论文部分内容阅读
针对新安江模型提出一种前期影响雨量(P_a)逐时段修正方法,以实测流量与预报流量过程的均方误差最小为目标,在场次洪水预报中采用粒子群算法对前期影响雨量进行逐时段修正。为进一步提升算法性能,通过混沌初始化提高初始种群质量,利用禁忌搜索策略扩大搜索空间,提高全局搜索避免早熟收敛,并在相邻修正时段间引入最优个体保留策略提高算法稳定性和运算效率。在我国湖南省双牌水电站的实际应用结果表明,所提方法能有效改善P_a的准确性,显著提高模型预报精度,是一种切实可行的方法。
Aiming at the Xin’anjiang model, a method of period-by-period correction of pre-impact rainfall (P_a) was proposed. The objective of this paper was to minimize the mean square error between measured flow and predicted flow. Particle swarm optimization . In order to further improve the performance of the algorithm, the initial population quality is improved by chaos initialization, the search space is expanded by using tabu search strategy, the global search is avoided to avoid premature convergence, and the optimal individual retention strategy is introduced between adjacent correction periods to improve the algorithm stability and computational efficiency. The practical application results of Shuangpai Hydropower Station in Hunan Province of China show that the proposed method can effectively improve the accuracy of P_a and improve the accuracy of model prediction remarkably, which is a feasible method.