节点文献

感潮河段潮位与洪水预报及上游水库控制模式研究

Study on the Flood and Tide Forecast of Tidal River and Upstream Reservoir Inference Operating Mode

【作者】 郭晓亮

【导师】 王本德; 王国利;

【作者基本信息】 大连理工大学 , 水文学及水资源, 2009, 博士

【摘要】 在世界范围内,感潮河段地处的三角洲是社会经济活动最为活跃的地区,并且由于同时受潮水和洪水的影响,也最有可能造成更大的洪灾损失。因此,研究减少感潮河段洪灾损失的工程和非工程措施也一直受到重视。现代监测、通信和预报技术的跨跃式进步,为采取更有效的非工程措施提供了前所未有的便利条件,也使得非工程措施越来越受到关注。为实现感潮河段防洪减灾目标,必须解决好两个关键问题:一是对感潮河段潮水和洪水的科学预测;二是基于该预测,通过对上游水库的控制运用进行错峰、错潮,从而实现防洪减灾的目标。本文即围绕上述关键问题开展研究。以作者参与的科研课题为支撑,以辽河油田的防洪减灾为背景开展研究,结合辽宁地区的实际洪水及辽河下游盘锦、营口地区的潮位特点,围绕潮水、洪水预报和上游水库实时水位和下泄流量控制对下游的防洪安全等问题进行研究,取得的主要研究与应用成果如下:1.对高潮潮位进行混沌相空间重构,初步证明营口站的高潮潮位具有混沌特性,为潮位的混沌分析预测奠定理论基础。相空间重建是一切混沌分析和研究的基础,首先对潮水动力系统的相空间重构,然后在重构的相空间中对潮水系统采用功率谱定性分析法,采用饱和关联维数法和Lyapunov指数定量分析法识别其混沌特性,通过验证营口站高潮潮位的混沌特性,为其混沌时间序列预测创造条件。2.建立了混沌时间序列的模糊支持向量机预测模型,提高了高潮潮位预报精度。在混沌相空间重构和混沌特性辨识的基础上,针对混沌时间序列固有的确定性和非线性特点,以传统的支持向量机预测模型为基础,通过引入模糊隶属函数来考虑时间序列随时间推移引起的内在变化,建立了混沌时间序列的模糊支持向量机预测模型,并在实例计算中,讨论了模糊隶属度参数变化对预报结果的影响。对营口高潮潮位时间序列实例计算表明,该方法具有较好的实用性。同时,基于成因分析引用模糊隶属度函数,更好的反映了近期数据对预测结果的影响,也使得外推预测更可信。研究还表明,该方法在水文月径流时间序列的预测中同样有效。3.采用了改进相应涨差模型及Kalman滤波实时校正方法,预报感潮河段洪水和无遥测降雨资料水库洪水,为水库实时控制运用提供了坚实的信息基础。感潮河段的洪水预报及上游水库的入库洪水预报,是上游水库实时控制运用的信息基础和流域防洪减灾的关键。首先,本文研究了在相应涨差模型计算中,引入Kalman滤波实时校正方法。在感潮河段的洪水预报中,采用了改进相应涨差模型的水位形式,验证了此方法的有效性。在无遥测降雨资料水库洪水预报中引入改进相应涨差模型,并对其在水库中的应用作了进一步改进。在流域上游大伙房水库的计算实例中,使用了改进相应涨差模型的流量形式作预报,同时在实例中结合Kalman在水文预报中初值和参数难以选取的问题,分别对Kalman滤波的初值P(0|0)和X(0|0)及噪声协方差阵Q和R给予讨论,并给出确定方法。实例证明该方法是有效的。4.基于贝叶斯理论分析了上游水库考虑降雨预报的风险,并建立了水库水位及泄量控制模式,其运行结果既可增加洪水资源利用量,又能保障感潮河段及周边区域的防洪安全。基于水文现象的不确定性和对预报信息的利用,上游水库的实时控制运用决策中必须进行相应的风险分析。本文应用贝叶斯定理,在互斥完备事件条件下,统计推断不同设计量级暴雨事件发生的概率,证明了作者所在的科研团队在水利部重大项目研究中风险分析基本结论:“设计标准洪水是预报无雨或小雨条件下的不可能事件;设计标准洪水是中雨或以上量级降雨预报漏报无条件下的可能事件”新理念及假定的合理性,进一步将假定更确切的表征为“设计标准暴雨事件E是各级降雨预报漏报条件下的随机事件,在‘无雨’预报的条件下发生漏报的概率趋于零,属实际不可能事件;在大雨以上量级预报条件下发生的概率极高,为防洪安全计,实时调度中应视为可能事件”,并为构建水位与泄量实时控制推理新模式提供理论依据。最后,以盘锦、营口上游的葠窝水库为例,建立的水库水位与泄量实时控制推理新模式,以该模式指导葠窝水库的实时控制,既可增加洪水资源利用率,又能为流域下游油田安全提供保障。5.成果的实用化—感潮河段洪水潮水预报预警系统的设计与开发。理论的价值在于应用。本文在深入分析辽河油田潮水洪水预报预警特性和计算机网络现状的基础上,选择恰当的软件体系结构,设计开发了实用化的“辽河油田潮水洪水预报预警系统”。系统既结合了传统C/S软件体系结构的交互性强,适宜大量的数据处理,便于集中管理等优势,又引进新近发展起来的B/S软件体系,发挥了该体系结构的访问方便,可扩展性强、易维护等优势。采用集成式的设计与开发方法,将GIS组件技术以及其它工具软件的功能与系统的设计开发有效集成,既缩短了开发周期,又提高了成果的实用性。该系统实现了实时数据、空间数据、属性数据的交互计算、统计、查询、预测等功能,结合“辽河油田潮水洪水预报预警系统”详细阐述了各关键技术及其应用。最后对全文做了总结,并对有待于进一步研究的问题进行了展望。

【Abstract】 World widely, great parts of socio-economic activities happen in deltas, where tides and floods affect more frequently. As the results, tides and floods may cause much more damage to deltas than other places. To eliminate flood and tide damage, both engineering and non-engineering measures were studied in the past decades. Modern progress in monitoring, communications and forecasting and computer technology provide a hopeful alternative for non-engineering measures playing an mort important role.To successfully control flood in deltas, two key issues must be resolved. One is the forecast of tidal and flood. The other is forecast-based upper stream reservoir operation. Thus, focusing on the key issues, this paper studied tide and flood forecast methods, risk analysis, and reservoir inference dynamic operating mode. Taking Panjin, Yingkou as backgrounds, this paper studied related key issues, and the main progress is summarized as follows:1. Chaos phase space of YingKou high tidal level is reconstructed and its chaotic characteristics are validated, which serves as the foundation for high tidal forecasting.Phase space reconstruction is the foundation of the chaos analysis and forecasting. Firstly, the phase space of high tidal is reconstructed; Then, the power spectrum mothed for qualitative analysis is studied; thirdly, G-P method, Lyapunov Exponents for quantitative analysis are used for validation of the YingKou high tidal level. These steps provide theoretic possibility for high tidal level chaotic system forecasts.2. The fuzzy SVM of chaotic time series model is developed, and the forecasting precision of high tidal level chaotic system is improved.Analyzing the characteristic of certainty and nonlinear of chaotic system, the prediction model of chaotic flow time series using fuzzy support vector machine is developed. This model introduces fuzzy membership function for considering the inner mechanism changed with time. Effects of fuzzy membership function parameter are discussed in the example. The results of Yingkou high tidal level forecaste show that the model is of good performance. Moreover, the introduction of fuzzy membership function reflects the effect of the recent data and makes the forecasting results credible. On runoff time series forecasting, this method shows a good result, too.3. Corresponding rising difference model and Kalman filter real-time correction method are used for tidal river flood forecasting and remote sensing rainfall data absence reservoirs, this paper studied tide and flood forecast method. The results play an important role for real-time reservoir operation decision making.Flood forecaste in the tidal river and upstream reservoir is the foundation of the real-time reservoir operation and basin wide flood disaster elimination. Firstly, Kalman filter real-time correction method is introduced into Corresponding rising difference model. The developed new model is applied for the flood forecast in the tidal river, and the validity of this approach is confirmed. The model is used in the telemetry rainfall absence reservoirs, and its application in the reservoir was further improved. For Dahuofang reservoir where part remote censoring data is absent, flood forecast is studied. While forecasting, initial value P(0|0) and X(0|0) of Kalman filter and noise covariance matrix Q and R are confirmed. The forecast results prove the model to be effective.4. Dynamic upstream reservoir real-time operating mode based on Bayes theorem can not only raise floodwater development, but also eliminate flood damage.Based on the uncertainty of hydrology and use of forecasted message, the risk analysis in real-time upstream reservoir operation must be taken account into. This paper applies Bayes theorem, on the condition of mutually full fledged events, statistical computes the probability of different design rainstorm flood event, and the new idea and hypothesis". This proved the conclusion, studied by my research group in an national key research project. The proved conclusion is that it is impossible to happen designed rain while forecasted rain is no rain or little rain; And the designed rain possibly happen only while forecasted rain is moderate or heavy rain. This conclusion gave an further confirm for the assumption, which is that design rainstorm flood event E is random events of rainfall forecasting omissions, when ’no rain’ is forecasted, the probability is extremely small, tends to zero. And the probability is high under the condition of the heavy rain forecasting" these conclusion made from a National key program is certificated. The above risk analysis results provide important data for reservoir real-time operating decision making. At last, the real-time reservoir elevation and discharge mode is developed. Taking ShenWo reservoir as an example, an inference operating mode is developed, which play an important role for flood resources development and flood control for PanJin and YingKou and LiaoHe Oilfield.5. Application - design and development of the flood and tide forecast and warning system.The value of theory lies in its application. Based on the analysis of the LiaoHe Oilfield flood forecast and early warning systems, the present progress of computer network, etc. "LiaoHe Oilfield Flood and Tide Forecasting and Warning System" is designed and developed. Combining the interactive, suitable for large data processing, facilitate centralized management and other advantages, traditional C/S (Client/Server) software architecture is used. Meanwhile, B/S (Browser/Server) software architecture is introduced, which is convenient, good scalability and maintainability etc. Using integrated design and development method, the system integrates ComGIS, ActiveX. This step can shorten software development period and efficiency increases the practicality value of the achievements. At last, taking "LiaoHe Oilfield Flood and Tide Forecasting and Warning System" as an example, we gave a detailed description of the application of the critical techniques.Finally, a summary is given and some remained problems are discussed.

  • 【分类号】TV124;TV697.13
  • 【被引频次】1
  • 【下载频次】426
  • 攻读期成果
节点文献中: 

本文链接的文献网络图示:

本文的引文网络