矩法参数估计研究  

Expected moments algorithm method for parameters estimation with consideration of historical flood

作　　者：王俊珍[1] 宋松柏[1] 

机构地区：[1]西北农林科技大学水利与建筑工程学院,陕西杨凌712100

出　　处：《水力发电学报》2014年第2期8-18,共11页Journal of Hydroelectric Engineering

基　　金：国家自然科学基金项目(51179160;50879070;51079037);高等学校博士学科点专项科研基金(20110204110017);西北农林科技大学大学生创新;创业训练项目(1210712076)

摘　　要：为了有效地利用历史洪水资料,提高设计洪水值的估计精度,研究期望矩法(Expected Moments Algorithm,EMA)在具有历史洪水资料序列中的应用。本文在介绍EMA法估算P-III型分布参数估计的基础上,应用蒙特卡洛试验进行EMA法统计历史洪水资料长度效率检验。结合传统参数估计法,以陕西省4个测站的年洪峰流量序列为例,研究EMA法的实用性。结果表明:EMA法利用历史洪水资料的效率较传统估计方法高,且随着历史洪水期长度的增加,EMA法在有效性和不偏性上的优势更为明显。在具有历史洪水资料的情形下,EMA法也能有效地提高P-III型分布参数估计精度,以期为现行的设计洪水计算提供依据。To make better use of historical flood information and improve precision of design flood, an expected moments algorithm （EMA） for estimation of flood parameters is presented. Monte-Carlo experiments with P-III distribution were conducted to verify this algorithm and its efficiency in using historical flood information. It was applied to an analysis of the annual peak discharge data for four stations in Shaanxi province and compared with other traditional methods for parameters estimation. Its applicability to flood frequency analysis was also examined. The results show a higher efficiency of EMA in the use of historical flood information, particularly its advantages in effectiveness and unbiasedness for long periods of historical flood records. EMA effectively improves the estimation precision of P-III distribution parameters for the cases of historical flood information available, and thus it would provide a basis for design flood calculation.

关 键 词：水文学 参数估计 期望矩法 P-III型分布 

分 类 号：P333.9[天文地球—水文科学]