Improving Of Design Of Experiment In Centrifugal Pump Multi-operation Design

Design of experiment is a mathematical statistical method by arranging experiment and analyzing experiment results. It can be established that the design goals of mathematical models between the impeller geometry parameter through experiments design. Quantitative value of impeller geometric parameters that meet design requirements obtain by solving the mathematical model. The method of numerical simulation applicant to forecast the performance of centrifugal pump and it has been widely used about the method of test verification. Therefore, there is an error in the prediction of the numerical simulation under small flow and high flow operations. Then it causes inaccurate about the mathematical model from Design of Experimental. Therefore, the accuracy of the performance predictive by numerical simulation has a great of significance for centrifugal pump design. In addition to, the choice of factors, the number of factors levels and the interaction effects between factors has great impact on model accuracy in the process of establishing the mathematical model. When the independent variable factor in the mathematical model is complex and it has more design goals, it will be become very difficult though the direct solution of a mathematical model to calculate the values of the impeller geometry. In this paper, the above problems have been studied.The main contents as follows1) With three special head design requirements of multi-stage pump, the hydraulic model of impeller was designed by Design of Experiment. The mathematical model that head under 0.3QBEP and QBEP and 1.2QBEP and main impeller geometry parameters is established. According to the mathematical model and the influence of impeller geometric parameters on the performance, the impeller is constantly optimized. The performance of best impeller hydraulic model was verified by test. Summarizing the problems in the process of design,the head fails to meet the design requirements under QBEP, 1.2 QBEP flow because of the accuracy of the numerical simulation to predict pump head and the accuracy of the established mathematical model.2) According to the choice of turbulence model, the research has carried out about different turbulence model impacting on the pump performance prediction accuracy. The Reynolds stress and eddy-viscosity category turbulence model’s features and applicable conditions and deficiencies in the performance prediction of centrifugal pump were described. The main discuss distinguish and prediction accuracy of Reynolds stress and eddy-viscosity turbulence category model for centrifugal pump head. The performance forecast of pump head under different discharge, the method has been proposed which choose effective turbulence model to ensure the performance prediction results more accurate. The selection method was verified.3) 3 levels by a factor of three full factorial experiments with 2 Level 3 factor full factorial experimental study by comparing methods for improving the accuracy of the mathematical model of the method. The accuracy of the mathematical model was verified by experiment. The method of establishing mathematical model and term factors square and interactions between factors affect the accuracy of the establishing mathematical model.4) In the mode FRONTIER optimized design platform, combining with established mathematical model based on the DOE method, the pump under multioperating conditions were optimized. By way of example to explain the mathematical model that was has been established. The main impeller geometric parameters to be optimized for achieving steep flow- head curve. With the design requirements as constraint conditions, the impeller geometric parameters satisficing the design requirements can obtain by automatic assignment for the impeller blade number and exit width and outlet Angle.