Research On Flow Field Characteristics,Monitoring And Identification Of High-speed Pump Under Small Flowrate Conditions

High-speed centrifugal pumps with characteristics of small flowrate and high head are widely used in petrochemical and aerospace fields.However,due to its high speed and complex flowing channels,its internal flow profiles are quite different from common centrifugal pumps,of which its fluid excitation force in off-design points has been a hot topic in relative fields.In this paper,a typical high-speed pump with flow rate Q_d=3m~3/h and rotation speed n=8500rpm is taken as the research object.The three-dimensional numerical simulation and unsteady pressure pulsation calculation were carried out by CFX software platform to study the characteristics of the internal flow field.A mathematic model was established to correlate the vibration signal with the internal flow field using the fuzzy nerve network,and an interactive software was designed to monitor the state of the high-speed pump equipment.The main research works are as follows:(1)The internal flow field research on the small flowrate point of high speed pump by CFD.Performing unstructured meshing of high-speed pump inlet pipes,induction wheels,impellers,volutes and other flow components,using CFX software to numerically simulate the high-speed pump from 0.2Q to 1.0Q to obtain the internal flow characteristics of pressure field,velocity field and vorticity.(1)The internal flow field research on the small flowrate point of high speed pump was performed by CFD.Hybrid mesh generation was fine carried out,with unstructured meshing in inducer,impellers,volutes and other flow components with complicated flowing boundaries,and structured meshing in inlet and outlet pipes and other small gap regions related with leakages.CFX software was used to numerically simulate the high-speed pump under conditions from 0.2Q to 1.0Q to obtain the internal flow characteristics.(3)The analysis of vibration features and reorganization of pump working status.The vibration feature was extracted by wavelet entropy and time domain variation.Based on the fuzzy neural network,a state recognition model of high-speed pump was established.The vibration characteristic data was driven to identify the steady state of the flow field in the high-speed pump.Design and development of the high-speed pump monitoring identification software.The system can collect and display the status data of the equipment in real time,and store the historical state data through database design.Meanwhile,it has the function of vibration and performance analysis,even condition status recognition.