Internal Fow Characteristics Analysis And Structure Optimization Of A Double Suction Multistage Split Case Centrifugal Pump

The double suction multi-stage split case centrifugal pump is a centrifugal pump with large flow rate and high head,which is widely used in long-distance water transfer and large-scale water lifting irrigation projects.However,the internal flow of the double suction multi-stage split case centrifugal pump is complex,and there are a lot of vortical structures,which will produce complex hydraulic excitation,cause the unstable operation of the pump,and aggravate the energy loss,so that the overall operation efficiency is low.Through the vortex identification technology,the fine vortical structure and its distribution and evolution law in the double suction multi-stage split case centrifugal pump can be obtained,and then the unsteady flow characteristics in the pump can be more clearly understood,which provides a new idea for the design optimization of the double suction multi-stage split case centrifugal pump.In this thesis,a BB1T200-480 double suction multi-stage split case centrifugal pump is taken as the research object.The internal flow characteristics of the prototype pump are analyzed and studied through the combination of experiment and numerical simulation,and the design of replacing the original interstage with the return channel is proposed.On the basis of the preliminary design of the return channel,artificial neural network and particle swarm optimization are used to optimize it.The main work and innovation of this thesis are as follows:（1）In this thesis,ICEM is used to generate high-precision hexahedral structured grid for the prototype pump,and the convergence of the grid is analyzed by using the GCI criterion based on Richardson extrapolation method.Then,the unsteady numerical simulation of the prototype pump is carried out by using the Detached Eddy Simulation approach,and the grid quality is verified to ensure that the Detached Eddy Simulation can fully analyze the turbulence fluctuation.Finally,the reliability of the simulation is verified by the performance test.（2）Based on the unsteady numerical simulation results of the Detached Eddy Simulation approach,the velocity field distribution characteristics and pressure fluctuation propagation law in the double suction multi-stage split case centrifugal pump under different working conditions are studied.The results reveal that:under the condition of small flow rate,the low efficiency of the pump is due to the existence of a large number of unstable flow structures in the first stage impeller;the distribution of streamlines is disordered and there are a large number of vortical structures in the back channel,which lead to the poor inflow quality of the secondary impeller and seriously affect the performance of the pump;the areas with high pressure pulsation intensity are mainly concentrated in the outlet of the impeller,tongue and its downstream areas.the pressure fluctuation spectrum near the trailing edge of the pressure surface of the impeller shows the characteristics of broadband frequency,and the broadband bandwidth decreases with the increase of flow rate;the blade passing frequency is dominant in the interstage and the volute.In addition,the energy loss in the pump is analyzed quantitatively based on the entropy production theory:when analyzing the energy loss in the pump,the entropy production caused by the wall effect can not be ignored;the loss mainly occures in the volute and the interstage.（3）For the first time,Q criterion,λ₂ criterion and?method are used to capture the vortical structure in the double suction multi-stage split case centrifugal pump.Compared with the traditional Q criterion andλ₂ criterion,the?method can avoid the influence of the wall shear layer and obtain the clearer vortical structure in the pump without the influence of the threshold.Based on the?method,the unsteady evolution characteristics of the vortical structure in the pump are analyzed.The shedding process of the wake vortex near the outlet of the first stage impeller and its impact with the tongue are mainly described.The results show that the rotor-stator interaction between the first stage impeller and the interstage mainly comes from the periodic shedding of the wake vortex.（4）The original two-stage pump is changed into three-stage pump to improve the head of the pump.The numerical simulation shows that the preliminary design can meet the requirements of practical engineering for hydraulic performance.Nine design parameters of the return channel are selected as optimization variables.Based on double-layer feedforward neural network,the approximate model between pump efficiency and optimization variables is established.Finally,the optimal design parameters combination of the return channel is obtained by global optimization based on particle swarm optimization algorithm.The results show that:the double-layer feedforward neural network can accurately reflect the relationship between the pump efficiency and the optimization variables,the deviation between the optimized predicted value and the calculated value is 0.51%;compared with the original model,the efficiency of the optimized model under 1.0Q_d is increased by 3.15%,the head is increased by 4.61m,and the flow situation in the return channel is improved;compared with the two-stage pump,the head is increased by 12.61m under 1.0Q_d,the efficiency is increased by 4.48%,5.65%and 6.55%respectively under 0.6Q_d,1.0Q_d and 1.4Q_d.