Study On Hydraulic Optimization Of The Centrifugal Pump's Impeller Based On Theories Of Boundary Vorticity Dynamics

Centrifugal pump is widely used in the national economic production and life.However,the theory or methods on the hydraulic design of centrifugal pump's impellers are old.There exist some common problems.So,the study of optimization design of impellers is of great significance.Based on the graduate education innovation project of Jiangsu province “Study on the Internal Flow Diagnosis and Hydraulic Optimization of the Centrifugal Pump Impeller Based on the Boundary Vorticity Flux”,this paper applied theories of Boundary Vorticity Dynamics(BVD)to analyze the internal flow of the centrifugal pump and studied the connection between Boundary Vorticity Flux(BVF)and centrifugal pump hydraulic performance parameters.The internal flow parameters are taken as the objectives to study the reliability of Back Propagation Network(BP)and Radial Basis Function Network(RBF)on the prediction of the internal diagnostic parameters of centrifugal pump's impeller.Then,the Genetic Algorithm(GA)and Particle Swarm Optimization Algorithm(PSO)are used respectively to optimize the results gotten from the ANN with single objective.The optimization strategy of multi-objective genetic algorithm applied in this paper is determined.The main contents and conclusions of this paper are as follows:1.Relation Between BVF and Inner Flow Field and Hydraulic Performance of the Centrifugal Pump.Firstly,a double-channel centrifugal pump is used as the object of study,and the local flow details of the internal flow field of the pump are obtained by numerical simulations.The adverse flow conditions near the blade pressure surface and the suction surface are emphatically analyzed.The relationship between BVF distribution and blade flow separation,vortex generation and dissipation,as well as the hydraulic performance is revealed by analyzing the BVF,frictional line and vortex distribution on the pressure surface and suction surface of the blade with BVD.The results show that a better flow condition will be achieved,the flow separation will be suppressed and a higher impeller head and efficiency will be obtained with the lower peak and mean of the BVF on the inner surface of the impeller,the higher the BVF distribution uniform index.2.Application of ANN in Hydraulic Performance Prediction of Centrifugal Pump.Based on the MATLAB platform,the influence on the prediction accuracy of BP neural network and RBF neural network is discussed respectively when the hidden layers of BP and the radial basis function expansion speed of RBF are set differently.And then the error norm of the predicted values and the calculated values is used as the criterion to judge the performance of the two kinds of neural networks,and the optimal ANN is finally selected.The results show that when the hidden layer is 18,the prediction error of BP neural network is the smallest,and the neural network structure is the best.On the other hand,when the radial basis function expansion speed is 0.3,the prediction error of RBF network is the smallest and the neural network structure is the best.Comparing the results gotten from the two kinds of ANN,the result shows that the prediction error of RBF neural network is smaller,the running time is shorter and the running stability is higher.3.Research on Hydraulic Optimization Design of Centrifugal Pump Impeller Based on Intelligent Optimization Algorithm.The mathematical model for the hydraulic optimization problems of the centrifugal pump impeller,taking the internal flow parameter BVF as the optimal object,is studied and established.The corresponding constraint conditions and optimization variables are determined.The range of optimization target and variables are determined,as well as the coding and decoding scheme.Comparing the applicability of GA and PSO in the hydraulic optimization of centrifugal pump impeller,the algorithm is improved to obtain the optimal algorithm suitable for hydraulic optimization of centrifugal pump impeller,and an efficient and reliable optimization strategy is established.The results show that taking BVF as the optimal object and using GA and PSO to solve hydraulic optimization problem,the result of GA is better than that of PSO.But taking only the BVF peak value as the optimal object,the uniform index couldn't be guaranteed to be the best.So,the multi-objective genetic algorithm is used to solve the problem.And the internal flow of the optimized impeller 3 and the impeller's BVF uniform index are both improved.Also,it's better than the relevant parameters of the impeller 2,the pump's head and efficiency are correspondingly improved.The multi-objective genetic algorithm,combined with RBF neural network,has strong global optimization ability,its running time is short,the result has high precision accuracy as well.