| Vertical inline pump is a single-suction single-stage centrifugal pump with an elbow-shaped inlet pipe,its characteristics are small size,large flow rate ranges and convenient installation;hence it is widely applied in areas where installation space is a constraint,such as urban water supply systems and ship transportation.There is however a bent pipe section called the inlet pipe which has different angles and cross sections that results in terrible flow conditions before the impeller.Meanwhile,there is a requirement of wide operating area for inline pumps in actural work,to improve the efficiency of different working conditions,especially for part-load condition,which is of great significance for the energy-saving and emission reduction.The common optimization methods,such as test method or DOE method,which are inefficient or high-cost.Therefore,a multi-condition optimization method of centrifugal pumps with low cost and high precision was designed to broaden the efficient operating area of the inline pump,which are of great academic significance and good engineering value.At the same time,the internal flow conditions of the optimized cases and the original case were compared and analyzed,the design parameters of the excellent inline pumps were summarized,which could be used for reference of the forward design of the vertical inline pump and was of good engineering significance.This thesis was supported by the National Natural Science Foundation of China(Grant No.51879121,51409123).A multi-objective optimization on the curved inlet pipe and impeller of the inline pump was carried out based on the automatic simulation program which established by MATLAB and WorkBench.The main works of the thesis are shown as follow.1.The popular hydraulic optimization methods of centrifugal pumps and the multi-optimization theory were summarized from the current literatures,and the advantages and disadvantiages of different approaches were listed.For different optimization problems,selecting suitable optimizatiom method according to the required accuracy can improve the efficiency of the optimization process and reduce the costs.2.The fundamental principle and mathematical model of multi-objective particle swarm optimization were introduced in this thesis.And the improvement strategies of this algorithm were introduced.Based on the modified parameter control methods and population topology,a modified multiobjective particle swarm optimization algorithm based on particle classification strategy was proposed in this thesis.The population size of modified MOPSO was determined by the attribution of the multiobjective problems,and the velocity and position updating strategies were selected according to the characteristics of particles.In this thesis,two groups of multi-objective problems with different features were applied to test the modified algorithm,and the results were compared with the NSGA-II algorithm which is usually adopted to optimize pumps.The results showed that the modified MOPSO performed better in term of convergence speed and accuracy,and it has better stability for different problems.3.Based on the URANS methods,the flow analysis for the elbow inlet pipe of the inline pump was carried out,and a validation experiment was adopted to ensure the reliability of the numerical results.The results showed that the flow separation occurred at the end of the first bend under the norminal and over-load conditions.,and the separation area developed along the lower side of the inlet pipe towards the outlet,which resulted in complexed secondary flow and the decline of the uniformity of the outflow.For the part-load condition,serious backflow was observed at the outer side near the outlet,which caused a large vortex near the upper side of the inlet pipe.The passage was blocked in by the vortex,which exacerbated the hydraulic losses in the inlet pipe.Meanwhile,the recirculation also leaded to an obvious rise of velocity at the outer side near the outlet,which tended to cause cavitation,hence influenced the stability of the inline pump.4.In this thesis,a multi-objective optimization design method based on the modified multi-objective particle swarm optimization and artificial neural network model for the inlet pipe of a vertical inlin pump was first proposed.Though the optimization process,11 design varibales of the inlet pipe were selected as the optimization variables,and the efficiencies of the inline pump at part-load,norminal and over-load conditions were choosed as the objective functions.Specifically,a program for automatical modeling,meshing and simulation was established using MATLAB code and WorkBench;150 group of sample data were generated by Latin Hypercube Sampling to train high-quality artificial neural networks.The surrogate models were finally sovled for the Pareto solutions by the modified MOPSO.The study found that the design of inlet pipe had only little effect on the performance of the inline pump under overload condition,and the efficiency at the norminal condition decreased along with the increase of the efficiency of part-load condition in the optimal schemes.The maximum increase of efficiency under the condition of small flow,design flow and large flow were 4.96%,2.45% and 0.79%,respectively,and the flow in the inline pump was effectively improved after optimization.5.In this thesis,a multi-objective two-component optimization method based on the modified multi-objective particle swarm optimization algorithm for the inlet pipe and the impeller of a vertical inline pimp was first proposed.Based on MATLAB Code and WorkBench platform,an automatic and direct optimization program for the inlet pipe and the impeller of the inline pump was established.In the optimization process,40 design variables of two components were used as optimization variables,and the efficiencies of the inline pump at part-load and norminal conditions were taken as the objective functions.After optimization,the maximum increase of efficiencies at part-load condition and norminal condition were 8.61% and 7.95%,respectively The results showed that the horizonal length of the inlet pipe was longer when the design has better performance under the condition of small flow.The design with better performance under the design flow rate usually has the inlet pipe with shorter transverse length and the impeller with small outlet blade angle,and the multi-blade design scheme(the number of blades is greater than 6)has better comprehensive performance.The inlet bends with longer transverse length and lower curvature can effectively prevent the backflow vortex under the part-load condition,reduce the hydraulic losses of the inlet pipe,and increase the uniformity of the impeller inflow. |
PDF Full Text Request