海港工程设计潮位计算方法

针对G um be l分布,考虑了连续样本及非连续样本两种情况,采用M on te-C arlo模拟分析技术,对海港工程设计潮位计算中的矩法(M OM)、最小二乘法(L-S)以及目前流行的线性矩法(L-M)进行了比较研究,同时采用沿海9个潮位站的资料进行了验证计算。结果表明,与M OM法和L-S法相比,L-M法具有最优的统计性能,建议为工程设计所采用;现有规范推荐的最小二乘法偏于安全,但对不连续系列或当总体分布参数Cv较小时,亦能提供较合理的设计成果。

Study on Calculation Methods of Design Tide Level for Sea Harbour Engineering

For the Gumbel probability distribution,three calculation methods of design tide level,namely the me-thods of moment(MOM),the least squares(L-S) and the linear moments(L-M) are studied by means of Monte-Carlo technique,among which two kinds of samples are considered including the so called"continuous sample"and"non-continuous sample".Samples of extreme tides observed at 9 tide stations are also applied to verify the conclusion of the Monte-Carlo analysis.It is shown that,for design tide level analysis,L-M is the best method comparing to MOM and L-S,as it always offers the most reasonable estimation if valued by the statistical unbiasness and effectiveness.The method of L-S which is recommended by the current "Code of Hydrology for Sea Harbour"for the tide level design usually offers estimation with positive bias,which implies the possibility of a over safe design value may be obtained for engineering design.But it is also indicated that for the case of non-continuous samples or samples with smaller coefficient of variation(C_V),L-S is able to provide reasonable estimation for the design tide level.