Numerical Simulation Of Selective Withdrawal Of Deep Reservoir Based On CFD

Selective withdrawal widely used in deep reservoir, theoretical and experimentalstudy results show that the withdrawal layer thickness are expressed by general formula,only the coefficient of C3is different, and the velocity distribution expression withinselective layer thickness are obtained by theoretical analysis and experimental data,respectively. This paper utilized the FLUENT software to custom buoyant turbulent heattransfer model, to determine the simulation method, to simulate the U.S. Army EngineerWaterways Experiment, the error between simulation and theory is0.78%.Expanded the model for actual reservoir, the simulative result of Jin-pen reservoirshow that: when selective withdrawal, the flow can be divided into three regions,suction, inertia—buoyancy and viscous—convection area, the achievement by theoryand experiment is applicable only to inertia-buoyancy region, not to other two region;under the condition of the stability of the flow, withdrawal layer thickness is23.80m, itis recommended that the C3=1.5. By analyzing the different linear temperaturedistribution, selective withdrawal characteristic exist differences; the change of orificesize (within the scope of the simulation of orifice size) is no substantive effect forselective withdrawal.Selective withdrawal has been widely used in managing reservoir waterwithdrawal. However, theoretical studies have been confined to using a single, lineardistribution of the density of the water body and have not addressed other circumstancesin detail. In this paper, using the Wood’s analysis method, the thicknesses of theselective withdrawal layer using a bilinear density distribution was determined, and thecoefficient C3was obtained. By FLUENT of CFD software, the vertical domain of selective withdrawal using a simulated bilinear density, the coefficient C3and therelevant velocity distribution were obtained. The results show that the thickness of theselective withdrawal layer was similar to the corresponding theoretical value, with thescope of the analysis of distance, the error value is between0.3%~13.4%, it isrecommended that the C3=1.656. The velocity analysis results also agreed with the testresults of selective withdrawal experiments conducted by the waterway station of theU.S. Army Corps of Engineers.Water intake with fixed height limits the application of selective withdrawaltechnology in a certain degree. This study proposes a technological idea to installbaffles and improved baffles on water intake. Through the rotation of upper and lowerbaffle, poor water layer can be blocked. A computational fluid mechanics (CFD) modelfor the upper baffle on water intake is constructed. The results show that the baffleinstalled on the upper part of orifice can reduce the withdrawal layer thickness and flowon the upper part of orifice centre. Thereby, the withdrawal flow on lower part ca n beindirectly increased, the8m upper baffle decline-30°, the distribution between upperand lower is maximum, it is1:4.3, but the4m baffle is most economic and suitable, thedistribution between upper and lower is1.393;4m upper improved baffle， the biggestdistribution is1:4.78, for the lower baffle the biggest distribution is3.97:1. While baffleand improved baffle, inclining angle are the important factors to influence thewithdrawal layer thickness and flow distribution. Compared with baffle, it shows thatimproved baffle in reducing withdrawal layer thickness is superior to baffle, and inaspect of flux, upper and lower flow distribution ratio would be further enhance.Therefore, the adjusting range of selective withdrawal can be economically enhanced byinstalling baffles and improved baffle on water intake.