| This thesis was supported by National Natural Science Foundation of China(Grant No.51879121)and Natural Science Foundation of Jiangsu Province(Grant No.BK20190851).This thesis selected the vertical inline pump commonly used in fields such as high-rise building pressurized water supply,urban heating system circulation,fire pressurization and equipment matching as the research object.The rotor-stator interaction between the impeller and the volute is an important reason for the pressure pulsation in the centrifugal pump.The optimization of geometric parameters of the impeller can effectively reduce the pressure pulsation caused by the rotor-stator interaction.Due to the excessive geometric parameters of the impeller and long unsteady calculation period,the optimization method aimed at reducing the pump pressure pulsation is not mature.The range of pump parameters is mostly continuous.The optimization method based on discrete value is more targeted for experienced designers and can effectively reduce calculation time.The rounding of pump design parameters is conducive to mechanical processing and manufacturing.Therefore,it is important to optimize the unsteady characteristics of pumps with multiple discrete variables.At present,there are relatively mature solutions for linear discrete optimization problems,but there is still a lack of effective solutions for discrete optimization problems with a high degree of nonlinearity.Therefore,it is necessary to research an optimization method with discrete variables for the unsteady characteristics of the centrifugal pump and reduce the pressure pulsation of the centrifugal pump,which has important academic significance and engineering value.The main research contents and innovations of this thesis are shown as follows.(1)The research status of unsteady flow and hydraulic optimization design methods of pumps were summarized from the current literatures,and the mathematical model and improvement strategy of the standard genetic algorithm were introduced.A modified discrete genetic algorithm MDGA for the optimization of discrete variables was proposed.The main improvement focused on the position of the discrete variable value and the adaptive genetic probability.The modified MDGA was compared with the classical genetic algorithm,and was tested and verified by four classical test functions.The results showed that the modified MDGA accurately found the global optimal value and performed better in terms of convergence accuracy and speed.(2)An open test rig was built to measure the pump performance and pressure pulsation of the vertical inline pump.The frequency domain characteristics of pressure signals obtained by the experiment and numerical simulation were analyzed based on the Fast Fourier Transform,and the time-frequency domain characteristics of the experimental pressure signals were analyzed based on the Wavelet Transform.The results showed that in the inlet pipe,the pressure pulsation was dominated by 2 times of the shaft frequency under low flow and design flow conditions,and the shaft frequency was dominated under the large flow condition.The main frequency of pressure pulsation in the volute was mainly the blade-passing frequency,and the pressure pulsation amplitude at the blade-passing frequency was relatively continuous.(3)The impeller of the vertical inline pump was optimized by combining the approximate model and the modified MDGA.Six geometric parameters of the impeller were selected as optimization variables,the efficiency under the design condition of the pump was the optimization objective,and the head was the constraint condition.90 group of sample data were generated by Latin Hypercube Sampling to train high-quality response surface model.The surrogate model was finally solved by the modified MDGA.The study found that the response surface model could accurately fit the mathematical relationship between efficiency and geometric parameters of the impeller.The efficiency of the pump and the efficiency of the impeller were both significantly improved after optimization.The maximum increase of pump efficiency under 0.6Q_d,1.0Q_d and1.4Q_d conditions was 7.05%,5.98%and 3.38%,respectively.The input power of the pump and the energy loss in the impeller and volute were both reduced.After optimization,the flow loss area and the overall flow loss were significantly reduced.The blade outlet angle,the impeller outlet diameter,the impeller outlet width,and the blade wrapping angle had a greater impact on the performance of the vertical inline pump.(4)The pressure pulsation intensity coefficient at the volute tongue of the pump was selected as the optimization objective,and the efficiency and head were set as the constraint conditions.40group of sample data were generated by Latin Hypercube Sampling.The feedforward artificial neural network was adopted to fit the relationship between the pressure pulsation intensity coefficient and the geometric variables of the impeller,and the response surface model was used to fit the relationship between the constraints and the design variables.Based on the modified MDGA,the three approximate models were optimized in the discrete space,and the optimal scheme was verified by numerical simulation.As a result,it was found that the efficiency under all flow conditions were improved significantly after optimization.The efficiency increase was the largest under the design flow condition,which is 2.97%.Under the design flow and large flow conditions,the pressure pulsation intensity coefficient at the volute tongue were significantly reduced,which improved the flow separation in the impeller and the volute.The pressure distribution became more uniform,and the gradient of the pressure pulsation intensity was also reduced. |
PDF Full Text Request