Research On Collaborative Optimization Of Impeller And Casing Profile Of Liquid-Ring Vacuum Pump

Liquid ring pump is a kind of fluid machinery used for pumping gas.Under the action of high-speed rotation of the impeller,the gas and liquid two phase flow in the pump approximately presents a separation state.The periodic change of the volume of the closed space composed of the gas-liquid interface and the impeller and the casing makes the pump constantly inhale and discharge the gas.Impeller and casing profiles have an important influence on the internal flow and hydraulic performance of the pump.Reasonable design of impeller and casing profiles can effectively improve the internal flow structure of the liquid ring pump,improve efficiency and suction performance.In this paper,a multi-objective collaborative optimization analysis of impeller and casing profiles of liquid ring vacuum pump is proposed by combining experiment and numerical simulation.Parametric control of impeller and casing profiles is carried out by Bezier curve method and direct free-form surface deformation method(DFFD),and Latin hypercube(LHS)experimental design method is adopted.A radial basis function(RBF)neural network surrogate model was constructed for the relationship between impeller and casing profiles parameters and hydraulic performance of liquid ring vacuum pump.The multi-objective optimization of liquid ring vacuum pump performance was carried out based on NSGA-II algorithm,and then the internal flow and performance of liquid ring vacuum pump were compared before and after impeller and casing profiles optimization,which provided a reference for the design of high performance liquid ring pump.The main contents and conclusions of this study are as follows:(1)The internal flow field and performance of the liquid ring pump were studied by combining the experimental method and the numerical simulation method.The results show that the experimental results are similar to the numerical simulation results.Therefore,the numerical simulation method can be used to carry out the collaborative optimization analysis of the impeller and casing profiles of the liquid ring pump.LHS sampling method was used to sample the parameters of impeller and casing profiles of liquid ring vacuum pump.According to the numerical simulation results of each sample,the RBF neural network surrogate model was established between optimization variables of impeller and casing of liquid ring vacuum pump with efficiency and inlet vacuum.(2)According to the RBF surrogate model,the single factor analysis was carried out to obtain the influence law of each variable on the liquid ring pump performance.It was found that the impeller and suction section optimization variables had a greater impact on the liquid ring pump performance,while the exhaust section profile had a smaller impact on the liquid ring pump performance.By using the correlation analysis method,the results show that the efficiency is positively correlated with the impeller outlet blade angleβand the radial displacement R_Fof the suction section,and negatively correlated with the radial displacement R_Hof the exhaust section.Inlet vacuum degree is positively correlated with blade wrap angleφand suction segment radial displacement R_G.According to the sample point classification analysis method,the high efficiency sample point blade types mainly include large wrap angle and large outlet blade angle,while small outlet blade angle will reduce pump efficiency.The sample point blade types with high inlet vacuum degree are mainly large wrap angle.The high efficiency casing profiles is mainly inspiratory contraction type,while the high vacuum casing profiles is mainly inspiratory expansion type.(3)Combined with the RBF neural network agent model and NSGA-II multi-objective genetic algorithm,the collaborative optimization of liquid ring vacuum pump impeller and casing profiles was carried out.The results show that the multi-objective collaborative optimization significantly improves the performance level of liquid ring vacuum pump.Under different inlet flow rates,the efficiency of the optimal model J is increased by 2.7%,4.1%and2.5%compared with the initial model O,and the inlet vacuum of the optimal model K is increased by 7.8%,11.2%and 11.7%compared with the initial model O.(4)The influence of impeller and casing profiles of liquid ring vacuum pump on the flow inside the pump is analyzed.The results show that the change of impeller and casing profiles will cause the change of the flow field in the pump.The larger outlet blade angle,the shrinking suction section profile and the expanding exhaust section profile can reduce the outlet reflux,thus improving the efficiency.The larger blade wrap angle and casing radial size as well as the expanded suction section casing profiles are helpful to improve the inlet vacuum degree of the pump.(5)The coupling optimization,independent optimization and step by step optimization of impeller and casing profiles were analyzed.The results show that coupling optimization is better than independent optimization.In the impeller optimization,the performance is improved by changing the state of the impeller and the working medium at its outlet,while the casing optimization indirectly affects the performance by changing the overall shape of the liquid ring.The two can further improve the optimization effect through coupling.In addition,under the same conditions of the surrogate model,the surrogate model will make the same response to the same disturbance,so the coupling optimization results of impeller and casing are basically consistent with the step by step optimization results.