基于时间序列法与马尔可夫链的多沙水库动态运用研究

Study on dynamic operation of reservoirs on sediment-laden rivers based on time series method and Markov chain

为了更好解决多沙河流水库蓄水和排沙间的矛盾,在传统水库运用模式的基础上,结合时间序列法和马尔可夫链原理建立多沙水库的动态运用模型。该模型首先运用时间序列法对历史水文资料进行展延,生成长系列入库水沙数据,然后根据马尔可夫链原理分析生成的水沙系列,确定不同频率洪水在不同条件下发生的概率,最后结合水库当前已发生的洪水预测即将发生的洪水的概率,对初始的水库调度参数进行调整,得到最终的水库运用参数,从而弥补水库运用参数与实时入库水沙条件不匹配的不足。将该方法应用于实际水库的控制运用,结果表明,该水库在供水量变化不大的情况下,较原有的运用方案具有更大的剩余库容。

A dynamic operation approach of reservoirs is proposed in this paper using time series method and Markov chain to resolve the conflicts between water storage and sediment release of sediment-laden river reservoirs.In this approach,time series method is used to generate a long series of reservoir inflow and sediments through extension of the historical data,Markov chain technique is adopted to analyze this long series and obtain the flood probability of different frequencies under different conditions,and then the probability of impending flood is forecasted using the current flood so that a dynamical adjustment of the parameters of reservoir operation can be implemented.Through adjustment,the operation parameters can better match the variations of inflow and incoming sediments.Application to the reservoir on Heihe River in Shaanxi province verified this approach,and it reveals that under nearly the same water supply rates,a greater remaining capacity can be maintained than that by the classical method.