| Sedimentation in the reservoir affects the normal service life of the reservoir and restricts the important function of the reservoir to "promote water conservancy and eliminate water damage".The analysis of the variation characteristics of the inflow and sediment and its prediction research are the premise and basis for evaluating the degree of influence of sediment deposition on the reservoir and the risk evaluation and selection of the dispatching plan.The research on these four aspects is closely related and is an important prerequisite for reservoir managers to solve the problem of sediment deposition.The Kizil Reservoir is located in Baicheng County,Aksu City,Xinjiang,China.It is a mountain-type reservoir built in the upper reaches of the Weigan River Basin.Since the reservoir was built and put into use in 1992 for 30 years,the reservoir area has been seriously affected by sediment deposition.The loss of storage capacity deposition has exceeded the expected value.The problem of sediment deposition has become a major problem that the reservoir is currently facing.In order to effectively put forward the scheduling scheme of reducing sediment deposition in the reservoir,this paper studies from four aspects: the variation characteristics of reservoir water and sediment,the prediction model of water and sediment,the impact assessment of sediment deposition and the risk assessment of the reservoir scheduling scheme.The specific conclusions are as follows:(1)Both the inflow runoff and the sediment load at the dam site of the Kizil Reservoir show an increasing trend.The years of sudden change in water and sediment were 1994 and 1992,respectively.There are three time scales for inflow runoff,namely 6a-9a,11a-15 a,26a-30 a,and their quasi-periods are 7a,13 a,28a;there are also three time scales for annual sediment transport,which are 2a-5a,10a-13 a,25a-28 a respectively,the quasi-periods are 4a,12 a,26a.The 13 a and 28 a cycles of annual runoff are closely related to solar activity,while the 7a cycle has little relationship with solar activity.(2)The influence of precipitation changes and human activities on runoff and sediment transport was analyzed by the double cumulative curve method.The results showed that after the runoff abrupt year,the measured cumulative runoff increased by 82 million cubic meters.The contribution rates of the increase in runoff are-385.37% and 485.37%,respectively.after the sudden change in sediment transport,the measured cumulative sediment transport increased by 5.859 million tons,and the contribution rates of human activities and precipitation to the increase in sediment transport are 26.43% and 73.57%,respectively.(3)There is a telecorrelation relationship between Xinjiang precipitation and the ENSO phenomenon,and the water and sediment in the Kizil Reservoir are indirectly affected by the ENSO phenomenon.The land use type and vegetation coverage in the upstream area of the reservoir fluctuated greatly in both the base period and the change period.At the same time,the water resource utilization mode changed,the groundwater recharge decreased,and the natural vegetation around the oasis was degraded.These changes reflect the obvious influence of human activities from the side.As the global climate warms,the melting rate of glaciers accelerates,the river flow increases,and the frequency of floods and the peak flow of floods also gradually increase.At the same time,the soil quality in this area is relatively loose,and the soil types with a large proportion are very prone to soil erosion.The comprehensive influence of various factors leads to the increase of the inflow runoff and sediment transport of the reservoir to a certain extent year by year.(4)The two combined prediction models of GA-SVR and PSO-SVR optimize the parameters C and g of SVR on the basis of SVR.Combined with the LSSVR algorithm and the SVR algorithm,the four prediction models are applied to the prediction of the annual runoff sequence and the annual sediment transport sequence of the Heizi Reservoir hydrological station.The advantages and disadvantages of the constructed model are evaluated by four performance evaluation indicators,such as root mean square error,average percentage error,correlation coefficient and Nash coefficient.The results show that the LSSVR algorithm has advantages in predicting annual runoff and annual sediment transport,reducing the prediction error and improving the prediction accuracy.The LSSVR algorithm has achieved good results in the forecast of annual runoff and annual sediment transport at the Heizi Reservoir hydrological station.(5)In the reservoir sedimentation impact evaluation model based on the improved cloud model,the evaluation results of the index layer are obtained and the degree of reservoir sedimentation impact is accurately and specifically described.It can deeply and specifically analyze each sediment deposition impact index,and find out the main factors that affect the deposition.Combining the cloud model and genetic algorithm in the computer field to solve the problem of reservoir sedimentation impact evaluation,the accuracy of the calculation results is higher than that of the traditional method,which is more in line with the actual needs of the project.The proposed improved cloud model multi-level fuzzy comprehensive evaluation model of reservoir sedimentation impact is applied to the evaluation of sedimentation impact of Kizil Reservoir.The level of siltation impact is severe.Reservoir managers should solve the problem of reservoir siltation from three aspects: reducing the sediment content of the upstream water,reducing the sedimentation of river sediment in the reservoir,and clearing the existing sediment in the reservoir.(6)The analytic hierarchy process is used to construct the index layer,the comment layer,the influence layer,the factor layer and the state layer from the top to the bottom to form the risk evaluation system of the reservoir dispatching scheme.The constructed multi-level fuzzy comprehensive evaluation model analyzes the risk level of the reservoir dispatching scheme,ignoring the randomness and discreteness in the evaluation system.The cloud model theory was introduced to overcome the defects of the evaluation model,and an improved multi-level fuzzy comprehensive evaluation model of reservoir dispatching scheme risk based on cloud model was established,which provided a simple and effective new method for qualitatively analyzing the risk level of reservoir dispatching scheme.Through the calculation of the constructed model,it is determined that the risk level of the conventional dispatching scheme of Kizil Reservoir is general risk,and the risk level of the optimal dispatching scheme to be adopted is slight risk,which provides reference value for the later management and implementation of the optimal dispatching scheme. |
PDF Full Text Request