Uncertainty Analysis Of Blade Stagger Angle Based On Compressed Sensing

In the actual processing,installation and assembly of the turbine blade,due to the process level and other reasons,there will be some error between the actual stagger angle of the blade and the design stagger angle,the change of the blade stagger angle will cause the change of the attack angle of the incoming flow,which will directly affect the aerodynamic performance of the turbine cascade,which will deviate from the design condition to some extent.In the CFD simulation,we usually use the deterministic parameter as the boundary condition,ignoring the influence of the uncertainty of the stagger angle,which will cause the calculation result to deviate from the actual working condition.Based on the background of this technology,this paper studies the uncertainty quantification of the influence of blade stagger angle on the aerodynamic performance of the cascade.However,the research results at home and abroad are less.The results of this paper can be used for blade design guidance and impeller mechanical operation reference.In this paper,the compressed sensing theory is applied to the polynomial chaotic expansion method.An effective algorithm for uncertainty quantization calculation is proposed.The polynomial chaotic expansion method based on compressed sensing is used,and the weight of one-dimensional signal and test function verification algorithm is adopted.Compared with the traditional Monte Carlo method and polynomial chaotic expansion,the method requires less sample points and can achieve higher reconstruction accuracy.On the basis of numerical verification,the method of compressive sensing-based polynomial chaos expansion is applied to the influence of blade stagger variation on the aerodynamic performance of single-row cascade.It is assumed that the blade installation angle obeys a specific distribution and passes the reconstruction algorithm.The model between the input parameters and the output response is established,and the statistical characteristics of the output response are obtained.Based on what we have done,the research dimension is raised from one dimension to high dimension,and the simulation results are compared.The statistical distribution of the aerodynamic parameters of the cascade is found and the prototype simulation results are obtained.The matching is perfect and has good periodicity.The transmission direction of the uncertainty in the single row of cascades is directed from the suction surface to the pressure surface,and the influence range is from one blade in the opposite direction of the propagation direction to two blades downstream in the propagation direction and The fl ow path is composed,wherein the suction surface of the blade is more sensitive to the influence of uncertainty,and the pressure uncertainty is developed from the maximum deflection of the suction surface to the rear of the pressure surface and the vicinity of the trailing edge,and is continuously reduced in the process.In the quantitative study of multi-dimensional blade uncertainty,the uncertainty of different source acting on the same object will overlay their effects.Considering that in the turbine machinery,the rotors and stators usually are matched in stages,the research object will be changed from a single row of cascades to a first-stage dynamic cascade.By comparing the results of the study of single-row cascades,it is found that the direction of the transmission between the moving and stationary leaves is pointed by the s,In the rotor area,the main influence is on the four blades and their adjacent flow paths centered at the exit of the stationary error blade,the influence on the aerodynamic parameters of the moving blades is less regular,and the scope of influence is large;In the static blade area,due to the "feedback" phenomenon of the rotors,the aerodynamic parameters of the blades in the static zone are non-periodic,the pressure variance value of the runner in the static zone is reduced overall,the coverage of the corresponding variance interval becomes larger,the propagation direction of the feedback phenomenon is directed by the moving blade to the static blade along the axial component.The direction along the frontal line is the same as the direction of rotation of the blade.For the quantitative calculation of multiple blade uncertainties,uncertainty from different sources will still have a superimposed effect on the same object.When a single stagger angle distribution satisfies the standard normal distribution,the expansion ratio varies by ±0.016.When the three stagger angle distributions satisfy the standard normal distribution,Its expansion ratio varies by ±0.0258.