基于双向指数平滑的水位数据修复方法

Water Level Data Restoration Method Based on Bi-directional Exponential Smoothing

受制于水位站数据采集方法、传输线路等问题,水位数据的完备性在工程中难以得到有效保障.针对长江航道水位数据缺失问题,提出一种基于指数平滑法的水位数据修复方法.采用二次曲线拟合的方法计算初始值与平滑系数,进而计算了各期平滑值,并采用双向预测方法对实验结果进行了修正.经实际数据验证,双向指数平滑修复方法可以一定程度上中和预测结果的随机误差,提高修复结果的稳定性,将相对平均误差降低至0.2m之内.方法简单有效,可应用于长江等内河航道水位数据修复,提升内河航道要素感知与服务水平.

The completeness of water level data cannot be effectively guaranteed in engineering due to the problems of data acquisition methods and transmission lines of water level stations.In order to solve the problem of missing water level data in the Yangtze River Channel, a water level data restoration method based on exponential smoothing method was proposed. The initial value and smoothing coefficient were calculated by the quadratic curve fitting method, and then the smoothing values of each period were calculated. The experimental results were corrected by the bi-directional prediction method. Through the actual data verification, the bi-directional exponential smoothing repair method can neutralize the random error of the predicted result to a certain extent, which improves the stability of the repair result and reduces the relative average error to within 0.2m.The proposed method is simple and effective, which can be used to repair the water level data of inland waterways such as the Yangtze River and improve the perception and service level of inland waterway elements.