Optimization Design And Numerical Simulation Of Gas-liquid Mixing Pump

Multiphase transportation technology is an efficient and economical way ofoilfeld development in recent years. As the core equipment of multiphasetransportation technology, multiphase flow pump is a set of conventional liquidpump and gas compressor performance in one of the multiphase transportationdevice. As the core component of the multiphase pump, the impeller impact thestability of fluid flow in the pump and the pump efficiency directly. The impeller’sperformance influence the efficiency of the pump directly. Therefore, it is of greatsignificance to improve the multiphase pump’s efficiency through do optimizitionon impeller blade and improve its work performance.The thesis is a optimized design of multiphase pump blades which based on theCFD numerical computing, neural networks and optimization algorithms combining.In this method, do parametric design on the impeller blade’s work face and back,select12control points which has been parameterized as optimization variables,determine3values of optimization variables, and achieve the purpose of change theblade profile by changing the values of optimization variables. After determine theoptimized variables and their values, take the optimization variable as factors andthe3values as level, use orthogonal experimental design table L27₃₁3design aset of27experimental program. Modeling each group of program, using fluentsoftware to predict the efficiency of the group of programs. The27-group programas sample data of the neural network,20of them as a network of training samples,the rest are as the prediction samples. Take12optimization variables as the networkinput and computational efficiency as the network output. Take the trained neuralnetwork as the fitness function of genetic algorithms and particle swarmoptimization. Using the global optimization capability of genetic algorithms andparticle swarm optimization to optimize the parameters of the multiphase pumpblades. Do modeling and simulations on the optimal solution of two optimizationalgorithms, compare the results, and from the comparison results, we can see theoutput of the genetic algorithm solution is more better, so we take the output of thegenetic algorithm solution as the optimal model of the multiphase pump blades. Donumerical simulations on before and after multiphase pump impeller optimization atthe air rate of0.1,0.3,0.5,0.7,0.9, and contrast to the optimization results.From the optimization results, we can see the efficiency and head of theoptimized multiphase pump are increased at each air rate. The optimization effect ismost obvious at the design conditions, and the optimized computational efficiencyimproved1.91%. This also proved the optimized design method has a certain degreeof reliability to some extent. The optimal design method is for a wide range, strongrobustness, and strong engineering applications.This thesis also based on cascade theory, velocity triangle, multiphase flowtheory, preliminary diseussion the corresponding relationship between the importantstructual parameters and performance parameters of the multiphase pump. Derivedthe expression of the multiphase pump flow, head, and efficiency. Provid atheoretical basis to futher improve the multiphase pump performance.