Research On Loss Model Of A Marine Compressor

With the continuous improvement of aero-engine performance,higher requirements are put forward for aero-engine.As a key component,compressor plays an important role.In the aerodynamic design of axial compressor,loss model is an important part,and the loss model is a continuous development process.The key content of improving the compressor design system is to modify the existing loss model and even propose a new loss model.Firstly,the flow structure in typical cascades is studied by numerical method,and the characteristic flow and loss characteristics of three types of cascades are analyzed.The results show that with the increase of the angle of attack,the separation form of the corner region changes from open separation to closed separation,and the concentrated shedding vortices and the corner vortices on the suction surface increase in the corner region.The low-energy fluid and the vortex structure interact in the corner region,resulting in the increase of aerodynamic loss.Secondly,the relative momentum loss thickness,shape factor,profile loss in reference state and profile loss in non reference state are modified by data regression based on the previous research methods of calculating profile loss in reference state and non reference state.In addition,this paper proposes a method to calculate the secondary flow loss and its distribution by decomposing the secondary flow loss into boundary layer loss and vortex loss.A new secondary flow loss calculation model is developed by refining the key parameters of the curve and fitting the key parameters in a certain range.Finally,the feasibility of the calculation method is verified by numerical calculation,The applicability of the new secondary flow loss calculation model is calibrated.Finally,with the help of neural network tools,based on a large number of blade geometry parameters and aerodynamic parameters data accumulated by numerical methods,a blade loss mapping model suitable for a certain range of parameters is established.Through numerical verification,the model has good loss prediction accuracy.