【作者】 朱永英；

【导师】 周惠成；

【作者基本信息】 大连理工大学 ， 水文学及水资源， 2008， 博士

【摘要】 我国不仅洪涝灾害频繁,而且水资源短缺和水环境、水生态恶化问题突出。依赖科技进步,充分发挥水利工程尤其是水库的作用,在确保防洪安全的前提下,充分发挥水资源的社会、经济与生态环境综合效益,是缓解我国水资源危机的重要措施。水库预报调度作为水库控制运用的重要非工程措施之一,在确保防洪安全,提高洪水资源利用率,充分发挥水库防洪与兴利综合效益等方面发挥了重要的作用。但目前水库的调度方式多数只重视了水库的防洪与兴利效益,往往忽略了水库下游河道生态环境需水的现状。随着社会经济的发展,生态环境问题越来越受到人们的关注,兴利用水与生态环境用水的矛盾将成为水资源开发利用的主要矛盾之一。在现有水利工程状况下,如何通过调整水库调度方式来改善下游河道的生态环境,维持河流系统的健康、可持续发展,是当前急需解决的重要课题。因此,充分利用预报信息及先进的调度决策方法,进行考虑生态环境目标的水库调度方式研究是水库调度理论发展的必然趋势。本文结合“十一五”国家科技支撑计划重点项目“东北地区水资源全要素优化配置与安全保障技术体系研究”,针对目前多数水库调度方式中只注重防洪与兴利目标而忽略生态环境目标的现状,研究将生态环境目标纳入水库调度综合目标之中,充分利用径流预报信息,改善水库现行的调度方式,在不影响或较小影响水库原社会经济效益的前提下,发挥水库的生态环境功能,增加水库的综合利用效益。为此,首先研究径流中长期分级预报及定量预报方法,以提高预报精度;然后研究水库调度多目标模糊优选决策方法中规格化公式及优劣方案的选取;最后,基于预报信息及多目标调度决策方法,进行考虑生态环境目标的水库兴利预报调度方式研究。主要内容及成果如下:(1)水文、气象预报信息利用的可行性分析。在分析我国水资源利用现状和目前水资源供需矛盾的基础上,对水文、气象预报信息利用的必要性和可行性进行分析,重点研究洪水预报和短中期降雨预报信息应用于水库调度中的可行性。(2)基于粗集—模糊推理的径流多因素分级预报方法研究与应用。针对多因素中长期分级预报中因子选择的难题,建立粗集—模糊推理径流多因素分级预报模型。通过属性重要性及相对分类精度概念有效地筛选因子并确定因子分级,采用粗集理论的约简算法生成最小决策规则集,利用特征展开近似模糊推理法进行推理预报,为解决径流分级预报中因子选择的难题进行了有益的探讨,并提高了径流分级预报精度。(3)基于多因子小波网络的径流预测方法研究。为提高小波网络径流中长期预测精度,以径流趋势和径流量分开预测的思想为指导,提出了基于多因子小波网络的径流预测模型。将原始径流序列及初选影响因子作小波多尺度分解,在小波多尺度空间,量化多个相关因子对径流的影响,使用多因子相关系数修正局部相似延拓方式,实现径流趋势的预测;用模糊增强算法解决小波网络预测中径流突变点处偏差较大的问题,实现径流量的预测。最后,经小波网络确定模型参数,以多次迭代后的输出作为最终的预测结果。大伙房水库年、月径流预测实例应用表明,该模型能同时达到较好的模拟精度和预测精度。(4)多目标模糊优选模型的规格化公式及优劣方案的选取研究。综合利用水库调度决策,属于复杂的水资源系统多目标模糊决策问题,应综合考虑社会、经济、生态、环境等因素,构建指标评价体系,应用多目标模糊优选模型评价调度决策方案。重点研究模糊优选多目标评价模型中的规格化公式及优劣方案的选取两个关键问题,解决模糊优选评价结果的不稳定性问题,以优选满意调度决策方案。(5)针对目前水库调度方式中多数忽略生态目标的现状,以大伙房水库为例,研究了基于径流分级预报信息并考虑生态环境目标的水库兴利预报调度方式。全面分析水库下游河道生态环境现状及生态环境需水特点,确定最小和适宜生态环境需水量。在最小和适宜生态环境需水量的基础上,以径流分级预报信息确定水库的生态供水量;利用径流分级预报信息,制定水库的预报调度决策;应用多目标模糊优选模型,确定基于径流分级预报生态用水的水库兴利预报调度方式,并分析预报误差对水库兴利调度的影响以及预报误差的修正方法。结果表明,利用径流分级预报信息进行考虑生态环境目标的大伙房水库兴利预报调度,可减小考虑生态环境目标对水库其他兴利部门用水的影响,充分发挥水库的综合效益。最后对全文作以总结,并对有待进一步研究的问题进行展望。更多还原

【Abstract】 Frequent floods disaster, inadequate water resources and deteriorative water ecological environment are great threat to China. Relying on scientific and technical progress, making the best use of water resources socially, economically and ecologically is an important measure to ease Chinese water crisis.As an important non-engineering measure, reservoir forecast operation can ensure flood control safety to a certian extent. On the other hand, it also can increase water resources utilization rate and maximize the reservoir benefit. At present, most of the traditional operation modes only focus on the benefit of flood control and water supply, but eco-environmental water demand of downstream river is often neglected. With the development of society and economy, eco-environment problem attracts more attention, and the contradiction between benefical water supply and eco-environmental water demand has become more observable. So it is urgent to adjust the current reservoir operation mode to improve the eco-environment of downstream river under the existing hydraulic engineering condition.In this paper, eco-environmental objective is incorporated into reservoir operation goals. Firstly, medium and long-term hydrological classified forecast and quantitative forecast are studied to improve forecast accuracy. Secondly, the choice method of normalization formulae and the ideal and non-ideal plans is studied in the multi-objective decision-making method. Finally, based on the runoff forecast information and the multi-objective decision-making method, reservoir benefical forecast operation mode is studied considering the eco-environment objective. Main contents and results are as follows.(1) The feasibility study of utilizing hydrometeorological information.Firstly, the contradiction between flood control, water supply and ecological water requirement is studied. Secondly, the necessity and feasibility of forecast information in multi-objective reservoir operation is analyzed. The feasibility study of applying flood forecast information, short-term and intermediate-term rainfall forecast information to reservoir operation is researched with emphasis.(2) Study on multi-factor classified runoff forecast model based on rough-fuzzy inference method.To solve the problem of complex factors’ selection in classified runoff forecast, a multi-factor classified runoff forecast model based rough-fuzzy inference method is created. Firstly, factors are selected and classified with the attributes’ importance and relative classification accuracy. Secondly, reduction solution based on rough set theory is used to generate the minimum decision-making rule set. The results indicate that the model can solve the forecast problem related to complex factors’ selection effectively.(3) Study on runoff forecast method based on multi-factor wavelet network.In order to improve accuracy of wavelet network long-term hydrological forecast, a runoff forecast method based on multi-factor wavelet network is proposed. Firstly, the wavelet multi-scale decomposition is carried on the original data of runoff and factors. Secondly, the relative relation between factors and runoff is quantified in the wavelet multi-scale space, and multiparty correlation coefficient is used to amend similar extension mode. Fuzzy enhancement algorithm is also proposed to modify large deviation at mutation point in runoff forecast. Finally, parameters are determined with the wavelet network model and the forecast results are gotten with the multiple iterative calculation. The annual and monthly runoff forecast examples of Dahuofang reservoir show the efficiency in the medium and long-term forecast.(4) Study on choice method of normalization formulae and the ideal and non-ideal plans in the multi-objective fuzzy optimization model.Reservoir operation is a fuzzy multi-objective decision problem. So economic, social, environmental and ecological factors should be considered altogether to establish the multi-objective fuzzy optimization model. The choice of normalization formulae and the ideal and non-ideal plans is analyzed and dispersal indices are established to make the approximate choice criterion for normalization formulae to optimize the satisfying decision alternative.(5) Study on reservoir benefical forecast operation mode considering eco-environment objective.This paper analyzes the current eco-environment situation of reservoir downstream river and the characteristic of ecological water requirement. Hydrologic and weather forecast information, reservoir downstream river eco-environment information and the multi-objective decision method are taken into account in reservoir multi-objective benefical forecast operation research. The case study of Dahuofang reservoir shows that the effect of ecological water requirement on the original water supply can be decreased with the benefical forecast operation mode considering ecological objectives.Finally, a summary is given and some problems to be further studied are discussed.更多还原