【摘要】：隨著社會的進(jìn)步和人民生活水平的提高,人們對水資源的需求也越來越高；作為城市供水的重要水源,河流、水庫的健康生存發(fā)展至關(guān)重要。泥沙一直以來都是我國水庫運(yùn)用中的重要問題,而且泥沙對河道的危害也很大。泥沙預(yù)報(bào)可以提前了解泥沙的運(yùn)動規(guī)律,做到防患于未然,將一切損失降到最低,因此泥沙預(yù)報(bào)就顯得至關(guān)重要。 本文主要針對日益嚴(yán)重的水庫、河道淤積問題,以大凌河流域的白石水庫作為研究對象,對水庫的入庫沙量、水庫下游站的沙量及沙量過程進(jìn)行預(yù)報(bào)。主要研究內(nèi)容如下： (1)基于BP神經(jīng)網(wǎng)絡(luò)的水庫入庫沙量預(yù)報(bào)。針對水庫淤積日益嚴(yán)重的問題,對入庫場次洪水進(jìn)行分析研究。首先分析了影響入庫洪水的各個(gè)因素,次降雨、降雨過程、時(shí)段降雨以及洪水地區(qū)組成與入庫沙量之間的關(guān)系,然后建立入庫沙量預(yù)報(bào)模型,以最大4小時(shí)降雨量、降雨均勻性系數(shù)和前期影響雨量作為模型輸入變量,初始為隨機(jī)權(quán)重,經(jīng)過5000次運(yùn)算進(jìn)行模擬。結(jié)果表明該模型可以有效的對入庫沙量進(jìn)行預(yù)報(bào)。 (2)基于線性回歸和神經(jīng)網(wǎng)絡(luò)的下游河道沙量預(yù)報(bào)。對于河道淤積問題,水庫泄洪沖淤是一個(gè)常用的方法。利用建庫前的歷史資料,對水庫下游站義縣站的沙量進(jìn)行預(yù)報(bào),通過線性回歸和神經(jīng)網(wǎng)絡(luò)這兩種方法進(jìn)行模擬,兩種模型都以上游朝陽站和迷力營子站的洪峰、洪水總量和平均含沙量為輸入條件,結(jié)果表明,神經(jīng)網(wǎng)絡(luò)的預(yù)報(bào)結(jié)果精度更高,其優(yōu)于線性回歸方法。 (3)基于相似推理理論的下游河道沙量過程預(yù)報(bào)。通過對流域內(nèi)的歷史場次洪水進(jìn)行相似性分析,得到兩兩相似的洪水,用平均含沙量指標(biāo),由一場洪水含沙量過程去推求另外一場洪水的含沙量過程。首先進(jìn)行指標(biāo)相關(guān)性分析,根據(jù)累計(jì)貢獻(xiàn)率大于等于85%的前3個(gè)指標(biāo)從而得到新的指標(biāo)；然后以新的指標(biāo)對沙量進(jìn)行模擬。分析可得,由于流域歷史資料較少,大部分的洪水過程預(yù)測結(jié)果一般,對于流域資料全的流域,該方法結(jié)果應(yīng)當(dāng)會更好。
[Abstract]:With the progress of society and the improvement of people's living standard, people's demand for water resources is higher and higher. As an important source of urban water supply, the healthy survival and development of rivers and reservoirs are very important. Sediment has always been an important problem in the reservoir operation in China, and the harm of sediment to the river is also great. Sediment prediction can be used to understand the movement law of sediment in advance, so as to prevent trouble and minimize all losses, so it is very important to forecast sediment. Aiming at the increasingly serious problem of reservoir and river siltation, this paper takes Baishi Reservoir in Daling River Basin as the research object, forecasts the amount of sediment into reservoir, the quantity of sediment and the process of sediment quantity in downstream station of reservoir. The main contents are as follows: (1) the prediction of reservoir sediment based on BP neural network. Aiming at the increasingly serious problem of reservoir siltation, the flood of reservoir entry site is analyzed and studied. This paper first analyzes the relationship between the factors affecting the inflow flood, the secondary rainfall, the rainfall process, the rainfall during the period, and the composition of the flood area and the amount of sediment in the reservoir, and then establishes the prediction model of the amount of sediment entering the reservoir for the maximum rainfall of 4 hours. The rainfall uniformity coefficient and the early influence rainfall are taken as the input variables of the model, the initial weight is random, and the model is simulated by 5000 operations. The results show that the model can effectively predict the amount of sediment entering reservoir. (2) based on linear regression and neural network, the prediction of sediment quantity in downstream channel can be carried out. Reservoir flood discharge is a common method for river channel siltation. Based on the historical data before the reservoir was built, the sediment volume of Yixian station, a downstream station of the reservoir, was forecasted. The two models were simulated by linear regression and neural network. Both models were based on the Hong Feng of Chaoyang station and Mailiying sub-station in the upper reaches of the reservoir. The results show that the prediction accuracy of the neural network is higher than that of the linear regression method. (3) based on the similarity reasoning theory, the prediction of sediment volume in the downstream channel is predicted. Based on the similarity analysis of the historical floods in the basin, the similar floods are obtained. By using the average sediment content index, the sediment content process of another flood is derived from one flood sediment content process. Based on the first three indexes whose cumulative contribution rate is more than 85%, a new index is obtained, and the new index is used to simulate the sediment quantity. It can be concluded that because of the lack of historical data, most of the flood process prediction results are general, and the result of this method should be better for the basin with all the basin data.
【學(xué)位授予單位】：大連理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2014
【分類號】：TV122;TV145

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 鄧新民,李祚泳;流域年均含沙量的B-P網(wǎng)絡(luò)預(yù)測模型及其效果檢驗(yàn)[J];成都?xì)庀髮W(xué)院學(xué)報(bào);1997年02期

2 Horacio TONIOLO;;Numerical simulation of sedimentation processes in reservoirs as a function of outlet location[J];International Journal of Sediment Research;2009年03期

3 ;Control of coupling among three major factors for formation of high-efficiency gas reservoir——A case study on the oolitic beach gas reservoir in Feixianguan Formation in the northeast Sichuan Basin[J];Chinese Science Bulletin;2007年S1期

4 韓其為,何明民;泥沙運(yùn)動統(tǒng)計(jì)理論[J];科學(xué)通報(bào);1980年02期

5 韓其為,何明民;水庫淤積與河床演變的(一維)數(shù)學(xué)模型[J];泥沙研究;1987年03期

6 樊爾蘭;懸移質(zhì)瞬時(shí)輸沙單位線的探討[J];泥沙研究;1988年02期

7 繆鳳舉,丁六逸,錢意穎;“洪水排沙、平水發(fā)電”——三門峽水庫汛期發(fā)電運(yùn)用方式的研究[J];泥沙研究;2001年02期

8 唐政洪,蔡強(qiáng)國;侵蝕產(chǎn)沙模型研究進(jìn)展和GIS應(yīng)用[J];泥沙研究;2002年05期

9 秦毅;石寶;李楠;凌燕;劉超;;含沙量預(yù)報(bào)方法探討[J];泥沙研究;2010年01期

10 孫連成;黃毅峰;趙會民;;人工沙灘親水段泥沙淤積試驗(yàn)研究[J];泥沙研究;2011年04期

，

本文編號：2174712